The monitor change detection code in _gdk_x11_screen_size_changed() and
process_monitors_change() goes to some length to make sure its only emitted
when there is an actual change to the data visible via the GdkScreen monitors
api.
However, commit 662e69ad added some code that always emits "monitors-changed"
in _gdk_x11_screen_size_changed when we have randr13 and get a ConfigureNotify
on the root window (even though we may already have emitted it in the
RRScreenChangesNotify event!).
As far as I can tell this is due to a comment in the bug referenced by the
commit (https://bugzilla.gnome.org/show_bug.cgi?id=601712#c4) where it says:
This version of the patch changes GdkDisplay to emit "monitors-changed" when
the primary monitor changes (see the change in _gdk_x11_screen_size_changed).
And, if you remove this part of the change the signal is not emitted when just
the primary is changed. However, this is not really the right approach. We
should just also check for if the primary changes in process_monitors_change()
to avoid spurious signal emissions.
https://bugzilla.gnome.org/show_bug.cgi?id=643216
The last paragraph still seems to be out of place though, as if it
was a previous attempt at an overview or part of some older text
elsewhere.
This overuses the passive too.
If a level 1 key maps to a key value passed to
gdk_test_simulate_key(), raise the GDK_SHIFT_MASK flag so the reqested
key value is generated. Also add a regression test for this fix.
On Windows and OS X we want to prefer the native backends over the X11
backend.
On Linux, nothing changes as nobody is going to enable those backends
(and if they do, they'll know what they get).
Previously we weren't installing the device headers when compiling
without XINPUT support. But we would include them from gdkx.h, so
essentially the build was broken.
With this patch the types will exist but not do anything.
At the same time, change the library sonames for -3.0 to just -3.
This is necessary since the 2.99 releases installed libraries like
libgtk-3.0.so.0.9903.0, and we want to prevent the library version
number from jumping back. So 3.0 will have libgtk-3.so.0.0.0.
Gdk-3.0 is already included above via --include-uninstalled, so
don't also specify it in INCLUDES. Doing so breaks if it isn't
installed already, since we try to find the gdk-3.0.pc.
- replace GdkNativeWindow with HWND, remove type casts
- no more GdkDisplayClass::get_drag_protocol but GdkWindowImpl::get_drag_protocol
- remove *_client_message*()
This will trigger a repaint of the window, but it may be more efficient
to just copy back the old surface contents and let gtk+ just update the
changed parts.
Just a noop that is better than crashing in the case this is called
(it can be called for a toplevel GtkWindow that is parented into
another widget by setting gtk_widget_set_parent_window to an offscreen
window).
constructors which take an object of the same class as its first argument are
mis-detected as method call with "self" argument by the GIR scanner. Using the
new (constructor) annotation from bug 561264, mark some of them as proper
constuctors, so that you can call them with NULL as first argument from
bindings; in particular, this fixes gdk_window_new() and the
gtk_radio_button_new_with*() constructors.
The previous function gdk_drag_get_protocol_for_display() took native
window handles, so it had to be changed. Because it didn't do what it
was named to do (it didn't return a protocol even though it was named
get_protocol) and because it doesn't operate on the display anymore but
on the actual window, it's now called gdk_window_get_drag_protocol().
... and all APIs making use of it.
That code like it hasn't been touched in years, Google codesearch
didn't find any users and most importantly it's a horrendous API, so
let's just make it die instead of having to port it over to
non-GdkNativeWindow usage, which would be required for multi-backend
GDK.
http://mail.gnome.org/archives/gtk-devel-list/2011-January/msg00049.html
Use GdkWindow instead. This requires calling
gdk_x11_window_foreign_new_for_display(), so might cause a slight
performance penalty, but is required to be portable.
Only master devices must modify the associated device to separate
a pointer/keyboard pair, slave devices must only call
_gdk_device_remove_slave().
Fixes bug 639767 - password not accepted in gnome-screensaver dialog,
reported by Frederic Crozat. On VT-switch, the X server removes its
grab on HW devices, the effect on clients is that slave devices
disappear, and these were mistakenly mangling the master device
hierarchy. so gdk_device_get_associated_device() on the client
pointer wouldn't return the paired keyboard anymore.
The final effect is that gtkplug-x11 wasn't setting a keyboard to
its generated events.
slave devices don't have coordinates themselves, as they depend
on a master, this only changes if they have a grab in effect,
so only keep toplevel tracking enabled in such situation. Fixes
Bug #640313 - BadDevice X error when ungrabbing a SLAVE device,
noticed by Jesse van den Kieboom.
This function allocates the button mask, so free it after
use, or right before the next XIQueryPointer() call, as done
in gdk_x11_device_xi2_window_at_position().
mods_state->effective is not being set in XIQueryPointer() currently, so
use base|latched|locked instead, effective is nothing else than a shorthand
for these ORs, and these 3 values are set correctly anytime.
The state attribute is available in GdkEventMotion, GdkEventButton,
GdkEventScroll, GdkEventKey and GdkEventCrossing. This type annotation
fixes the wrapping of this attribute in the GI PyGObject bindings.
https://bugzilla.gnome.org/show_bug.cgi?id=639929
When copying allocated events, also copy the source device.
When synthesizing double or triple clicks, copy the original
button press event including device information.
https://bugzilla.gnome.org/show_bug.cgi?id=639822
This change does not introduce any functionality change, mostly
cosmtic cleanups, like re-linebreak when introduced annotations messed
up indentation or whitespace errors fixes.
-Update the project files to simplify them a bit after the seperation of
GDK-Pixbuf (move GDK-Pixbuf includes into the property sheet, move the
linking of Cairo/Pango/PangoCairo into the property sheet)--this is for
all DLL/EXE Projects (GDK/GTK/gtk-demo)
-Update the GDK-Win32 project as the source files have changed
significantly (especially as GDK3 was not compilable on Windows for a
while--thanks to Hans Breuer for the help in the process-Bug 639127)
-Made up for missed headers in the "install" stage and removed the removed
headers in the property sheet
-Updated GTK+ .def file generation as an extra macro is needed for that
-Updated gdk/Makefile.am for the generation of gdk.vcproj from gdk.vcprojin
Prevents an Xlib warning on Xnest, or Xorg with xinerama, or other
non-RANDR-capable xserver. Reintroduce a have_randr12 field in
GdkDisplayX11 to avoid having to call XRRQuery{Extension,Version} twice,
and don't select randr 1.2 events if that's false.
https://bugzilla.gnome.org/show_bug.cgi?id=634711
Signed-off-by: Julien Cristau <jcristau@debian.org>
When commenting out a binary, also comment out the related variables.
Don't include Makefile.decl in gtk-doc Makefile.am's as they disagree
on assigning to EXTRA_DIST.
There's no usecase for them, so remove them before we have to commit to
keeping an API.
Make the hooks private for now, actually removing them will come in
followup patches.
Its usecase was GERD - http://testbit.eu/~timj/historic/gerd/ - and that
project is long since dead.
I couldn't find any app using it after asking around and googling either.
Its usecase was GERD - http://testbit.eu/~timj/historic/gerd/ - and that
project is long since dead.
It has been superseded in GTK 2.2 by GdkDisplayPointerHooks anyway.
There are sure regressions but basic stuff seems to be working
again after all the API breakage done with comments like
"Win32 and Quartz need to be ported still."
We need to defer setting the default display until the
GdkDisplay is fully initialized. Also, short-circuit some
encoding conversions when creating windows, to avoid an
implicit dependency on the display being in the list of
displays yet.
This was an incomplete attempt to get rid of the custom free function.
Lets just keep it for now. Bug 637849, patch by Dan Winship.
Also add a test case for this function.
Normally HFS+ (the MacOSX file system) isn't case-sensitive, so having both
GtkQuartzWindow.h and gtkquartzwindow.h causes the latter to overwrite the
former during git pull, breaking the build.
An overlooked API change in the gdk-backend work: many of the
keymap functions used to accept NULL to mean 'default keymap'.
They no longer do, so update the docs to match the new behaviour.
