2005-10-25 Kristian Rietveld <kris@gtk.org>
Fixes#311955, reported by Markku Vire; and a lot more.
* gtk/gtktreemodelfilter.c: Made this thing work correctly.
Levels can now have invisible nodes cached, functions have been
rewritten/refactored to take this into account; also the logic in
some functions has been corrected.
Each referenced node now references all visible children or a single
invisible child, in order to be able to monitor all changes in the
child model required for correct operation.
When a virtual root is set, all nodes on the virtual root level are
referenced.
At last, a bunch of random bug fixes and comment updates.
2005-10-25 Matthias Clasen <mclasen@redhat.com>
* gdk/x11/gdkxftdefaults.c (init_xft_settings): Make the
initialization of screen_x11->xft_rgba more explicit. (#319627,
Bogdan Nicula)
2005-10-25 Matthias Clasen <mclasen@redhat.com>
* gtk/gtkfilechooser.c (gtk_file_chooser_class_init): Fix whitespace
in a blurb. (#319499, Francisco Javier F. Serrador)
2005-10-24 Matthias Clasen <mclasen@redhat.com>
Add settings to hide the input method and Unicode control character
submenus in the context menus of GtkEntry and GtkTextView.
* gdk/x11/gdkevents-x11.c: Map GTK+ settings to X settings.
* gtk/gtksettings.c (gtk_settings_class_init): Declare GTK+ settings here.
* gtk/gtktextview.c (popup_targets_received):
* gtk/gtkentry.c (popup_targets_received): Use the settings here.
2005-10-24 Kristian Rietveld <kris@gtk.org>
* gtk/gtktreemodelsort.c (gtk_tree_model_sort_row_inserted): emitting
row-has-child-toggled when the level is not in the cache does not
make sense. (This bug probably crept in during the early stages
of GtkTreeModelSort when there wasn't a correct row-has-child-toggled
implementation yet),
(gtk_tree_model_sort_real_unref_node): only get the child iter
when we will actually also use it.
2005-10-22 Michael Natterer <mitch@imendio.com>
* gtk/gtktreeview.c (gtk_tree_view_scroll_to_cell): check for the
widget being realized, in addition to being visible, to avoid
running into precondition check in gtk_tree_view_get_cell_area().
(approved by Kris).
* modules/engines/ms-windows/msw_style.c (setup_msw_rc_style):
Experimentally, scrollbar steppers can shrink to 8 pixels on
Win32. Reflect that in the theme.
* modules/engines/ms-windows/Theme/gtk-2.0/gtkrc: Experimentally,
there is a 1-pixel border between a scrollbar and its child
in ScrolledWindows on Win32. Reflect that in the theme.
2005-10-20 Matthias Clasen <mclasen@redhat.com>
* gtk/gtklabel.c (gtk_label_create_window): Use GDK_WINDOW_CHILD
for the selection window. (#318806, Alex Larsson)
2005-10-18 Federico Mena Quintero <federico@ximian.com>
Fixes bug #317999:
* tests/autotestfilechooser.c
(test_button_folder_states_for_action): Test that we have either
$cwd or the explicitly-set folder.
(test_reload_sequence): Likewise.
* gtk/gtkfilechooserdefault.c
(gtk_file_chooser_default_get_current_folder): If our reload_state
is RELOAD_EMPTY, return a GtkFilePath corresponding to $cwd.
2005-10-14 Tor Lillqvist <tml@novell.com>
* gdk/win32/gdkdisplay-win32.c: Remove the clipboard viewer code.
It didn't really do anything useful, and was just confusing and
incomplete. Comments claimed we don't do delayed rendering, but in
fact we do, for images. (The delayed rendering code has other
problems, though, see #168173.) The clipboard viewer code was
probably even buggy (the WM_CHANGECBCHAIN handled didn't propagate
the message when necessary). It was just test code, it said so in
a comment. Add something similar back later if necessary.
(_win32_on_clipboard_change,
_gdk_win32_register_clipboard_notification): Remove.
(gdk_display_supports_selection_notification,
gdk_display_request_selection_notification): Always just return
FALSE. We didn't generate any GDK_OWNER_CHANGE events anywhere.
2005-10-13 Tor Lillqvist <tml@novell.com>
* gdk/win32/gdkevents-win32.c (gdk_event_translate): Don't treat
Alt-Enter specially. It does not have any special meaning and
should be passed on to the application. (#318378, Tim Evans)
2005-10-13 Tor Lillqvist <tml@novell.com>
Set visual depth to 24 for 32 bits-per-pixel devices on
Win32. This allows gdk_drawable_real_draw_pixbuf() to use the
optimized composite_0888() function rather than the slower image
dithering functions to draw pixbufs (#313993, John Ehresman)
* gdk/win32/gdkimage-win32.c (_gdk_win32_new_image): Use
_gdk_windowing_get_bits_for_depth() to initialize
GdkImage::bits_per_pixel.
(_gdk_windowing_get_bits_for_depth): Return 32 bits for depth 24.
* gdk/win32/gdkpixmap-win32.c (gdk_pixmap_new): Use
_gdk_windowing_get_bits_for_depth() to initialize
BITMAPINFOHEADER::biBitCount.
* gdk/win32/gdkvisual-win32.c (_gdk_visual_init): Set
GdkVisual::depth to 24 even if GetDeviceCaps(BITSPIXEL) returns
32.
2005-10-11 Matthias Clasen <mclasen@redhat.com>
* gtk/gtktextbtree.c (_gtk_text_btree_delete): Try to match an off
toggle here with the matching on toggle if it immediately follows.
This is a common case, and handling it here prevents quadratic blowup
in cleanup_line() below. (#317125)
* gtk/gtktextsegment.h:
* gtk/gtktextsegment.c (_gtk_char_segment_new_from_two_strings): Pass
the character counts into this function instead of computing them
again.
2005-10-10 Tommi Komulainen <tommi.komulainen@iki.fi>
* gtk/gtkfilechooser.h (GtkFileChooserError): Add ALREADY_EXISTS error
code.
* gtk/gtkfilechooserdefault.c
(gtk_file_chooser_default_add_shortcut_folder): Return an error code
when the shortcut already exists in the sidebar. (#147521)
2005-10-07 Federico Mena Quintero <federico@ximian.com>
Fixes bug #317999:
* tests/autotestfilechooser.c (test_reload_sequence): Ensure that
the the result of gtk_file_chooser_get_folder() is NULL or
non-NULL at the right times. Log the tests.
(test_button_folder_states): New test for the initial states of
the current folder in GtkFileChooserButton.
(main): Test that the number of warnings/errors/critical errors is
zero.
* gtk/gtkfilechooser.c (gtk_file_chooser_get_current_folder):
Clarify the documentation on when this can return NULL.
(gtk_file_chooser_get_current_folder_uri): Likewise.
* gtk/gtkfilechooserbutton.c (struct
_GtkFileChooserButtonPrivate): Added a folder_has_been_set flag;
we use it to keep track of whether a folder has been set.
(gtk_file_chooser_button_map): Implement. If no folder has been
loaded before, we at least try to load $cwd here.
(gtk_file_chooser_button_constructor): If the construct-time
dialog already has a folder set, turn on our folder_has_been_set
flag.
(dialog_current_folder_changed_cb): Turn on our
folder_has_been_set flag.
2005-10-07 Michael Natterer <mitch@gimp.org>
* gtk/gtkfilechooserdefault.c: remove some explicit #defines of
PROFILE_FILE_CHOOSER because they break compilation after the
removal of <unistd.h>.
2005-10-05 Matthias Clasen <mclasen@redhat.com>
* gtk/gtkrc.c (gtk_rc_clear_realized_style): Readd the memory
leak fix, since the assertion failures were traced to
an xchat bug. (#314696)
2005-10-04 Federico Mena Quintero <federico@ximian.com>
* gtk/gtkfilechooserdefault.c (switch_to_shortcut): Take a
shortcut position, not a ShortcutsIndex, to make the function more
generic.
(home_folder_handler): Pass the resolved index to
switch_to_shortcut(), rather than just SHORTCUTS_HOME.
(desktop_folder_handler): Likewise for SHORTCUTS_DESKTOP.
(gtk_file_chooser_default_class_init): Add a "quick-bookmark"
binding signal, and bind it to Alt-1, Alt-2. ..., Alt-0. These
make the respective key switch to the first, second, ..., tenth
bookmark.
(quick_bookmark_handler): Implement.
2005-10-04 Federico Mena Quintero <federico@ximian.com>
* gtk/gtkpathbar.c (get_dir_name): Don't special-case Home and
Desktop; just use their real names on the file system for the
user-visible names.
* gtk/gtkfilechooserdefault.c
(gtk_file_chooser_default_class_init): Add a "desktop-folder"
binding signal, and bind it to M-d by default.
(desktop_folder_handler): Implement the default handler.
(home_folder_handler): Use a helper function.
(switch_to_shortcut): New helper function.
(shortcuts_append_home): Don't special-case the name of "Home";
just use the folder name.
2005-10-04 Tor Lillqvist <tml@novell.com>
* gtk/gtkcalendar.c (gtk_calendar_init): Make it compile without
HAVE__NL_TIME_FIRST_WEEKDAY. (#317910, Mathias Hasselmann)
2005-10-04 Matthias Clasen <mclasen@redhat.com>
* gtk/gtkrc.c (gtk_rc_clear_realized_style): Revert the change
from yesterday, since it leads to assertion failures. (#317879,
Sebastian Bacher)
2005-10-03 Matthias Clasen <mclasen@redhat.com>
* gtk/gtkrc.c (gtk_rc_clear_realized_style): Unref the style when
removing it from the hash table. (#314696, Benjamin Berg)
2005-10-01 Tor Lillqvist <tml@novell.com>
* gdk/win32/gdkdrawable-win32.c (blit_from_pixmap): In case
BitBlt() fails with ERROR_INVALID_HANDLE, the most probable cause
is that the the desktop isn't visible because the session has been
switched, the screen is locked, or a terminal server session
disconnected, so no error message necessary. (#137796)
It is of course remotely possible that BitBlt() failing with
ERROR_INVALID_HANDLE might also be caused by some other
problem. We could strive for perfection and track whether the
desktop is visible by using WTSRegisterSessionNotification() and
handling WM_WTSESSION_CHANGE. I think that's overdoing it just for
this issue, though. If we would track desktop visibility, we
should then avoid even trying to update the display at all while
the desktop isn't visible.
2005-09-30 Matthias Clasen <mclasen@redhat.com>
* gtk/gtkcalendar.c (gtk_calendar_init): Another attempt
to correct the calculations for the first week day. We may
just have to remove this code if too many locales turn out
to have broken data.
2005-09-30 Matthias Clasen <mclasen@redhat.com>
* gtk/gtksocket-x11.c (_gtk_socket_windowing_size_request):
Prevent overflow when storing size hints in an unsigned
short variable. Tracked down by Ray Strode and Søren Sandmann.
2005-09-28 Federico Mena Quintero <federico@ximian.com>
* tests/autotestfilechooser.c (test_reload_sequence): Oops, return
if the unmap test fails.
(test_reload_sequence): Use gtk_widget_show_now() instead of
gtk_widget_show() for the dialog.
Thanks to Tim Janik for noticing these.
2005-09-28 Federico Mena Quintero <federico@ximian.com>
* tests/testcellrenderertext.c: New file with a set of tests for
GtkCellRendererText. The idea is to run a text renderer through
most of its interesting property values to see that it works
correctly.
* tests/Makefile.am: Added testcellrenderertext.
* gtk/gtkfilechooserdefault.c (load_set_model): Mark the
entry/exit of this function for profiling.
(gtk_file_chooser_default_style_set): Mark the start/end of the
calls to the parent class and the the signal emission for profiling.
2005-09-28 Federico Mena Quintero <federico@ximian.com>
* tests/testcellrenderertext.c: New file with a set of tests for
GtkCellRendererText. The idea is to run a text renderer through
most of its interesting property values to see that it works
correctly.
* tests/Makefile.am: Added testcellrenderertext.
* gtk/gtkfilechooserdefault.c (load_set_model): Mark the
entry/exit of this function for profiling.
(gtk_file_chooser_default_style_set): Mark the start/end of the
calls to the parent class and the the signal emission for profiling.
2005-09-29 Tor Lillqvist <tml@novell.com>
* gtk/gtkmain.c (_gtk_get_localedir): The locale directory is
passed to bindtextdomain() which isn't UTF-8-aware, so convert to
system codepage using g_win32_locale_filename_from_utf8().
(#317457, Kazuki Iwamoto)
2005-09-28 Matthias Clasen <mclasen@redhat.com>
* gtk/gtktexttag.c (gtk_text_attributes_ref): Return the attributes
to make this function work as boxed copy function. (#317455,
Gustavo Carneiro)
2005-09-27 Federico Mena Quintero <federico@ximian.com>
Don't reload the current folder unnecessarily on ::map().
* gtk/gtkfilechooserprivate.h (ReloadState): New enum to represent
the reloading state.
(struct _GtkFileChooserDefault): Added a "reload_state" field.
* gtk/gtkfilechooserdefault.c (gtk_file_chooser_default_init):
Initialize impl->reload_state.
(gtk_file_chooser_default_map): Check the impl->reload_state; load
a default folder if no folder has been set, or reload the current
one only if we had been unmapped first.
(gtk_file_chooser_default_update_current_folder): Set the
reload_state to RELOAD_HAS_FOLDER.
(gtk_file_chooser_default_unmap): Implement, and set the
reload_state to RELOAD_WAS_UNMAPPED.
(shortcuts_model_create): Don't call shortcuts_add_bookmarks()
here; they'll get (re)loaded on ::map() anyway.
* gtk/gtkfilechooserwidget.c
(gtk_file_chooser_widget_constructor): Don't set a default folder here.
* tests/autotestfilechooser.c (test_action_widgets): Don't take in
a dialog; build it ourselves.
(test_reload): New test to ensure that we don't load the default
folder more than once, and that we reload it when
unmapping/remapping.
(get_impl_from_dialog): New utility function.
(test_widgets_for_current_action): Use get_impl_from_dialog().
