gtk4.git
12 years agoBump version
Matthias Clasen [Tue, 18 Mar 2014 00:21:28 +0000 (20:21 -0400)]
Bump version

12 years agoGtkPopover: Fix two doc comments
Matthias Clasen [Mon, 17 Mar 2014 23:13:53 +0000 (19:13 -0400)]
GtkPopover: Fix two doc comments

12 years agox11: Allow unmaximizing from tiled state too
Matthias Clasen [Mon, 17 Mar 2014 22:18:53 +0000 (18:18 -0400)]
x11: Allow unmaximizing from tiled state too

This lets us unmaximize a half-tiled csd window on a touch system.

12 years agoUpdate a11y test output
Matthias Clasen [Mon, 17 Mar 2014 22:14:53 +0000 (18:14 -0400)]
Update a11y test output

12 years agowayland: Remove cruft in set_keep_above / set_keep_below
Jasper St. Pierre [Mon, 17 Mar 2014 19:42:43 +0000 (15:42 -0400)]
wayland: Remove cruft in set_keep_above / set_keep_below

It's been decided: these will most likely never be supported on
Wayland, so remove the "stub" implementation of them.

12 years agowayland: Fix get_frame_extents
Jasper St. Pierre [Mon, 17 Mar 2014 19:41:08 +0000 (15:41 -0400)]
wayland: Fix get_frame_extents

We need to traverse up the hierarchy for windows here. Just use
our existing helper method for this.

12 years agoImplement get_root_origin generically for all backends
Jasper St. Pierre [Mon, 17 Mar 2014 19:39:39 +0000 (15:39 -0400)]
Implement get_root_origin generically for all backends

It seems that some backends implemented get_root_origin wrong
and returned the client window coordinates, not the frame window
coordinates. Since it's possible to implement generically for all
windows, let's do that instead of having a separate impl vfunc.

12 years agoUpdates
Matthias Clasen [Mon, 17 Mar 2014 19:45:21 +0000 (15:45 -0400)]
Updates

12 years agowayland: Fix "fake root" coords
Jasper St. Pierre [Mon, 17 Mar 2014 19:27:13 +0000 (15:27 -0400)]
wayland: Fix "fake root" coords

We were incorrectly summing up our own window over and over
rather than the coordinates of the parent windows.

12 years agowayland: Add a giant doc comment explaining "fake root" coordinate space
Jasper St. Pierre [Mon, 17 Mar 2014 19:25:23 +0000 (15:25 -0400)]
wayland: Add a giant doc comment explaining "fake root" coordinate space

12 years agowayland: Properly apply the fake root offset to event coordinates
Jasper St. Pierre [Mon, 17 Mar 2014 19:11:13 +0000 (15:11 -0400)]
wayland: Properly apply the fake root offset to event coordinates

GdkEvent's x_root and y_root values should be in the same "fake root
window" coordinate space as gdk_window_get_root_coords.

12 years agowayland: Fill in x_root / y_root of events properly
Jasper St. Pierre [Mon, 17 Mar 2014 18:58:37 +0000 (14:58 -0400)]
wayland: Fill in x_root / y_root of events properly

Lots of code, including dragging code in GtkWindow, use these
fields. Setting them to 0 causes lots of strange and weird bugs.

Use the same "hack" from query_device_state of just using
win_x / win_y for now. We'll convert this to the proper fake root
coordinate system used by get_root_coords in the next commit.

12 years agoUpdate documentation about link styling
Matthias Clasen [Mon, 17 Mar 2014 19:00:45 +0000 (15:00 -0400)]
Update documentation about link styling

The ::link-color and ::visited-link-color style properties
are ignored now. Document that.

12 years agoprinting: Fail nicely when /tmp is not writable
Marek Kasik [Fri, 14 Mar 2014 14:36:10 +0000 (15:36 +0100)]
printing: Fail nicely when /tmp is not writable

Don't crash when /tmp is not writable when printing to file.
Show that getting of printer details failed for CUPS printers.

https://bugzilla.gnome.org/show_bug.cgi?id=693200

12 years agotextview: Mind the lateral windows when setting handles/popovers positions
Carlos Garnacho [Mon, 17 Mar 2014 16:42:07 +0000 (17:42 +0100)]
textview: Mind the lateral windows when setting handles/popovers positions

If a textview had lateral windows that might displace the text window, the
handles and popovers would appear displaced. Those lateral windows aren't
affected by RTL/LTR settings, so just checking for left/top is ok here.

12 years agox11: Implement "drag to top to maximize" gesture on emulated window dragging
Carlos Garnacho [Thu, 13 Mar 2014 20:12:55 +0000 (21:12 +0100)]
x11: Implement "drag to top to maximize" gesture on emulated window dragging

And the counterpart to unmaximize when dragging a maximized window, if
touch devices aren't going to use EWMH moveresize, having this one at least
makes things feel a bit less awkward.

https://bugzilla.gnome.org/show_bug.cgi?id=709914

12 years agox11: Fallback to emulated window dragging for touch devices
Carlos Garnacho [Thu, 13 Mar 2014 19:19:26 +0000 (20:19 +0100)]
x11: Fallback to emulated window dragging for touch devices

Sadly, EWMH moveresize mechanism can't work with touch devices for two
reasons:

1) As a mutter implementation detail, the device is queried in order
to check whether the dragging button is still pressed. Touch devices
won't report the button 1 being pressed through pointer emulation.
2) Even bypassing that check, on X11 touch events are selected prior
to sequences being started, either through XISelectEvents or
XIGrabTouchBegin, no late registering through active grabs is allowed,
as WMs do on reaction to EWMH moveresize messages.

So for the time being, make touch devices fallback on emulated window
dragging, which at least allows for moving windows.

https://bugzilla.gnome.org/show_bug.cgi?id=709914

12 years agodialog: Avoid possible use-after-free
Bastien Nocera [Mon, 17 Mar 2014 08:27:36 +0000 (09:27 +0100)]
dialog: Avoid possible use-after-free

When the dialogue's titlebar was replaced, we were still trying to
update the label we constructed but that was now destroyed.

https://bugzilla.gnome.org/show_bug.cgi?id=726492

12 years agoFinnish translation update by Jiri Grönroos
Ville-Pekka Vainio [Sun, 16 Mar 2014 19:46:47 +0000 (21:46 +0200)]
Finnish translation update by Jiri Grönroos

12 years agoUpdated French translation
Luis Menina [Sun, 16 Mar 2014 16:06:52 +0000 (16:06 +0000)]
Updated French translation

12 years agoUpdated French translation
Claude Paroz [Sun, 16 Mar 2014 12:51:19 +0000 (12:51 +0000)]
Updated French translation

12 years agoUpdated French translation
Claude Paroz [Sun, 16 Mar 2014 11:05:02 +0000 (11:05 +0000)]
Updated French translation

12 years agoTest size constraints
Matthias Clasen [Fri, 14 Mar 2014 17:45:47 +0000 (13:45 -0400)]
Test size constraints

This adds some min/max size hints to testimage.

12 years agoShow dialog titles
Matthias Clasen [Fri, 14 Mar 2014 10:28:49 +0000 (06:28 -0400)]
Show dialog titles

We want to present a clean, rounded top when there is nothing
else to show, but many dialogs in applications rely on showing
information in their title, so add a label and show the title
when it is not empty.

12 years ago_gtk_widget_get_device_window: Fix for keyboard devices
Owen W. Taylor [Thu, 13 Mar 2014 14:30:05 +0000 (10:30 -0400)]
_gtk_widget_get_device_window: Fix for keyboard devices

Avoid a warning by returning NULL for keyboard devices, which can
never have an active pointer for a widget.

12 years agoxdg-shell: Update to latest state change mechanism
Jasper St. Pierre [Sun, 9 Feb 2014 16:39:01 +0000 (11:39 -0500)]
xdg-shell: Update to latest state change mechanism

12 years agoUpdate a11y test output
Matthias Clasen [Thu, 13 Mar 2014 03:43:28 +0000 (23:43 -0400)]
Update a11y test output

These harmless changes are a consequence of
66fae0330c5cdb68f50699ff2645ff0cf112911e.

12 years agoAdd a Since annotation to gdk_device_get_last_event_window
Matthias Clasen [Thu, 13 Mar 2014 03:04:37 +0000 (23:04 -0400)]
Add a Since annotation to gdk_device_get_last_event_window

12 years agoUse GDK's current window tracking when synthesizing events in GTK+
Owen W. Taylor [Wed, 12 Mar 2014 16:39:03 +0000 (12:39 -0400)]
Use GDK's current window tracking when synthesizing events in GTK+

Add gdk_device_get_last_event_window(), and use to implement the window
tracking we need for synthesizing crossing events for sensitivity changes
and gtk grabs, rather than keeping the information in qdata and updating
it based when GTK+ gets events.

https://bugzilla.gnome.org/show_bug.cgi?id=726187

12 years agodocumentation: clarify gtk_icon_theme_add_builtin_icon() usage
Balint Reczey [Mon, 3 Mar 2014 15:07:47 +0000 (16:07 +0100)]
documentation: clarify gtk_icon_theme_add_builtin_icon() usage

https://bugzilla.gnome.org/show_bug.cgi?id=726135

12 years agoGtkTreeView: stop ignoring crossing events from grabs
Owen W. Taylor [Wed, 12 Mar 2014 23:27:42 +0000 (19:27 -0400)]
GtkTreeView: stop ignoring crossing events from grabs

Stop ignoring various crossing events from grabs:

Enter events with type GRAB/GTK_GRAB/GTK_UNGRAB/STATE_CHANGED:
  Ignoring these events was added as a workaround for synthesized
  events not having the right coordinates (see bug 555109) but
  now they do have the right coordinates. (see bug 704456)

Leave events with types types GTK_GRAB/GTK_UNGRAB:
  Ignoring these events was added because since we were ignoring
  the enter events as above, ignoring the leave events meant we
  could lose the prelighted row in a grab-triggered leave/enter
  pair. (See bug 653676. It's also now impossible to
  reproduce the leave events that were reported in that bug as causing
  problems.)

Leave events of type GRAB.
  Ignoring these events was added without a ChangeLog entry in 2001,
  possibly to keep the prelight from flashing when activating menus.
  But ignoring these events could lead to stuck prelighting, and we don't
  do it for any other widgets.

https://bugzilla.gnome.org/show_bug.cgi?id=726209

The bug this patch is fixing is that currently if you have a GtkPopover in
clicking off the popover to dismiss it on a GtkTreeView (which triggers
a synthetic enter event on the GtkTreeView) will leave the GtkTreeView
in a confused state until the user moves the mouse again.

12 years agoWork around gcc stupidity
Matthias Clasen [Thu, 13 Mar 2014 02:47:10 +0000 (22:47 -0400)]
Work around gcc stupidity

With -Wc++-compat, gcc warns about 'and' even in contexts where it
poses absolutely no danger.

https://bugzilla.gnome.org/show_bug.cgi?id=726136

12 years agotreeview: Don't popup a search dialog for keypresses that close it
Benjamin Otte [Thu, 13 Mar 2014 00:57:29 +0000 (01:57 +0100)]
treeview: Don't popup a search dialog for keypresses that close it

That doesn't make sense.
And it causes issues, because when holding down the tab key, we
show/hide a lot of windows and cause a lot of map/unmap events that
stall the event pipeline.

12 years agotreeview: Split out a function
Benjamin Otte [Thu, 13 Mar 2014 00:43:27 +0000 (01:43 +0100)]
treeview: Split out a function

This is in preparation for the next patch.

12 years agoa11y: Don't use signals
Benjamin Otte [Fri, 7 Mar 2014 20:45:18 +0000 (21:45 +0100)]
a11y: Don't use signals

Instead, call functions directly. Fixes the fact that the signals
weren't disconnected even when the accessible was destroyed.

https://bugzilla.gnome.org/show_bug.cgi?id=725733

12 years agogdk_cairo_surface_create_from_pixbuf: Add missing (allow-none) annotation
Owen W. Taylor [Wed, 12 Mar 2014 15:49:14 +0000 (11:49 -0400)]
gdk_cairo_surface_create_from_pixbuf: Add missing (allow-none) annotation

The window is allowed to be %NULL

12 years agoUpdated Czech translation
Marek Černocký [Wed, 12 Mar 2014 15:06:14 +0000 (16:06 +0100)]
Updated Czech translation

12 years agocsd: Fix the northeast resize handle position
Matthias Clasen [Wed, 12 Mar 2014 04:28:26 +0000 (00:28 -0400)]
csd: Fix the northeast resize handle position

12 years agocsd: Fix invisible border/shadow confusion
Matthias Clasen [Wed, 12 Mar 2014 04:15:02 +0000 (00:15 -0400)]
csd: Fix invisible border/shadow confusion

We did not set an input shape on the window, so the region outside
the invisible border where we draw the outer edges of the shadow
were still part of the window, as far as clicks and cursors were
concerned. Fix this by setting an input shape that makes all clicks
outside of the resize borders go through to the underlying window.

https://bugzilla.gnome.org/show_bug.cgi?id=726125

12 years agobroadway: remove unused libcrypt
Ryan Lortie [Wed, 12 Mar 2014 04:10:03 +0000 (00:10 -0400)]
broadway: remove unused libcrypt

eb1ab0dac2f3d7a05cfe8b87e83b363d39279a38 removed support for authentication
based on crypt()-hashed passwords but it didn't remove the header.

Finish up with the removal.

This allows the broadway backend to build on FreeBSD (which has no
crypt.h).

https://bugzilla.gnome.org/show_bug.cgi?id=726149

12 years agoUpdated Portuguese translation
Duarte Loreto [Wed, 12 Mar 2014 01:37:52 +0000 (01:37 +0000)]
Updated Portuguese translation

12 years agoUpdated Portuguese translation
Duarte Loreto [Wed, 12 Mar 2014 01:14:44 +0000 (01:14 +0000)]
Updated Portuguese translation

12 years agotextview: Postpone text handles creation until when needed
Carlos Garnacho [Mon, 10 Mar 2014 21:01:42 +0000 (22:01 +0100)]
textview: Postpone text handles creation until when needed

This is only necessary for touch devices, so unnecessary on many setups.

12 years agotextview: Postpone magnifier creation until when needed
Carlos Garnacho [Mon, 10 Mar 2014 21:00:44 +0000 (22:00 +0100)]
textview: Postpone magnifier creation until when needed

This is only necessary for touch devices, so unlikely on many setups.

12 years agoentry: revert text handles to cursor mode after cut/paste
Carlos Garnacho [Mon, 10 Mar 2014 20:57:07 +0000 (21:57 +0100)]
entry: revert text handles to cursor mode after cut/paste

The current text selection goes away in that case, so reflect
that too in the handles.

12 years agoentry: postpone text handles creation until when needed
Carlos Garnacho [Mon, 10 Mar 2014 20:56:12 +0000 (21:56 +0100)]
entry: postpone text handles creation until when needed

This is only necessary for touch devices, so unlikely in many setups.

12 years agoentry: postpone magnifier creation until when needed
Carlos Garnacho [Mon, 10 Mar 2014 20:54:31 +0000 (21:54 +0100)]
entry: postpone magnifier creation until when needed

12 years agoa11y: Don't allocate a list just for counting widgets in GtkContainerAccessible
Carlos Garnacho [Mon, 10 Mar 2014 14:15:37 +0000 (15:15 +0100)]
a11y: Don't allocate a list just for counting widgets in GtkContainerAccessible

It's more straightforward if counting through gtk_container_foreach().

12 years agotestsuite: Update a11y/menubutton3.ui test expectations
Carlos Garnacho [Thu, 6 Mar 2014 23:11:17 +0000 (00:11 +0100)]
testsuite: Update a11y/menubutton3.ui test expectations

Popover is now minimally accessible.

https://bugzilla.gnome.org/show_bug.cgi?id=725864

12 years agoa11y: Add GtkPopopverAccessible
Carlos Garnacho [Thu, 6 Mar 2014 23:07:54 +0000 (00:07 +0100)]
a11y: Add GtkPopopverAccessible

And let GtkPopover use it as its GtkAccessible implementation, this
accessible sets the POPUP_FOR relationship to the relative-to widget,
and keeps track of changes there.

https://bugzilla.gnome.org/show_bug.cgi?id=725864

12 years agowindow: Emit a11y signals directly on popover added/removed
Carlos Garnacho [Mon, 10 Mar 2014 11:27:50 +0000 (12:27 +0100)]
window: Emit a11y signals directly on popover added/removed

As those are internal children, there's no signal that GtkWindowAccessible
could catch when those are added or removed, so make GtkWindow use the private
GtkContainerAccessible methods to add/remove the child accessible when that
happens.

https://bugzilla.gnome.org/show_bug.cgi?id=725864

12 years agoa11y: Add private GtkContainerAccessible functions to add/remove a child
Carlos Garnacho [Mon, 10 Mar 2014 15:12:05 +0000 (16:12 +0100)]
a11y: Add private GtkContainerAccessible functions to add/remove a child

This may be useful in container implementations, or for internal children
that trigger no signal emission.

https://bugzilla.gnome.org/show_bug.cgi?id=725864

12 years agoa11y: Make GtkWindowAccessible know about popovers
Carlos Garnacho [Thu, 6 Mar 2014 23:07:00 +0000 (00:07 +0100)]
a11y: Make GtkWindowAccessible know about popovers

https://bugzilla.gnome.org/show_bug.cgi?id=725864

12 years agowayland: Clean up a bit
Jasper St. Pierre [Mon, 10 Mar 2014 17:39:55 +0000 (13:39 -0400)]
wayland: Clean up a bit

12 years agowayland: Move some code around
Jasper St. Pierre [Mon, 10 Mar 2014 17:31:53 +0000 (13:31 -0400)]
wayland: Move some code around

12 years agowayland: Fix submenu positioning
Jasper St. Pierre [Mon, 10 Mar 2014 17:30:21 +0000 (13:30 -0400)]
wayland: Fix submenu positioning

window->x / window->y are in "root window coordinates", e.g. relative
to the topmost toplevel. However, the coordinates in get_xdg_popup are
relative to the passed-in surface, so we need to do the reverse
translation here.

12 years agomessage dialog: Fine-tune vertical spacing
Matthias Clasen [Mon, 10 Mar 2014 10:13:38 +0000 (06:13 -0400)]
message dialog: Fine-tune vertical spacing

Requested in
https://bugzilla.gnome.org/show_bug.cgi?id=723668

12 years agogtk3-demo: run button should be able to focus.
Phillip Wood [Fri, 7 Mar 2014 21:35:17 +0000 (21:35 +0000)]
gtk3-demo: run button should be able to focus.

https://bugzilla.gnome.org/show_bug.cgi?id=725141

12 years agoWindow: Focus custom titlebar with F10
Phillip Wood [Tue, 4 Mar 2014 11:06:30 +0000 (11:06 +0000)]
Window: Focus custom titlebar with F10

As discussed on desktop-devel-list [1], "There should be an intuitive,
consistent, immediate way to jump to the widgets that live in the
header bar." F10 has been suggested for this as it is already used to
active menubars.

F10 will focus the custom titlebar widget if the window has one and it
isn't already focused. If the titlebar widget doesn't exist or is
already focused then F10 focuses the menubar if there is one.

[1] https://mail.gnome.org/archives/desktop-devel-list/2014-February/msg00176.html

https://bugzilla.gnome.org/show_bug.cgi?id=725141

12 years agoUpdated Traditional Chinese translation(Hong Kong and Taiwan)
Chao-Hsiung Liao [Sun, 9 Mar 2014 04:55:17 +0000 (12:55 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)

12 years agoUpdated Latvian translation
Rūdolfs Mazurs [Sat, 8 Mar 2014 17:09:20 +0000 (19:09 +0200)]
Updated Latvian translation

12 years agoUpdated Latvian translation
Rūdolfs Mazurs [Sat, 8 Mar 2014 14:19:11 +0000 (16:19 +0200)]
Updated Latvian translation

12 years agoUpdated Chinese (China) translation
Wylmer Wang [Sat, 8 Mar 2014 13:18:19 +0000 (13:18 +0000)]
Updated Chinese (China) translation

12 years agoUpdated Chinese (China) translation
Wylmer Wang [Sat, 8 Mar 2014 13:16:19 +0000 (13:16 +0000)]
Updated Chinese (China) translation

12 years agoupdate Punjabi Translation 07March2014: Alam
A S Alam [Sat, 8 Mar 2014 02:43:50 +0000 (20:43 -0600)]
update Punjabi Translation 07March2014: Alam

12 years agoUpdated Korean translation
Changwoo Ryu [Fri, 7 Mar 2014 15:46:29 +0000 (00:46 +0900)]
Updated Korean translation

12 years agowindow: Turn popovers into internal children
Carlos Garnacho [Thu, 6 Mar 2014 22:06:40 +0000 (23:06 +0100)]
window: Turn popovers into internal children

It turns out popovers are already smart enough to cope with this
situation, so let popovers be internal children so things that rely
on gtk_container_forall(), like DnD, work without modifications.

https://bugzilla.gnome.org/show_bug.cgi?id=725727

12 years agoFix font size when gdk_x11_display_set_window_scale() is used
Owen W. Taylor [Wed, 5 Mar 2014 20:42:38 +0000 (15:42 -0500)]
Fix font size when gdk_x11_display_set_window_scale() is used

We have a hack in the XSETTINGS code to substitute gtk-xft-dpi
with gdk-unscaled-dpi unless the screen has a fixed window scale,
in which case we just use gtk-xft-dpi.

But if the screen is changed to have a fixed window scale, then
the substituted value of gdk-unscaled-dpi will stick around until
the next (coincidental) change to XSETTINGS. To fix this, force
an immediate reread of the XSETTINGS property when
gdk_x11_display_set_window_scale() is used.

https://bugzilla.gnome.org/show_bug.cgi?id=725754

12 years agoentry: Make DnD coordinate calculation compensate for entry icons
Carlos Garnacho [Fri, 7 Mar 2014 00:21:50 +0000 (01:21 +0100)]
entry: Make DnD coordinate calculation compensate for entry icons

This made DnD have effect farther on the left when dragging text over
any entry with icons in it.

https://bugzilla.gnome.org/show_bug.cgi?id=725866

12 years agoentry: use priv->dnd_position when rendering the DND cursor
Carlos Garnacho [Fri, 7 Mar 2014 00:19:50 +0000 (01:19 +0100)]
entry: use priv->dnd_position when rendering the DND cursor

This makes "cursor position" track the DnD point again, looks much
more intuitive than just rendering it on the pre-DnD position.

https://bugzilla.gnome.org/show_bug.cgi?id=725866

12 years agoImprove GtkStack documentation
Matthias Clasen [Fri, 7 Mar 2014 03:50:40 +0000 (22:50 -0500)]
Improve GtkStack documentation

As has been pointed out in
https://bugzilla.gnome.org/show_bug.cgi?id=725711
it is a little confusing that we have the concept of
a visible child and of the child being visible itself.

12 years agotests: fix testcalendar
Carlos Garnacho [Fri, 7 Mar 2014 01:48:07 +0000 (02:48 +0100)]
tests: fix testcalendar

Missing argument in gtk_style_context_get(), fooled by varargs.

12 years agocalendar: grab focus first, emit signals after
Carlos Garnacho [Fri, 7 Mar 2014 01:35:30 +0000 (02:35 +0100)]
calendar: grab focus first, emit signals after

It may be unusual, but handlers of day-selected may want to transfer
focus somewhere else, without getting it reset back right after by/to
the calendar. This makes popovers demo work on the calendar again, for
one...

12 years agoAllow selecting a dark variant with GTK_THEME
Matthias Clasen [Fri, 7 Mar 2014 00:27:37 +0000 (19:27 -0500)]
Allow selecting a dark variant with GTK_THEME

Dark themes need to be debugged, too.

Author:    Matthias Clasen <mclasen@redhat.com>

12 years agopopover: Track toplevel's focus widget when visible
Carlos Garnacho [Thu, 6 Mar 2014 15:58:06 +0000 (16:58 +0100)]
popover: Track toplevel's focus widget when visible

If the toplevel focus widget is forced out of the popover (eg. through
gtk_widget_grab_focus() anywhere else), then dismiss the popover.

12 years agopopover: Track toplevel focus changes
Carlos Garnacho [Thu, 6 Mar 2014 13:50:35 +0000 (14:50 +0100)]
popover: Track toplevel focus changes

Make the popover temporarily undo the GTK+ grab, so it remains modal
to its window, but does not attempt to steal focus on other non-modal
windows that get the focus.

This was most confusing with keyboard navigation, as the focus would
remain stuck on the popover, and not move to the newly focused window
after the popover was dismissed. It didn't have as much effect on
pointer operations as only the first click would be consumed in order
to hide the popover.

12 years agopopover: Only give back focus to drawable widgets
Carlos Garnacho [Thu, 6 Mar 2014 19:38:11 +0000 (20:38 +0100)]
popover: Only give back focus to drawable widgets

If the previous focus widget is not drawable anymore, don't even
bother in transmitting the focus grab to it.

12 years agopopover: dispose popovers when unmanaging them
Carlos Garnacho [Thu, 6 Mar 2014 17:59:40 +0000 (18:59 +0100)]
popover: dispose popovers when unmanaging them

This makes popovers get rid of dangling pointers at the time of destroying
the relative-to widget, just in case these are kept alive by a reference.

12 years agogtkentrycompletion: fix a critical warning
Claudio Saavedra [Thu, 6 Mar 2014 14:34:10 +0000 (16:34 +0200)]
gtkentrycompletion: fix a critical warning

There are early returns in this method before the completion timeout
is set later on, so set the source to 0 to avoid trying to remove it
later again.

https://bugzilla.gnome.org/show_bug.cgi?id=725824

12 years agoUpdated Brazilian Portuguese translation
Rafael Ferreira [Thu, 6 Mar 2014 10:03:41 +0000 (10:03 +0000)]
Updated Brazilian Portuguese translation

12 years agogtkmenu: displace popups by their css margin
Lars Uebernickel [Wed, 5 Mar 2014 11:59:24 +0000 (12:59 +0100)]
gtkmenu: displace popups by their css margin

Without a margin, the pointer is above the first (or last) menu item,
making it easy to accidentally activate that item.

https://bugzilla.gnome.org/show_bug.cgi?id=591258

12 years agoGtkIconViewAccessible: Don't access freed item
Marek Kasik [Tue, 4 Mar 2014 14:17:23 +0000 (15:17 +0100)]
GtkIconViewAccessible: Don't access freed item

Don't get index from deleted GtkIconViewItem.
Remove the item before traversing list of items.

https://bugzilla.gnome.org/show_bug.cgi?id=701884

12 years agoreftests: Remove unused argument from function
Benjamin Otte [Mon, 3 Mar 2014 13:21:02 +0000 (14:21 +0100)]
reftests: Remove unused argument from function

12 years agoseparatortoolitem: Make gcc stop complaining
Benjamin Otte [Sun, 2 Mar 2014 20:49:35 +0000 (21:49 +0100)]
separatortoolitem: Make gcc stop complaining

No, a ButtonEvent and a MotionEvent are not the same thing.

12 years agobroadway: Print a useful error message
Benjamin Otte [Sun, 2 Mar 2014 16:59:34 +0000 (17:59 +0100)]
broadway: Print a useful error message

.. instead of the generic "failed to open display".

12 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Wed, 5 Mar 2014 17:27:25 +0000 (18:27 +0100)]
Updated Norwegian bokmål translation

12 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Wed, 5 Mar 2014 17:25:51 +0000 (18:25 +0100)]
Updated Norwegian bokmål translation

12 years agoUpdated Spanish translation
Daniel Mustieles [Wed, 5 Mar 2014 15:40:04 +0000 (16:40 +0100)]
Updated Spanish translation

12 years agoUpdated Spanish translation
Daniel Mustieles [Wed, 5 Mar 2014 15:39:52 +0000 (16:39 +0100)]
Updated Spanish translation

12 years agopopover: Set pointer motion mask on the popover window
Carlos Garnacho [Wed, 5 Mar 2014 12:50:17 +0000 (13:50 +0100)]
popover: Set pointer motion mask on the popover window

This is not necessary for the popover itself, but helps tooltips
code confine the widget lookup within the popover if the pointer
is inside it, otherwise the widget lookup may turn out wrong for
motion events, starting the tooltip widget lookup from the toplevel
window, mistakenly triggering tooltips on the natural window
descendants (ie. the widget below the popover)

https://bugzilla.gnome.org/show_bug.cgi?id=724785

12 years agoicontheme: support emblems in symbolic icons
Cosimo Cecchi [Wed, 6 Nov 2013 02:32:28 +0000 (18:32 -0800)]
icontheme: support emblems in symbolic icons

We were failing to apply emblems in the symbolic icons code path.

https://bugzilla.gnome.org/show_bug.cgi?id=725683

12 years agoUpdated Traditional Chinese translation(Hong Kong and Taiwan)
Chao-Hsiung Liao [Tue, 4 Mar 2014 12:17:10 +0000 (20:17 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)

12 years agoBump version
Matthias Clasen [Tue, 4 Mar 2014 05:19:56 +0000 (00:19 -0500)]
Bump version

12 years ago3.11.8
Matthias Clasen [Tue, 4 Mar 2014 03:35:20 +0000 (22:35 -0500)]
3.11.8

12 years agoUpdate messagedialog a11y test
Matthias Clasen [Tue, 4 Mar 2014 04:37:37 +0000 (23:37 -0500)]
Update messagedialog a11y test

12 years agoFix up builder test
Matthias Clasen [Tue, 4 Mar 2014 04:30:55 +0000 (23:30 -0500)]
Fix up builder test

Dialogs now report two children. Thats alright, adapt
the test to deal with it.

12 years agoFix up signal deprecation
Matthias Clasen [Tue, 4 Mar 2014 04:07:50 +0000 (23:07 -0500)]
Fix up signal deprecation

12 years agoDeprecate GdkEventVisibility and ::visibility-notify-event
Matthias Clasen [Mon, 3 Mar 2014 23:31:06 +0000 (18:31 -0500)]
Deprecate GdkEventVisibility and ::visibility-notify-event

These don't really work anymore in a composited world.
Lets make it official.
https://bugzilla.gnome.org/show_bug.cgi?id=481073

12 years agoAllow applications to use gtk-menu-bar-accel in 3.12
Matthias Clasen [Mon, 23 Sep 2013 17:10:19 +0000 (13:10 -0400)]
Allow applications to use gtk-menu-bar-accel in 3.12

gnome-terminal is still using this setting, so we'll let
applications override it for another cycle. It is no longer
backed by a system-wide setting, though, and it will still
go away eventually.

This partically reverts b26c74e5dac0b5fd869d2c8cc7a0edb6b01aeebd

12 years agoAllow applications to use gtk-enable-mnemonics in 3.12
Matthias Clasen [Mon, 23 Sep 2013 17:01:49 +0000 (13:01 -0400)]
Allow applications to use gtk-enable-mnemonics in 3.12

gnome-terminal is still using this setting, so we'll let
applications override it for another cycle. It is no longer
backed by a system-wide setting, though, and it will still
go away eventually.

This partically reverts 7e3a494fac62140f466fc43f6cc061215953e9b1