Olivier Fourdan [Tue, 17 May 2016 13:02:01 +0000 (15:02 +0200)]
wayland: Add API to retrieve the Wayland output
In Wayland, surfaces get an enter/leave notification each time they
enter or leave an output.
Add an API to GdkWaylandWindow to retrieve the output the window has
last entered.
https://bugzilla.gnome.org/show_bug.cgi?id=766566
Olivier Fourdan [Tue, 17 May 2016 11:09:23 +0000 (13:09 +0200)]
display: Add vfunc for get_monitor_at_window
Some backends (namely Wayland) do not support global coordinates so
using the window position to determine the monitor will always fail on
such backends.
In such cases, the backend itself might be better suited to identify
the monitor a given window resides on.
Add a vfunc get_monitor_at_window() to the display class so that we can
use the backend to retrieve the monitor, if the backend implements it.
https://bugzilla.gnome.org/show_bug.cgi?id=766566
Matthias Clasen [Wed, 18 May 2016 16:59:39 +0000 (12:59 -0400)]
scale: Bring back marks-related style classes
It turns out that it is too hard (and in some cases, impossible)
to get this information from node positioning, so bring back the
.scale-has-marks-above/below style classes on the main node.
This should allow us to fix the 'pointy sliders'.
https://bugzilla.gnome.org/show_bug.cgi?id=766440
Matthias Clasen [Wed, 18 May 2016 11:48:26 +0000 (07:48 -0400)]
Revert "scrolledwindow: Better size requisition for GTK_SCROLL_NATURAL children"
This reverts commit
096bea3f0ea8e80fe4c9a2a1a860f32c8a94cc33.
This was accidentally pushed.
Carlos Garnacho [Tue, 17 May 2016 17:27:00 +0000 (19:27 +0200)]
x11/dnd: Ignore grab broken events from other devices than our DnD device.
The GdkDragContext should only listen to GDK_GRAB_BROKEN events sent to
its own pointer device. It turns out that the passive key grabs mistake
GDK into sending a GdkEventGrabBroken on the master keyboard, which the
DnD machinery mistakes as a signal to cancel the operation.
https://bugzilla.gnome.org/show_bug.cgi?id=766442
Carlos Garnacho [Thu, 12 May 2016 11:34:42 +0000 (13:34 +0200)]
wayland: Avoid spurious crossing events from master touch device
Only generate crossing events on wl_touch.down for the virtual master
device used for touch events, and only whenever this virtual device
actually moves across surfaces. This behavior resembles better what is
expected in X11, where the pointer is warped to the touch position
on XITouchBegin.
This avoids the double emission of leave events when the pointer
emulating touch is lifted, that crossing event will be instead
generated when/if the focus surface changes.
https://bugzilla.gnome.org/show_bug.cgi?id=766314
Matthew Waters [Fri, 18 Mar 2016 06:17:23 +0000 (17:17 +1100)]
gdk/wayland: use the multi-thread safe wayland API
This is required for proper integration with any other library/application that
may perform wayland API calls and poll() the wayland fd from multiple threads.
Using wl_display_dispatch{_queue}() is thread-safe if not mixed with custom
poll() usage, which GSource/GMainContext does.
Essentially, the problem is that multiple threads polling and reading
the same fd is extremely racy. Use the wayland provided API for allowing
concurrent access to the wayland display fd.
See the wayland man pages for wl_display_prepare_read(),
wl_display_cancel_read() and wl_display_read_events() for more details.
https://bugzilla.gnome.org/show_bug.cgi?id=763852
Debarshi Ray [Tue, 17 May 2016 14:28:20 +0000 (16:28 +0200)]
scrolledwindow: Better size requisition for GTK_SCROLL_NATURAL children
GtkScrolledWindow leans towards using the minimum size of its child
widget, unless the scrollbar policy is GTK_POLICY_NEVER. This is
probably fine for most GtkScrollable implementations out there.
Especially when using GTK_SCROLL_MINIMUM, which is the default for all
implementations inside gtk+.
However, this is not good for GTK_SCROLL_NATURAL children. eg.,
VteTerminal's minimum size is 1x1 and natural size is the number of
visible rows and columns requested by the user. We really want to use
the natural size unless the user has resized the window to change that.
https://bugzilla.gnome.org/show_bug.cgi?id=766569
Matthias Clasen [Tue, 17 May 2016 18:16:39 +0000 (14:16 -0400)]
Update cursor docs
Include new images for context-menu and for no-drop.
Matthias Clasen [Tue, 17 May 2016 17:40:52 +0000 (13:40 -0400)]
gtk-demo: Update no-drop cursor image
Adwaita is using a different cursor for this now.
Matthias Clasen [Tue, 17 May 2016 17:39:18 +0000 (13:39 -0400)]
gtk3-demo: Add context-menu cursor
The Adwaita cursor theme covers this one now.
Debarshi Ray [Mon, 16 May 2016 19:50:05 +0000 (21:50 +0200)]
scrolledwindow: Remove redundant use of MAX
This code tries to add the minimum content size, if one is set, to the
GtkScrolledWindow's size requisition. This is obvious from the check
for non-negative values of min-content-height and min-content-width.
Using MAX needlessly makes the code harder to read by implying that
there is more to it when there actually isn't.
Fall out from
0d9ebb501df60cf1803858efcd1c79542588abd8
https://bugzilla.gnome.org/show_bug.cgi?id=766569
Baurzhan Muftakhidinov [Tue, 17 May 2016 10:18:52 +0000 (10:18 +0000)]
Updated Kazakh translation
(cherry picked from commit
d1585a7a60d760a0c98934f970c4403023a399c3)
Debarshi Ray [Mon, 16 May 2016 19:13:14 +0000 (21:13 +0200)]
scrolledwindow: Fix typo in get_preferred_height calculation
When we are beginning to calculate the height, if the vscrollbar_policy
is not GTK_POLICY_NEVER, and there is no min-content-height, then we
need some small non-zero value to get started. The idea is to always
ask for at least enough to fit the horizontal scrollbar.
Simply put, this should be the mirror image of the corresponding width
calculation code.
Those who got used to the buggy behaviour might notice that their
GtkScrolledWindows are not as tall as they used to be.
Fall out from
55196a705f00564a44647bfc97981db0a783369a
https://bugzilla.gnome.org/show_bug.cgi?id=766530
Matt Watson [Fri, 13 May 2016 07:22:20 +0000 (00:22 -0700)]
stack: always update window position at transition start
Sometimes on wayland we're seeing a frame draw before the first tick
callback, where the window position has not yet been updated.
With this change, the window is properly positioned as soon as
gtk_stack_start_transition returns.
https://bugzilla.gnome.org/show_bug.cgi?id=766405
Javier Jardón [Sat, 9 Jan 2016 18:35:34 +0000 (18:35 +0000)]
Use upstream gettext
https://bugzilla.gnome.org/show_bug.cgi?id=373745
Daniel Mustieles [Mon, 16 May 2016 15:13:24 +0000 (17:13 +0200)]
Updated Spanish translation
Tiago Santos [Sun, 15 May 2016 21:40:45 +0000 (21:40 +0000)]
Updated Portuguese translation
Tiago Santos [Sun, 15 May 2016 21:22:46 +0000 (21:22 +0000)]
Updated Portuguese translation
Matthias Clasen [Sun, 15 May 2016 18:18:09 +0000 (14:18 -0400)]
scale: Avoid a crash
gtk_scale_get_layout was not careful enough, leading to a crash
when opening the inspector on a scale.
Alban Browaeys [Sat, 14 May 2016 21:57:28 +0000 (23:57 +0200)]
widget: fix GtkLabelAccessible NULL links.
Fix testsuite/a11y/about.ui GtkAboutDialog :
"CRITICAL **: atk_hyperlink_get_start_index: assertion 'ATK_IS_HYPERLINK (link)' failed"
That is set widget->priv->accessible as soon as accessible object is generated.
When accessible object is created accessible->priv->widget is set,
if widget->priv->accessible is not , then _gtk_label_accessible_update_links
exits early, thus without creating the links on the accessible side.
(This as it checks for the widget to have the accessible set before proceeding).
https://bugzilla.gnome.org/show_bug.cgi?id=766458
Benjamin Otte [Sat, 14 May 2016 16:35:27 +0000 (18:35 +0200)]
widget: No longer postpone style-updated on unrealized widgets
GTK used to not emit GtkWidget::style-updated on widgets that weren't
realized. This sped up construction of complex widgetry in the early
days of GTK3 where we instantly invalidated on every change.
We don't do that anymore, so in theory (and in my limited testing with
widget-factory) this shouldn't be a prolem anymore.
What is a problem though is that postponing style-updated leads to 2
problems:
(1) Unrealized widgets will not emit style-updated which may cause them
to not properly update their state and return wrong values from
get_preferred_width/height() etc
(2) Emitting style-updated during realize can happen too late.
When a widget is not made child-visible by its parent (common
examples: notebook, paned) it will also not be realized when the
parent is initially shown. However, when they get realized later
(after a resize of the parent), they will emit style-updated (and
potentially queue a resize) during size-allocate.
https://bugzilla.gnome.org/show_bug.cgi?id=765700
Benjamin Otte [Sat, 14 May 2016 15:13:52 +0000 (17:13 +0200)]
scrolledwindow: destroy children in destroy()
If we don't do that, testsuite/gtk/templates starts failing.
Timm Bäder [Sat, 30 Apr 2016 16:46:22 +0000 (18:46 +0200)]
gdkwindow: Use GList less
Timm Bäder [Sat, 30 Apr 2016 10:10:14 +0000 (12:10 +0200)]
query-settings: Indicate deprecated settings
Timm Bäder [Fri, 26 Feb 2016 20:58:36 +0000 (21:58 +0100)]
stack: Don't underallocate the last_child
Timm Bäder [Fri, 26 Feb 2016 20:58:22 +0000 (21:58 +0100)]
stack: Remove some code duplication
Timm Bäder [Mon, 18 Apr 2016 18:48:53 +0000 (20:48 +0200)]
Remove useless casts from gdk_event_triggers_context_menu calls
Timm Bäder [Sat, 16 Apr 2016 11:57:32 +0000 (13:57 +0200)]
query-settings: Don't compare int to uint
Timm Bäder [Sat, 16 Apr 2016 11:57:08 +0000 (13:57 +0200)]
listbox: Don't compare int to uint
Timm Bäder [Wed, 11 May 2016 13:33:18 +0000 (15:33 +0200)]
colorbutton: Stop using state-changed
Timm Bäder [Tue, 10 May 2016 17:16:19 +0000 (19:16 +0200)]
toolbar: Limit item position to number of contained elements
Greater values don't make much sense.
https://bugzilla.gnome.org/show_bug.cgi?id=682080
Matthias Clasen [Thu, 12 May 2016 15:37:13 +0000 (11:37 -0400)]
wayland: Avoid unitialized memory reads
I didn't pay attention when I replaced g_new0 with g_newa. Oops.
Lapo Calamandrei [Thu, 12 May 2016 09:50:13 +0000 (11:50 +0200)]
Adwaita: brighter pushed button background
providing more contrast with the text color.
Lapo Calamandrei [Thu, 12 May 2016 09:45:34 +0000 (11:45 +0200)]
Adwaita: bold size label on font button
looks more consistent with other special buttons with icons.
Lapo Calamandrei [Thu, 12 May 2016 09:40:28 +0000 (11:40 +0200)]
Adwaita: hide separator on button.font and button.file
we used to do that, got missed at some point during 3.20 cycle.
Matthias Clasen [Wed, 11 May 2016 17:25:15 +0000 (13:25 -0400)]
Docs: Clarify life cycle issue
gtk_gesture_get_last_event() wasn't very clear about how long
it is safe to use the returned pointer.
Matthias Clasen [Wed, 11 May 2016 17:24:04 +0000 (13:24 -0400)]
DND: Avoid a use-after-free
Cancelling the gesture causes the last_event pointer to become
invalid. Make a copy of the event so we can keep using it
regardless of the gesture state.
Matthias Clasen [Wed, 11 May 2016 01:47:20 +0000 (21:47 -0400)]
places sidebar: Avoid markup confusion in tooltips
At one point, the sidebar was using gtk_treeview_set_tooltip_column,
which expects tooltips to be markup. With the listbox-based sidebar,
we don't do that anymore. So don't escape the tooltip text.
https://bugzilla.gnome.org/show_bug.cgi?id=766175
Matthias Clasen [Wed, 11 May 2016 01:29:10 +0000 (21:29 -0400)]
x11: Trap possible X error
XIGetClientPointer can generate X errors (e.g. when the X server
does not support XI2. Trap them and carry on.
https://bugzilla.gnome.org/show_bug.cgi?id=766233
Matthias Clasen [Tue, 10 May 2016 19:41:03 +0000 (15:41 -0400)]
Fix scale length
The vertical scales should really be of the same length, so
add an empty value to be drawn by the other scale.
Matthias Clasen [Tue, 10 May 2016 19:31:29 +0000 (15:31 -0400)]
Forgotten rename
Timm Bäder [Tue, 10 May 2016 16:33:37 +0000 (18:33 +0200)]
Scale: Set the correct value alignment
We're measuring both the width of the minimal and the maximal value for
the value gadget, but only give the value the minimal width for its
current value, resultnig in an always left-aligned value.
Fix this by assigning the width of the value gadget to the value layout
and letting pango align the text inside the layout.
https://bugzilla.gnome.org/show_bug.cgi?id=766120
Timm Bäder [Mon, 9 May 2016 16:30:14 +0000 (18:30 +0200)]
widget-factory: Let one of the scales draw a value
https://bugzilla.gnome.org/show_bug.cgi?id=766120
Matthias Clasen [Tue, 10 May 2016 19:22:35 +0000 (15:22 -0400)]
Document new api
Matthias Clasen [Tue, 10 May 2016 19:16:45 +0000 (15:16 -0400)]
Make gdk_event_get_pointer_emulated public
There is not strong reason to keep the getter private.
At the same time, strip _-prefixes from a few other GdkEvent
APIs. Update all callers.
Olivier Fourdan [Mon, 9 May 2016 06:56:56 +0000 (08:56 +0200)]
gtkmenu: ignore left/right scroll events
Menus are placed vertically by definition, it does not make much sense
to support horizontal axis for scrolling.
Use GDK_EVENT_STOP/GDK_EVENT_PROPAGATE instead of TRUE/FALSE and add a
default case to return GDK_EVENT_PROPAGATE for unhandled events.
https://bugzilla.gnome.org/show_bug.cgi?id=765939
Olivier Fourdan [Tue, 3 May 2016 14:44:14 +0000 (16:44 +0200)]
gtkmenu: ignore emulated scroll events
On X11, we get both smooth and emulated scroll events, whereas other
backends such as wayland will give smooth events only with touchpad
scrolling.
Discard emulated scroll events so that we get consistent behaviours
between backends.
Allow for both horizontal and vertical smooth events for scrolling so
that horizontal scrolling still works without emulated scroll events as
well, again for consistency between gdk backends.
https://bugzilla.gnome.org/show_bug.cgi?id=765939
Olivier Fourdan [Tue, 3 May 2016 14:15:07 +0000 (16:15 +0200)]
gdkevent: make _gdk_event_get_pointer_emulated() private
And not just internal to gdk, so we can use it in gtk as well, to
differentiate emulated scroll events from others.
https://bugzilla.gnome.org/show_bug.cgi?id=765939
Timm Bäder [Tue, 10 May 2016 10:50:42 +0000 (12:50 +0200)]
scale: Remove double nullable annotation
Being both (allow-none) and (nullable) at the same time is a bit much.
Was from
591e7f5ef8538982e227b2c2cefc536a33cafa6c.
Timm Bäder [Tue, 10 May 2016 10:31:56 +0000 (12:31 +0200)]
widget: Add Since annotation to gtk_widget_queue_allocate
Chun-wei Fan [Mon, 9 May 2016 08:37:39 +0000 (16:37 +0800)]
GTK: Fix build for pre-C99 compilers
Some compilers we support, such as pre-2013 Visual Studio, does not support
for INIFINITY, log2() and exp2(), so check for exp2() and log2() during
configure, and use fallbacks for them and INIFINTY if they are not found.
https://bugzilla.gnome.org/show_bug.cgi?id=766207
Chun-wei Fan [Tue, 10 May 2016 05:18:23 +0000 (13:18 +0800)]
Visual Studio builds: Refine "installation" process
Split out the part where we generate/update the caches for the GSchemas
and the icons, so that it is easier to ensure that things continue to
function correctly when we have GlibEtcInstallRoot != CopyDir.
Benjamin Otte [Mon, 9 May 2016 22:59:19 +0000 (00:59 +0200)]
scrolledwindow: Remove child before destroying self
Children tend to call back into the scrolled window while being removed
and that doesn't work too well if the scrolled window is destroyed
already as Christian Hergert found out.
Benjamin Otte [Mon, 9 May 2016 22:58:50 +0000 (00:58 +0200)]
cssimage: Make it possible to shrink builtin check/optionmarks
... below 7px of size.
Cédric Valmary [Mon, 9 May 2016 20:49:18 +0000 (20:49 +0000)]
Updated Occitan translation
Emmanuele Bassi [Mon, 9 May 2016 11:45:34 +0000 (12:45 +0100)]
docs: Update examples of matching key bindings
The selector for matching GtkEntry has changed to `entry` after 3.20.
https://bugzilla.gnome.org/show_bug.cgi?id=766166
Chun-wei Fan [Mon, 9 May 2016 07:42:37 +0000 (15:42 +0800)]
gdk/Makefile.am: Dist gdkmonitorprivate.h
This file needs to be distributed with the tarball as well...
Chun-wei Fan [Mon, 9 May 2016 07:35:26 +0000 (15:35 +0800)]
Visual Studio builds: Clean up projects a bit
Consolidate the .lib's that we need to link to for GDK and GTK+ into the
property sheets, and also link to setupapi.lib as that is needed by
the port to the new GdkMonitor API.
Chun-wei Fan [Mon, 9 May 2016 05:14:01 +0000 (13:14 +0800)]
gdk/win32/gdkmonitor-win32.c: Fix build on non-GCC
WINBOOL is MinGW-specific, so change it to BOOL, which is universally
available.
Also, Visua Studio is more picky on where __stdcall (WINAPI) is placed, so
fix that to be in-sync with what is done in the other sources.
Georges Basile Stavracas Neto [Sun, 8 May 2016 20:36:26 +0000 (17:36 -0300)]
placesview: update guiding list of protocols
The previous list of protocols in the guidance popover
was suboptimal and needed adjustments.
This commit updates this list to match the latest mockups [1].
[1] https://raw.githubusercontent.com/gnome-design-team/gnome-mockups/
4847c79ea1aaf75046b3ec5bb1b167ac604d86ed/nautilus/nautilus-next/connect-to-server-wire.png
https://bugzilla.gnome.org/show_bug.cgi?id=756570
Timm Bäder [Sun, 8 May 2016 19:33:53 +0000 (21:33 +0200)]
dialogs: remove 2px border width again
Aurimas Černius [Sun, 8 May 2016 15:30:49 +0000 (18:30 +0300)]
Updated Lithuanian translation
Rafael Fontenelle [Sun, 8 May 2016 14:25:31 +0000 (14:25 +0000)]
Updated Brazilian Portuguese translation
(cherry picked from commit
207b793316a90faff00d5ea6fbe483d5a184d066)
Timm Bäder [Sun, 8 May 2016 13:52:56 +0000 (15:52 +0200)]
tooltip: Include gtkwidgetprivate.h
That's where gtk_widget_query_tooltip is defined.
Timm Bäder [Sun, 8 May 2016 13:34:28 +0000 (15:34 +0200)]
gdk: Fix parameter name mismatch
Timm Bäder [Mon, 2 May 2016 06:59:40 +0000 (08:59 +0200)]
infobar: Fix typo
Style class names are prefixed with a '.'
Matthias Clasen [Sat, 7 May 2016 20:22:13 +0000 (16:22 -0400)]
gtk-demo: Fix revealer demo
This only used by luck before. We are changing a property from the
::notify handler for that property. Now that GtkRevealer is notifying
the property when it stops animations on unmap, we end up in a life
lock situation where we never make it out of the notify queue.
Fix this by not restarting the animation if the widget is unmapped.
Ray Strode [Sat, 7 May 2016 02:29:31 +0000 (22:29 -0400)]
widget: add missing detail to ::query-tooltip emission
This was an oversight in commit
dfb368e29bb58b4313b578f0ce75cfc8ead9a1.
Matthias Clasen [Sat, 7 May 2016 00:51:25 +0000 (20:51 -0400)]
tooltip: Add a missing *
This was an oversight in the previous commit.
Matthias Clasen [Sat, 7 May 2016 00:50:38 +0000 (20:50 -0400)]
gtk3-demo: Add a close button to the markup window
Matthias Clasen [Sat, 7 May 2016 00:47:49 +0000 (20:47 -0400)]
gtk-demo: Don't duplicate text in markup example
I forgot to clear the buffer before inserting the new markup. Oops.
Matthieu Gautier [Wed, 4 May 2016 10:42:40 +0000 (12:42 +0200)]
revealer: emit notify::child-revealed when animation stops
Depending of float rounding during target calculation, the size of the
GtkRevealer can be set to zero will the animation is not finished.
If the GtkRevealer is in a GtkPaned, it will be hidden and so the animation
will be stopped before it is finished.
In this case, force the emission of the child-revealed signal to let
client code know the animation is finished.
https://bugzilla.gnome.org/show_bug.cgi?id=765973
Matthias Clasen [Fri, 6 May 2016 17:45:46 +0000 (13:45 -0400)]
Simplify our templates
This is applying the new property name simplification.
Matthias Clasen [Fri, 6 May 2016 17:45:09 +0000 (13:45 -0400)]
Clean up builder parser data after parsing
No need to have these linger around in qdata.
Matthias Clasen [Fri, 6 May 2016 17:42:30 +0000 (13:42 -0400)]
builder tool: Canonicalize property names
When simplifying, convert property names to canonical form,
so we don't waste time while looking them up later.
Cédric Valmary [Fri, 6 May 2016 18:31:42 +0000 (18:31 +0000)]
Updated Occitan translation
Lapo Calamandrei [Fri, 6 May 2016 17:34:12 +0000 (19:34 +0200)]
Adwaita: darken treeview exapanders
Lapo Calamandrei [Fri, 6 May 2016 17:01:19 +0000 (19:01 +0200)]
Adwaita: cosmetic change
Lapo Calamandrei [Fri, 6 May 2016 16:50:37 +0000 (18:50 +0200)]
Adwaita: relocate a selector
relocate treview acceleditor > label in the treeview section and
add a comment for a testcase.
Matthias Clasen [Fri, 6 May 2016 14:12:14 +0000 (10:12 -0400)]
Avoid emitting ::style-set by name
GtkStyle is deprecated, but we still emit ::style-set quite
a bit, so lets at least not be slow while doing it.
Matthias Clasen [Fri, 6 May 2016 13:54:24 +0000 (09:54 -0400)]
Don't emit ::query-tooltip by name
This signal is emitted quite a bit, and we can easily avoid it.
Matthias Clasen [Fri, 6 May 2016 13:38:18 +0000 (09:38 -0400)]
Stop storing has-tooltip in qdata
This is queried quite a bit, and we have room for an extra
bit in GtkWidgetPrivate.
Matthias Clasen [Fri, 6 May 2016 11:54:22 +0000 (07:54 -0400)]
Don't use qdata for getting the default css style
There's no point in using qdata for looking this up, since
we only ever set it on the default screen anyway.
Lapo Calamandrei [Fri, 6 May 2016 13:36:26 +0000 (15:36 +0200)]
Adwaita: tidy treeview style properties...
...removing a double definition in the process.
Matthias Clasen [Fri, 6 May 2016 11:08:56 +0000 (07:08 -0400)]
settings: Fix font names *again*
This is the last time, I hope.
Matthias Clasen [Fri, 6 May 2016 10:42:43 +0000 (06:42 -0400)]
widget: Store accessible in GtkWidgetPrivate
Every widget may have one of these, and they are accessed somewhat
frequently.
Matthias Clasen [Fri, 6 May 2016 04:46:17 +0000 (00:46 -0400)]
Ensure we initialize a quark first
The style function here can also be called before the corresponding
class_init had a chance to create the quark. Fix this.
Matthias Clasen [Fri, 6 May 2016 04:45:05 +0000 (00:45 -0400)]
tooltip: Ensure our quark is not 0
Many of the tooltip functions can be called before a tooltip
object is created, and thus initializing the quark in class_init
is not good enough.
Matthias Clasen [Fri, 6 May 2016 04:21:08 +0000 (00:21 -0400)]
Use a cheaper way to store settings per display
This was another very frequent use of qdata. Since we typically
have only one or two display objects, storing the display-settings
association in a simple array is faster than using object data
or a hash table.
Matthias Clasen [Thu, 5 May 2016 22:51:30 +0000 (18:51 -0400)]
Use a cheaper way to store debug flags
This was the most frequent use of qdata *by far*.
Matthias Clasen [Thu, 5 May 2016 23:21:27 +0000 (19:21 -0400)]
Fix indentation mishap
Matthias Clasen [Thu, 5 May 2016 23:17:31 +0000 (19:17 -0400)]
Fix a typo
Christian Hergert [Fri, 6 May 2016 08:31:41 +0000 (11:31 +0300)]
wayland: use g_signal_handler_disconnect()
Use of g_signal_handlers_disconnect_by_func() needs to do more work than
necessary to find all the matching handlers. Instead, just hold on to the
signal identifier and remove it directly so we hit the fast path.
Not terribly ground breaking in terms of performance gains, but its done
enough to be worthwhile.
https://bugzilla.gnome.org/show_bug.cgi?id=766049
Matthias Clasen [Thu, 5 May 2016 19:37:21 +0000 (15:37 -0400)]
Update color name css style test
Verify rebeccapurple too.
Matthias Clasen [Thu, 5 May 2016 19:17:56 +0000 (15:17 -0400)]
scale: Fix mark positioning at the left/top edge
We were failing to take the widget allocation.x/y into account
when deciding whether we need to push in the mark.
https://bugzilla.gnome.org/show_bug.cgi?id=765922
Matthias Clasen [Thu, 5 May 2016 18:13:03 +0000 (14:13 -0400)]
Generalize the previous commit
Update all xsettings when we get a screen, to prevent similar
problems from occurring in the future.
Matthias Clasen [Thu, 5 May 2016 17:30:55 +0000 (13:30 -0400)]
settings: Avoid a crash
Christian Hergert reported seeing webkit crashes with recent
GTK+. The stacktrace points at the CSS machinery calling into
GtkSettings to get the font name, and then getting surprised
by a property notification that triggers style validation.
To avoid this, query the font name xsetting right away when
we get set a screen.
Carlos Garnacho [Thu, 5 May 2016 17:00:28 +0000 (19:00 +0200)]
x11: Rely on passive keyboard grabs with XI2 for DnD
The active keyboard grab can be spared then. This way the passive
key grabs allow other key combinations (eg. alt-tab) that are not
mandatory to grab here.
Lapo Calamandrei [Thu, 5 May 2016 16:36:03 +0000 (18:36 +0200)]
Adwaita: more flexible headerbar_fill mixin
add an additional background layer as a parameter.