gtk2/gdk/win32/gdkselection-win32.c

2865 lines
92 KiB
C
Raw Normal View History

/* GDK - The GIMP Drawing Kit
* Copyright (C) 1995-1997 Peter Mattis, Spencer Kimball and Josh MacDonald
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
* Copyright (C) 1998-2002 Tor Lillqvist
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
2012-02-27 13:01:10 +00:00
* License along with this library. If not, see <http://www.gnu.org/licenses/>.
*/
/*
* Modified by the GTK+ Team and others 1997-2000. See the AUTHORS
* file for a list of people on the GTK+ Team. See the ChangeLog
* files for a list of changes. These files are distributed with
* GTK+ at ftp://ftp.gtk.org/pub/gtk/.
*/
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/*
GTK+ selection works like this:
There are three selections that matter - GDK_SELECTION_CLIPBOARD,
GDK_SELECTION_PRIMARY and DND. Primary selection is only handled
internally by GTK+ (it's not portable to Windows). DND is actually
represented by two selections - LOCAL and OLE2, one for each DnD protocol,
but they work the same way.
"Target" is a GdkAtom describing a clipboard format.
For Clipboard:
GTK+ calls gtk_clipboard_set_contents(), which first ensures the
clipboard is owned by the clipboard widget (which also indirectly
causes an SelectionRequest xevent to be sent to it), then clears the old
supported targets from the clipboard, then adds all the
targets it's given to the clipboard. No data is sent anywhere.
gtk_clipboard_set_contents() is also given a callback to invoke when
the actual data is needed. This callback is implemented by the widget
from which the data can be put into clipboard.
GTK+ might also call gtk_clipboard_set_can_store(), which sets the
targets for which the data can be put into system clipboard, so that
it remains usable even if the application is no longer around. Usually
all data formats are storable, except for the shortcut formats, which
refer to actual widgets directly, and are thus only working while
the application is alive.
("C:" means clipboard client (requestor), "S:" means clipboard server (provider))
When something needs to be obtained from clipboard, GTK+ calls
C: gtk_selection_convert().
That function has a shortcut where it directly gets the selection contents by calling
S: gtk_selection_invoke_handler(),
asking the widget to provide data, and then calling
C: gtk_selection_retrieval_report()
to report the data back to the caller.
If that shortcut isn't possible (selection is owned by another process),
gtk_selection_convert() calls
C:gdk_selection_convert() (_gdk_x11_display_convert_selection())
On X11 gdk_selection_convert() just calls
C:XConvertSelection(),
which sends SelectionRequest xevent to the window that owns the selection.
The client gives its clipboard window as the requestor for that event,
and gives the property as GDK_SELECTION.
Server-side GTK+ catches SelectionRequest in a
S:_gtk_selection_request()
event handler, which calls
S:gtk_selection_invoke_handler()
to get the data, and then calls
S:gdk_property_change() (_gdk_x11_window_change_property())
to submit the data, by setting the property given by the message sender
(GDK_SELECTION) on the requestor window (our client clipboard window).
On X11 data submission takes from of
S:XChangeProperty()
call, which causes SelectionNotify (and PropertyNotify for INCR)
xevent to be sent, which client-side GTK+ catches and handles in
C:_gtk_selection_notify()
(and
C:_gtk_selection_property_notify(),
for INCR)
event handler, which calls
C:gtk_selection_retrieval_report()
to report back to the caller. The caller gets the property
data from the window, and returns it up the stack.
On X11 the "TARGETS" target might be given in a SelectionRequest xmessage to request
all supported targets for a selection.
If data must be stored on the clipboard, because the application is quitting,
GTK+ will call
S:gdk_clipboard_store() -> gdk_display_store_clipboard() (gdk_x11_display_store_clipboard())
on all the clipboards it owns.
X11 gdk_display_store_clipboard() puts a list of storeable targets into GDK_SELECTION
property of the clipboard window, then calls
S:XConvertSelection()
on the clipboard manager window (retrieved from the CLIPBOARD_MANAGER atom),
and the clipboard manager responds by requesting all these formats and storing the data,
then responds with SelectionNotify xevent to allow the application to quit.
When clipboard owner changes, the old owner receives SelectionClear xevent,
GTK+ handles it by clearing the clipboard object on its own level, GDK
is not involved.
On Windows:
Clipboard is opened by OpenClipboard(), emptied by EmptyClipboard() (which also
makes the window the clipboard owner), data is put into it by SetClipboardData().
Clipboard is closed with CloseClipboard().
If SetClipboardData() is given a NULL data value, the owner will later
receive WM_RENDERFORMAT message, in response to which it must call
SetClipboardData() with the provided handle and the actual data this time.
This way applications can avoid storing everything in the clipboard
all the time, only putting the data there as it is requested by other applications.
At some undefined points of time an application might get WM_RENDERALLFORMATS
message, it should respond by opening the clipboard and rendering
into it all the data that it offers, as if responding to multiple WM_RENDERFORMAT
messages.
On GDK-Win32:
GTK+ calls gtk_clipboard_set_contents(), which first ensures the
clipboard is owned by the clipboard widget (calls OpenClipboard(),
then EmptyClipboard() to become the owner, then
sends a TARGETS GDK_SELECTION_REQUEST to itself, without closing the clipboard),
then clears the old supported targets from the clipboard, then adds all the
targets it's given to the clipboard. No data is sent anywhere.
gtk_clipboard_set_contents() is also given a callback to invoke when
the actual data is needed. This callback is implemented by the widget
from which the data can be put into clipboard.
GTK+ might also call gtk_clipboard_set_can_store(), which sets the
targets for which the data can be put into system clipboard, so that
it remains usable even if the application is no longer around. Usually
all data formats are storable, except for the shortcut formats, which
refer to actual widgets directly, and are thus only working while
the application is alive.
("C:" means clipboard client (requestor), "S:" means clipboard server (provider))
("transmute" here means "change the format of some data"; this term is used here
instead of "convert" to avoid clashing with g(t|d)k_selection_convert(), which
is completely unrelated)
When something needs to be obtained from clipboard, GTK+ calls
C: gtk_selection_convert().
That function has a shortcut where it directly gets the selection contents by calling
S: gtk_selection_invoke_handler(),
asking the widget to provide data, and then calling
C: gtk_selection_retrieval_report()
to report the data back to the caller.
If that shortcut isn't possible (selection is owned by another process),
gtk_selection_convert() calls
C:gdk_selection_convert() (_gdk_win32_display_convert_selection())
On GDK-Win32 gdk_selection_convert() just calls
C:OpenClipboard()
to open clipboard (if that fails, it shedules a timeout to regularly
try to open clipboard for the next 30 seconds, and do the actions
outlined below once the clipboard is opened, or notify about
conversion failure after 30 seconds),
C:EnumClipboardFormats() (2000+)
to get the list of supported formats, figures out the format it should
use to request the data (first it looks for supported formats with names
that match the target name, then looks through compatibility
formats for the target and checks whether these are supported).
Note that it has no list of supported targets at hand,
just the single requested target, and thus it might have
to do some transmutation between formats; the caller up the stack
either only supports just one format that it asks for,
or supports multiple formats and asks for them in sequence (from
the most preferred to the least preferred), until one call succeeds,
or supports multiple formats and asks for the TARGETS format first,
and then figures out what to ask for - GDK can't know that.
Either way, GDK has to call
C:GetClipboardData()
to get the data (this causes WM_RENDERFORMAT to be sent to the owner,
if the owner uses delayed rendering for the requested format, otherwise
it just picks the data right from the OS)
Server-side GDK catches WM_RENDERFORMAT, figures out a target
to request (this one is easier, as it has the list of supported
targets saved up), and posts a GDK_SELECTION_REQUEST event, then runs the main loop,
while GTK+ catches the event in a
S:_gtk_selection_request()
event handler, which calls
S:gtk_selection_invoke_handler()
to get the data, and then calls
S:gdk_property_change() (_gdk_win32_window_change_property())
to submit the data, by first transmuting it to the format actually requested
by the sender of WM_RENDERFORMAT, and then by returning thedata back up the stack,
to the WM_RENDERFORMAT handler, which then calls
S:SetClipboardData()
with the handle provided by the sender.
Meanwhile, the client code, still in
C:_gdk_win32_display_convert_selection(),
gets the data in response to GetClipboardData(),
transmutes it (if needed) to the target format, sets the requested
window property to that data (unlike change_property!),
calls
C:CloseClipboard() (if there are no more clipboard opeartions
scheduled)
and posts a GDK_SELECTION_NOTIFY event, which GTK+ catches in
C:_gtk_selection_notify()
event handler, which calls
C:gtk_selection_retrieval_report()
to report back to the caller. The caller gets the property
data from the window, and returns it up the stack.
On GDK-Win32 the "TARGETS" target might be given in a GDK_SELECTION_REQUEST to request
all supported targets for a selection.
Note that this server side -
client side should call gdk_selection_convert() -> gdk_selection_convert() with "TARGETS" target
to get the list of targets offered by the clipboard holder. It never causes GDK_SELECTION_REQUEST
to be generated, just queries the system clipboard.
On server side GDK_SELECTION_REQUEST is only generated internally:
in response to WM_RENDERFORMAT (it renders a target),
in response to idataobject_getdata() (it renders a target),
after DnD ends (with a DELETE target, this is caught by GTK to make it delete the selection),
and in response to owner change, with TARGETS target, which makes it register its formats by calling
S:SetClipboardData(..., NULL)
If data must be stored on the clipboard, because the application is quitting,
GTK+ will call
S:gdk_clipboard_store() -> gdk_display_store_clipboard() (gdk_win32_display_store_clipboard())
on all the clipboards it owns.
GDK-Win32 gdk_display_store_clipboard() sends WM_RENDERALLFORMATS to the window,
then posts a GDK_SELECTION_NOTIFY event allow the application to quit.
When clipboard owner changes, the old owner receives WM_DESTROYCLIPBOARD message,
GDK handles it by posting a GDK_SELECTION_CLEAR event, which
GTK+ handles by clearing the clipboard object on its own level.
Any operations that require OpenClipboard()/CloseClipboard() combo (i.e.
everything, except for WM_RENDERFORMAT handling) must be put into a queue,
and then a once-per-second-for-up-to-30-seconds timeout must be added.
The timeout function must call OpenClipboard(),
and then proceed to perform the queued actions on the clipboard, once it opened,
or return and try again a second later, as long as there are still items in the queue,
and remove the queue items that are older than 30 seconds.
Once the queue is empty, the clipboard is closed.
DND:
GDK-Win32:
S:idataobject_getdata()
sends a GDK_SELECTION_REQUEST event, which results in a call to
S:_gdk_win32_window_change_property()
which passes clipboard data back via the selection singleton.
GDK-Win32 uses delayed rendering for all formats, even text.
GTK+ will call
C:gtk_selection_convert() -> gdk_selection_convert() (_gdk_win32_display_convert_selection())
to get the data associated with the drag, when GTK+ apps want to inspect the data,
but with a OLE2_DND selection instead of CLIPBOARD selection.
_gdk_win32_display_convert_selection() queries the droptarget global variable,
which should already contain a matched list of supported formats and targets,
picks a format there, then queries it from the IDataObject that the droptarget kept around.
Then optionally transmutes the data, and sets the property. Then posts GDK_SELECTION_NOTIFY.
GTK+ catches that event and processes it, causeing "selection-received" signal to
be emitted on the selection widget, and its handler is
C:gtk_drag_selection_received(),
which emits the "drag-data-received" signal for the app.
*/
#include "config.h"
#include <string.h>
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
#include <stdlib.h>
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* For C-style COM wrapper macros */
#define COBJMACROS
/* for CIDA */
#include <shlobj.h>
Large changes to the Win32 backend, partially made necessary by the 2000-05-02 Tor Lillqvist <tml@iki.fi> Large changes to the Win32 backend, partially made necessary by the changes to the backend-independent internal structures. Attempts to implement similar backing store stuff as on X11. The current (CVS) version of the Win32 backend is *not* as stable as it was before the no-flicker branch was merged. A zipfile with that version is available from http://www.gimp.org/win32/. That should be use by "production" code until this CVS version is usable. (But note, the Win32 backend has never been claimed to be "production quality".) * README.win32: Add the above comment about versions. * gdk/gdkwindow.c: Don't use backing store for now on Win32. * gdk/gdk.def: Update. * gdk/gdkfont.h: Declare temporary Win32-only functions. Will presumably be replaced by some more better mechanism as 1.4 gets closer to release shape. * gdk/makefile.{cygwin,msc}: Update. * gdk/win32/*.c: Correct inclusions of the backend-specific and internal headers. Change code according to changes in these. Use gdk_drawable_*, not gdk_window_* where necessary. * gdk/win32/gdkdnd-win32.c: Use MISC selector for GDK_NOTE, not our old DND. * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text): Don't try to interpret single characters as UTF-8. Thanks to Hans Breuer. Use correct function name in warning messages. * gdk/win32/gdkevents-win32.c: Use correct parameter lists for the GSourceFuncs gdk_event_prepare and gdk_event_check. (gdk_event_get_graphics_expose): Do implement, use PeekMessage. Thanks to Hans Breuer. (event_mask_string): Debugging function to print an GdkEventMask. (gdk_pointer_grab): Use it. * gdk/win32/gdkfont-win32.c: The Unicode subrange that the (old) book I used claimed was Hangul actually is CJK Unified Ideographs Extension A. Also, Hangul Syllables were missing. Improve logging. * gdk/win32/gdkgc-win32.c: Largish changes. * gdk/win32/gdkim-win32.c (gdk_set_locale): Use g_win32_getlocale() from GLib, and not setlocale() to get current locale name. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkwin32.h: Move stuff from gdkprivate-win32.h to gdkwin32.h, similarily as in the X11 backend. * gdk/win32/gdkwindow-win32.c (gdk_propagate_shapes): Bugfix, assignment was used instead of equals in if test. Thanks to Hans Breuer. * gdk/win32/makefile.{cygwin,msc} * gtk/makefile.{cygwin,msc}: Updates. Better kludge to get the path to the Win32 headers that works also with the mingw compiler. * gtk/gtkstyle.c: Include <string.h>.
2000-05-01 22:06:49 +00:00
#include "gdkproperty.h"
Changes multihead reorganizing code for win32 support, mostly from a patch Wed Jun 5 18:34:47 2002 Owen Taylor <otaylor@redhat.com> Changes multihead reorganizing code for win32 support, mostly from a patch by Hans Breuer. * gdk/gdkcolor.c gdk/x11/gdkcolor-x11.c gdk/gdkcursor.c gdk/x11/gdkcursor-x11.c gdk/gdkevents.c gdk/x11/gdkevents-x11.c gdk/gdkfont.c gdk/x11/gdkfont-x11.c gdk/gdkkeys.c gdk/x11/gdkkeys-x11.c gdk/gdkimage.c gdk/x11/gdkimage-x11.c gdk/gdkscreen.c gdk/x11/gdkmain-x11.c gdk/gdkdisplay.c gdk/gdkevents-x11.c gdk/gdkpango.c gdk/x11/gdkpango-x11.c gdk/gdkselection.c gdk/x11/gdkselection-x11.c gdk/gdkwindow.c gdk/x11/gdkwindow-x11.c gdk/gdkvisual.c gdk/x11/gdkvisual-x11.c: Move port-independent singlehead wrapper functions into port-independent part of GDK. (#80009) * gdk/win32/gdkcolor-win32.c gdk/win32/gdkcursor-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkfont-win32.c gdk/win32/gdkimage-win32.c gdk/win32/gdkkeys-win32.c gdk/win32/gdkmain-win32.c gdk/win32/gdkproperty-win32.c gdk/win32/gdkselection-win32.c gdk/win32/gkwindow-win32.c: Turn singlehead functions into "multihead" functions that ignore their GdkDisplay or GdkScreen arguments. * gdk/win32/gdkdrawable-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkinput-win32.c gdk/win32/gdkprivate-win32.h: Misc multihead-compatibility changes. * gtk/gtk.def gdk/gdk.def: Update for multihead functions. * gdk/gdkcolormap.h gdk/gdkvisual.h gdk/x11/gdkcolormap-x11.c gdk/x11/gdkvisual-x11.c: Remove the screen fields from the public parts of the colormap/visual structures, add accessors instead. * gdk/gdkpixbuf-render.c gdk/gdkpixmap.c gdk/gdkrgb.c gdk/x11/gdkcolormap-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkprivate-x11.h gtk/gtkgc.c gtk/gtkstyle.c gtk/gtkwidget.c: Use accessors to get the screen for colormaps, visuals; move the fields into the private structures for the x11 backend. * gdk/gdkdisplay.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/gdkscreen.[ch] gdk/x11/gdkscreen-x11.c: Remove virtualization of screen and display functions. (#79990, patch from Erwann Chenede) * gdk/win32/gdkdisplay-x11.c gdk/win32/gdkscreen-win32.c gdk/win32/{Makefile.am, makefile.msc, makefile.mingw}: New files containing stub implementations of Display, Screen functions. * gdk/x11/gdkscreen-x11.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/x11/gdkx.h: Clean up function exports and what headers they are in. (#79954) * gdk/x11/gdkx.h: Fix macro that was referring to a non-existant screen->screen_num. (In the patch for #79972, Erwann Chenede) * gdk/gdkscreen.c gdk/gdkwindow.c gdk/x11/gdkinternals.h gdk/x11/gdkscreen-x11.c: Fix gdk_screen_get_window_at_pointer() to use window hooks. (#79972, patch partly from Erwann Chenede) * gdk/x11/gdkdisplay-x11.c gdk/x11/gdkevents-x11.c: Fix some warnings.
2002-06-06 00:26:42 +00:00
#include "gdkdisplay.h"
Large changes to the Win32 backend, partially made necessary by the 2000-05-02 Tor Lillqvist <tml@iki.fi> Large changes to the Win32 backend, partially made necessary by the changes to the backend-independent internal structures. Attempts to implement similar backing store stuff as on X11. The current (CVS) version of the Win32 backend is *not* as stable as it was before the no-flicker branch was merged. A zipfile with that version is available from http://www.gimp.org/win32/. That should be use by "production" code until this CVS version is usable. (But note, the Win32 backend has never been claimed to be "production quality".) * README.win32: Add the above comment about versions. * gdk/gdkwindow.c: Don't use backing store for now on Win32. * gdk/gdk.def: Update. * gdk/gdkfont.h: Declare temporary Win32-only functions. Will presumably be replaced by some more better mechanism as 1.4 gets closer to release shape. * gdk/makefile.{cygwin,msc}: Update. * gdk/win32/*.c: Correct inclusions of the backend-specific and internal headers. Change code according to changes in these. Use gdk_drawable_*, not gdk_window_* where necessary. * gdk/win32/gdkdnd-win32.c: Use MISC selector for GDK_NOTE, not our old DND. * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text): Don't try to interpret single characters as UTF-8. Thanks to Hans Breuer. Use correct function name in warning messages. * gdk/win32/gdkevents-win32.c: Use correct parameter lists for the GSourceFuncs gdk_event_prepare and gdk_event_check. (gdk_event_get_graphics_expose): Do implement, use PeekMessage. Thanks to Hans Breuer. (event_mask_string): Debugging function to print an GdkEventMask. (gdk_pointer_grab): Use it. * gdk/win32/gdkfont-win32.c: The Unicode subrange that the (old) book I used claimed was Hangul actually is CJK Unified Ideographs Extension A. Also, Hangul Syllables were missing. Improve logging. * gdk/win32/gdkgc-win32.c: Largish changes. * gdk/win32/gdkim-win32.c (gdk_set_locale): Use g_win32_getlocale() from GLib, and not setlocale() to get current locale name. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkwin32.h: Move stuff from gdkprivate-win32.h to gdkwin32.h, similarily as in the X11 backend. * gdk/win32/gdkwindow-win32.c (gdk_propagate_shapes): Bugfix, assignment was used instead of equals in if test. Thanks to Hans Breuer. * gdk/win32/makefile.{cygwin,msc} * gtk/makefile.{cygwin,msc}: Updates. Better kludge to get the path to the Win32 headers that works also with the mingw compiler. * gtk/gtkstyle.c: Include <string.h>.
2000-05-01 22:06:49 +00:00
#include "gdkprivate-win32.h"
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
#include "gdkselection-win32.h"
#include "gdk/gdkdndprivate.h"
#include "gdkwin32dnd-private.h"
#include "gdkwin32.h"
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
typedef struct _GdkWin32ClipboardQueueInfo GdkWin32ClipboardQueueInfo;
typedef enum _GdkWin32ClipboardQueueAction GdkWin32ClipboardQueueAction;
enum _GdkWin32ClipboardQueueAction
{
GDK_WIN32_CLIPBOARD_QUEUE_ACTION_CONVERT = 0,
GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS
};
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
struct _GdkWin32ClipboardQueueInfo
{
GdkDisplay *display;
GdkWindow *requestor;
GdkAtom selection;
GdkAtom target;
guint32 time;
/* Number of seconds since we started our
* attempts to open clipboard.
*/
guint32 idle_time;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* What to do once clipboard is opened */
GdkWin32ClipboardQueueAction action;
};
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* List of GdkWin32ClipboardQueueInfo slices */
static GList *clipboard_queue = NULL;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
#define HIDA_GetPIDLFolder(pida) (LPCITEMIDLIST)(((LPBYTE)pida)+(pida)->aoffset[0])
#define HIDA_GetPIDLItem(pida, i) (LPCITEMIDLIST)(((LPBYTE)pida)+(pida)->aoffset[i+1])
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
G_DEFINE_TYPE (GdkWin32Selection, gdk_win32_selection, G_TYPE_OBJECT)
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
static void
gdk_win32_selection_class_init (GdkWin32SelectionClass *klass)
{
}
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
void
_gdk_win32_selection_init (void)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
_win32_selection = GDK_WIN32_SELECTION (g_object_new (GDK_TYPE_WIN32_SELECTION, NULL));
}
static void
gdk_win32_selection_init (GdkWin32Selection *win32_selection)
{
GArray *atoms;
GArray *cfs;
GSList *pixbuf_formats;
GSList *rover;
int i;
GArray *comp;
GdkSelTargetFormat fmt;
win32_selection->ignore_destroy_clipboard = FALSE;
win32_selection->clipboard_opened_for = INVALID_HANDLE_VALUE;
win32_selection->dnd_target_state = GDK_WIN32_DND_NONE;
win32_selection->dnd_source_state = GDK_WIN32_DND_NONE;
win32_selection->dnd_data_object_target = NULL;
win32_selection->property_change_format = 0;
win32_selection->property_change_data = NULL;
win32_selection->property_change_target_atom = 0;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
atoms = g_array_sized_new (FALSE, TRUE, sizeof (GdkAtom), GDK_WIN32_ATOM_INDEX_LAST);
g_array_set_size (atoms, GDK_WIN32_ATOM_INDEX_LAST);
cfs = g_array_sized_new (FALSE, TRUE, sizeof (UINT), GDK_WIN32_CF_INDEX_LAST);
g_array_set_size (cfs, GDK_WIN32_CF_INDEX_LAST);
win32_selection->known_atoms = atoms;
win32_selection->known_clipboard_formats = cfs;
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_GDK_SELECTION) = g_intern_static_string ("GDK_SELECTION");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CLIPBOARD_MANAGER) = g_intern_static_string ("CLIPBOARD_MANAGER");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_WM_TRANSIENT_FOR) = g_intern_static_string ("WM_TRANSIENT_FOR");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_TARGETS) = g_intern_static_string ("TARGETS");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_DELETE) = g_intern_static_string ("DELETE");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_SAVE_TARGETS) = g_intern_static_string ("SAVE_TARGETS");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_UTF8_STRING) = g_intern_static_string ("UTF8_STRING");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_TEXT) = g_intern_static_string ("TEXT");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_COMPOUND_TEXT) = g_intern_static_string ("COMPOUND_TEXT");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_TEXT_URI_LIST) = g_intern_static_string ("text/uri-list");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_TEXT_HTML) = g_intern_static_string ("text/html");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_PNG) = g_intern_static_string ("image/png");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_JPEG) = g_intern_static_string ("image/jpeg");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_BMP) = g_intern_static_string ("image/bmp");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_GIF) = g_intern_static_string ("image/gif");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_LOCAL_DND_SELECTION) = g_intern_static_string ("LocalDndSelection");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_DROPFILES_DND) = g_intern_static_string ("DROPFILES_DND");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_OLE2_DND) = g_intern_static_string ("OLE2_DND");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_PNG)= g_intern_static_string ("PNG");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_JFIF) = g_intern_static_string ("JFIF");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_GIF) = g_intern_static_string ("GIF");
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* These are a bit unusual. It's here to allow GTK+ applications
* to actually support CF_DIB and Shell ID List clipboard formats on their own,
* instead of allowing GDK to use them internally for interoperability.
*/
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CF_DIB) = g_intern_static_string ("CF_DIB");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CFSTR_SHELLIDLIST) = g_intern_static_string (CFSTR_SHELLIDLIST);
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CF_UNICODETEXT) = g_intern_static_string ("CF_UNICODETEXT");
_gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CF_TEXT) = g_intern_static_string ("CF_TEXT");
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* MS Office 2007, at least, offers images in common file formats
* using clipboard format names like "PNG" and "JFIF". So we follow
* the lead and map the GDK target name "image/png" to the clipboard
* format name "PNG" etc.
*/
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_PNG) = RegisterClipboardFormatA ("PNG");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_JFIF) = RegisterClipboardFormatA ("JFIF");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_GIF) = RegisterClipboardFormatA ("GIF");
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_UNIFORMRESOURCELOCATORW) = RegisterClipboardFormatA ("UniformResourceLocatorW");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_CFSTR_SHELLIDLIST) = RegisterClipboardFormatA (CFSTR_SHELLIDLIST);
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_HTML_FORMAT) = RegisterClipboardFormatA ("HTML Format");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_TEXT_HTML) = RegisterClipboardFormatA ("text/html");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_PNG) = RegisterClipboardFormatA ("image/png");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_JPEG) = RegisterClipboardFormatA ("image/jpeg");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_BMP) = RegisterClipboardFormatA ("image/bmp");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_GIF) = RegisterClipboardFormatA ("image/gif");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_TEXT_URI_LIST) = RegisterClipboardFormatA ("text/uri-list");
_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_UTF8_STRING) = RegisterClipboardFormatA ("UTF8_STRING");
win32_selection->sel_prop_table = g_hash_table_new (NULL, NULL);
win32_selection->sel_owner_table = g_hash_table_new (NULL, NULL);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
pixbuf_formats = gdk_pixbuf_get_formats ();
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_selection->n_known_pixbuf_formats = 0;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
for (rover = pixbuf_formats; rover != NULL; rover = rover->next)
{
gchar **mime_types =
gdk_pixbuf_format_get_mime_types ((GdkPixbufFormat *) rover->data);
gchar **mime_type;
for (mime_type = mime_types; *mime_type != NULL; mime_type++)
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_selection->n_known_pixbuf_formats++;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_selection->known_pixbuf_formats = g_new (GdkAtom, win32_selection->n_known_pixbuf_formats);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
i = 0;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
for (rover = pixbuf_formats; rover != NULL; rover = rover->next)
{
gchar **mime_types =
gdk_pixbuf_format_get_mime_types ((GdkPixbufFormat *) rover->data);
gchar **mime_type;
for (mime_type = mime_types; *mime_type != NULL; mime_type++)
win32_selection->known_pixbuf_formats[i++] = g_intern_string (*mime_type);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
}
g_slist_free (pixbuf_formats);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_selection->dnd_selection_targets = g_array_new (FALSE, FALSE, sizeof (GdkSelTargetFormat));
win32_selection->clipboard_selection_targets = g_array_new (FALSE, FALSE, sizeof (GdkSelTargetFormat));
win32_selection->compatibility_formats = g_hash_table_new_full (NULL, NULL, NULL, (GDestroyNotify) g_array_unref);
/* GTK+ actually has more text formats, but it's unlikely that we'd
* get anything other than UTF8_STRING these days.
* GTKTEXTBUFFERCONTENTS format can potentially be converted to
* W32-compatible rich text format, but that's too complex to address right now.
*/
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 3);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_UTF8_STRING);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_UTF8_STRING);
fmt.transmute = FALSE;
g_array_append_val (comp, fmt);
fmt.format = CF_UNICODETEXT;
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
fmt.format = CF_TEXT;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_formats, fmt.target, comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 3);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_PNG);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_PNG);
fmt.transmute = FALSE;
g_array_append_val (comp, fmt);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_PNG);
g_array_append_val (comp, fmt);
fmt.format = CF_DIB;
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_formats, fmt.target, comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 4);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_JPEG);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_JPEG);
fmt.transmute = FALSE;
g_array_append_val (comp, fmt);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_JFIF);
g_array_append_val (comp, fmt);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_PNG);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
fmt.format = CF_DIB;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_formats, fmt.target, comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 4);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_GIF);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_GIF);
fmt.transmute = FALSE;
g_array_append_val (comp, fmt);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_GIF);
g_array_append_val (comp, fmt);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_PNG);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
fmt.format = CF_DIB;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_formats, fmt.target, comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 2);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_BMP);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_IMAGE_BMP);
fmt.transmute = FALSE;
g_array_append_val (comp, fmt);
fmt.format = CF_DIB;
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_formats, fmt.target, comp);
/* Not implemented, but definitely possible
comp = g_array_sized_new (FALSE, FALSE, 2);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_TEXT_URI_LIST);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_TEXT_URI_LIST);
fmt.transmute = FALSE;
g_array_append_val (comp, fmt);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_CFSTR_SHELLIDLIST);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_formats, fmt.target, comp);
*/
win32_selection->compatibility_targets = g_hash_table_new_full (NULL, NULL, NULL, (GDestroyNotify) g_array_unref);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 2);
fmt.format = CF_TEXT;
fmt.transmute = FALSE;
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CF_TEXT);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_UTF8_STRING);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (CF_TEXT), comp);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 2);
fmt.format = CF_UNICODETEXT;
fmt.transmute = FALSE;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CF_UNICODETEXT);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_UTF8_STRING);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (CF_UNICODETEXT), comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 3);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_PNG);
fmt.transmute = FALSE;
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_PNG);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_PNG);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_BMP);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_PNG)), comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 4);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_JFIF);
fmt.transmute = FALSE;
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_JFIF);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_JPEG);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_PNG);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_BMP);
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_JFIF)), comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 4);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_GIF);
fmt.transmute = FALSE;
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_GIF);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_GIF);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_PNG);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_BMP);
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_GIF)), comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 3);
fmt.format = CF_DIB;
fmt.transmute = FALSE;
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CF_DIB);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_IMAGE_BMP);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (CF_DIB), comp);
comp = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), 3);
fmt.format = _gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_CFSTR_SHELLIDLIST);
fmt.transmute = FALSE;
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_CFSTR_SHELLIDLIST);
g_array_append_val (comp, fmt);
fmt.target = _gdk_atom_array_index (atoms, GDK_WIN32_ATOM_INDEX_TEXT_URI_LIST);
fmt.transmute = TRUE;
g_array_append_val (comp, fmt);
g_hash_table_replace (win32_selection->compatibility_targets, GINT_TO_POINTER (_gdk_cf_array_index (cfs, GDK_WIN32_CF_INDEX_CFSTR_SHELLIDLIST)), comp);
}
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
/* The specifications for COMPOUND_TEXT and STRING specify that C0 and
* C1 are not allowed except for \n and \t, however the X conversions
* routines for COMPOUND_TEXT only enforce this in one direction,
* causing cut-and-paste of \r and \r\n separated text to fail.
* This routine strips out all non-allowed C0 and C1 characters
* from the input string and also canonicalizes \r, and \r\n to \n
*/
static gchar *
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
sanitize_utf8 (const gchar *src,
gint length)
{
GString *result = g_string_sized_new (length + 1);
const gchar *p = src;
const gchar *endp = src + length;
while (p < endp)
{
if (*p == '\r')
{
p++;
if (*p == '\n')
p++;
g_string_append_c (result, '\n');
}
else
{
gunichar ch = g_utf8_get_char (p);
char buf[7];
gint buflen;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
if (!((ch < 0x20 && ch != '\t' && ch != '\n') || (ch >= 0x7f && ch < 0xa0)))
{
buflen = g_unichar_to_utf8 (ch, buf);
g_string_append_len (result, buf, buflen);
}
p = g_utf8_next_char (p);
}
}
g_string_append_c (result, '\0');
return g_string_free (result, FALSE);
}
static gchar *
_gdk_utf8_to_string_target_internal (const gchar *str,
gint length)
{
GError *error = NULL;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
gchar *tmp_str = sanitize_utf8 (str, length);
gchar *result = g_convert_with_fallback (tmp_str, -1,
"ISO-8859-1", "UTF-8",
NULL, NULL, NULL, &error);
if (!result)
{
g_warning ("Error converting from UTF-8 to STRING: %s",
error->message);
g_error_free (error);
}
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
g_free (tmp_str);
return result;
}
static void
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
selection_property_store (GdkWindow *owner,
GdkAtom type,
gint format,
guchar *data,
gint length)
{
GdkSelProp *prop;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
prop = g_hash_table_lookup (win32_sel->sel_prop_table, GDK_WINDOW_HWND (owner));
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
if (prop != NULL)
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
{
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_free (prop->data);
g_free (prop);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_hash_table_remove (win32_sel->sel_prop_table, GDK_WINDOW_HWND (owner));
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
}
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
prop = g_new (GdkSelProp, 1);
prop->data = data;
prop->length = length;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
prop->bitness = format;
prop->target = type;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_hash_table_insert (win32_sel->sel_prop_table, GDK_WINDOW_HWND (owner), prop);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
}
void
_gdk_dropfiles_store (gchar *data)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
if (data != NULL)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_assert (win32_sel->dropfiles_prop == NULL);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->dropfiles_prop = g_new (GdkSelProp, 1);
win32_sel->dropfiles_prop->data = (guchar *) data;
win32_sel->dropfiles_prop->length = strlen (data) + 1;
win32_sel->dropfiles_prop->bitness = 8;
win32_sel->dropfiles_prop->target = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TEXT_URI_LIST);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
}
else
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->dropfiles_prop != NULL)
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_free (win32_sel->dropfiles_prop->data);
g_free (win32_sel->dropfiles_prop);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->dropfiles_prop = NULL;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
}
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
static void
generate_selection_notify (GdkWindow *requestor,
GdkAtom selection,
GdkAtom target,
GdkAtom property,
guint32 time)
{
GdkEvent tmp_event;
memset (&tmp_event, 0, sizeof (tmp_event));
tmp_event.selection.type = GDK_SELECTION_NOTIFY;
tmp_event.selection.window = requestor;
tmp_event.selection.send_event = FALSE;
tmp_event.selection.selection = selection;
tmp_event.selection.target = target;
tmp_event.selection.property = property;
tmp_event.selection.requestor = 0;
tmp_event.selection.time = time;
gdk_event_put (&tmp_event);
}
void
_gdk_win32_clear_clipboard_queue ()
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GList *tmp_list, *next;
GdkWin32ClipboardQueueInfo *info;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GDK_NOTE (DND, g_print ("Clear clipboard queue\n"));
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
for (tmp_list = clipboard_queue; tmp_list; tmp_list = next)
{
info = (GdkWin32ClipboardQueueInfo *) tmp_list->data;
next = g_list_next (tmp_list);
clipboard_queue = g_list_remove_link (clipboard_queue, tmp_list);
g_list_free (tmp_list);
switch (info->action)
{
case GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS:
break;
case GDK_WIN32_CLIPBOARD_QUEUE_ACTION_CONVERT:
generate_selection_notify (info->requestor, info->selection, info->target, NULL, info->time);
break;
}
g_clear_object (&info->requestor);
g_slice_free (GdkWin32ClipboardQueueInfo, info);
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->targets_request_pending = FALSE;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Send ourselves a selection request message with
* the TARGETS target, we will do multiple SetClipboarData(...,NULL)
* calls in response to announce the formats we support.
*/
static void
send_targets_request (guint time)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWindow *owner;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
GdkEvent tmp_event;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->targets_request_pending)
return;
Changes multihead reorganizing code for win32 support, mostly from a patch Wed Jun 5 18:34:47 2002 Owen Taylor <otaylor@redhat.com> Changes multihead reorganizing code for win32 support, mostly from a patch by Hans Breuer. * gdk/gdkcolor.c gdk/x11/gdkcolor-x11.c gdk/gdkcursor.c gdk/x11/gdkcursor-x11.c gdk/gdkevents.c gdk/x11/gdkevents-x11.c gdk/gdkfont.c gdk/x11/gdkfont-x11.c gdk/gdkkeys.c gdk/x11/gdkkeys-x11.c gdk/gdkimage.c gdk/x11/gdkimage-x11.c gdk/gdkscreen.c gdk/x11/gdkmain-x11.c gdk/gdkdisplay.c gdk/gdkevents-x11.c gdk/gdkpango.c gdk/x11/gdkpango-x11.c gdk/gdkselection.c gdk/x11/gdkselection-x11.c gdk/gdkwindow.c gdk/x11/gdkwindow-x11.c gdk/gdkvisual.c gdk/x11/gdkvisual-x11.c: Move port-independent singlehead wrapper functions into port-independent part of GDK. (#80009) * gdk/win32/gdkcolor-win32.c gdk/win32/gdkcursor-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkfont-win32.c gdk/win32/gdkimage-win32.c gdk/win32/gdkkeys-win32.c gdk/win32/gdkmain-win32.c gdk/win32/gdkproperty-win32.c gdk/win32/gdkselection-win32.c gdk/win32/gkwindow-win32.c: Turn singlehead functions into "multihead" functions that ignore their GdkDisplay or GdkScreen arguments. * gdk/win32/gdkdrawable-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkinput-win32.c gdk/win32/gdkprivate-win32.h: Misc multihead-compatibility changes. * gtk/gtk.def gdk/gdk.def: Update for multihead functions. * gdk/gdkcolormap.h gdk/gdkvisual.h gdk/x11/gdkcolormap-x11.c gdk/x11/gdkvisual-x11.c: Remove the screen fields from the public parts of the colormap/visual structures, add accessors instead. * gdk/gdkpixbuf-render.c gdk/gdkpixmap.c gdk/gdkrgb.c gdk/x11/gdkcolormap-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkprivate-x11.h gtk/gtkgc.c gtk/gtkstyle.c gtk/gtkwidget.c: Use accessors to get the screen for colormaps, visuals; move the fields into the private structures for the x11 backend. * gdk/gdkdisplay.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/gdkscreen.[ch] gdk/x11/gdkscreen-x11.c: Remove virtualization of screen and display functions. (#79990, patch from Erwann Chenede) * gdk/win32/gdkdisplay-x11.c gdk/win32/gdkscreen-win32.c gdk/win32/{Makefile.am, makefile.msc, makefile.mingw}: New files containing stub implementations of Display, Screen functions. * gdk/x11/gdkscreen-x11.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/x11/gdkx.h: Clean up function exports and what headers they are in. (#79954) * gdk/x11/gdkx.h: Fix macro that was referring to a non-existant screen->screen_num. (In the patch for #79972, Erwann Chenede) * gdk/gdkscreen.c gdk/gdkwindow.c gdk/x11/gdkinternals.h gdk/x11/gdkscreen-x11.c: Fix gdk_screen_get_window_at_pointer() to use window hooks. (#79972, patch partly from Erwann Chenede) * gdk/x11/gdkdisplay-x11.c gdk/x11/gdkevents-x11.c: Fix some warnings.
2002-06-06 00:26:42 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
owner = _gdk_win32_display_get_selection_owner (gdk_display_get_default (),
GDK_SELECTION_CLIPBOARD);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (owner == NULL)
return;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->clipboard_opened_for == INVALID_HANDLE_VALUE)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (OpenClipboard (GDK_WINDOW_HWND (owner)))
{
win32_sel->clipboard_opened_for = GDK_WINDOW_HWND (owner);
GDK_NOTE (DND, g_print ("Opened clipboard for 0x%p @ %s:%d\n", win32_sel->clipboard_opened_for, __FILE__, __LINE__));
}
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GDK_NOTE (DND, g_print ("... sending GDK_SELECTION_REQUEST to ourselves\n"));
memset (&tmp_event, 0, sizeof (tmp_event));
tmp_event.selection.type = GDK_SELECTION_REQUEST;
tmp_event.selection.window = owner;
tmp_event.selection.send_event = FALSE;
tmp_event.selection.selection = GDK_SELECTION_CLIPBOARD;
tmp_event.selection.target = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TARGETS);
tmp_event.selection.property = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GDK_SELECTION);
tmp_event.selection.requestor = owner;
tmp_event.selection.time = time;
win32_sel->property_change_target_atom = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TARGETS);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
gdk_event_put (&tmp_event);
win32_sel->targets_request_pending = TRUE;
}
#define CLIPBOARD_IDLE_ABORT_TIME 30
static const gchar *
predefined_name (UINT fmt)
{
#define CASE(fmt) case fmt: return #fmt
switch (fmt)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
CASE (CF_TEXT);
CASE (CF_BITMAP);
CASE (CF_METAFILEPICT);
CASE (CF_SYLK);
CASE (CF_DIF);
CASE (CF_TIFF);
CASE (CF_OEMTEXT);
CASE (CF_DIB);
CASE (CF_PALETTE);
CASE (CF_PENDATA);
CASE (CF_RIFF);
CASE (CF_WAVE);
CASE (CF_UNICODETEXT);
CASE (CF_ENHMETAFILE);
CASE (CF_HDROP);
CASE (CF_LOCALE);
CASE (CF_DIBV5);
CASE (CF_MAX);
CASE (CF_OWNERDISPLAY);
CASE (CF_DSPTEXT);
CASE (CF_DSPBITMAP);
CASE (CF_DSPMETAFILEPICT);
CASE (CF_DSPENHMETAFILE);
#undef CASE
default:
return NULL;
}
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
gchar *
_gdk_win32_get_clipboard_format_name (UINT fmt,
gboolean *is_predefined)
{
gint registered_name_w_len = 1024;
wchar_t *registered_name_w = g_malloc (registered_name_w_len);
gchar *registered_name = NULL;
int gcfn_result;
const gchar *predef = predefined_name (fmt);
/* FIXME: cache the result in a hash table */
do
{
gcfn_result = GetClipboardFormatNameW (fmt, registered_name_w, registered_name_w_len);
if (gcfn_result > 0 && gcfn_result < registered_name_w_len)
{
registered_name = g_utf16_to_utf8 (registered_name_w, -1, NULL, NULL, NULL);
g_clear_pointer (&registered_name_w, g_free);
if (!registered_name)
gcfn_result = 0;
else
*is_predefined = FALSE;
break;
}
/* If GetClipboardFormatNameW() used up all the space, it means that
* we probably need a bigger buffer, but cap this at 1 kilobyte.
*/
if (gcfn_result == 0 || registered_name_w_len > 1024 * 1024)
{
gcfn_result = 0;
g_clear_pointer (&registered_name_w, g_free);
break;
}
registered_name_w_len *= 2;
registered_name_w = g_realloc (registered_name_w, registered_name_w_len);
gcfn_result = registered_name_w_len;
} while (gcfn_result == registered_name_w_len);
if (registered_name == NULL &&
predef != NULL)
{
registered_name = g_strdup (predef);
*is_predefined = TRUE;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
return registered_name;
}
static GArray *
get_compatibility_formats_for_target (GdkAtom target)
{
GArray *result = NULL;
gint i;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
result = g_hash_table_lookup (win32_sel->compatibility_formats, target);
if (result != NULL)
return result;
for (i = 0; i < win32_sel->n_known_pixbuf_formats; i++)
{
if (target != win32_sel->known_pixbuf_formats[i])
continue;
/* Any format known to gdk-pixbuf can be presented as PNG or BMP */
result = g_hash_table_lookup (win32_sel->compatibility_formats,
_gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_PNG));
break;
}
return result;
}
static GArray *
_gdk_win32_selection_get_compatibility_targets_for_format (UINT format)
{
GArray *result = NULL;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
result = g_hash_table_lookup (win32_sel->compatibility_targets, GINT_TO_POINTER (format));
if (result != NULL)
return result;
/* TODO: reverse gdk-pixbuf conversion? We have to somehow
* match gdk-pixbuf format names to the corresponding clipboard
* format names. The former are known only at runtime,
* the latter are presently unknown...
* Maybe try to get the data and then just feed it to gdk-pixbuf,
* see if it knows what it is?
*/
return result;
}
void
_gdk_win32_add_format_to_targets (UINT format,
GArray *array,
GList **list)
{
gboolean predef;
gchar *format_name = _gdk_win32_get_clipboard_format_name (format, &predef);
GdkAtom target_atom;
GdkSelTargetFormat target_selformat;
GArray *target_selformats;
gint i,j;
if (format_name != NULL)
{
target_atom = g_intern_string (format_name);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GDK_NOTE (DND, g_print ("Maybe add as-is format %s (0x%p)\n", format_name, target_atom));
g_free (format_name);
if (array && target_atom != 0)
{
for (j = 0; j < array->len; j++)
if (g_array_index (array, GdkSelTargetFormat, j).target == target_atom)
break;
if (j == array->len)
{
target_selformat.format = format;
target_selformat.target = target_atom;
target_selformat.transmute = FALSE;
g_array_append_val (array, target_selformat);
}
}
if (list && target_atom != 0 && g_list_find (*list, target_atom) == NULL)
*list = g_list_prepend (*list, target_atom);
}
target_selformats = _gdk_win32_selection_get_compatibility_targets_for_format (format);
if (array && target_selformats != NULL)
for (i = 0; i < target_selformats->len; i++)
{
target_selformat = g_array_index (target_selformats, GdkSelTargetFormat, i);
for (j = 0; j < array->len; j++)
if (g_array_index (array, GdkSelTargetFormat, j).target == target_selformat.target &&
g_array_index (array, GdkSelTargetFormat, j).format == target_selformat.format)
break;
if (j == array->len)
g_array_append_val (array, target_selformat);
}
if (list && target_selformats != NULL)
for (i = 0; i < target_selformats->len; i++)
{
target_selformat = g_array_index (target_selformats, GdkSelTargetFormat, i);
if (g_list_find (*list, target_selformat.target) == NULL)
*list = g_list_prepend (*list, target_selformat.target);
}
}
static void
transmute_cf_unicodetext_to_utf8_string (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
wchar_t *ptr, *p, *q;
gchar *result;
glong wclen, u8_len;
/* Strip out \r */
ptr = (wchar_t *) data;
p = ptr;
q = ptr;
wclen = 0;
while (p < ptr + length / 2)
{
if (*p != L'\r')
{
*q++ = *p;
wclen++;
}
p++;
}
result = g_utf16_to_utf8 (ptr, wclen, NULL, &u8_len, NULL);
if (result)
{
*set_data = (guchar *) result;
if (set_data_length)
*set_data_length = u8_len + 1;
if (set_data_destroy)
*set_data_destroy = (GDestroyNotify) g_free;
}
}
static void
transmute_utf8_string_to_cf_unicodetext (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
glong wclen;
GError *err = NULL;
glong size;
gint i;
wchar_t *wcptr, *p;
wcptr = g_utf8_to_utf16 ((char *) data, length, NULL, &wclen, &err);
if (err != NULL)
{
g_warning ("Failed to convert utf8: %s", err->message);
g_clear_error (&err);
return;
}
wclen++; /* Terminating 0 */
size = wclen * 2;
for (i = 0; i < wclen; i++)
if (wcptr[i] == L'\n' && (i == 0 || wcptr[i - 1] != L'\r'))
size += 2;
*set_data = g_malloc0 (size);
if (set_data_length)
*set_data_length = size;
if (set_data_destroy)
*set_data_destroy = (GDestroyNotify) g_free;
p = (wchar_t *) *set_data;
for (i = 0; i < wclen; i++)
{
if (wcptr[i] == L'\n' && (i == 0 || wcptr[i - 1] != L'\r'))
*p++ = L'\r';
*p++ = wcptr[i];
}
g_free (wcptr);
}
static int
wchar_to_str (const wchar_t *wstr,
char **retstr,
UINT cp)
{
char *str;
int len;
int lenc;
len = WideCharToMultiByte (cp, 0, wstr, -1, NULL, 0, NULL, NULL);
if (len <= 0)
return -1;
str = g_malloc (sizeof (char) * len);
lenc = WideCharToMultiByte (cp, 0, wstr, -1, str, len, NULL, NULL);
if (lenc != len)
{
g_free (str);
return -3;
}
*retstr = str;
return 0;
}
static void
transmute_utf8_string_to_cf_text (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
glong rlen;
GError *err = NULL;
glong size;
gint i;
char *strptr, *p;
wchar_t *wcptr;
wcptr = g_utf8_to_utf16 ((char *) data, length, NULL, NULL, &err);
if (err != NULL)
{
g_warning ("Failed to convert utf8: %s", err->message);
g_clear_error (&err);
return;
}
if (wchar_to_str (wcptr, &strptr, CP_ACP) != 0)
{
g_warning ("Failed to convert utf-16 %S to ACP", wcptr);
g_free (wcptr);
return;
}
rlen = strlen (strptr);
g_free (wcptr);
rlen++; /* Terminating 0 */
size = rlen * sizeof (char);
for (i = 0; i < rlen; i++)
if (strptr[i] == '\n' && (i == 0 || strptr[i - 1] != '\r'))
size += sizeof (char);
*set_data = g_malloc0 (size);
if (set_data_length)
*set_data_length = size;
if (set_data_destroy)
*set_data_destroy = (GDestroyNotify) g_free;
p = (char *) *set_data;
for (i = 0; i < rlen; i++)
{
if (strptr[i] == '\n' && (i == 0 || strptr[i - 1] != '\r'))
*p++ = '\r';
*p++ = strptr[i];
}
g_free (strptr);
}
static int
str_to_wchar (const char *str,
wchar_t **wretstr,
UINT cp)
{
wchar_t *wstr;
int len;
int lenc;
len = MultiByteToWideChar (cp, 0, str, -1, NULL, 0);
if (len <= 0)
return -1;
wstr = g_malloc (sizeof (wchar_t) * len);
lenc = MultiByteToWideChar (cp, 0, str, -1, wstr, len);
if (lenc != len)
{
g_free (wstr);
return -3;
}
*wretstr = wstr;
return 0;
}
static void
transmute_cf_text_to_utf8_string (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
char *ptr, *p, *q;
gchar *result;
glong wclen, u8_len;
wchar_t *wstr;
/* Strip out \r */
ptr = (gchar *) data;
p = ptr;
q = ptr;
wclen = 0;
while (p < ptr + length / 2)
{
if (*p != '\r')
{
*q++ = *p;
wclen++;
}
p++;
}
if (str_to_wchar (ptr, &wstr, CP_ACP) < 0)
return;
result = g_utf16_to_utf8 (wstr, -1, NULL, &u8_len, NULL);
if (result)
{
*set_data = (guchar *) result;
if (set_data_length)
*set_data_length = u8_len + 1;
if (set_data_destroy)
*set_data_destroy = (GDestroyNotify) g_free;
}
g_free (wstr);
}
static void
transmute_cf_dib_to_image_bmp (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
/* Need to add a BMP file header so gdk-pixbuf can load
* it.
*
* If the data is from Mozilla Firefox or IE7, and
* starts with an "old fashioned" BITMAPINFOHEADER,
* i.e. with biSize==40, and biCompression == BI_RGB and
* biBitCount==32, we assume that the "extra" byte in
* each pixel in fact is alpha.
*
* The gdk-pixbuf bmp loader doesn't trust 32-bit BI_RGB
* bitmaps to in fact have alpha, so we have to convince
* it by changing the bitmap header to a version 5
* BI_BITFIELDS one with explicit alpha mask indicated.
*
* The RGB bytes that are in bitmaps on the clipboard
* originating from Firefox or IE7 seem to be
* premultiplied with alpha. The gdk-pixbuf bmp loader
* of course doesn't expect that, so we have to undo the
* premultiplication before feeding the bitmap to the
* bmp loader.
*
* Note that for some reason the bmp loader used to want
* the alpha bytes in its input to actually be
* 255-alpha, but here we assume that this has been
* fixed before this is committed.
*/
BITMAPINFOHEADER *bi = (BITMAPINFOHEADER *) data;
BITMAPFILEHEADER *bf;
gpointer result;
gint data_length = length;
gint new_length;
gboolean make_dibv5 = FALSE;
BITMAPV5HEADER *bV5;
guchar *p;
guint i;
if (bi->biSize == sizeof (BITMAPINFOHEADER) &&
bi->biPlanes == 1 &&
bi->biBitCount == 32 &&
bi->biCompression == BI_RGB &&
#if 0
/* Maybe check explicitly for Mozilla or IE7?
*
* If the clipboard format
* application/x-moz-nativeimage is present, that is
* a reliable indicator that the data is offered by
* Mozilla one would think. For IE7,
* UniformResourceLocatorW is presumably not that
* uniqie, so probably need to do some
* GetClipboardOwner(), GetWindowThreadProcessId(),
* OpenProcess(), GetModuleFileNameEx() dance to
* check?
*/
(IsClipboardFormatAvailable
(RegisterClipboardFormatA ("application/x-moz-nativeimage")) ||
IsClipboardFormatAvailable
(RegisterClipboardFormatA ("UniformResourceLocatorW"))) &&
#endif
TRUE)
{
/* We turn the BITMAPINFOHEADER into a
* BITMAPV5HEADER before feeding it to gdk-pixbuf.
*/
new_length = (data_length +
sizeof (BITMAPFILEHEADER) +
(sizeof (BITMAPV5HEADER) - sizeof (BITMAPINFOHEADER)));
make_dibv5 = TRUE;
}
else
{
new_length = data_length + sizeof (BITMAPFILEHEADER);
}
result = g_try_malloc (new_length);
if (result == NULL)
return;
bf = (BITMAPFILEHEADER *) result;
bf->bfType = 0x4d42; /* "BM" */
bf->bfSize = new_length;
bf->bfReserved1 = 0;
bf->bfReserved2 = 0;
*set_data = result;
if (set_data_length)
*set_data_length = new_length;
if (set_data_destroy)
*set_data_destroy = g_free;
if (!make_dibv5)
{
bf->bfOffBits = (sizeof (BITMAPFILEHEADER) +
bi->biSize +
bi->biClrUsed * sizeof (RGBQUAD));
if (bi->biCompression == BI_BITFIELDS && bi->biBitCount >= 16)
{
/* Screenshots taken with PrintScreen or
* Alt + PrintScreen are found on the clipboard in
* this format. In this case the BITMAPINFOHEADER is
* followed by three DWORD specifying the masks of the
* red green and blue components, so adjust the offset
* accordingly. */
bf->bfOffBits += (3 * sizeof (DWORD));
}
memcpy ((char *) result + sizeof (BITMAPFILEHEADER),
bi,
data_length);
return;
}
bV5 = (BITMAPV5HEADER *) ((char *) result + sizeof (BITMAPFILEHEADER));
bV5->bV5Size = sizeof (BITMAPV5HEADER);
bV5->bV5Width = bi->biWidth;
bV5->bV5Height = bi->biHeight;
bV5->bV5Planes = 1;
bV5->bV5BitCount = 32;
bV5->bV5Compression = BI_BITFIELDS;
bV5->bV5SizeImage = 4 * bV5->bV5Width * ABS (bV5->bV5Height);
bV5->bV5XPelsPerMeter = bi->biXPelsPerMeter;
bV5->bV5YPelsPerMeter = bi->biYPelsPerMeter;
bV5->bV5ClrUsed = 0;
bV5->bV5ClrImportant = 0;
/* Now the added mask fields */
bV5->bV5RedMask = 0x00ff0000;
bV5->bV5GreenMask = 0x0000ff00;
bV5->bV5BlueMask = 0x000000ff;
bV5->bV5AlphaMask = 0xff000000;
((char *) &bV5->bV5CSType)[3] = 's';
((char *) &bV5->bV5CSType)[2] = 'R';
((char *) &bV5->bV5CSType)[1] = 'G';
((char *) &bV5->bV5CSType)[0] = 'B';
/* Ignore colorspace and profile fields */
bV5->bV5Intent = LCS_GM_GRAPHICS;
bV5->bV5Reserved = 0;
bf->bfOffBits = (sizeof (BITMAPFILEHEADER) +
bV5->bV5Size);
p = ((guchar *) result) + sizeof (BITMAPFILEHEADER) + sizeof (BITMAPV5HEADER);
memcpy (p, ((char *) bi) + bi->biSize,
data_length - sizeof (BITMAPINFOHEADER));
for (i = 0; i < bV5->bV5SizeImage/4; i++)
{
if (p[3] != 0)
{
gdouble inverse_alpha = 255./p[3];
p[0] = p[0] * inverse_alpha + 0.5;
p[1] = p[1] * inverse_alpha + 0.5;
p[2] = p[2] * inverse_alpha + 0.5;
}
p += 4;
}
}
static void
transmute_cf_shell_id_list_to_text_uri_list (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
guint i;
CIDA *cida = (CIDA *) data;
guint number_of_ids = cida->cidl;
GString *result = g_string_new (NULL);
PCIDLIST_ABSOLUTE folder_id = HIDA_GetPIDLFolder (cida);
wchar_t path_w[MAX_PATH + 1];
for (i = 0; i < number_of_ids; i++)
{
PCUIDLIST_RELATIVE file_id = HIDA_GetPIDLItem (cida, i);
PIDLIST_ABSOLUTE file_id_full = ILCombine (folder_id, file_id);
if (SHGetPathFromIDListW (file_id_full, path_w))
{
gchar *filename = g_utf16_to_utf8 (path_w, -1, NULL, NULL, NULL);
if (filename)
{
gchar *uri = g_filename_to_uri (filename, NULL, NULL);
if (uri)
{
g_string_append (result, uri);
g_string_append (result, "\r\n");
g_free (uri);
}
g_free (filename);
}
}
ILFree (file_id_full);
}
*set_data = (guchar *) result->str;
if (set_data_length)
*set_data_length = result->len;
if (set_data_destroy)
*set_data_destroy = g_free;
g_string_free (result, FALSE);
}
void
transmute_image_bmp_to_cf_dib (const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length,
GDestroyNotify *set_data_destroy)
{
gint size;
guchar *ptr;
g_return_if_fail (length >= sizeof (BITMAPFILEHEADER));
/* No conversion is needed, just strip the BITMAPFILEHEADER */
size = length - sizeof (BITMAPFILEHEADER);
ptr = g_malloc (size);
memcpy (ptr, data + sizeof (BITMAPFILEHEADER), size);
*set_data = ptr;
if (set_data_length)
*set_data_length = size;
if (set_data_destroy)
*set_data_destroy = g_free;
}
static void
transmute_selection_format (UINT from_format,
GdkAtom to_target,
const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length)
{
if ((to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_PNG) &&
from_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_PNG)) ||
(to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_JPEG) &&
from_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_JFIF)) ||
(to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GIF) &&
from_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_GIF)))
{
/* No transmutation needed */
*set_data = g_memdup (data, length);
*set_data_length = length;
}
else if (to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_UTF8_STRING) &&
from_format == CF_UNICODETEXT)
{
transmute_cf_unicodetext_to_utf8_string (data, length, set_data, set_data_length, NULL);
}
else if (to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_UTF8_STRING) &&
from_format == CF_TEXT)
{
transmute_cf_text_to_utf8_string (data, length, set_data, set_data_length, NULL);
}
else if (to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_BMP) &&
(from_format == CF_DIB || from_format == CF_DIBV5))
{
transmute_cf_dib_to_image_bmp (data, length, set_data, set_data_length, NULL);
}
else if (to_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TEXT_URI_LIST) &&
from_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_CFSTR_SHELLIDLIST))
{
transmute_cf_shell_id_list_to_text_uri_list (data, length, set_data, set_data_length, NULL);
}
else
{
g_warning ("Don't know how to transmute format 0x%x to target 0x%p", from_format, to_target);
}
}
void
transmute_selection_target (GdkAtom from_target,
UINT to_format,
const guchar *data,
gint length,
guchar **set_data,
gint *set_data_length)
{
if ((from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_PNG) &&
to_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_PNG)) ||
(from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_JPEG) &&
to_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_JFIF)) ||
(from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GIF) &&
to_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_GIF)))
{
/* No conversion needed */
*set_data = g_memdup (data, length);
*set_data_length = length;
}
else if (from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_UTF8_STRING) &&
to_format == CF_UNICODETEXT)
{
transmute_utf8_string_to_cf_unicodetext (data, length, set_data, set_data_length, NULL);
}
else if (from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_UTF8_STRING) &&
to_format == CF_TEXT)
{
transmute_utf8_string_to_cf_text (data, length, set_data, set_data_length, NULL);
}
else if (from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_BMP) &&
to_format == CF_DIB)
{
transmute_image_bmp_to_cf_dib (data, length, set_data, set_data_length, NULL);
}
else if (from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_IMAGE_BMP) &&
to_format == CF_DIBV5)
{
transmute_image_bmp_to_cf_dib (data, length, set_data, set_data_length, NULL);
}
/*
else if (from_target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TEXT_URI_LIST) &&
to_format == _gdk_win32_selection_cf (GDK_WIN32_CF_INDEX_CFSTR_SHELLIDLIST))
{
transmute_text_uri_list_to_shell_id_list (data, length, set_data, set_data_length, NULL);
}
*/
else
{
g_warning ("Don't know how to transmute from target 0x%p to format 0x%x", from_target, to_format);
}
}
static GdkAtom
convert_clipboard_selection_to_targets_target (GdkWindow *requestor)
{
gint fmt;
int i;
int format_count = CountClipboardFormats ();
GArray *targets = g_array_sized_new (FALSE, FALSE, sizeof (GdkSelTargetFormat), format_count);
for (fmt = 0; 0 != (fmt = EnumClipboardFormats (fmt)); )
_gdk_win32_add_format_to_targets (fmt, targets, NULL);
GDK_NOTE (DND, {
g_print ("... ");
for (i = 0; i < targets->len; i++)
{
const char *atom_name = (const char *)g_array_index (targets, GdkSelTargetFormat, i).target;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_print ("%s", atom_name);
if (i < targets->len - 1)
g_print (", ");
}
g_print ("\n");
});
if (targets->len > 0)
{
gint len = targets->len;
GdkAtom *targets_only = g_new0 (GdkAtom, len);
for (i = 0; i < targets->len; i++)
targets_only[i] = g_array_index (targets, GdkSelTargetFormat, i).target;
g_array_free (targets, TRUE);
selection_property_store (requestor, GDK_SELECTION_TYPE_ATOM,
32, (guchar *) targets_only,
len * sizeof (GdkAtom));
return _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GDK_SELECTION);
}
else
{
g_array_free (targets, TRUE);
return NULL;
}
}
static GdkAtom
convert_clipboard_selection_to_target (GdkWindow *requestor,
GdkAtom target)
{
UINT format;
HANDLE hdata;
guchar *ptr;
gint length;
gboolean transmute = FALSE;
GdkAtom result = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GDK_SELECTION);
gboolean found;
const char *atom_name;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
atom_name = (const char *)target;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
for (format = 0, found = FALSE;
!found && 0 != (format = EnumClipboardFormats (format));
)
{
gboolean predef;
gchar *format_name = _gdk_win32_get_clipboard_format_name (format, &predef);
if (format_name == NULL)
continue;
found = g_strcmp0 (format_name, atom_name) == 0;
g_free (format_name);
}
if (format == 0)
{
gint i;
GArray *compat_formats = get_compatibility_formats_for_target (target);
for (i = 0; compat_formats != NULL && i < compat_formats->len; i++)
{
if (!IsClipboardFormatAvailable (g_array_index (compat_formats, GdkSelTargetFormat, i).format))
continue;
format = g_array_index (compat_formats, GdkSelTargetFormat, i).format;
transmute = g_array_index (compat_formats, GdkSelTargetFormat, i).transmute;
break;
}
}
if (format == 0)
return NULL;
if ((hdata = GetClipboardData (format)) == NULL)
return NULL;
if ((ptr = GlobalLock (hdata)) != NULL)
{
guchar *data = NULL;
gint data_len = 0;
length = GlobalSize (hdata);
GDK_NOTE (DND, g_print ("... format 0x%x: %d bytes\n", format, length));
if (transmute)
{
transmute_selection_format (format, target, ptr, length, &data, &data_len);
}
else
{
data = g_memdup (ptr, length);
data_len = length;
}
if (data)
selection_property_store (requestor, target,
8, data, data_len);
else
result = NULL;
GlobalUnlock (hdata);
}
return result;
}
static GdkAtom
convert_selection_with_opened_clipboard (GdkDisplay *display,
GdkWindow *requestor,
GdkAtom target,
guint32 time)
{
GdkAtom property;
if (target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TARGETS))
property = convert_clipboard_selection_to_targets_target (requestor);
else
property = convert_clipboard_selection_to_target (requestor, target);
return property;
}
static void
announce_delayrendered_targets_with_opened_clipboard (GdkWin32Selection *win32_sel)
{
gint i;
/* Announce the formats we support, but don't actually put any data out there.
* Other processes will send us WM_RENDERFORMAT to get the data.
*/
for (i = 0; i < win32_sel->clipboard_selection_targets->len; i++)
{
GdkSelTargetFormat *fmt = &g_array_index (win32_sel->clipboard_selection_targets, GdkSelTargetFormat, i);
/* Some calls here may be duplicates, but we don't really care */
if (fmt->format != 0)
SetClipboardData (fmt->format, NULL);
}
}
static gboolean
open_clipboard_timeout (gpointer data)
{
GList *tmp_list, *next;
GdkWin32ClipboardQueueInfo *info;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
GDK_NOTE (DND, g_print ("Open clipboard timeout ticks\n"));
/* Clear out old and invalid entries */
for (tmp_list = clipboard_queue; tmp_list; tmp_list = next)
{
info = (GdkWin32ClipboardQueueInfo *) tmp_list->data;
next = g_list_next (tmp_list);
if (GDK_WINDOW_DESTROYED (info->requestor) ||
info->idle_time >= CLIPBOARD_IDLE_ABORT_TIME)
{
clipboard_queue = g_list_remove_link (clipboard_queue, tmp_list);
g_list_free (tmp_list);
switch (info->action)
{
case GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS:
break;
case GDK_WIN32_CLIPBOARD_QUEUE_ACTION_CONVERT:
generate_selection_notify (info->requestor, info->selection, info->target, NULL, info->time);
break;
}
g_clear_object (&info->requestor);
g_slice_free (GdkWin32ClipboardQueueInfo, info);
}
}
if (clipboard_queue == NULL)
{
GDK_NOTE (DND, g_print ("Stopping open clipboard timer\n"));
if (win32_sel->clipboard_opened_for != INVALID_HANDLE_VALUE)
{
API_CALL (CloseClipboard, ());
win32_sel->clipboard_opened_for = INVALID_HANDLE_VALUE;
GDK_NOTE (DND, g_print ("Closed clipboard @ %s:%d\n", __FILE__, __LINE__));
}
return FALSE;
}
for (tmp_list = clipboard_queue; tmp_list; tmp_list = next)
{
GdkAtom property;
info = (GdkWin32ClipboardQueueInfo *) tmp_list->data;
next = g_list_next (tmp_list);
/* CONVERT works with any opened clipboard,
* but TARGETS needs to open the clipboard with the hande of the
* owner window.
*/
if (info->action == GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS &&
win32_sel->clipboard_opened_for == NULL)
{
GDK_NOTE (DND, g_print ("Need to re-open clipboard, closing\n"));
API_CALL (CloseClipboard, ());
win32_sel->clipboard_opened_for = INVALID_HANDLE_VALUE;
}
if (win32_sel->clipboard_opened_for == INVALID_HANDLE_VALUE)
{
if (!OpenClipboard (GDK_WINDOW_HWND (info->requestor)))
{
info->idle_time += 1;
continue;
}
win32_sel->clipboard_opened_for = GDK_WINDOW_HWND (info->requestor);
GDK_NOTE (DND, g_print ("Opened clipboard for 0x%p @ %s:%d\n", win32_sel->clipboard_opened_for, __FILE__, __LINE__));
}
clipboard_queue = g_list_remove_link (clipboard_queue, tmp_list);
g_list_free (tmp_list);
switch (info->action)
{
case GDK_WIN32_CLIPBOARD_QUEUE_ACTION_CONVERT:
property = convert_selection_with_opened_clipboard (info->display,
info->requestor,
info->target,
info->time);
generate_selection_notify (info->requestor,
GDK_SELECTION_CLIPBOARD,
info->target,
property,
info->time);
break;
case GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS:
announce_delayrendered_targets_with_opened_clipboard (win32_sel);
break;
default:
g_assert_not_reached ();
}
g_clear_object (&info->requestor);
g_slice_free (GdkWin32ClipboardQueueInfo, info);
}
if (clipboard_queue != NULL)
return TRUE;
if (win32_sel->clipboard_opened_for != INVALID_HANDLE_VALUE)
{
API_CALL (CloseClipboard, ());
win32_sel->clipboard_opened_for = INVALID_HANDLE_VALUE;
GDK_NOTE (DND, g_print ("Closed clipboard @ %s:%d\n", __FILE__, __LINE__));
}
GDK_NOTE (DND, g_print ("Stopping open clipboard timer\n"));
return FALSE;
}
static void
queue_open_clipboard (GdkWin32ClipboardQueueAction action,
GdkDisplay *display,
GdkWindow *requestor,
GdkAtom target,
guint32 time)
{
guint id;
GList *tmp_list, *next;
GdkWin32ClipboardQueueInfo *info;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
for (tmp_list = clipboard_queue; tmp_list; tmp_list = next)
{
info = (GdkWin32ClipboardQueueInfo *) tmp_list->data;
next = g_list_next (tmp_list);
if (info->action == action &&
info->requestor == requestor)
return;
}
info = g_slice_new (GdkWin32ClipboardQueueInfo);
info->display = display;
g_set_object (&info->requestor, requestor);
info->selection = GDK_SELECTION_CLIPBOARD;
info->target = target;
info->idle_time = 0;
info->time = time;
info->action = action;
GDK_NOTE (DND, g_print ("Queueing open clipboard\n"));
if (win32_sel->clipboard_opened_for == INVALID_HANDLE_VALUE &&
clipboard_queue == NULL)
{
id = gdk_threads_add_timeout_seconds (1, (GSourceFunc) open_clipboard_timeout, NULL);
g_source_set_name_by_id (id, "[gdk-win32] open_clipboard_timeout");
GDK_NOTE (DND, g_print ("Started open clipboard timer\n"));
}
clipboard_queue = g_list_append (clipboard_queue, info);
}
gboolean
_gdk_win32_display_set_selection_owner (GdkDisplay *display,
GdkWindow *owner,
GdkAtom selection,
guint32 time,
gboolean send_event)
{
HWND hwnd;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
g_return_val_if_fail (selection != NULL, FALSE);
GDK_NOTE (DND, {
const char *sel_name = (const char *)selection;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_print ("gdk_selection_owner_set_for_display: %p %s\n",
(owner ? GDK_WINDOW_HWND (owner) : NULL),
sel_name);
});
if (selection != GDK_SELECTION_CLIPBOARD)
{
if (owner != NULL)
g_hash_table_insert (win32_sel->sel_owner_table, selection, GDK_WINDOW_HWND (owner));
else
g_hash_table_remove (win32_sel->sel_owner_table, selection);
return TRUE;
}
/* Rest of this function handles the CLIPBOARD selection */
if (owner != NULL)
{
if (GDK_WINDOW_DESTROYED (owner))
return FALSE;
hwnd = GDK_WINDOW_HWND (owner);
}
else
hwnd = NULL;
if (win32_sel->clipboard_opened_for != hwnd &&
win32_sel->clipboard_opened_for != INVALID_HANDLE_VALUE)
{
API_CALL (CloseClipboard, ());
win32_sel->clipboard_opened_for = INVALID_HANDLE_VALUE;
GDK_NOTE (DND, g_print ("Closed clipboard @ %s:%d\n", __FILE__, __LINE__));
}
if (!OpenClipboard (hwnd))
{
if (GetLastError () != ERROR_ACCESS_DENIED)
WIN32_API_FAILED ("OpenClipboard");
return FALSE;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->clipboard_opened_for = hwnd;
GDK_NOTE (DND, g_print ("Opened clipboard for 0x%p @ %s:%d\n", win32_sel->clipboard_opened_for, __FILE__, __LINE__));
win32_sel->ignore_destroy_clipboard = TRUE;
GDK_NOTE (DND, g_print ("... EmptyClipboard()\n"));
if (!API_CALL (EmptyClipboard, ()))
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->ignore_destroy_clipboard = FALSE;
API_CALL (CloseClipboard, ());
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->clipboard_opened_for = INVALID_HANDLE_VALUE;
GDK_NOTE (DND, g_print ("Closed clipboard @ %s:%d\n", __FILE__, __LINE__));
return FALSE;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->ignore_destroy_clipboard = FALSE;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Any queued clipboard operations were just made pointless
* by EmptyClipboard().
*/
_gdk_win32_clear_clipboard_queue ();
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* This is kind of risky, but we don't close the clipboard
* to ensure that it's still open when GDK_SELECTION_REQUEST
* is handled.
*/
if (owner == NULL)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (!API_CALL (CloseClipboard, ()))
return FALSE;
GDK_NOTE (DND, g_print ("Closed clipboard @ %s:%d\n", __FILE__, __LINE__));
win32_sel->clipboard_opened_for = INVALID_HANDLE_VALUE;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
send_targets_request (time);
return TRUE;
}
GdkWindow*
_gdk_win32_display_get_selection_owner (GdkDisplay *display,
GdkAtom selection)
{
GdkWindow *window;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
HWND selection_owner;
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
2017-11-15 17:13:31 +00:00
g_return_val_if_fail (selection != NULL, NULL);
Changes multihead reorganizing code for win32 support, mostly from a patch Wed Jun 5 18:34:47 2002 Owen Taylor <otaylor@redhat.com> Changes multihead reorganizing code for win32 support, mostly from a patch by Hans Breuer. * gdk/gdkcolor.c gdk/x11/gdkcolor-x11.c gdk/gdkcursor.c gdk/x11/gdkcursor-x11.c gdk/gdkevents.c gdk/x11/gdkevents-x11.c gdk/gdkfont.c gdk/x11/gdkfont-x11.c gdk/gdkkeys.c gdk/x11/gdkkeys-x11.c gdk/gdkimage.c gdk/x11/gdkimage-x11.c gdk/gdkscreen.c gdk/x11/gdkmain-x11.c gdk/gdkdisplay.c gdk/gdkevents-x11.c gdk/gdkpango.c gdk/x11/gdkpango-x11.c gdk/gdkselection.c gdk/x11/gdkselection-x11.c gdk/gdkwindow.c gdk/x11/gdkwindow-x11.c gdk/gdkvisual.c gdk/x11/gdkvisual-x11.c: Move port-independent singlehead wrapper functions into port-independent part of GDK. (#80009) * gdk/win32/gdkcolor-win32.c gdk/win32/gdkcursor-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkfont-win32.c gdk/win32/gdkimage-win32.c gdk/win32/gdkkeys-win32.c gdk/win32/gdkmain-win32.c gdk/win32/gdkproperty-win32.c gdk/win32/gdkselection-win32.c gdk/win32/gkwindow-win32.c: Turn singlehead functions into "multihead" functions that ignore their GdkDisplay or GdkScreen arguments. * gdk/win32/gdkdrawable-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkinput-win32.c gdk/win32/gdkprivate-win32.h: Misc multihead-compatibility changes. * gtk/gtk.def gdk/gdk.def: Update for multihead functions. * gdk/gdkcolormap.h gdk/gdkvisual.h gdk/x11/gdkcolormap-x11.c gdk/x11/gdkvisual-x11.c: Remove the screen fields from the public parts of the colormap/visual structures, add accessors instead. * gdk/gdkpixbuf-render.c gdk/gdkpixmap.c gdk/gdkrgb.c gdk/x11/gdkcolormap-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkprivate-x11.h gtk/gtkgc.c gtk/gtkstyle.c gtk/gtkwidget.c: Use accessors to get the screen for colormaps, visuals; move the fields into the private structures for the x11 backend. * gdk/gdkdisplay.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/gdkscreen.[ch] gdk/x11/gdkscreen-x11.c: Remove virtualization of screen and display functions. (#79990, patch from Erwann Chenede) * gdk/win32/gdkdisplay-x11.c gdk/win32/gdkscreen-win32.c gdk/win32/{Makefile.am, makefile.msc, makefile.mingw}: New files containing stub implementations of Display, Screen functions. * gdk/x11/gdkscreen-x11.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/x11/gdkx.h: Clean up function exports and what headers they are in. (#79954) * gdk/x11/gdkx.h: Fix macro that was referring to a non-existant screen->screen_num. (In the patch for #79972, Erwann Chenede) * gdk/gdkscreen.c gdk/gdkwindow.c gdk/x11/gdkinternals.h gdk/x11/gdkscreen-x11.c: Fix gdk_screen_get_window_at_pointer() to use window hooks. (#79972, patch partly from Erwann Chenede) * gdk/x11/gdkdisplay-x11.c gdk/x11/gdkevents-x11.c: Fix some warnings.
2002-06-06 00:26:42 +00:00
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
if (selection == GDK_SELECTION_CLIPBOARD)
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
selection_owner = GetClipboardOwner ();
else
selection_owner = g_hash_table_lookup (win32_sel->sel_owner_table, selection);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (selection_owner)
window = gdk_win32_window_lookup_for_display (display,
selection_owner);
else
window = NULL;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, {
const char *sel_name = (const char *)selection;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_print ("gdk_selection_owner_get: %s = %p\n",
sel_name,
(window ? GDK_WINDOW_HWND (window) : NULL));
});
return window;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
static GdkAtom
convert_dnd_selection_to_target (GdkAtom target,
GdkWindow *requestor)
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
UINT format;
gint i, with_transmute;
guchar *ptr;
gint length;
gboolean transmute = FALSE;
GdkWin32DragContext *context_win32;
FORMATETC fmt;
STGMEDIUM storage;
HRESULT hr;
GdkAtom result = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_OLE2_DND);
g_assert (win32_sel->target_drag_context != NULL);
g_assert (win32_sel->dnd_data_object_target != NULL);
context_win32 = GDK_WIN32_DRAG_CONTEXT (win32_sel->target_drag_context);
fmt.ptd = NULL;
fmt.dwAspect = DVASPECT_CONTENT;
fmt.lindex = -1;
fmt.tymed = TYMED_HGLOBAL;
/* We rely on GTK+ applications to synthesize the DELETE request
* for themselves, since they do know whether a DnD operation was a
* move and whether was successful. Therefore, we do not need to
* actually send anything here. Just report back without storing
* any data.
*/
if (target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_DELETE))
return result;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
for (format = 0, with_transmute = 0; format == 0 && with_transmute < 2; with_transmute++)
{
for (i = 0;
i < context_win32->droptarget_format_target_map->len;
i++)
{
GdkSelTargetFormat selformat = g_array_index (context_win32->droptarget_format_target_map, GdkSelTargetFormat, i);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (selformat.target != target ||
selformat.transmute != (with_transmute == 0 ? FALSE : TRUE))
continue;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
fmt.cfFormat = selformat.format;
hr = IDataObject_QueryGetData (win32_sel->dnd_data_object_target, &fmt);
if (!SUCCEEDED (hr) || hr != S_OK)
continue;
format = selformat.format;
transmute = selformat.transmute;
break;
}
}
if (format == 0)
return NULL;
hr = IDataObject_GetData (win32_sel->dnd_data_object_target, &fmt, &storage);
if (!SUCCEEDED (hr) || hr != S_OK)
return NULL;
if ((ptr = GlobalLock (storage.hGlobal)) != NULL)
{
guchar *data = NULL;
gint data_len = 0;
SetLastError (0);
length = GlobalSize (storage.hGlobal);
if (GetLastError () == NO_ERROR)
{
if (transmute)
{
transmute_selection_format (format, target, ptr, length, &data, &data_len);
}
else
{
data = g_memdup (ptr, length);
data_len = length;
}
if (data)
selection_property_store (requestor, target, 8,
data, data_len);
else
result = NULL;
}
else
result = NULL;
GlobalUnlock (storage.hGlobal);
}
else
result = NULL;
ReleaseStgMedium (&storage);
return result;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
}
void
_gdk_win32_display_convert_selection (GdkDisplay *display,
GdkWindow *requestor,
GdkAtom selection,
GdkAtom target,
guint32 time)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
GdkAtom property = _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GDK_SELECTION);
2017-11-15 17:13:31 +00:00
g_return_if_fail (selection != NULL);
g_return_if_fail (requestor != NULL);
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
if (GDK_WINDOW_DESTROYED (requestor))
return;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, {
const char *sel_name = (const char *)selection;
const char *tgt_name = (const char *)target;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_print ("gdk_selection_convert: %p %s %s\n",
GDK_WINDOW_HWND (requestor),
sel_name, tgt_name);
});
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (selection == GDK_SELECTION_CLIPBOARD)
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->clipboard_opened_for != INVALID_HANDLE_VALUE ||
OpenClipboard (GDK_WINDOW_HWND (requestor)))
{
if (win32_sel->clipboard_opened_for == INVALID_HANDLE_VALUE)
{
win32_sel->clipboard_opened_for = GDK_WINDOW_HWND (requestor);
GDK_NOTE (DND, g_print ("Opened clipboard for 0x%p @ %s:%d\n", win32_sel->clipboard_opened_for, __FILE__, __LINE__));
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
queue_open_clipboard (GDK_WIN32_CLIPBOARD_QUEUE_ACTION_CONVERT, display, requestor, target, time);
open_clipboard_timeout (NULL);
}
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
else
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
{
queue_open_clipboard (GDK_WIN32_CLIPBOARD_QUEUE_ACTION_CONVERT, display, requestor, target, time);
/* Do not generate a selection notify message */
return;
}
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else if (selection == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_DROPFILES_DND))
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* This means he wants the names of the dropped files.
* gdk_dropfiles_filter already has stored the text/uri-list
* data temporarily in dropfiles_prop.
*/
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->dropfiles_prop != NULL)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
selection_property_store
(requestor, win32_sel->dropfiles_prop->target, win32_sel->dropfiles_prop->bitness,
win32_sel->dropfiles_prop->data, win32_sel->dropfiles_prop->length);
g_clear_pointer (&win32_sel->dropfiles_prop, g_free);
}
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else if (selection == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_OLE2_DND))
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
property = convert_dnd_selection_to_target (target, requestor);
}
else
property = NULL;
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Generate a selection notify message so that we actually fetch the
* data (if property == GDK_SELECTION) or indicating failure (if
* property == NULL).
*/
generate_selection_notify (requestor, selection, target, property, time);
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
void
_gdk_win32_selection_property_change (GdkWin32Selection *win32_sel,
GdkWindow *window,
GdkAtom property,
GdkAtom type,
gint format,
GdkPropMode mode,
const guchar *data,
gint nelements)
{
if (property == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_GDK_SELECTION) &&
win32_sel->property_change_target_atom == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TARGETS))
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
{
win32_sel->property_change_target_atom = 0;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->clipboard_opened_for == INVALID_HANDLE_VALUE &&
OpenClipboard (GDK_WINDOW_HWND (window)))
{
win32_sel->clipboard_opened_for = GDK_WINDOW_HWND (window);
GDK_NOTE (DND, g_print ("Opened clipboard for 0x%p @ %s:%d\n", win32_sel->clipboard_opened_for, __FILE__, __LINE__));
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (win32_sel->clipboard_opened_for == INVALID_HANDLE_VALUE)
{
queue_open_clipboard (GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS, NULL, window, type, GDK_CURRENT_TIME);
return;
}
else
{
queue_open_clipboard (GDK_WIN32_CLIPBOARD_QUEUE_ACTION_TARGETS, NULL, window, type, GDK_CURRENT_TIME);
open_clipboard_timeout (NULL);
}
}
else if (property == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_OLE2_DND) &&
mode == GDK_PROP_MODE_REPLACE &&
win32_sel->property_change_target_atom == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_DELETE))
{
/* no-op on Windows */
win32_sel->property_change_target_atom = 0;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else if (mode == GDK_PROP_MODE_REPLACE &&
(win32_sel->property_change_target_atom == 0 ||
win32_sel->property_change_data == NULL ||
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
win32_sel->property_change_format == 0))
{
g_warning ("Setting selection property with 0x%p == NULL or 0x%x == 0 or 0x%p == 0",
win32_sel->property_change_data,
win32_sel->property_change_format,
win32_sel->property_change_target_atom);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
}
else if (mode == GDK_PROP_MODE_REPLACE &&
win32_sel->property_change_data != NULL &&
win32_sel->property_change_format != 0)
{
guchar *set_data = NULL;
gint set_data_length = 0;
gint byte_length = format / 8 * nelements;
if (win32_sel->property_change_transmute)
transmute_selection_target (type,
win32_sel->property_change_format,
data,
byte_length,
&set_data,
&set_data_length);
else
{
set_data_length = byte_length;
set_data = g_memdup (data, set_data_length);
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (set_data != NULL && set_data_length > 0)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
HGLOBAL hdata;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if ((hdata = GlobalAlloc (GMEM_MOVEABLE, set_data_length)) != 0)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
gchar *ucptr;
win32_sel->property_change_data->tymed = TYMED_HGLOBAL;
win32_sel->property_change_data->pUnkForRelease = NULL;
win32_sel->property_change_data->hGlobal = hdata;
ucptr = GlobalLock (hdata);
memcpy (ucptr, set_data, set_data_length);
GlobalUnlock (hdata);
}
else
{
WIN32_API_FAILED ("GlobalAlloc");
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_free (set_data);
}
win32_sel->property_change_format = 0;
win32_sel->property_change_data = 0;
win32_sel->property_change_target_atom = 0;
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else
{
const char *prop_name = (const char *)property;
const char *type_name = (const char *)type;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
gchar *datastring = _gdk_win32_data_to_string (data, MIN (10, format*nelements/8));
g_warning ("Unsupported property change on window 0x%p, %s property %s, %d-bit, target 0x%x of %d bytes: %s",
window,
(mode == GDK_PROP_MODE_REPLACE ? "REPLACE" :
(mode == GDK_PROP_MODE_PREPEND ? "PREPEND" :
(mode == GDK_PROP_MODE_APPEND ? "APPEND" :
"???"))),
prop_name,
format,
type_name,
nelements,
datastring);
g_free (datastring);
}
}
gint
_gdk_win32_display_get_selection_property (GdkDisplay *display,
GdkWindow *requestor,
guchar **data,
GdkAtom *ret_type,
gint *ret_format)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
GdkSelProp *prop;
g_return_val_if_fail (requestor != NULL, 0);
g_return_val_if_fail (GDK_IS_WINDOW (requestor), 0);
if (GDK_WINDOW_DESTROYED (requestor))
return 0;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, g_print ("gdk_selection_property_get: %p",
GDK_WINDOW_HWND (requestor)));
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
prop = g_hash_table_lookup (win32_sel->sel_prop_table, GDK_WINDOW_HWND (requestor));
if (prop == NULL)
{
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, g_print (" (nothing)\n"));
*data = NULL;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
return 0;
}
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
*data = g_malloc (prop->length + 1);
(*data)[prop->length] = '\0';
if (prop->length > 0)
memmove (*data, prop->data, prop->length);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, {
const char *type_name = (const char *)prop->target;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_print (" %s format:%d length:%"G_GSIZE_FORMAT"\n", type_name, prop->bitness, prop->length);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_free (type_name);
});
if (ret_type)
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
*ret_type = prop->target;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
if (ret_format)
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
*ret_format = prop->bitness;
return prop->length;
}
void
_gdk_selection_property_delete (GdkWindow *window)
{
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, g_print ("_gdk_selection_property_delete: %p (no-op)\n",
GDK_WINDOW_HWND (window)));
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
#if 0
prop = g_hash_table_lookup (sel_prop_table, GDK_WINDOW_HWND (window));
if (prop != NULL)
{
g_free (prop->data);
g_free (prop);
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_hash_table_remove (sel_prop_table, GDK_WINDOW_HWND (window));
}
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
#endif
}
void
_gdk_win32_display_send_selection_notify (GdkDisplay *display,
2011-11-02 15:49:55 +00:00
GdkWindow *requestor,
GdkAtom selection,
GdkAtom target,
GdkAtom property,
guint32 time)
{
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, {
const char *sel_name = (const char *)selection;
const char *tgt_name = (const char *)target;
const char *prop_name = (const char *)property;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_print ("gdk_selection_send_notify_for_display: %p %s %s %s (no-op)\n",
requestor, sel_name, tgt_name, prop_name);
});
}
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
/* It's hard to say whether implementing this actually is of any use
* on the Win32 platform? gtk calls only
* gdk_text_property_to_utf8_list_for_display().
*/
gint
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
gdk_text_property_to_text_list_for_display (GdkDisplay *display,
GdkAtom encoding,
gint format,
Start implementing display/screen closing scheme; keep a flag for whether Thu Aug 1 11:26:03 2002 Owen Taylor <otaylor@redhat.com> * gdk/gdkdisplay.[ch] gdk/gdkscreen.[ch] gdkinternals.h: Start implementing display/screen closing scheme; keep a flag for whether displays and screens are closed, call g_object_run_dispose(). Remove public gdk_screen_close(). * gdk/x11/gdkdisplay-x11.c gdk/x11/gdkscreen-x11.c: Add dispose() methods; move appropriate parts of the finalize there. * gdk/x11/gdkcolor-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkmain-x11.c gdk/x11/gdkpango-x11.c gdk/x11/gdkpixmap-x11.c gdk/x11/gdkproperty-x11.c gdk/x11/gdkselection-x11.c gdk/x11/gdkwindow-x11.c: Start of making everything correctly ignore operations when a display has been closed. * gdk/x11/gdkwindow-x11.c (gdk_window_get_decorations): Handle decorations == NULL. * gdk/x11/gdkcolor-x11.c (gdk_colormap_remove): Remove unnecessary hash table creation. * gdk/x11/gdkinput.c gdk/x11/gdkinput-x11.c gdk/win32/gdkinput.c Fix up gdk_device_get_history - handle events, n_events == NULL, etc. * gdk/x11/gdkproperty-x11.c (gdk_property_get): Handle failure better. * gdk/x11/gdkselection-x11.c (gdk_selection_property_get): Handle failure better, handle data == NULL, move docs here, remove an excess round trip by asking for all selection data at once. * gdk/gdkselection.c gdk/win32/{x11,win32}/gdkselection-{x11,win32}.c gdk/{x11,win32}/gdkmain-{x11,win32}.c gdk/gdkdisplay.c: Move gdk_text_property_to_text_list(), gdk_string_to_compound_text(), gdk_display_set_sm_client_id() to display-independent part of GDK. * gdk/Makefile.am (gdk_c_sources): Sort gdkdisplay/screen.[ch] into the right place.
2002-08-01 15:28:40 +00:00
const guchar *text,
gint length,
gchar ***list)
{
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
gchar *result;
const gchar *charset;
gchar *source_charset;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, {
const char *enc_name = (const char *)encoding;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_print ("gdk_text_property_to_text_list_for_display: %s %d %.20s %d\n",
enc_name, format, text, length);
});
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
if (!list)
return 0;
if (encoding == g_intern_static_string ("STRING"))
source_charset = g_strdup ("ISO-8859-1");
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else if (encoding == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_UTF8_STRING))
source_charset = g_strdup ("UTF-8");
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
else
source_charset = g_strdup ((const char *)encoding);
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
g_get_charset (&charset);
result = g_convert ((const gchar *) text, length, charset, source_charset,
NULL, NULL, NULL);
g_free (source_charset);
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
if (!result)
return 0;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
*list = g_new (gchar *, 1);
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
**list = result;
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
return 1;
}
void
gdk_free_text_list (gchar **list)
{
g_return_if_fail (list != NULL);
Apply the same fixes and improvements as to the gtk-1-3-win32-production 2002-01-10 Tor Lillqvist <tml@iki.fi> Apply the same fixes and improvements as to the gtk-1-3-win32-production branch: Bug fixes and cleanup of selection and DND functionality. Still doesn't work as well as the win32-production branch, though, but getting closer. After this, need to add Archaeopteryx Software's OLE2 DND support. * gdk/win32/gdkselection-win32.c (gdk_selection_owner_set, gdk_selection_send_notify, generate_selection_notify): Don't use SendMessage() to generate events for the same app, instead use gdk_event_put(). * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c * gdk/win32/gdkevents-win32.c: Thus, remove declaration, definition, initialisation and handling of gdk_selection_notify_msg, gdk_selection_request_msg and gdk_selection_clear_msg. * gdk/win32/gdkselection-win32.c (gdk_text_property_to_text_list, gdk_free_text_list, gdk_string_to_compound_text, gdk_free_compound_text): Implement trivially, witrh a text_list always having a single element, and a compound text always consisting of just a single (UTF-8!) string. Let's see how well this works. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Fix non-ASCII paste from the clipboard: Try getting the same formats from the Windows clipboard that gdk_property_change() puts there: CF_UNICODETEXT, UTF8_STRING or CF_TEXT+CF_LOCALE. * gdk/win32/gdkproperty-win32.c (gdk_property_change): When storing text on the clipboard, handle non-ASCII text correctly. The logic is as follows: If we have only ASCII characters, use CF_TEXT. Else, if we are on NT, use CF_UNICODETEXT. Else (we are on Win9x), if all the characters are present in the code page of some installed locale, use CF_TEXT and also set CF_LOCALE to that locale. Else (still on Win9x) store as RTF. We use a very simple RTF string, just the text, no fonts or other crap, with the non-ASCII characters as Unicode \uN keywords. Additionally, also store the UTF-8 string as such, under the format "UTF8_STRING", so that GDK can also paste from the Clipboard what it has copied there. (Thus no need to implement any RTF parser.) (find_common_locale): New function, implements the search for a locale for case 3 above. * gdk/win32/gdkglobals-win32.c: New global variables compound_text, text_uri_list, utf8_string, cf_rtf and cf_utf8_string. * gdk/win32/gdkim-win32.c (_gdk_ucs2_to_utf8): New function, converts from a wchar_t string to UTF-8. (_gdk_utf8_to_ucs2): Rename from _gdk_win32_nmbstowchar_ts. (_gdk_utf8_to_wcs): Rename from gdk_nmbstowchar_ts. * gdk/win32/gdkevents-win32.c (build_keypress_event): Use _gdk_ucs2_to_utf8(). * gdk/win32/gdkselection-win32.c: Remove some unnecessary logging. * gdk/win32/gdkdnd-win32.c: Plug memory leaks, the gdk_drag_context_ref() was called unnecessarily in a couple of places, meaning drag contexts were never freed. The same memory leaks seem to be present in gdk/linux-fb/gdkselection-fb.c, BTW. (gdk_drop_reply): For WIN32_DROPFILES drops, free the temporarily stored file list. * gdk/win32/gdkselection-win32.c: Clarify the use of the sel_prop_table. Now it is used only for storing the GDK_SELECTION "properties". The file names dropped with WM_DROPFILES -style DND is stored temporarily (between the drop and the target picking them up) in a separate place. Have a separate hash table to map selection atoms to owner windows. This used to be quite mixed up. (_gdk_dropfiles_store): New function, to store the dropped file list for the drop target to possibly fetch, and clear it afterwards, from gdk_drop_reply(). (gdk_selection_owner_get): Much simplified now.
2002-01-10 00:53:39 +00:00
g_free (*list);
g_free (list);
}
static gint
make_list (const gchar *text,
gint length,
gboolean latin1,
gchar ***list)
{
GSList *strings = NULL;
gint n_strings = 0;
gint i;
const gchar *p = text;
const gchar *q;
GSList *tmp_list;
GError *error = NULL;
while (p < text + length)
{
gchar *str;
q = p;
while (*q && q < text + length)
q++;
if (latin1)
{
str = g_convert (p, q - p,
"UTF-8", "ISO-8859-1",
NULL, NULL, &error);
if (!str)
{
g_warning ("Error converting selection from STRING: %s",
error->message);
g_error_free (error);
}
}
else
str = g_strndup (p, q - p);
if (str)
{
strings = g_slist_prepend (strings, str);
n_strings++;
}
p = q + 1;
}
if (list)
*list = g_new (gchar *, n_strings + 1);
(*list)[n_strings] = NULL;
i = n_strings;
tmp_list = strings;
while (tmp_list)
{
if (list)
(*list)[--i] = tmp_list->data;
else
g_free (tmp_list->data);
tmp_list = tmp_list->next;
}
g_slist_free (strings);
return n_strings;
}
gint
_gdk_win32_display_text_property_to_utf8_list (GdkDisplay *display,
GdkAtom encoding,
gint format,
const guchar *text,
gint length,
gchar ***list)
{
g_return_val_if_fail (text != NULL, 0);
g_return_val_if_fail (length >= 0, 0);
Changes multihead reorganizing code for win32 support, mostly from a patch Wed Jun 5 18:34:47 2002 Owen Taylor <otaylor@redhat.com> Changes multihead reorganizing code for win32 support, mostly from a patch by Hans Breuer. * gdk/gdkcolor.c gdk/x11/gdkcolor-x11.c gdk/gdkcursor.c gdk/x11/gdkcursor-x11.c gdk/gdkevents.c gdk/x11/gdkevents-x11.c gdk/gdkfont.c gdk/x11/gdkfont-x11.c gdk/gdkkeys.c gdk/x11/gdkkeys-x11.c gdk/gdkimage.c gdk/x11/gdkimage-x11.c gdk/gdkscreen.c gdk/x11/gdkmain-x11.c gdk/gdkdisplay.c gdk/gdkevents-x11.c gdk/gdkpango.c gdk/x11/gdkpango-x11.c gdk/gdkselection.c gdk/x11/gdkselection-x11.c gdk/gdkwindow.c gdk/x11/gdkwindow-x11.c gdk/gdkvisual.c gdk/x11/gdkvisual-x11.c: Move port-independent singlehead wrapper functions into port-independent part of GDK. (#80009) * gdk/win32/gdkcolor-win32.c gdk/win32/gdkcursor-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkfont-win32.c gdk/win32/gdkimage-win32.c gdk/win32/gdkkeys-win32.c gdk/win32/gdkmain-win32.c gdk/win32/gdkproperty-win32.c gdk/win32/gdkselection-win32.c gdk/win32/gkwindow-win32.c: Turn singlehead functions into "multihead" functions that ignore their GdkDisplay or GdkScreen arguments. * gdk/win32/gdkdrawable-win32.c gdk/win32/gdkevents-win32.c gdk/win32/gdkinput-win32.c gdk/win32/gdkprivate-win32.h: Misc multihead-compatibility changes. * gtk/gtk.def gdk/gdk.def: Update for multihead functions. * gdk/gdkcolormap.h gdk/gdkvisual.h gdk/x11/gdkcolormap-x11.c gdk/x11/gdkvisual-x11.c: Remove the screen fields from the public parts of the colormap/visual structures, add accessors instead. * gdk/gdkpixbuf-render.c gdk/gdkpixmap.c gdk/gdkrgb.c gdk/x11/gdkcolormap-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkimage-x11.c gdk/x11/gdkprivate-x11.h gtk/gtkgc.c gtk/gtkstyle.c gtk/gtkwidget.c: Use accessors to get the screen for colormaps, visuals; move the fields into the private structures for the x11 backend. * gdk/gdkdisplay.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/gdkscreen.[ch] gdk/x11/gdkscreen-x11.c: Remove virtualization of screen and display functions. (#79990, patch from Erwann Chenede) * gdk/win32/gdkdisplay-x11.c gdk/win32/gdkscreen-win32.c gdk/win32/{Makefile.am, makefile.msc, makefile.mingw}: New files containing stub implementations of Display, Screen functions. * gdk/x11/gdkscreen-x11.[ch] gdk/x11/gdkdisplay-x11.[ch] gdk/x11/gdkx.h: Clean up function exports and what headers they are in. (#79954) * gdk/x11/gdkx.h: Fix macro that was referring to a non-existant screen->screen_num. (In the patch for #79972, Erwann Chenede) * gdk/gdkscreen.c gdk/gdkwindow.c gdk/x11/gdkinternals.h gdk/x11/gdkscreen-x11.c: Fix gdk_screen_get_window_at_pointer() to use window hooks. (#79972, patch partly from Erwann Chenede) * gdk/x11/gdkdisplay-x11.c gdk/x11/gdkevents-x11.c: Fix some warnings.
2002-06-06 00:26:42 +00:00
if (encoding == g_intern_static_string ("STRING"))
{
return make_list ((gchar *)text, length, TRUE, list);
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else if (encoding == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_UTF8_STRING))
{
return make_list ((gchar *)text, length, FALSE, list);
}
else
{
const char *enc_name = (const char *)encoding;
g_warning ("gdk_text_property_to_utf8_list_for_display: encoding %s not handled\n", enc_name);
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
if (list)
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
*list = NULL;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
return 0;
}
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
gchar *
_gdk_win32_display_utf8_to_string_target (GdkDisplay *display,
const gchar *str)
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
return _gdk_utf8_to_string_target_internal (str, strlen (str));
}
void
gdk_win32_selection_clear_targets (GdkDisplay *display,
GdkAtom selection)
{
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (selection == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_OLE2_DND) ||
selection == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_LOCAL_DND_SELECTION))
{
g_array_set_size (win32_sel->dnd_selection_targets, 0);
}
else if (selection == GDK_SELECTION_CLIPBOARD)
{
g_array_set_size (win32_sel->clipboard_selection_targets, 0);
}
else if (selection == GDK_SELECTION_PRIMARY)
{
/* Do nothing */
}
else
{
const char *sel_name = (const char *)selection;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_warning ("Unsupported generic selection %s (0x%p)", sel_name, selection);
}
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
gint
_gdk_win32_add_target_to_selformats (GdkAtom target,
GArray *array)
{
gint added_count = 0;
const char *target_name;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
wchar_t *target_name_w;
GdkSelTargetFormat fmt;
gint i;
GArray *compatibility_formats;
gint starting_point;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
for (i = 0; i < array->len; i++)
if (g_array_index (array, GdkSelTargetFormat, i).target == target)
break;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Don't put duplicates into the array */
if (i < array->len)
return added_count;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_TARGETS) ||
target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_COMPOUND_TEXT) || target == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_SAVE_TARGETS))
{
/* Add the "we don't really support transferring that to
* other processes" format, just to keep the taget around.
*/
fmt.target = target;
fmt.format = 0;
fmt.transmute = FALSE;
g_array_append_val (array, fmt);
added_count += 1;
return added_count;
}
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Only check the newly-added pairs for duplicates,
* all the ones that exist right now have different targets.
*/
starting_point = array->len;
target_name = (const char *)target;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
target_name_w = g_utf8_to_utf16 (target_name, -1, NULL, NULL, NULL);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (target_name_w == NULL)
return added_count;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
fmt.format = RegisterClipboardFormatW (target_name_w);
GDK_NOTE (DND, g_print ("Registered clipboard format %S as 0x%x\n", target_name_w, fmt.format));
g_free (target_name_w);
fmt.target = target;
fmt.transmute = FALSE;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Add the "as-is" format */
g_array_append_val (array, fmt);
added_count += 1;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
compatibility_formats = get_compatibility_formats_for_target (target);
for (i = 0; compatibility_formats != NULL && i < compatibility_formats->len; i++)
{
gint j;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
fmt = g_array_index (compatibility_formats, GdkSelTargetFormat, i);
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Don't put duplicates into the array */
for (j = starting_point; j < array->len; j++)
if (g_array_index (array, GdkSelTargetFormat, j).format == fmt.format)
break;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (j < array->len)
continue;
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Add a compatibility format */
g_array_append_val (array, fmt);
added_count += 1;
}
gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) 2004-07-07 Tor Lillqvist <tml@iki.fi> * gdk/win32/gdkdrawable-win32.c (gdk_win32_draw_text) * gdk/win32/gdkfont-win32.c (gdk_text_extents) * gdk/win32/gdkproperty-win32.c (find_common_locale, gdk_property_change) * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Use g_utf8_to_utf16() instead of the removed _gdk_utf8_to_ucs2() (see below). * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkmain-win32.c (_gdk_windowing_init) * gdk/win3/gdkprivate-win32.h: Add a variable for the TARGETS atom. Initialize it. Declare it. Drop the variable for the COMPOUND_TEXT atom. * gdk/win32/gdkim-win32.c (gdk_wcstombs): Don't return UTF-8. This function is supposed to return the string in the locale's charset and encoding. Use g_convert(). (gdk_mbstowcs): Similarily, don't take an UTF-8 string, but a string in the locale's charset. Use g_convert(). (_gdk_ucs2_to_utf8, _gdk_utf8_to_wcs, _gdk_utf8_to_ucs2): Delete. The UCS-2 functions didn't handle surrogates anyway. Use GLib's UTF-16 functions instead. Windows uses UTF-16. * gdk/win32/gdkprivate-win32.h: Remove declarations of the deleted functions mentioned above. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Use CF_TEXT also if the string is of type STRING, i.e. ISO-8859-1, and the current codepage is 1252, and contains no C1 chars. Accept also UTF8_STRING. * gdk/win32/gdkselection-win32.c (_gdk_selection_property_store): Mark as static. When storing STRING data, convert to Latin-1. (#140537) (gdk_selection_owner_set_for_display): Now that STRING is always ISO-8859-1, use UTF8_STRING when sending the selection request to ourselves. (gdk_selection_convert): Handle also UTF8_STRING. (#140537, John Ehresman) (gdk_text_property_to_text_list_for_display): Make work more like X11 version. Do obey the encoding parameter. (gdk_string_to_compound_text_for_display, gdk_utf8_to_compound_text_for_display): Don't even pretend supporting COMPOUND_TEXT. (gdk_utf8_to_string_target): Convert to ISO-8859-1, like on X11. (sanitize_utf8): Zero-terminate string.
2004-07-07 00:10:03 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
return added_count;
}
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
/* This function is called from gtk_selection_add_target() and
* gtk_selection_add_targets() in gtkselection.c. It is this function
* that takes care of setting those clipboard formats for which we use
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
* delayed rendering (that is, all formats, as we use delayed rendering
* for everything). This function only registers the formats, but does
* not announce them as supported. That is handled as a special case
* in gdk_window_property_change().
*
* Implementation detail:
* This function will be called repeatedly, every time the PRIMARY selection changes.
* It will also be called immediately before the CLIPBOARD selection changes.
* We let GTK+ handle the PRIMARY selection internally and do nothing here
* (therefore it's not possible to middle-click-paste between processes,
* unless one process deliberately puts PRIMARY selection contents into
* CLIPBOARD selection, and the other process does paste on middle-click).
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
*/
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
void
gdk_win32_display_add_selection_targets (GdkDisplay *display,
GdkWindow *owner,
GdkAtom selection,
GdkAtom *targets,
guint ntargets)
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GdkWin32Selection *win32_sel = _gdk_win32_selection_get ();
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
gint i;
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
GDK_NOTE (DND, {
const char *sel_name = (const char *)selection;
g_print ("gdk_win32_selection_add_targets: %p: %s: ",
owner ? GDK_WINDOW_HWND (owner) : NULL,
sel_name);
for (i = 0; i < ntargets; i++)
{
const char *tgt_name = (const char *)targets[i];
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_print ("%s", tgt_name);
if (i < ntargets - 1)
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
g_print (", ");
}
g_print ("\n");
Bug 559408 - Transparency lost when images are copied between GTK+ and the 2009-01-29 Tor Lillqvist <tml@novell.com> Bug 559408 - Transparency lost when images are copied between GTK+ and the clipboard on Windows The code in gdkselection-win32.c is simplified quite a lot. When an image is stored on the clipboard from GTK+, only the "PNG" and CF_DIB formats are advertised. PNG is our preferred format because it can losslessly represent any image that gdk-pixbuf can handle, even with alpha, unambiguously. CF_DIB is also advertised because of its ubiquitous support in Windows software. Delayed rendering is used as before, so for other data than text nothing is actually stored on the clipboard until some app asks for it. It's pointless to advertise images formats like "image/png" or "image/x-MS-bmp" that no other Windows application is going to look for anyway. Likewise it's pointless to store text under formats like "text/plain" or "UTF8_STRING". Just store CF_UNICODETEXT. CF_TEXT will be available as a synthesized format anyways. Office 2007 offers and asks for images in PNG format using the registered clipboard format name "PNG", so we do likewise. If the "PNG" format is available that is preferred when copying to GTK+ from the clipboard. Unfortunately Word 2007 still uses CF_DIB without handling alpha. But PowerPoint 2007 uses PNG if available and handles alpha correctly. Both Word and Powerpoint offer PNG with alpha nicely. Firefox and IE7 offer images that contain alpha as 32-bit version 3 BI_RGB DIBs with the undocumented "extra" bytes actually being alpha. Also, alpha is premultiplied into the RGB bytes, presumably because that is how AlphaBlend() wants such DIBs. That is also taken care of. At least for Firefox it easy to be sure that a CF_DIB on the clipboard is from Firefox. Also some general stylistic cleanup, comment improvements, and improvements of debugging printout especially in the clipboard handling. Those are not detailled below. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c * gdk/win32/gdkgmain-win32.c: Move some globals that were used only in gdkselection-win32.c to be local to that file. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Don't bother checking if text to be placed on the clipboard consists of only ASCII. * gdk/win32/gdkselection-win32.c: Add static variables for a list of GdkPixbuf-supported formats, well-known registered clipboard formats for images, and for GdkAtoms for well-known image and text formats. (_gdk_win32_selection_init): Initialize above static variables. (selection_property_store) (gdk_selection_property_get) (_gdk_selection_property_delete): Don't use a FIFO of GdkSelProps for a window after all, it breaks testtext. See bug #163844. (gdk_selection_convert): When converting to the TARGETS format, i.e. when the caller wants to know what clipboard formats are available, if PNG is available we report just that and skip CF_DIB, JPEG and GIF even if advertised. If CF_UNICODETEXT is available, report only UTF8_STRING. When converting to the UTF8_STRING format, i.e. when the caller wants text from the clipboard, try just CF_UNICODETEXT. There is no point in trying CF_TEXT as Windows will synthesize CF_UNICODETEXT from CF_TEXT anyway, if some app has stored just CF_TEXT. When converting to the image/bmp format, i.e. when the caller wants an CF_DIB image from the clipboard, we check if the DIB is a version 3 32-bit BI_RGB one that is likely to actually contain alpha in the "extra" bytes. Such a DIB is likely to have premultiplied alpha even, at least in the case of Firefox 3 and IE7. We then edit the DIB in-place into a version 5 one in BI_BITFIELDS format and undo the alpha premultiplication. In any case, prepend a BMP file header before letting go of the data which will be fed to the gdk-pixbuf bmp loader by upper levels. (gdk_win32_selection_add_targets): If some kind of pixmap image format is being added, actually advertise just PNG and CF_DIB. Note that alpha won't be stored on the clipboard through CF_DIB. This is because gdk-pixbuf's bmp loader doesn't save alpha. Furthermore, few if any non-GTK+ Windows apps seem to understand a version 5 DIB with proper alpha anyway. (_gdk_win32_selection_convert_to_dib): Simplified muchly. svn path=/trunk/; revision=22255
2009-01-30 01:16:47 +00:00
});
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (selection == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_OLE2_DND) ||
selection == _gdk_win32_selection_atom (GDK_WIN32_ATOM_INDEX_LOCAL_DND_SELECTION) ||
selection == GDK_SELECTION_CLIPBOARD)
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
GArray *fmts = NULL;
gint added_count = 0;
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
if (selection == GDK_SELECTION_CLIPBOARD)
fmts = win32_sel->clipboard_selection_targets;
else
fmts = win32_sel->dnd_selection_targets;
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
for (i = 0; i < ntargets; i++)
added_count += _gdk_win32_add_target_to_selformats (targets[i], fmts);
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Re-announce our list of supported formats */
if (added_count > 0)
send_targets_request (GDK_CURRENT_TIME);
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else if (selection == GDK_SELECTION_PRIMARY)
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
{
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
/* Do nothing */
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
}
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
else
{
const char *sel_name = (const char *)selection;
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
W32: Massive W32 DnD fix Massive changes to OLE2 DnD protocol, which was completely broken before: * Keep GdkDragContext and OLE2 objects separate (don't ref/unref them together, don't necessarily create them together). * Keep IDataObject formats in the object itself, not in a global variable. * Fix getdata() to look up the request target in its format list, not in the global hash table * Create target GdkDragContext on each drag_enter, destroy it on drag_leave, whereas IDropTarget is created when a window becomes a drag destination and is re-used indefinitely. * Query the source IDataObject for its supported types, cache them in the target (!) context. This is how GTK+ works, honestly. * Remember current_src_object when we initiate a drag, to be able to detect later on that the data object is ours and use a shortcut when querying targets * Make sure GDK_DRAG_MOTION is only sent when something changes * Support GTK drag cursors * Ensure that exotic GTK clipboard formats are registered (but try to avoid registering formats that can't be used between applications). * Don't enumerate internal formats * Ensure that DnD indicator window can't accept drags or receive any kind of input (use WS_EX_TRANSPARENT). * Remove unneeded indentation in _gdk_win32_dnd_do_dragdrop() * Fix indentation in gdk_win32_drag_context_drop_finish() * Remove obsolete comments in _gdk_win32_window_register_dnd() * Check for DnD in progress when processing WM_KILLFOCUS, don't emit a grab break event in such cases (this allows alt-tabbing while DnD is in progress, though there may be lingering issues with focus after dropping...) * Support Shell ID List -> text/uri-list conversion, now it's possible to drop files (dragged from Explorer) on GTK+ applications * Explicitly use RegisterClipboardFormatA() when we know that the string is not in unicode. Otherwise explicitly use RegisterClipboardFormatW() with a UTF8->UTF16 converted string * Fix _gdk_win32_display_get_selection_owner() to correctly bail when selection owner HWND is NULL (looking up GdkWindow for NULL HWND always succeeds and returns the root window - not the intended effect) * More logging * Send DROP_FINISHED event after DnD loop ends * Send STATUS event on feedback * Move GetKeyboardState() and related code into _gdk_win32_window_drag_begin(), so that it's closer to the point where last_pt and start_pt are set * Use & 0x80 to check for the key being pressed. Windows will set low-order bit to 1 for all mouse buttons to indicate that they are toggled, so simply checking for the value not being 0 is not enough anymore. This is probably a new thing in modern W32 that didn't exist before (OLE2 DnD code is old). * Fixed (hopefully) and simplified HiDPI parts of the code. Also adds managed DnD implementation for W32 GDK backend (for both OLE2 and LOCAL protocols). Mostly a copy of the X11 backend code, but there are some minor differences: * doesn't use drag_window field in GdkDragContext, uses the one in GdkWin32DragContext exclusively * subtracts hotspot offset from the window coordinates when showing the dragback animation * tries to consistently support scaling and caches the scale in the context * Some keynav code is removed (places where grabbing/ungrabbing should happen is marked with TODOs), and the rest is probably inert. Also significantly changes the way selection (and clipboard) is handled (as MSDN rightly notes, the handling for DnD and Clipboard formats is virtually the same, so it makes sense to handle both with the same code): * Don't spam GDK_OWNER_CHANGE, send them only when owner actually changes * Open clipboard when our process becomes the clipboard owner (we are doing it anyway, to empty the clipboard and *become* the owner), and then don't close it until a scheduled selection request event (with TARGETS target) is received. Process that event by announcing all of our supported formats (by that time add_targets() should have been called up the stack, thus the formats are known; just in case, add_targets() will also schedule a selection request, if one isn't scheduled already, so that late-coming formats can still be announced). * Allow clipboard opening for selection_convert() to be delayed if it fails initially. * The last two points above should fix all the bugs about GTK+ rising too much ruckus over OpenClipboard() failures, as owner change *is allowed* to fail (though not all callers currently handle that case), and selection_convert() is asynchronous to begin with. Still, this is somewhat risky, as there's a possibility that the code will work in unexpected ways and the clipboard will remain open. There's now logging to track the clipboard being opened and closed, and a number of failsafes that try to ensure that it isn't kept open for no reason. * Added copious notes on the way clipboard works on X11, Windows and GDK-W32, also removed old comments in DnD implementation, replaced some of them with the new ones * A lot of crufty module-global variables are stuffed into a singleton object, GdkWin32Selection. It's technically possible to make it a sub-object of the Display object (the way Wayland backend does), but since Display object on W32 is a singleton anyway... why bother? * Fixed the send_change_events() a bit (was slightly broken in one of the previous iterations) * Ensure that there's no confusion between selection conversion (an artifact term from X11) and selection transmutation (changing the data to be W32-compatible) * Put all the transmutation code and format-target-matching code into gdkselection-win32.c, now this code isn't spread across multiple files. * Consequently, moved some code away from gdkproperty-win32.c and gdkdnd-win32.c * Extensive format transmutation checks for OLE2 DnD and clipboard. We now keep track of which format mappings are for transmutations, and which aren't (for example, when formats are passed as-is, or when a registered name is just an alias) * Put transmutation code into separate functions * Ensure that drop target keeps a format->target map for supported formats, this is useful when selection_convert() is called, as it only receives a single target and no hints on the format from which the data should be transmuted into this target. * Add clear_targets() on W32, to de called by GTK * Use g_set_object() instead of g_ref_object() where it is allowed. * Fix indentation (and convert tabs to spaces), remove unused variables (This commit is cherry-picked from the gtk-3-22 branch) https://bugzilla.gnome.org/show_bug.cgi?id=786509
2017-08-19 12:06:27 +00:00
g_warning ("Unsupported generic selection %s (0x%p)", sel_name, selection);
}
New debugging function, to log a clipboard format name symbolically. 2005-04-04 Tor Lillqvist <tml@novell.com> * gdk/win32/gdkmain.c (_gdk_win32_cf_to_string): New debugging function, to log a clipboard format name symbolically. (_gdk_win32_data_to_string): Also new, to log random data bytes. Implement delayed rendering on Win32, specifically for transfering images through the clipboard from GTK+ apps to other apps (#168173, implementation by Ivan Wong): * gdk/win32/gdkevents-win32.c (gdk_event_translate): Handle WM_RENDERFORMAT. * gdk/win32/gdkprivate-win32.h * gdk/win32/gdkglobals-win32.c: Add _format_atom_table, _delayed_rendering_data and _image_bmp. * gdk/win32/gdkmain-win32.c: Initialize _image_bmp. * gdk/win32/gdkproperty-win32.c (gdk_property_change): Accept formats other than GDK_TARGET_STRING or _utf8_string, and assume they are handled through delayed rendering. * gdk/win32/gdkselection-win32.c (gdk_selection_convert): Return all available formats (including those registered by GTK+ apps) on request_targets. (gdk_selection_property_get): We should append a zero byte like X11 does. (gdk_win32_selection_add_targets): New function, for gtkselection's use. Win32 requires that the clipboard owner registers all valid formats even if the owner wants delayed rendering. (_gdk_win32_selection_convert_to_dib): New function. Convert images to DIB using gdk-pixbuf. * gdk/win32/gdkwin32.h: Declare gdk_win32_selection_add_targets(). * gtk/gtkselection.c (gtk_selection_add_target, gtk_selection_add_targets): Call gdk_win32_selection_add_targets() to register target formats. * gdk/gdk.symbols: Add gdk_win32_selection_add_targets().
2005-04-04 00:12:26 +00:00
}