Johan Dahlin [Sun, 8 Jul 2007 03:22:18 +0000 (03:22 +0000)]
Parse <accessibility> but don't do anything yet. Do not strip
2007-07-08 Johan Dahlin <jdahlin@async.com.br>
* gtk/gtkwidget.c (gtk_widget_buildable_custom_tag_start):
Parse <accessibility> but don't do anything yet.
* gtk/gtk-builder-convert (GtkBuilderConverter._parse):
Do not strip atk/accessibility tags
* tests/buildertest.c (test_widget):
Test this
Richard Hult [Sat, 7 Jul 2007 16:19:40 +0000 (16:19 +0000)]
]) ([GdkQuartzWindow -windowDidBecomeMain:]): Keep a stack of main windows
2007-07-07 Richard Hult <richard@imendio.com>
* gdk/quartz/GdkQuartzWindow.c ([GdkQuartzWindow -windowDidResignMain:])
([GdkQuartzWindow -windowDidBecomeMain:]):
* gdk/quartz/gdkwindow-quartz.c (_gdk_quartz_window_did_resign_main)
(_gdk_quartz_window_did_become_main, gdk_window_hide)
(_gdk_windowing_window_destroy): Keep a stack of main windows and
select the most recent one when hiding/closing the current one.
Johan Dahlin [Sat, 7 Jul 2007 15:44:30 +0000 (15:44 +0000)]
(GtkBuilderConverter): Refactor; Optimize widget lookups by not
traversing the whole tree.
Use a common api to create sizegroups, actiongroups and uimanagers
to avoid name conflicts.
Output newly created objects sorted by id before the old roots under
<interface>
Richard Hult [Fri, 6 Jul 2007 19:49:42 +0000 (19:49 +0000)]
Fix (de)miniaturizing transient windows, by (un)setting the parent before
2007-07-06 Richard Hult <richard@imendio.com>
* gdk/quartz/gdkprivate-quartz.h:
* gdk/quartz/gdkwindow-quartz.c:
* gdk/quartz/GdkQuartzWindow.c: Fix (de)miniaturizing transient
windows, by (un)setting the parent before and after miniaturizing.
Tor Lillqvist [Wed, 4 Jul 2007 08:20:18 +0000 (08:20 +0000)]
On Win32 call gtk_status_icon_button_press() in an idle callback and not
2007-07-04 Tor Lillqvist <tml@novell.com>
* gtk/gtkstatusicon.c: On Win32 call
gtk_status_icon_button_press() in an idle callback and not
directly from the window procedure to avoid "g_main_loop_run():
called recursively from within a source's check() or prepare()
member, iteration not possible" warnings.
Chris Wilson [Wed, 4 Jul 2007 07:41:57 +0000 (07:41 +0000)]
Remove a surplus GDK_THREADS_LEAVE() and avoid a potential use after free
2007-07-04 Chris Wilson <chris@chris-wilson.co.uk>
* gtk/gtkrecentchoosermenu.c (idle_populate_func),
(idle_populate_clean_up), (gtk_recent_chooser_menu_populate):
Remove a surplus GDK_THREADS_LEAVE() and avoid a potential use after
free in the source destroy notify. (#453033)
Johan Dahlin [Wed, 4 Jul 2007 03:49:33 +0000 (03:49 +0000)]
pack children from the start instead of the end, this should probably be
2007-07-04 Johan Dahlin <jdahlin@async.com.br>
* gtk/gtkcelllayout.c (_gtk_cell_layout_buildable_add_child):
pack children from the start instead of the end, this should
probably be made configurable at some point.
Tor Lillqvist [Tue, 3 Jul 2007 16:58:42 +0000 (16:58 +0000)]
Handle GDK_PIXBUF_EXTRA_LIBS like GDK_EXTRA_LIBS, i.e. clear it if
2007-07-03 Tor Lillqvist <tml@novell.com>
* configure.in: Handle GDK_PIXBUF_EXTRA_LIBS like GDK_EXTRA_LIBS,
i.e. clear it if enable_explicit_deps isn't on. If we build with
--with-included-loaders and --enable-explicit-deps=no we don't
want -ltiff -ljpeg etc in the gdk-pixbuf-2.0.pc file.
Tor Lillqvist [Tue, 3 Jul 2007 16:53:17 +0000 (16:53 +0000)]
Implement. Just return FALSE for now. What should this function actually
2007-07-03 Tor Lillqvist <tml@novell.com>
* gdk/win32/gdkkeys-win32.c (gdk_keymap_have_bidi_layouts):
Implement. Just return FALSE for now. What should this function
actually do? Does keyboards layouts being "in use" mean that such
layouts can be switched to on the fly? If so we need to actually
check that.
Johan Dahlin [Tue, 3 Jul 2007 03:26:07 +0000 (03:26 +0000)]
Mimic the behavior of gtk_dialog_add_action_widget closer, to make
2007-07-03 Johan Dahlin <jdahlin@async.com.br>
* gtk/gtkdialog.c (gtk_dialog_buildable_custom_finished): Mimic the behavior
of gtk_dialog_add_action_widget closer, to make clicking on the buttons
actually work.
new function to check whether a rubber banding operation is currently
2007-07-02 Kristian Rietveld <kris@gtk.org>
* gtk/gtk.symbols:
* gtk/gtktreeview.[ch] (gtk_tree_view_is_rubber_banding_active): new
function to check whether a rubber banding operation is currently
active. (#393579, Christian Neumair).
* gdk-pixbuf.c (gdk_pixbuf_get_option): Document the
"orientation" option.
* gdk-pixbuf.symbols:
* gdk-pixbuf-core.h:
* gdk-pixbuf-util.c (gdk_pixbuf_apply_embedded_orientation):
New function to handle Exif orientation information in
tiff and jpeg images. (#439567, Michael Chudobiak)
Richard Hult [Sat, 30 Jun 2007 21:50:06 +0000 (21:50 +0000)]
Continue the event handling cleanup:
2007-06-30 Richard Hult <richard@imendio.com>
Continue the event handling cleanup:
* gdk/quartz/gdkevents-quartz.c: (synthesize_crossing_events):
Don't do anything if the old and new windows are the same,
simplifies the callers.
(find_mouse_window_for_ns_event): Always return a window (root
instead of NULL). Remove unneccessary check for the current mouse
window being NULL, that was a workaround for a now fixed
bug. Convert to root coordinates if we don't find a window.
(synthesize_crossing_events_for_ns_event): Simplify and add
comment.
Richard Hult [Sat, 30 Jun 2007 19:50:31 +0000 (19:50 +0000)]
No need to special case the root window now that it has the right size.
2007-06-30 Richard Hult <richard@imendio.com>
* gdk/quartz/gdkevents-quartz.c: (create_crossing_event),
(synthesize_crossing_events), (find_mouse_window_for_ns_event),
(synthesize_crossing_events_for_ns_event): No need to special case
the root window now that it has the right size.
Richard Hult [Sat, 30 Jun 2007 19:39:45 +0000 (19:39 +0000)]
Improve comments in those files. Add a debugging helper to print out
2007-06-30 Richard Hult <richard@imendio.com>
* gdk/quartz/gdkwindow-quartz.c:
* gdk/quartz/gdkevents-quartz.c: Improve comments in those
files. Add a debugging helper to print out information about a
window.
Richard Hult [Sat, 30 Jun 2007 08:56:50 +0000 (08:56 +0000)]
Improve the tracking rect updating: - No need to override setFrame. - Use
2007-06-30 Richard Hult <richard@imendio.com>
* gdk/quartz/GdkQuartzView.c: Improve the tracking rect updating:
- No need to override setFrame.
- Use our own size instead of the bounds from for the view as it's
not always updated before the tracking rect.
- Only reset the tracking rect if the new window is nil).
* gdk/quartz/GdkQuartzWindow.c: Update the bounds of the content
view when the window size changes.
Johan Dahlin [Sat, 30 Jun 2007 07:03:34 +0000 (07:03 +0000)]
Move type-func to <object> instead of <child>, add a test to make sure
2007-06-30 Johan Dahlin <jdahlin@async.com.br>
* gtk/gtkbuilderparser.c: (_get_type_by_symbol), (parse_object),
(parse_child):
* tests/buildertest.c: (test_types):
Move type-func to <object> instead of <child>, add a test to
make sure that it works as desired, #452463
Matthias Clasen [Sat, 30 Jun 2007 01:43:04 +0000 (01:43 +0000)]
New function to determine if keyboard layouts for both LTR and LTR
2007-06-29 Matthias Clasen <mclasen@redhat.com>
* gdk/gdk.symbols:
* gdk/gdkkeys.h:
* gdk/x11/gdkkeys-x11.c (gdk_keymap_have_bidi_layouts): New
function to determine if keyboard layouts for both LTR and LTR
languages are in use. Refactor the direction caching code to
make this information available. (#451575, Behnam Esfahbod)
Ryan Lortie [Fri, 29 Jun 2007 16:02:24 +0000 (16:02 +0000)]
Connect the "destroy" signal so that the 'window' pointer is set back to
2007-06-29 Ryan Lortie <desrt@desrt.ca>
* tests/testgtk.c (create_composited_window): Connect the "destroy"
signal so that the 'window' pointer is set back to NULL (like for the
other examples).
* gdk/x11/gdkwindow-x11.c (gdk_window_impl_x11_finalize): Don't
destroy the Damage here. By finalize() XDestroyWindow has already
been called (and took the Damage with it).
Emmanuele Bassi [Fri, 29 Jun 2007 14:51:13 +0000 (14:51 +0000)]
Do not cache the icon pixbufs but rely on the GtkIconTheme not to hit the
2007-06-29 Emmanuele Bassi <ebassi@gnome.org>
* gtk/gtkrecentmanager.c (get_icon_for_mime_type): Do not
cache the icon pixbufs but rely on the GtkIconTheme not to
hit the disk cache too much (should tentatively fix #426987).
Johan Dahlin [Thu, 28 Jun 2007 15:15:02 +0000 (15:15 +0000)]
Prettify the output by running it through xmllint --format if xmllint is
2007-06-28 Johan Dahlin <jdahlin@async.com.br>
* gtk/gtk-builder-convert (_indent): Prettify the output by
running it through xmllint --format if xmllint is available
which also requires the subprocess module only available
in python 2.4 or later
Christian Persch [Thu, 28 Jun 2007 14:32:58 +0000 (14:32 +0000)]
Make the volume button use range 0.0 to 1.0 by default, and fix tooltips
2007-06-28 Christian Persch <chpe@gnome.org>
* gtk/gtkvolumebutton.c: (gtk_volume_button_class_init),
(gtk_volume_button_init), (gtk_volume_button_dispose),
(gtk_volume_button_update_tooltip): Make the volume button use range
0.0 to 1.0 by default, and fix tooltips code to cope with different ranges.
Bug #354181.