gtk4.git
13 years agowayland: Return NULL as the keyval name for 0
Matthias Clasen [Sat, 6 Apr 2013 14:35:26 +0000 (10:35 -0400)]
wayland: Return NULL as the keyval name for 0

This is what the X11 backend does, and it makes the "NoSymbol"
disappear from menuitems that don't have an accelerator.

13 years agoUpdate Simplified Chinese translation
Aron Xu [Sat, 6 Apr 2013 14:30:15 +0000 (22:30 +0800)]
Update Simplified Chinese translation

13 years agoUpdate Simplified Chinese translation
Wylmer Wang [Sat, 6 Apr 2013 14:18:38 +0000 (22:18 +0800)]
Update Simplified Chinese translation

13 years agoGtkRecentChooserMenu: avoid accessing a NULL action
Tristan Van Berkom [Sat, 6 Apr 2013 09:42:19 +0000 (18:42 +0900)]
GtkRecentChooserMenu: avoid accessing a NULL action

When the GtkActivatable sets the related action, the activatable
properties are sync'ed. In this case the null check is important
to avoid updating from a NULL action.

13 years agogtk_recent_chooser_set_filter: Allow NULL value to unset the filter.
Tristan Van Berkom [Sat, 6 Apr 2013 08:56:19 +0000 (17:56 +0900)]
gtk_recent_chooser_set_filter: Allow NULL value to unset the filter.

While GtkRecentChooser implementations accept for the "filter" property
to unset the current filter, the API fires an unneeded assertion.

This patch removes the assertion.

13 years agoreftests: Add a test for window default size
Benjamin Otte [Sat, 6 Apr 2013 08:46:00 +0000 (10:46 +0200)]
reftests: Add a test for window default size

https://bugzilla.gnome.org/show_bug.cgi?id=697275

13 years agotests: Remove multiscreen code from teststusicon
Benjamin Otte [Thu, 4 Apr 2013 14:54:45 +0000 (16:54 +0200)]
tests: Remove multiscreen code from teststusicon

13 years agotests: Remove testmultiscreen example
Benjamin Otte [Thu, 4 Apr 2013 14:38:19 +0000 (16:38 +0200)]
tests: Remove testmultiscreen example

There aren't multiple screens anymore, so no need to test that.

Multiple displays can be tested with gtk-demo and testgtk.

13 years agotestgtk: Get rid of screen part in display & screen test
Benjamin Otte [Thu, 4 Apr 2013 14:36:09 +0000 (16:36 +0200)]
testgtk: Get rid of screen part in display & screen test

13 years agogdk: Deprecate gdk_display_get_n_screens()
Benjamin Otte [Thu, 4 Apr 2013 14:22:48 +0000 (16:22 +0200)]
gdk: Deprecate gdk_display_get_n_screens()

13 years agox11: Simplify code for single-screen case
Benjamin Otte [Thu, 4 Apr 2013 14:22:12 +0000 (16:22 +0200)]
x11: Simplify code for single-screen case

13 years agogtk: Simplify code for reduced screen count
Benjamin Otte [Thu, 4 Apr 2013 14:21:50 +0000 (16:21 +0200)]
gtk: Simplify code for reduced screen count

13 years agoicontheme: Remove unused function
Benjamin Otte [Thu, 4 Apr 2013 14:19:36 +0000 (16:19 +0200)]
icontheme: Remove unused function

13 years agogtk-demo: Remove multiscreen support
Benjamin Otte [Tue, 2 Apr 2013 12:38:24 +0000 (14:38 +0200)]
gtk-demo: Remove multiscreen support

There's only one screen per display, so no need to let users select from
a list of that 1 screen.

13 years agoGdkOffscreenWindow: Implement gdk_window_get_frame_extents()
Tristan Van Berkom [Sat, 6 Apr 2013 08:09:27 +0000 (17:09 +0900)]
GdkOffscreenWindow: Implement gdk_window_get_frame_extents()

This avoids crashes in gtk_window_get_position() whenever the GdkWindow
is offscreen.

13 years agoreftests/Makefile.am: remove whitespace following trailing backslash
Thomas Hindoe Paaboel Andersen [Fri, 5 Apr 2013 19:39:41 +0000 (21:39 +0200)]
reftests/Makefile.am: remove whitespace following trailing backslash

13 years agoUpdated kn translations
Shankar Prasad [Fri, 5 Apr 2013 18:54:52 +0000 (00:24 +0530)]
Updated kn translations

13 years agobroadway: Unsure we clear the grab on window hide
Alexander Larsson [Fri, 5 Apr 2013 15:06:09 +0000 (17:06 +0200)]
broadway: Unsure we clear the grab on window hide

13 years agobroadway: Fix warning
Alexander Larsson [Fri, 5 Apr 2013 14:56:35 +0000 (16:56 +0200)]
broadway: Fix warning

Don't pass NULL source_device in grab/ungrab events

13 years agoUpdated kn translations
Shankar Prasad [Fri, 5 Apr 2013 13:50:14 +0000 (19:20 +0530)]
Updated kn translations

13 years agoobjects-finalize test: Blacklist gdk_pixbuf_simple_anim_iter_get_type()
Tristan Van Berkom [Fri, 5 Apr 2013 11:44:12 +0000 (20:44 +0900)]
objects-finalize test: Blacklist gdk_pixbuf_simple_anim_iter_get_type()

13 years agowayland: Implement more keymap functions
Matthias Clasen [Fri, 5 Apr 2013 11:21:40 +0000 (07:21 -0400)]
wayland: Implement more keymap functions

This commit implements gdk_keymap_get_entries_for_keyval
and gdk_keymap_lookup_key.

13 years agowayland: Implement gdk_keymap_get_entries_for_keycode
Matthias Clasen [Fri, 5 Apr 2013 01:52:12 +0000 (21:52 -0400)]
wayland: Implement gdk_keymap_get_entries_for_keycode

13 years agoAdded test case to assert all object types finalize properly.
Tristan Van Berkom [Fri, 5 Apr 2013 07:32:56 +0000 (16:32 +0900)]
Added test case to assert all object types finalize properly.

The test case uses a weak ref to assert objects can finalize,
and then spins the main loop shortly after finalizing to assert
that the finalized object did not leak GSources into the main context
causing latent crashes.

13 years agowayland: Always initialize directions
Matthias Clasen [Fri, 5 Apr 2013 01:10:18 +0000 (21:10 -0400)]
wayland: Always initialize directions

I forgot to initialize directionm in gdk_wayland_keymap_new,
leading to crash.

13 years agobuild: bump required GLib version to 2.37.0
Cosimo Cecchi [Thu, 4 Apr 2013 21:17:05 +0000 (17:17 -0400)]
build: bump required GLib version to 2.37.0

13 years agoUpdated Hungarian translation
Balázs Úr [Thu, 4 Apr 2013 19:49:01 +0000 (21:49 +0200)]
Updated Hungarian translation

13 years agobloatpad: add a test for GApplication's busy state
Cosimo Cecchi [Wed, 3 Apr 2013 18:19:18 +0000 (14:19 -0400)]
bloatpad: add a test for GApplication's busy state

https://bugzilla.gnome.org/show_bug.cgi?id=672018

13 years agoFix gtk_window_set_opacity()
Alexander Larsson [Thu, 4 Apr 2013 13:38:40 +0000 (15:38 +0200)]
Fix gtk_window_set_opacity()

We *do* want to call gdk_window_set_opacity() on toplevels, because
this particular case does work.

https://bugzilla.gnome.org/show_bug.cgi?id=697263

13 years agowayland: Emit ::direction-changed when appropriate
Matthias Clasen [Thu, 4 Apr 2013 03:52:52 +0000 (23:52 -0400)]
wayland: Emit ::direction-changed when appropriate

We always emit direction-changed when we get a new keymap, but
for state changes, we compare old and new direction and only
emit the signal when the direction actually changes.

13 years agowayland: Implement keymap direction
Matthias Clasen [Thu, 4 Apr 2013 03:45:54 +0000 (23:45 -0400)]
wayland: Implement keymap direction

This is very similar to the X11 implementation.

13 years agowayland: display: Fix memory leak
Pavel Vasin [Thu, 28 Mar 2013 11:21:11 +0000 (15:21 +0400)]
wayland: display: Fix memory leak

https://bugzilla.gnome.org/show_bug.cgi?id=696767

13 years agoGtkHeaderBar: Fix memory leak
Pavel Vasin [Thu, 28 Mar 2013 11:17:54 +0000 (15:17 +0400)]
GtkHeaderBar: Fix memory leak

https://bugzilla.gnome.org/show_bug.cgi?id=696767

13 years agoGtkWindow: fix notify of :attached-to
Sébastien Wilmet [Wed, 3 Apr 2013 17:08:16 +0000 (19:08 +0200)]
GtkWindow: fix notify of :attached-to

https://bugzilla.gnome.org/show_bug.cgi?id=697200

13 years agobroadway: Ensure we allocate shm surface space early
Alexander Larsson [Wed, 3 Apr 2013 13:50:54 +0000 (15:50 +0200)]
broadway: Ensure we allocate shm surface space early

This is nicer then getting random sigbus later

13 years agoBroadway: Fix shm names
Alexander Larsson [Wed, 3 Apr 2013 13:49:54 +0000 (15:49 +0200)]
Broadway: Fix shm names

We were sometimes getting zero chars in the name, making them
shorter, due to an off-by-zero in the size.

13 years agobroadway: destroy old surface before creating new one
Alexander Larsson [Wed, 3 Apr 2013 13:46:21 +0000 (15:46 +0200)]
broadway: destroy old surface before creating new one

We were not using the old one anyway, and this may in some cases
use less memory (although in most cases the server has a ref to the
surface anyway).

13 years agowindow: don't schedule a display of mnemonics if window has no modifier
Cosimo Cecchi [Tue, 2 Apr 2013 20:36:30 +0000 (16:36 -0400)]
window: don't schedule a display of mnemonics if window has no modifier

When the window has no mnemonics modifier set, as in the case of a
GtkMenu, never schedule a display of mnemonics on focus-in.

Previously, for those windows, the GdkModifierType mask fetched from the
device would typically have been zero, leading to the

mnemonic_modifier == (mask & gtk_accelerator_get_default_mod_mask ())

check to succeed, so we would always trigger a display for popup menus.

https://bugzilla.gnome.org/show_bug.cgi?id=697144

13 years agowindow: refactor a function
Cosimo Cecchi [Tue, 2 Apr 2013 20:33:33 +0000 (16:33 -0400)]
window: refactor a function

Instead of having maybe_set_mnemonics_visible(), separate the checks
from the actual scheduling of mnemonics display.

https://bugzilla.gnome.org/show_bug.cgi?id=697144

13 years agowindow: rename functions dealing with mnemonic delayed display
Cosimo Cecchi [Tue, 2 Apr 2013 19:56:31 +0000 (15:56 -0400)]
window: rename functions dealing with mnemonic delayed display

Don't mention "auto mnemonics", since those methods are purely about
scheduling a delayed display, and that makes understanding the code a
bit harder.

https://bugzilla.gnome.org/show_bug.cgi?id=697144

13 years agowindow: reset timeout source id before setting property
Cosimo Cecchi [Tue, 2 Apr 2013 19:54:02 +0000 (15:54 -0400)]
window: reset timeout source id before setting property

gtk_window_set_mnemonics_visible() will try to g_source_remove() it
otherwise, which seems harmless, but conceptually wrong.

https://bugzilla.gnome.org/show_bug.cgi?id=697144

13 years agoUpdated Polish translation
Wojciech Szczęsny [Tue, 2 Apr 2013 14:05:11 +0000 (16:05 +0200)]
Updated Polish translation

13 years agogdkdisplay: Warn if asked for a non-zero screen not a zero screen
Rob Bradford [Tue, 2 Apr 2013 11:04:22 +0000 (12:04 +0100)]
gdkdisplay: Warn if asked for a non-zero screen not a zero screen

Following on from a6b29d73 this commit correctly warn if you try and use
deprecated multiple screen behaviour.

13 years agostyle: Add missing deprecation markers
Benjamin Otte [Tue, 2 Apr 2013 09:44:08 +0000 (11:44 +0200)]
style: Add missing deprecation markers

Some functions in gtkstyle.h were overlooked when we added the
GDK_DEPRECATED macros.

Also add IGNORE_DEPRECATIONS to the few remaining callers of those
functions.

13 years agowindow: Remove call to gtk_widget_reset_rc_styles()
Benjamin Otte [Tue, 2 Apr 2013 09:42:54 +0000 (11:42 +0200)]
window: Remove call to gtk_widget_reset_rc_styles()

First of all, that call is deprecated. Second, we don't have RC styles
anymore. Third, what that function does today is invalidate style
contexts, but that happens automatically when setting the screen on the
style context later.

So this function is completely unnecessary.

13 years agobroadway: Mark deprecated function call
Benjamin Otte [Mon, 1 Apr 2013 13:38:19 +0000 (15:38 +0200)]
broadway: Mark deprecated function call

... to avoid gcc warning us. Ideally, we'd not call a deprecated
function here, but I'm lazy.

13 years agobroadway: Don't use deprecated API
Benjamin Otte [Mon, 1 Apr 2013 13:38:04 +0000 (15:38 +0200)]
broadway: Don't use deprecated API

13 years agocelrendererspinner: Don't use deprecated API
Benjamin Otte [Mon, 1 Apr 2013 13:37:38 +0000 (15:37 +0200)]
celrendererspinner: Don't use deprecated API

13 years agowidget: Get rid of deprecation warnings
Benjamin Otte [Mon, 1 Apr 2013 13:37:20 +0000 (15:37 +0200)]
widget: Get rid of deprecation warnings

13 years agotestsensitive: Remove deprecated part
Benjamin Otte [Mon, 1 Apr 2013 13:37:04 +0000 (15:37 +0200)]
testsensitive: Remove deprecated part

13 years agogdkdisplay: Remove get_screen() and get_n_screens() vfuncs
Benjamin Otte [Mon, 1 Apr 2013 12:47:17 +0000 (14:47 +0200)]
gdkdisplay: Remove get_screen() and get_n_screens() vfuncs

13 years agoMention display:screen change in release notes
Matthias Clasen [Tue, 2 Apr 2013 03:05:51 +0000 (23:05 -0400)]
Mention display:screen change in release notes

13 years agotests: improve gtkmenu testcase
Ryan Lortie [Sat, 23 Mar 2013 20:15:04 +0000 (16:15 -0400)]
tests: improve gtkmenu testcase

Don't just create a menushell and populate it with random data -- verify that
the resulting menu layout is actually correct.

This is introduced in a separate commit because the old code was failing this
part of the test.

https://bugzilla.gnome.org/show_bug.cgi?id=696468

13 years agoIntroduce GtkMenuTracker
Ryan Lortie [Fri, 22 Mar 2013 21:24:46 +0000 (17:24 -0400)]
Introduce GtkMenuTracker

GtkMenuTracker folds a nested structure of sections in a GMenuModel into
a single linear menu, which it expresses to its user by means of 'insert
item at position' and 'remove item at position' callbacks.

The logic for where to insert separators and how to handle action
namespaces is contained within the tracker, removing the need to have
this logic duplicated in the 3 or 4 places that consume GMenuModel.

In comparison with the previous code, the tracker no longer completely
destroys and rebuilds menus every time a single change occurs.  As a
result, the new gtkmenu testcase now runs in approximately 3 seconds
instead of ~60 before.

https://bugzilla.gnome.org/show_bug.cgi?id=696468

13 years agotests: add a test for gtk_menu_shell_bind()
Ryan Lortie [Fri, 22 Mar 2013 21:20:43 +0000 (17:20 -0400)]
tests: add a test for gtk_menu_shell_bind()

Borrow the RandomMenu code from the GLib testsuite and hook it up to
gtk_menu_shell_bind().

https://bugzilla.gnome.org/show_bug.cgi?id=696468

13 years agogtktextattributes: move comment above relevant struct
Sébastien Wilmet [Mon, 1 Apr 2013 19:09:14 +0000 (21:09 +0200)]
gtktextattributes: move comment above relevant struct

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agoGtkTextTagTable: simplify a bit the code
Sébastien Wilmet [Mon, 1 Apr 2013 14:38:41 +0000 (16:38 +0200)]
GtkTextTagTable: simplify a bit the code

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agoGtkTextView: remove some dead code
Sébastien Wilmet [Sat, 30 Mar 2013 22:09:34 +0000 (23:09 +0100)]
GtkTextView: remove some dead code

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agogtktextattributes: explain what is "pg_bg"
Sébastien Wilmet [Sun, 31 Mar 2013 17:08:49 +0000 (19:08 +0200)]
gtktextattributes: explain what is "pg_bg"

I didn't know what "pg" stands for.

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agogtktextattributes: include the right headers
Sébastien Wilmet [Fri, 29 Mar 2013 13:09:06 +0000 (14:09 +0100)]
gtktextattributes: include the right headers

Some function prototypes in gtktexttagprivate.h are implemented in
gtktextattributes.c.

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agogtktexttypes: remove inline_byte_begins_utf8_char()
Sébastien Wilmet [Thu, 28 Mar 2013 20:50:08 +0000 (21:50 +0100)]
gtktexttypes: remove inline_byte_begins_utf8_char()

The function was used only in gtk_text_byte_begins_utf8_char().

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agoDoc: fix dead link in the text widget overview
Sébastien Wilmet [Mon, 1 Apr 2013 19:00:04 +0000 (21:00 +0200)]
Doc: fix dead link in the text widget overview

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agoDoc: move the GtkTextAttributes struct
Sébastien Wilmet [Thu, 28 Mar 2013 22:16:20 +0000 (23:16 +0100)]
Doc: move the GtkTextAttributes struct

It's more logical to have the GtkTextAttributes and GtkTextAppearance
structs together. And it creates a separation between gtk_text_tag and
gtk_text_attributes functions.

https://bugzilla.gnome.org/show_bug.cgi?id=697048

13 years agobroadway: Use wss: on https: uris
Alexander Larsson [Mon, 1 Apr 2013 15:29:06 +0000 (17:29 +0200)]
broadway: Use wss: on https: uris

13 years agoDeprecate gdk_get_display()
Benjamin Otte [Sun, 25 Nov 2012 20:11:48 +0000 (21:11 +0100)]
Deprecate gdk_get_display()

13 years agogdk: Shuffle code around
Benjamin Otte [Sun, 25 Nov 2012 20:09:58 +0000 (21:09 +0100)]
gdk: Shuffle code around

Make implementation, header and documentation be at the correct place.

13 years agox11: Make display:screen relation 1:1
Benjamin Otte [Sun, 25 Nov 2012 19:17:30 +0000 (20:17 +0100)]
x11: Make display:screen relation 1:1

Only ever open the default screen.

If apps need support for multiple screens, they need to
gdk_display_open() them manually.

13 years agowayland: Remove unused variable
Benjamin Otte [Fri, 29 Mar 2013 12:36:01 +0000 (13:36 +0100)]
wayland: Remove unused variable

13 years agoRemove an unused filed in GdkWindowImplWayland
Matthias Clasen [Mon, 1 Apr 2013 02:46:31 +0000 (22:46 -0400)]
Remove an unused filed in GdkWindowImplWayland

The unused toplevel_window_type field was copied over from
the X backend.

13 years agoRemove an unused field in GdkWindowX11
Matthias Clasen [Mon, 1 Apr 2013 02:44:59 +0000 (22:44 -0400)]
Remove an unused field in GdkWindowX11

The toplevel_window_type field was only ever set to -1, and
never used. The actual, used toplevel_window_type field lives
in GdkWindow.

13 years agoFinnish translation update by Jiri Grönroos
Ville-Pekka Vainio [Sun, 31 Mar 2013 17:17:23 +0000 (20:17 +0300)]
Finnish translation update by Jiri Grönroos

13 years ago[l10n] Updated Catalan (Valencian) translation
Carles Ferrando [Sun, 31 Mar 2013 15:03:03 +0000 (17:03 +0200)]
[l10n] Updated Catalan (Valencian) translation

13 years agowayland: Set a wayland-client log handler to print errors we get
Kristian Høgsberg [Sat, 30 Mar 2013 18:18:54 +0000 (14:18 -0400)]
wayland: Set a wayland-client log handler to print errors we get

13 years agowayland: Don't error out on G_IO_HUP until we've read G_IO_IN data
Kristian Høgsberg [Sat, 30 Mar 2013 18:17:07 +0000 (14:17 -0400)]
wayland: Don't error out on G_IO_HUP until we've read G_IO_IN data

We can get G_IO_HUP and G_IO_IN at the same time, if the compositor writes
data to us and then closes our connection.  Make sure that we dispatch events
always if we have G_IO_IN and then error out if we get G_IO_HUP after that.

13 years ago[l10n] Updated Catalan (Valencian) translation
Carles Ferrando [Sat, 30 Mar 2013 15:24:03 +0000 (16:24 +0100)]
[l10n] Updated Catalan (Valencian) translation

13 years ago[l10n] Update Catalan translation
Pau Iranzo [Sat, 30 Mar 2013 15:22:48 +0000 (16:22 +0100)]
[l10n] Update Catalan translation

13 years agoUpdated Slovenian translation
Matej Urbančič [Sat, 30 Mar 2013 10:13:50 +0000 (11:13 +0100)]
Updated Slovenian translation

13 years agoUpdated Slovenian translation
Matej Urbančič [Sat, 30 Mar 2013 10:09:59 +0000 (11:09 +0100)]
Updated Slovenian translation

13 years agoUpdated Slovenian translation
Matej Urbančič [Sat, 30 Mar 2013 10:07:17 +0000 (11:07 +0100)]
Updated Slovenian translation

13 years agoUpdated Slovenian translation
Matej Urbančič [Sat, 30 Mar 2013 08:47:45 +0000 (09:47 +0100)]
Updated Slovenian translation

13 years agobroadway: Reuse surfaces passed to server
Alexander Larsson [Fri, 29 Mar 2013 08:50:59 +0000 (09:50 +0100)]
broadway: Reuse surfaces passed to server

This way we don't have to reopen all the time for pure updates,
and we can immediately unlink the shm file to avoid "leaking" them
on improper shutdown.

13 years agobroadway: Move surface open to server
Alexander Larsson [Fri, 29 Mar 2013 08:38:34 +0000 (09:38 +0100)]
broadway: Move surface open to server

This way we can cache the last opened surface

13 years agobroadway: Make chrome work again
Alexander Larsson [Thu, 28 Mar 2013 23:23:41 +0000 (00:23 +0100)]
broadway: Make chrome work again

13 years agobroadway: Add support for password authentication
Alexander Larsson [Thu, 28 Mar 2013 21:57:33 +0000 (22:57 +0100)]
broadway: Add support for password authentication

13 years agobroadway: Limit window resize based on window geometry
Alexander Larsson [Thu, 28 Mar 2013 19:03:13 +0000 (20:03 +0100)]
broadway: Limit window resize based on window geometry

13 years agobroadway: Sync surface updates with paint clock
Alexander Larsson [Thu, 28 Mar 2013 18:22:42 +0000 (19:22 +0100)]
broadway: Sync surface updates with paint clock

We now only update surface data after we have painted. Before we painted
in an idle, which meant we might send black data some times if we e.g.
resized the window and had not painted yet. Also, it means we're updating
less often to the daemon, saving resources.

We still have to queue a flush in the idle for non-draw operations,
otherwise e.g. resize of a toplevel will never be flushed if the clock
is frozen (e.g. during toplevel resize).

13 years agobroadway: Ensure window sizes are synched right
Alexander Larsson [Thu, 28 Mar 2013 18:00:23 +0000 (19:00 +0100)]
broadway: Ensure window sizes are synched right

We don't want to update the window size on configure event, only
the position, as the size is client side controlled. We were
updating to an old size during resizes which causes us to send
surfaces of the wrong size to the daemon.

13 years agobroadway: Clean up moveresize code
Alexander Larsson [Thu, 28 Mar 2013 16:38:44 +0000 (17:38 +0100)]
broadway: Clean up moveresize code

13 years agobroadway: Remove leftover cruft that caused resize to fail
Alexander Larsson [Thu, 28 Mar 2013 16:38:01 +0000 (17:38 +0100)]
broadway: Remove leftover cruft that caused resize to fail

getTransientToplevel() doesn't exist anymore so this broke.

13 years agobroadway: Fix mouse position typo
Alexander Larsson [Thu, 28 Mar 2013 15:27:49 +0000 (16:27 +0100)]
broadway: Fix mouse position typo

13 years agoreftests: Add a reftest for color resolving errors
Benjamin Otte [Thu, 28 Mar 2013 22:17:32 +0000 (23:17 +0100)]
reftests: Add a reftest for color resolving errors

This checks the fixes in the last two commits.

13 years agocssvalue: At cycle detection to color resolving
Benjamin Otte [Thu, 28 Mar 2013 22:13:43 +0000 (23:13 +0100)]
cssvalue: At cycle detection to color resolving

The following CSS would infloop:

@define-color self @self

as it would infinitely lookup the color named "self" and try to resolve
it. This patch adds detection of such cycles to the resolve function by
keeping a list of currently resolving colors in the cycle_list variable.

13 years agocssvalue: Handle named colors not resolving
Benjamin Otte [Thu, 28 Mar 2013 21:52:30 +0000 (22:52 +0100)]
cssvalue: Handle named colors not resolving

If a named color references a nonexistant named color, we didn't catch
that error and ended up crashing on a NULL-dereference.
This crashed Boxes, because its CSS referenced values from the theme
that didn't exist in any theme.

13 years agomenuitem: Remove old themeing enhancement
Benjamin Otte [Tue, 26 Mar 2013 12:27:45 +0000 (13:27 +0100)]
menuitem: Remove old themeing enhancement

CSS doesn't allow themeing parents based on the state of their children,
so this code just causes extra redraws today.

13 years agomenushell: Remove unused header
Benjamin Otte [Tue, 26 Mar 2013 12:15:20 +0000 (13:15 +0100)]
menushell: Remove unused header

13 years agomenu: Move enum declaration to correct header
Benjamin Otte [Tue, 26 Mar 2013 12:14:05 +0000 (13:14 +0100)]
menu: Move enum declaration to correct header

That enum is used in GtkMenuShellClass, so put it there.

13 years agotearoffmenuitem: Use public API, don't poke GtkMenuPrivate
Benjamin Otte [Tue, 26 Mar 2013 12:02:59 +0000 (13:02 +0100)]
tearoffmenuitem: Use public API, don't poke GtkMenuPrivate

13 years agowayland: Implement animatable cursors
Chris Cummins [Thu, 28 Mar 2013 12:08:43 +0000 (12:08 +0000)]
wayland: Implement animatable cursors

Cursor animations are handled on a per-device basis, with
GdkWaylandDevice updating the pointer surface for each frame.

https://bugzilla.gnome.org/show_bug.cgi?id=696429

13 years agowayland: Do not free wayland cursor buffers
Chris Cummins [Tue, 26 Mar 2013 14:44:24 +0000 (14:44 +0000)]
wayland: Do not free wayland cursor buffers

The cursor buffer is only non-null when a cursor is created from pixbuf,
so it is not necessary to keep track of whether to free this buffer on
finalize.

13 years agowayland: Differentiate wayland and pixbuf cursors
Chris Cummins [Tue, 26 Mar 2013 14:19:37 +0000 (14:19 +0000)]
wayland: Differentiate wayland and pixbuf cursors

By keeping a pointer to the wl_cursor struct in GdkWaylandCursor, it is
no longer necessary to duplicate cursor data (width, height, hotspots,
etc.) between wl_cursor and GdkWaylandCursor.