gtk4.git
10 years agowindow: fix initial shadow width for maximized/fullscreen windows
Cosimo Cecchi [Mon, 13 Apr 2015 22:31:10 +0000 (15:31 -0700)]
window: fix initial shadow width for maximized/fullscreen windows

To calculate the shadow width, we look at the value of priv->fullscreen
and priv->maximized.
Those fields will have the actual value only after GTK receives back a
window state event though, so they will be wrong in _realize(). Look at
priv->fullscreen_initially and priv->maximize_initially too, to avoid
the size changing right after realize, which would make the window
flicker if maximized at startup.

https://bugzilla.gnome.org/show_bug.cgi?id=747808

10 years agoAdwaita: fix outline radius for dialog-action-area buttons
Sebastian Keller [Thu, 16 Apr 2015 23:07:03 +0000 (01:07 +0200)]
Adwaita: fix outline radius for dialog-action-area buttons

https://bugzilla.gnome.org/show_bug.cgi?id=748013

10 years agoUpdated Brazilian Portuguese translation
Rafael Ferreira [Thu, 16 Apr 2015 21:05:04 +0000 (21:05 +0000)]
Updated Brazilian Portuguese translation

10 years agoscrolledwindow: Do not round dx/dy to int
Carlos Garnacho [Thu, 16 Apr 2015 20:45:54 +0000 (22:45 +0200)]
scrolledwindow: Do not round dx/dy to int

Libinput will use 0.0f on the "scrolling finished" event, so check for this
instead of rounding (<1 values are sort of frequent on touchpads). This
impedes bug #745315 to resurface after commit d563b943ed3.

10 years agoscrolledwindow: Ensure the animation is cancelled on arriving scroll events
Carlos Garnacho [Thu, 16 Apr 2015 16:40:32 +0000 (18:40 +0200)]
scrolledwindow: Ensure the animation is cancelled on arriving scroll events

When the scrolledwindow receives scroll events, it ensures the timeout to
maybe start the "snap back to edges" animation is reset, but it does nothing
about the animation source. It must be reset just the same, to maybe be
started after the timeout fires up.

10 years agoAdwaita: colorpicker fixes.
Lapo Calamandrei [Thu, 16 Apr 2015 07:57:23 +0000 (09:57 +0200)]
Adwaita: colorpicker fixes.

Added another :not() to fix the first and last swatches in the
first section of colorpicker.
Dodge the artifacts due to stacked anti-aliased rounded corners by
increasing the radius of the base object by 1px.

10 years agoHighContrast: sync color picker changes with Adwaita
Jakub Steiner [Wed, 15 Apr 2015 22:47:40 +0000 (00:47 +0200)]
HighContrast: sync color picker changes with Adwaita

10 years agoUpdated Slovenian translation
Matej Urbančič [Wed, 15 Apr 2015 20:22:07 +0000 (22:22 +0200)]
Updated Slovenian translation

10 years agoAdwaita: clearer comment for colorbutton hack
Lapo Calamandrei [Wed, 15 Apr 2015 19:53:11 +0000 (21:53 +0200)]
Adwaita: clearer comment for colorbutton hack

10 years agoAdwaita: colorpicker and colorbutton fixes.
Lapo Calamandrei [Wed, 15 Apr 2015 19:43:59 +0000 (21:43 +0200)]
Adwaita: colorpicker and colorbutton fixes.

With nasty gross ugly code, no other ways, sorry.

10 years agoAdwaita: GtkPlacesSidebar insensitive items.
Lapo Calamandrei [Wed, 15 Apr 2015 18:20:52 +0000 (20:20 +0200)]
Adwaita: GtkPlacesSidebar insensitive items.

Set the right colors there.

10 years agoHighContrast: don't need to work around ffox
Jakub Steiner [Wed, 15 Apr 2015 12:55:35 +0000 (14:55 +0200)]
HighContrast: don't need to work around ffox

No need for the workaround for ffox 36, works in 37

https://bugzilla.gnome.org/show_bug.cgi?id=747827

10 years agoconfigure: specify that mir backend needs gio-unix
Daniel van Vugt [Wed, 15 Apr 2015 11:19:14 +0000 (13:19 +0200)]
configure: specify that mir backend needs gio-unix

10 years agoFix window button updates in split headers
Matthias Clasen [Tue, 14 Apr 2015 21:05:34 +0000 (17:05 -0400)]
Fix window button updates in split headers

We were only updating window buttons when the headerbar was
a direct child of the window. That is not the case in more
complicated situations, such as the split headers in gedit
or polari. To fix such cases, make the headerbar itself listen
for state changes on its toplevel.

https://bugzilla.gnome.org/show_bug.cgi?id=747805

10 years agoUpdated Danish translation
Ask Hjorth Larsen [Mon, 13 Apr 2015 17:31:03 +0000 (19:31 +0200)]
Updated Danish translation

10 years agoUpdated Danish translation
Ask Hjorth Larsen [Mon, 13 Apr 2015 17:24:02 +0000 (19:24 +0200)]
Updated Danish translation

10 years agogesturesingle: Ignore motion/updates from unhandled sequences
Carlos Garnacho [Mon, 13 Apr 2015 14:15:11 +0000 (16:15 +0200)]
gesturesingle: Ignore motion/updates from unhandled sequences

This used to do the right thing, practically ignoring those event
if they went all through the event handler, unless this motion event
triggered the cancellation of this gesture from the event handler
of another controllers.

In this case, the controller would be reset, but the motion event
would still go through its handler, setting again
current_button/sequence as the motion event has buttons in its
modifiers. This leaves GtkGestureSingle with inconsistent data
that may interfere with future runs.

https://bugzilla.gnome.org/show_bug.cgi?id=747469

10 years agoscrolledwindow: Show scrollbars on tablet devices
Carlos Garnacho [Mon, 13 Apr 2015 13:29:57 +0000 (15:29 +0200)]
scrolledwindow: Show scrollbars on tablet devices

The code managing scrollbars visibility was too pervasively checking for
mouse devices, leaving pen/eraser/cursor devices with no scrollbars at
all. Relax these checks a bit, and actually toggle full-width scrollbars
on pen/eraser devices, so it is an easier target.

https://bugzilla.gnome.org/show_bug.cgi?id=747608

10 years agoHC: fix selected list-rows
Jakub Steiner [Mon, 13 Apr 2015 15:07:16 +0000 (17:07 +0200)]
HC: fix selected list-rows

https://bugzilla.gnome.org/show_bug.cgi?id=747689

10 years agox11: Relax requirements for setting ParentRelative
Evangelos Foutras [Sat, 11 Apr 2015 08:20:56 +0000 (11:20 +0300)]
x11: Relax requirements for setting ParentRelative

XSetWindowBackgroundPixmap() will throw BadMatch only in the case of a
different parent window depth. Different visuals are fine and actually
expected in Gtk+ 3.16 (since we don't stick to the system default visual
but try to pick a better one).

https://bugzilla.gnome.org/show_bug.cgi?id=747524

10 years agoplug: Remove size_allocate handler
Carlos Garnacho [Thu, 9 Apr 2015 12:09:42 +0000 (14:09 +0200)]
plug: Remove size_allocate handler

Since windows can be intermediate containers this is somewhat redundant,
plus gtk_window_size_allocate() will do additional things like allocating
the popovers, which was mistakenly skipped by local plugs not chaining up.

https://bugzilla.gnome.org/show_bug.cgi?id=747553

10 years agoUpdated Slovenian translation
Matej Urbančič [Sun, 12 Apr 2015 19:18:45 +0000 (21:18 +0200)]
Updated Slovenian translation

10 years agoUpdate Arabic translation
Khaled Hosny [Sat, 11 Apr 2015 23:54:18 +0000 (01:54 +0200)]
Update Arabic translation

10 years agoUpdate Arabic translation
Khaled Hosny [Sat, 11 Apr 2015 23:39:53 +0000 (01:39 +0200)]
Update Arabic translation

10 years agoUpdated Hebrew translation
Yosef Or Boczko [Sat, 11 Apr 2015 21:47:26 +0000 (00:47 +0300)]
Updated Hebrew translation

10 years agoUpdated Swedish translation
Anders Jonsson [Sat, 11 Apr 2015 20:54:09 +0000 (20:54 +0000)]
Updated Swedish translation

10 years agoUpdated Swedish translation
Anders Jonsson [Sat, 11 Apr 2015 20:51:26 +0000 (20:51 +0000)]
Updated Swedish translation

10 years agoResolve redefinition of typedef GtkCssNode
William Hua [Thu, 9 Apr 2015 13:28:30 +0000 (09:28 -0400)]
Resolve redefinition of typedef GtkCssNode

This causes a build failure under Mac OS because GtkCssNode is defined
in both gtk/gtkcssnodeprivate.h and gtk/gtkcsstypesprivate.h.

https://bugzilla.gnome.org/show_bug.cgi?id=747560

10 years agoRemove no longer needed in-tree editing
Matthias Clasen [Fri, 10 Apr 2015 01:03:13 +0000 (21:03 -0400)]
Remove no longer needed in-tree editing

10 years agoUse a popover for creating new folders
Matthias Clasen [Fri, 10 Apr 2015 00:56:54 +0000 (20:56 -0400)]
Use a popover for creating new folders

10 years agoHighContrast: fix scrollbar slider 'padding'
Jakub Steiner [Fri, 10 Apr 2015 18:19:42 +0000 (20:19 +0200)]
HighContrast: fix scrollbar slider 'padding'

10 years agoHighContrast: insensitive backdrop list rows have a border
Jakub Steiner [Fri, 10 Apr 2015 13:13:12 +0000 (15:13 +0200)]
HighContrast: insensitive backdrop list rows have a border

https://bugzilla.gnome.org/show_bug.cgi?id=747288

10 years agoAdwaita: reset .list-row.button props more aggressively...
Lapo Calamandrei [Fri, 10 Apr 2015 14:39:35 +0000 (16:39 +0200)]
Adwaita: reset .list-row.button props more aggressively...

...otherwise we run into problem with `.button` inherited style
properties, see https://bugzilla.gnome.org/show_bug.cgi?id=747288
which applies to Adwaita as well.

10 years agoa11y: Only emit signals when cells change; not upon creation
Joanmarie Diggs [Thu, 9 Apr 2015 00:08:25 +0000 (20:08 -0400)]
a11y: Only emit signals when cells change; not upon creation

This is a workaround for atspi-atk behaviour.

atspi-atk uses signal emission hooks. So it to already catches
signal emissions on creation of objects, before anyone could even
think of g_signal_connect()ing.

https://bugzilla.gnome.org/show_bug.cgi?id=746706

10 years agonotebook: prevent tab closebutton :hover to change metric
Jakub Steiner [Fri, 10 Apr 2015 14:27:40 +0000 (16:27 +0200)]
notebook: prevent tab closebutton :hover to change metric

10 years agoGtkApplication: Prevent more crashes around shutdown
Matthias Clasen [Fri, 10 Apr 2015 12:03:24 +0000 (08:03 -0400)]
GtkApplication: Prevent more crashes around shutdown

gtk_application_shutdown clears the impl member, so all
callbacks, signal handlers, etc that might still be triggered
between a shutdown call and the return from the mainloop
better be prepared to deal with impl being NULL.

10 years agogtk/inspector/css-node-tree.c: Fix Build
Chun-wei Fan [Fri, 10 Apr 2015 07:48:59 +0000 (15:48 +0800)]
gtk/inspector/css-node-tree.c: Fix Build

strcasecmp() is unfortunately not universally available, along with
strings.h.  Fix the build by replacing strcasecmp() with
g_ascii_strcasecmp(), and remove the strings.h include.

https://bugzilla.gnome.org/show_bug.cgi?id=747604

10 years agoVisual Studio Builds: Fix Header "Installation"
Chun-wei Fan [Fri, 10 Apr 2015 08:19:49 +0000 (16:19 +0800)]
Visual Studio Builds: Fix Header "Installation"

gtksidebar.h was renamed as gtkstacksidebar.h, so fix that in the header
"installation".

10 years agopopover: Remove signal handlers in dispose
Matthias Clasen [Thu, 9 Apr 2015 00:20:38 +0000 (20:20 -0400)]
popover: Remove signal handlers in dispose

If we don't, they may get triggered at an inopportune
time later, causing us to crash.

See
https://bugzilla.redhat.com/show_bug.cgi?id=1203499

https://bugzilla.gnome.org/show_bug.cgi?id=747539

10 years agowayland: Wait for output and seat information
Matthias Clasen [Thu, 9 Apr 2015 04:14:20 +0000 (00:14 -0400)]
wayland: Wait for output and seat information

When we open the connection, we get the wl_output object,
but we return before all the information such as monitor
geometry has arrived, which causes us to misinform early
users of this information. Do a roundtrip here that causes
us to wait until the information is complete. Do the same
for seats, just in case.
https://bugzilla.gnome.org/show_bug.cgi?id=747471

10 years agowayland: Add more debug output
Matthias Clasen [Thu, 9 Apr 2015 04:13:37 +0000 (00:13 -0400)]
wayland: Add more debug output

10 years agoMake gtk-update-icon-cache not fall over leftover temp files
Matthias Clasen [Thu, 9 Apr 2015 00:30:26 +0000 (20:30 -0400)]
Make gtk-update-icon-cache not fall over leftover temp files

This is a followup to 0fd185fa6de2a89c11b4a28. There is no
good reason to only try again if --force is passed. Do it
always.

See
https://bugzilla.redhat.com/show_bug.cgi?id=1194957

10 years agotree view: Avoid a crash with rubberbanding
Matthias Clasen [Wed, 8 Apr 2015 22:54:45 +0000 (18:54 -0400)]
tree view: Avoid a crash with rubberbanding

When a treeview is destroyed while rubberbanding is going
on, we crash because the rb tree is nuked before we want
to access it to stop the rubberbanding. To avoid this crash
end the rubberbanding early in destroy().

See
https://bugzilla.redhat.com/show_bug.cgi?id=1173904

10 years agoGtkApplication: avoid a crash
Matthias Clasen [Wed, 8 Apr 2015 19:44:21 +0000 (15:44 -0400)]
GtkApplication: avoid a crash

It can apparently happen that we get focus in events
on windows after gtk_application_shutdown() has been
called. Avoid an unnecessary crash in this case.

See
https://bugzilla.redhat.com/show_bug.cgi?id=1176339

10 years agoRevert "image: Optimize non-resize changes"
Matthias Clasen [Wed, 8 Apr 2015 18:13:20 +0000 (14:13 -0400)]
Revert "image: Optimize non-resize changes"

This reverts commit bae97a4c6bcb1c82d4a5f296c3d428173e31c6e4.

We're still seeing regressions from there, see
http://bugzilla.redhat.com/show_bug.cgi?id=1208183

10 years agotestgtk: Respect display cursor size limitations
Matthias Clasen [Tue, 7 Apr 2015 22:39:10 +0000 (18:39 -0400)]
testgtk: Respect display cursor size limitations

Instead of hardcoding 64, ask the display about the
maximal supported cursor size.

10 years agomir: Fix build failure due to changed API
William Hua [Tue, 7 Apr 2015 14:33:35 +0000 (10:33 -0400)]
mir: Fix build failure due to changed API

mir_surface_set_event_handler() now takes three arguments.

10 years agoFix indicator proximity checks
Matthias Clasen [Tue, 7 Apr 2015 01:27:17 +0000 (21:27 -0400)]
Fix indicator proximity checks

The coordinate translations here were not working properly
for window widgets inside the scrolled window, as can be
seen e.g. for the horizontal scrollbar of the 'Tree View'
example in gtk3-demo.

https://bugzilla.gnome.org/show_bug.cgi?id=747406

10 years agoMake indicators pop out when needed
Matthias Clasen [Mon, 6 Apr 2015 13:16:03 +0000 (09:16 -0400)]
Make indicators pop out when needed

When moving over a non-expanded indicator from the outside, we were
not expanding it, due to on_scrollbar being true. This can be seen
e.g. when moving from the content pane over to the sidebar indicator
in gtk3-demo. We must still ensure that the indicator is expanded
when receiving motion events over the indicator.

https://bugzilla.gnome.org/show_bug.cgi?id=747407

10 years agoUpdated breton translation
Denis Arnaud [Mon, 6 Apr 2015 13:00:01 +0000 (15:00 +0200)]
Updated breton translation

10 years agodocs: Note the side effects of GtkLabel setters
Emmanuele Bassi [Sat, 4 Apr 2015 12:54:34 +0000 (13:54 +0100)]
docs: Note the side effects of GtkLabel setters

The gtk_label_set_text() and gtk_label_set_markup() functions have
various side effects that ought to be documented, especially for
non-C developers using properties directly.

https://bugzilla.gnome.org/show_bug.cgi?id=747121

10 years agofile chooser: Ensure a folder is loaded when browsing
Matthias Clasen [Mon, 6 Apr 2015 02:27:23 +0000 (22:27 -0400)]
file chooser: Ensure a folder is loaded when browsing

Just showing an empty list when the user hits Ctrl-l in
recent files is not very useful. Load the home directory
instead.

10 years agofile chooser: Open location entry more easily
Matthias Clasen [Mon, 6 Apr 2015 02:12:46 +0000 (22:12 -0400)]
file chooser: Open location entry more easily

Show the location entry also when alphanumerics are typed
into the list.

10 years agotestfilechooser: Improve filter setup
Matthias Clasen [Mon, 6 Apr 2015 02:09:22 +0000 (22:09 -0400)]
testfilechooser: Improve filter setup

It is not ideal when the default filter filters away
all files, so change things to select the 'all files'
filter by default.

10 years agoFormatting fix
Matthias Clasen [Mon, 6 Apr 2015 02:09:04 +0000 (22:09 -0400)]
Formatting fix

10 years agoFormatting fix
Matthias Clasen [Mon, 6 Apr 2015 02:07:59 +0000 (22:07 -0400)]
Formatting fix

10 years agotestgtk: Set initial cursor
Matthias Clasen [Sun, 5 Apr 2015 23:10:39 +0000 (19:10 -0400)]
testgtk: Set initial cursor

Arrange for the entry and the cursor area to both start out
with the arrow cursor.

10 years agotestgtk: Redo the cursor example
Matthias Clasen [Sun, 5 Apr 2015 23:00:44 +0000 (19:00 -0400)]
testgtk: Redo the cursor example

Use cursor names instead of font cursors, so we can also show
cursors that are not represented in the X cursor font and thus
don't have a value in the GdkCursorType enumeration.

10 years agowayland: Don't mix free and g_strdup
Matthias Clasen [Sun, 5 Apr 2015 22:59:51 +0000 (18:59 -0400)]
wayland: Don't mix free and g_strdup

While it works most of the time, this is technically
not correct.

10 years agotestgtk: Update cursor theme on size change
Matthias Clasen [Sun, 5 Apr 2015 17:28:26 +0000 (13:28 -0400)]
testgtk: Update cursor theme on size change

We were connecting to the wrong signal, so we were not picking
up all changes to the spin button value.

10 years agoGtkListBox: Make model binding introspectable
Matthias Clasen [Sun, 5 Apr 2015 12:13:24 +0000 (08:13 -0400)]
GtkListBox: Make model binding introspectable

Introspection can't deal with bare pointers, so annotate
the callback as taking objects. This is the outcome of
discussion in
https://bugzilla.gnome.org/show_bug.cgi?id=746893

10 years agogtktextbuffer: small cleanup in clipboard handling
Paolo Borelli [Sun, 29 Mar 2015 10:22:42 +0000 (12:22 +0200)]
gtktextbuffer: small cleanup in clipboard handling

Make code shorter and also more efficient since we move the
selection check out of the loop

https://bugzilla.gnome.org/show_bug.cgi?id=747096

10 years agox11/gl: Fix use after free
Olivier Fourdan [Fri, 3 Apr 2015 15:51:57 +0000 (17:51 +0200)]
x11/gl: Fix use after free

https://bugzilla.gnome.org/show_bug.cgi?id=747305

Signed-off-by: Olivier Fourdan <fourdan@xfce.org>
10 years agoAdwaita: stronger entry focus indication
Lapo Calamandrei [Thu, 2 Apr 2015 16:49:33 +0000 (18:49 +0200)]
Adwaita: stronger entry focus indication

See https://bugzilla.gnome.org/show_bug.cgi?id=746202 for details.

10 years agoUpdated Slovak translation
Dušan Kazik [Thu, 2 Apr 2015 08:57:09 +0000 (08:57 +0000)]
Updated Slovak translation

10 years agoscrolledwindow: Ignore 0/0 scroll events when possibly cancelling animation
Carlos Garnacho [Tue, 31 Mar 2015 18:24:07 +0000 (20:24 +0200)]
scrolledwindow: Ignore 0/0 scroll events when possibly cancelling animation

These should be used eventually to start kinetic scrolling, so should definitely
be ignored on cancellation.

https://bugzilla.gnome.org/show_bug.cgi?id=747133

10 years agoUpdated Russian translation
Stas Solovey [Wed, 1 Apr 2015 12:15:49 +0000 (12:15 +0000)]
Updated Russian translation

10 years agoUpdated Russian translation
Stas Solovey [Wed, 1 Apr 2015 11:31:10 +0000 (11:31 +0000)]
Updated Russian translation

10 years agoUdated Czech translation
Marek Černocký [Wed, 1 Apr 2015 09:55:02 +0000 (11:55 +0200)]
Udated Czech translation

10 years agoUpdated Slovenian translation
Matej Urbančič [Tue, 31 Mar 2015 19:09:21 +0000 (21:09 +0200)]
Updated Slovenian translation

10 years agoscrolledwindow: Remove needless "dragging" field from Indicator struct
Carlos Garnacho [Tue, 31 Mar 2015 11:13:50 +0000 (13:13 +0200)]
scrolledwindow: Remove needless "dragging" field from Indicator struct

The "over" state already stays set while scrollbar dragging happens, there's
no need to double track that.

https://bugzilla.gnome.org/show_bug.cgi?id=746961

10 years agoscrolledwindow: Check the event widget on captured motion events
Carlos Garnacho [Tue, 31 Mar 2015 11:01:34 +0000 (13:01 +0200)]
scrolledwindow: Check the event widget on captured motion events

This path is only intended to be triggered on events directed towards the
child of the scrolledwindow, so make it explicitly so. This avoids scrollbar
"over" state flashing when dragging finishes within the slider.

https://bugzilla.gnome.org/show_bug.cgi?id=746961

10 years agowayland: Set a more correct xdg_surface application id
Jonas Ådahl [Mon, 23 Mar 2015 05:08:09 +0000 (13:08 +0800)]
wayland: Set a more correct xdg_surface application id

The "app_id" of a xdg_surface should be the ID that can potentially be
used to get the DBUS name or the .desktop file.

For GtkApplication programs this is often the ID passed when creating the
GtkApplication object, so when available lets use that.

As fallbacks, first try g_get_prgname as it often corresponds to the
basename part of the .dektop file for non-GtkApplication programs.
Otherwise use gdk_get_program_class, even though that string usually
doesn't conform to the expectations of xdg_surface.set_application_id.

https://bugzilla.gnome.org/show_bug.cgi?id=746435

10 years agoUpdated Russian translation
Stas Solovey [Mon, 30 Mar 2015 22:40:20 +0000 (22:40 +0000)]
Updated Russian translation

10 years agoRevert "Adwaita: desaturate selected color for backdrop"
Jakub Steiner [Mon, 30 Mar 2015 12:54:25 +0000 (14:54 +0200)]
Revert "Adwaita: desaturate selected color for backdrop"

This reverts commit 6f478db2facd8744714b7ab374d2687d777653a3.

As per discussion on bug 746202, the real solution to the case
in bug 704352 is what the reporter suggests.

10 years agoUpdated Ukrainian translation
Daniel Korostil [Mon, 30 Mar 2015 05:52:44 +0000 (08:52 +0300)]
Updated Ukrainian translation

10 years agogtkdnd: Restore widget opacity after fading out
Carlos Garnacho [Sat, 28 Mar 2015 15:22:09 +0000 (16:22 +0100)]
gtkdnd: Restore widget opacity after fading out

Otherwise the widget is left fully transparent, and will start as such if
the widget is reused.

10 years agogtkdnd: Set the icon helper from the drag site onto the context
Carlos Garnacho [Sat, 28 Mar 2015 13:16:45 +0000 (14:16 +0100)]
gtkdnd: Set the icon helper from the drag site onto the context

If the icon happened to come from the drag site icon helper (ie. set through
gtk_drag_source_set_icon*), it would be referenced on the GtkDragInfo,
but not hooked into its context. This results on non visible drag windows,
until set_icon_helper() happened to be called on some path.

10 years agoroundedbox: Fix a copy/paste error
Benjamin Otte [Sun, 29 Mar 2015 20:16:14 +0000 (22:16 +0200)]
roundedbox: Fix a copy/paste error

This caused shadows to not be drawn in certain situations when overlayed
via a popover animation.

https://bugzilla.gnome.org/show_bug.cgi?id=746862

10 years agoUpdated Czech translation
Marek Černocký [Sun, 29 Mar 2015 11:49:43 +0000 (13:49 +0200)]
Updated Czech translation

10 years agoradio button: Fix directional navigation
Matthias Clasen [Sun, 29 Mar 2015 04:24:35 +0000 (00:24 -0400)]
radio button: Fix directional navigation

GtkRadioButton had a deficient copy of the focus sort code
in GtkContainer, causing focus to jump over the next button
in the list. Just use _gtk_container_focus_sort() here,
which fixes the bug _and_ saves 80 lines of code.

https://bugzilla.gnome.org/show_bug.cgi?id=746817

10 years agoFix separator in testsplitheaders
Matthias Clasen [Sun, 29 Mar 2015 02:32:12 +0000 (22:32 -0400)]
Fix separator in testsplitheaders

10 years agogtkprogressbar: fix size allocation
Lars Uebernickel [Tue, 24 Mar 2015 13:25:49 +0000 (14:25 +0100)]
gtkprogressbar: fix size allocation

As of 74405cc, progress bars use a new design with values drawn on top
(or to the left) of the through instead of inside of it. This change
brought a number of regressions: the min-horizontal-bar-height and
min-vertical-bar-width style properties are not respected anymore. For
vertical progress bars, the value was drawn too close to the bar and not
centered vertically.

Fix this by respecting the style properties and drawing the value label
at the correct position.

Also, the xspacing and yspacing properties didn't server any apparent
purpose. Change their semantics to mean "the spacing between the label
and the bar". Hence, they only need to be added to the size request when
showing the label. Since we are changing semantics anyway, reduce their
default values from 7 to 2, to avoid and excessive gap.

https://bugzilla.gnome.org/show_bug.cgi?id=746688

10 years agomake libcanberra support optional
Antoine Jacoutot [Fri, 27 Mar 2015 18:04:20 +0000 (19:04 +0100)]
make libcanberra support optional

Make it possible to control libcanberra support so we don't risk ending up with
a cyclic dependency when using packages: gtk+ -> libcanberra-gtk3 -> gtk+

https://bugzilla.gnome.org/show_bug.cgi?id=746904

10 years agoUpdated Slovenian translation
Matej Urbančič [Sat, 28 Mar 2015 21:42:42 +0000 (22:42 +0100)]
Updated Slovenian translation

10 years agoUpload Catalan translation
Jordi Mas [Sat, 28 Mar 2015 12:36:25 +0000 (13:36 +0100)]
Upload Catalan translation

10 years agoGtkListBox: fix model binding refcount issue
Ryan Lortie [Fri, 27 Mar 2015 15:57:38 +0000 (11:57 -0400)]
GtkListBox: fix model binding refcount issue

As it is, GtkListBox model binding will work nicely as long as your
create_widget_func returns a floating reference on the newly-created
widget.

If you try to return a full reference (as any higher-level language
would do) then you will leak that reference.

Fix that up by converting any floating references into full references
and then unconditionally releasing the full reference after adding to
the box.

https://bugzilla.gnome.org/show_bug.cgi?id=746893

10 years agoscrolledwindow: Extend trackpoint heuristic to ALPS trackpoints
Hans de Goede [Thu, 26 Mar 2015 18:46:56 +0000 (19:46 +0100)]
scrolledwindow: Extend trackpoint heuristic to ALPS trackpoints

These have a different characteristic device name, look for it too
in order to enable trackpoint features.

10 years agodocs: Add initialization example for GtkGLArea
Emmanuele Bassi [Thu, 26 Mar 2015 17:40:40 +0000 (17:40 +0000)]
docs: Add initialization example for GtkGLArea

Show how to safely check for errors when initializing the OpenGL
pipeline.

10 years agox11/gl: Do not use GLX_EXT_texture_from_pixmap if not available
Emmanuele Bassi [Thu, 26 Mar 2015 17:26:29 +0000 (17:26 +0000)]
x11/gl: Do not use GLX_EXT_texture_from_pixmap if not available

If the GLX_EXT_texture_from_pixmap extension is not available when we
did the extensions check, then there's no point in using the backend
specific code paths that rely on it.

10 years agoW32: Only override window miximized position for CSDed windows
Руслан Ижбулатов [Thu, 26 Mar 2015 17:01:59 +0000 (17:01 +0000)]
W32: Only override window miximized position for CSDed windows

This fixes an issue with non-CSDed windows being improperly positioned
when maximized, which was introduced by the previous commit.

https://bugzilla.gnome.org/show_bug.cgi?id=746821

10 years agoW32: Correctly set maximized window position
Руслан Ижбулатов [Thu, 26 Mar 2015 15:22:07 +0000 (15:22 +0000)]
W32: Correctly set maximized window position

Use screen workarea to *also* set the position of a maximized window,
not just its size. Without this the window position defaults to 0:0
(the topleft corner), which is wrong when taskbar is position along the
top or left edge of the screen.

https://bugzilla.gnome.org/show_bug.cgi?id=746821

10 years agobbox: correct documentation typo
Will Thompson [Sun, 23 Jun 2013 14:35:21 +0000 (15:35 +0100)]
bbox: correct documentation typo

10 years agoFix an AtkStateType emitted for GtkBooleanCellAccessible
Joanmarie Diggs [Wed, 25 Mar 2015 23:21:37 +0000 (19:21 -0400)]
Fix an AtkStateType emitted for GtkBooleanCellAccessible

An apparent copy-and-paste error is resulting in ATs being notified of
changes in ATK_STATE_CHECKED when the cell's "sensitive" state changes.

https://bugzilla.gnome.org/show_bug.cgi?id=746784

10 years agotextview: Fix "Since" versions after backport
Rico Tzschichholz [Wed, 25 Mar 2015 19:52:43 +0000 (20:52 +0100)]
textview: Fix "Since" versions after backport

Fix for 28063ee2e42e7ce47b7bd5326f2d53875a377d57 which got cherry-picked
to 3.16

10 years agodemo/glarea: Check errors on GtkGLArea
Emmanuele Bassi [Wed, 25 Mar 2015 15:20:36 +0000 (15:20 +0000)]
demo/glarea: Check errors on GtkGLArea

We should not call OpenGL API if GtkGLArea is in an error state.

https://bugzilla.gnome.org/show_bug.cgi?id=746746

10 years agoglarea: Check that the context is not NULL
Emmanuele Bassi [Wed, 25 Mar 2015 15:19:30 +0000 (15:19 +0000)]
glarea: Check that the context is not NULL

The public API should perform more checks when it comes to an NULL
context.

https://bugzilla.gnome.org/show_bug.cgi?id=746746

10 years agox11/gl: Do not try to call MakeCurrent with a NULL context
Emmanuele Bassi [Wed, 25 Mar 2015 15:18:44 +0000 (15:18 +0000)]
x11/gl: Do not try to call MakeCurrent with a NULL context

This should not happen, but better safe than sorry.

https://bugzilla.gnome.org/show_bug.cgi?id=746746

10 years agox11/gl: Trap GL context creation
Emmanuele Bassi [Wed, 25 Mar 2015 14:34:42 +0000 (14:34 +0000)]
x11/gl: Trap GL context creation

Avoid an X11 error in case the context creation fails, so we can avoid a
crash and pick it up in the upper layers, and display an error.

https://bugzilla.gnome.org/show_bug.cgi?id=746746

10 years agofilechooser: Use same priority everywhere
Benjamin Otte [Wed, 25 Mar 2015 03:27:33 +0000 (04:27 +0100)]
filechooser: Use same priority everywhere

I managed to stall recent files today while trying to save a GTK
testcase in glade that contained enough spinning spinners that the CPU
was saturated just redrawing things.

I had to navigate the filesystem!