gtk+3.0.git
7 years agoMerge branch 'gtk-3-22' into 'gtk-3-22'
Matthias Clasen [Mon, 19 Feb 2018 17:57:04 +0000 (17:57 +0000)]
Merge branch 'gtk-3-22' into 'gtk-3-22'

gtkprintbackendcups.c: fix \n at end of a debugging note

See merge request GNOME/gtk!24

7 years agoMerge branch 'gtk-debug-flags-with-open-display' into 'gtk-3-22'
Matthias Clasen [Mon, 19 Feb 2018 17:55:12 +0000 (17:55 +0000)]
Merge branch 'gtk-debug-flags-with-open-display' into 'gtk-3-22'

gtk_init: Fix debug flags handling when a display is already open

See merge request GNOME/gtk!26

7 years agoUpdated Lithuanian translation
Aurimas Černius [Sun, 18 Feb 2018 19:43:14 +0000 (21:43 +0200)]
Updated Lithuanian translation

7 years agoUpdate German translation
Mario Blättermann [Sun, 18 Feb 2018 12:48:31 +0000 (12:48 +0000)]
Update German translation

7 years agoMerge branch 'gtk-3-22' into 'gtk-3-22'
Christoph Reiter [Sun, 18 Feb 2018 10:45:52 +0000 (10:45 +0000)]
Merge branch 'gtk-3-22' into 'gtk-3-22'

Fix annotation for gtk_tree_view_is_blank_at_pos()

See merge request GNOME/gtk!12

7 years agogtk_init: Fix debug flags handling when a display is already open
Christoph Reiter [Sun, 18 Feb 2018 09:34:52 +0000 (10:34 +0100)]
gtk_init: Fix debug flags handling when a display is already open

In PyGObject gdk_init() is called before gtk_init() and thus there is
already a default display open when gtk_init() is called.
The code assigning the display to the debug_flags struct gets only
called when the default display changes, which never happens
when there already is one. As a result GTK_DEBUG=interactive
doesn't do anyting with Python apps.

This makes it call the change callback in case a display is already
there.

See https://gitlab.gnome.org/GNOME/pygobject/issues/166

7 years agoFix incorrect optimization in find_builtin_icon().
John Lindgren [Sun, 18 Feb 2018 03:03:27 +0000 (22:03 -0500)]
Fix incorrect optimization in find_builtin_icon().

When we found an icon with exactly the requested size, we'd stop
searching immediately (good), but we'd neglect to set the returned
min_difference to 0 (bad).  This caused theme_lookup_icon() to
prefer other, potentially much worse, matches over the exact one.

7 years agodocs: Add proper deprecation text for GtkStatusIcon
Emmanuele Bassi [Tue, 13 Feb 2018 11:12:26 +0000 (11:12 +0000)]
docs: Add proper deprecation text for GtkStatusIcon

We link to the HowDoI for GNotification in the class description, but we
should be more verbose in the deprecation notices for each function of
the GtkStatusIcon class.

https://bugzilla.gnome.org/show_bug.cgi?id=743975

7 years agogtkprintbackendcups.c: fix \n at end of a debugging note
Ivan Zakharyaschev [Sat, 17 Feb 2018 03:19:56 +0000 (06:19 +0300)]
gtkprintbackendcups.c: fix \n at end of a debugging note

(cherry picked from commit 7461ceebe34bfc6f882145e89e1ba2f187d08303)

7 years agoUpdate Polish translation
Piotr Drąg [Sat, 17 Feb 2018 16:18:59 +0000 (17:18 +0100)]
Update Polish translation

7 years agoUpdate Croatian translation
gogo [Fri, 16 Feb 2018 23:01:50 +0000 (23:01 +0000)]
Update Croatian translation

7 years agoUpdate Croatian translation
gogo [Fri, 16 Feb 2018 22:57:44 +0000 (22:57 +0000)]
Update Croatian translation

7 years agoUpdate Brazilian Portuguese translation
Rafael Fontenelle [Fri, 16 Feb 2018 16:20:06 +0000 (16:20 +0000)]
Update Brazilian Portuguese translation

7 years agoUpdate Polish translation
Piotr Drąg [Thu, 15 Feb 2018 16:12:30 +0000 (17:12 +0100)]
Update Polish translation

7 years agoUpdate Swedish translation
Anders Jonsson [Thu, 15 Feb 2018 09:47:55 +0000 (09:47 +0000)]
Update Swedish translation

7 years agogtkplacessidebar: Rename favorites to starred
Carlos Soriano [Thu, 15 Feb 2018 09:10:18 +0000 (10:10 +0100)]
gtkplacessidebar: Rename favorites to starred

To have more consistency in the name. The same was done in Nautilus in
commit https://gitlab.gnome.org/GNOME/nautilus/commit/27b039b37c1149fc8abbef6a683dd939e6f2bf1f.

7 years agoUpdate POTFILES.in
Piotr Drąg [Thu, 15 Feb 2018 03:31:13 +0000 (04:31 +0100)]
Update POTFILES.in

7 years agoMerge branch 'wip/carlosg/imwayland-3-22' into 'gtk-3-22'
Matthias Clasen [Thu, 15 Feb 2018 03:03:02 +0000 (03:03 +0000)]
Merge branch 'wip/carlosg/imwayland-3-22' into 'gtk-3-22'

modules: Add wayland IM implementation

See merge request GNOME/gtk!5

7 years ago3.22.28
Matthias Clasen [Wed, 14 Feb 2018 17:48:01 +0000 (12:48 -0500)]
3.22.28

7 years agomodules: Add wayland IM implementation
Carlos Garnacho [Thu, 30 Nov 2017 17:53:26 +0000 (18:53 +0100)]
modules: Add wayland IM implementation

This IM context implementation goes through the gtk-text-input protocol,
leaving up to the compositor the actual interaction with IM engines. If
the protocol is not offered by the compositor, GTK+ will fallback to the
IMs as specified through GtkSettings.

7 years agogdk/wayland: Add GdkDisplay call to query available globals
Carlos Garnacho [Tue, 13 Feb 2018 13:28:44 +0000 (14:28 +0100)]
gdk/wayland: Add GdkDisplay call to query available globals

The internal known_globals hashtable is used to carry accounting for
interfaces that depend on others (as ordering is not guaranteed), extend
its usage so it also keeps track of unimplemented interfaces (here at
least).

The API call will then use this to allow querying the globals offered by
the compositor, it will be useful to determine whether we can use
text-input protocols or should fallback to other IMs.

7 years agoFix irregular gdk_frame_clock_get_frame_time
Daniel van Vugt [Fri, 15 Sep 2017 09:49:12 +0000 (17:49 +0800)]
Fix irregular gdk_frame_clock_get_frame_time

This fixes stuttering in animations that rely on the regularity of
gdk_frame_clock_get_frame_time.

https://bugzilla.gnome.org/show_bug.cgi?id=787665

BEFORE
gdkgears:
58 FPS and visibly stuttering
gnome-maps on a 59.95Hz monitor:
"paint" g_get_monotonic_time +17278μs, gdk_frame_clock_get_frame_time +17278μs
"paint" g_get_monotonic_time +17449μs, gdk_frame_clock_get_frame_time +17426μs
"paint" g_get_monotonic_time +17620μs, gdk_frame_clock_get_frame_time +17600μs

AFTER
gdkgears:
60 FPS and smoother
gnome-maps on a 59.95Hz monitor:
"paint" g_get_monotonic_time +18228μs, gdk_frame_clock_get_frame_time +16680μs
"paint" g_get_monotonic_time +15010μs, gdk_frame_clock_get_frame_time +16680μs
"paint" g_get_monotonic_time +17134μs, gdk_frame_clock_get_frame_time +16680μs

7 years agoFix annotation for gtk_tree_view_is_blank_at_pos()
Gniourf [Wed, 14 Feb 2018 12:51:22 +0000 (13:51 +0100)]
Fix annotation for gtk_tree_view_is_blank_at_pos()

The annotation (allow-none) is wrong. Since
gtk_tree_view_is_blank_at_pos() also calls
gtk_tree_view_get_path_at_pos(), the same fields should have the same
annotations.

7 years agoAdd missing private header to the build
Emmanuele Bassi [Wed, 14 Feb 2018 10:36:24 +0000 (10:36 +0000)]
Add missing private header to the build

Closes: #26
        https://bugzilla.gnome.org/show_bug.cgi?id=793442

7 years agoUpdate Estonian translation
Mart Raudsepp [Tue, 13 Feb 2018 22:20:41 +0000 (22:20 +0000)]
Update Estonian translation

7 years ago3.22.27
Matthias Clasen [Tue, 13 Feb 2018 21:09:13 +0000 (16:09 -0500)]
3.22.27

7 years agoUpdate Estonian translation
Mart Raudsepp [Tue, 13 Feb 2018 20:14:22 +0000 (20:14 +0000)]
Update Estonian translation

7 years agoMerge branch 'gtk-3-22' into 'gtk-3-22'
Matthias Clasen [Tue, 13 Feb 2018 13:27:05 +0000 (13:27 +0000)]
Merge branch 'gtk-3-22' into 'gtk-3-22'

calendar: Use the new "%OB" format if supported

See merge request GNOME/gtk!2

7 years agocalendar: Use the new "%OB" format if supported
Rafal Luzynski [Sat, 10 Feb 2018 13:07:56 +0000 (14:07 +0100)]
calendar: Use the new "%OB" format if supported

Due to the recent changes introduced in glibc 2.27 "%OB" is the
correct format to obtain a month name as used in the calendar
header.  The same rule has been working in BSD family (including
OS X) since 1990s.  This simple hack checks whether "%OB" is supported
at runtime and uses it if it is, falls back to the old "%B" otherwise.

Closes: #9
7 years agox11: Don't call XInput API for core events
Benjamin Otte [Mon, 12 Feb 2018 20:09:50 +0000 (21:09 +0100)]
x11: Don't call XInput API for core events

Fixes emacs crashing with XMing.

https://bugzilla.redhat.com/show_bug.cgi?id=1483942

7 years agoUpdate Estonian translation
Mart Raudsepp [Mon, 12 Feb 2018 00:38:59 +0000 (00:38 +0000)]
Update Estonian translation

7 years agoUpdate Greek translation
Efstathios Iosifidis [Sun, 11 Feb 2018 23:38:57 +0000 (23:38 +0000)]
Update Greek translation

7 years agodocs: Use the appropriate deprecation for properties
Emmanuele Bassi [Sun, 11 Feb 2018 16:45:53 +0000 (16:45 +0000)]
docs: Use the appropriate deprecation for properties

The GtkFontButton:font-name property has been deprecated, but it's using
a non-standard annotation in its documentation.

7 years agogtk: Cast g_object_ref() to the right type
Emmanuele Bassi [Sun, 11 Feb 2018 16:22:49 +0000 (16:22 +0000)]
gtk: Cast g_object_ref() to the right type

In GLib 2.56, g_object_ref() will check that you're assigning the return
value to a variable of the same type you're passing in.

7 years agoDo not use deprecated error trap API
Emmanuele Bassi [Sun, 11 Feb 2018 16:16:05 +0000 (16:16 +0000)]
Do not use deprecated error trap API

Use the X11-specific API instead.

7 years agotoolbar: Use gdk_display_flush()
Emmanuele Bassi [Sun, 11 Feb 2018 15:38:20 +0000 (15:38 +0000)]
toolbar: Use gdk_display_flush()

Instead of the deprecated gdk_flush().

7 years agoselection: Disable deprecation warnings
Emmanuele Bassi [Sun, 11 Feb 2018 15:37:22 +0000 (15:37 +0000)]
selection: Disable deprecation warnings

We are using a lot of deprecated API, and we know it.

Since the selection code is going to be replaced in GTK 4.0, there's no
real point in keeping the warnings enabled in 3.22.

7 years agoplacessidebar: Move declaration inside conditional block
Emmanuele Bassi [Sun, 11 Feb 2018 15:14:43 +0000 (15:14 +0000)]
placessidebar: Move declaration inside conditional block

The `sidebar` variable is only used by the conditional block for
libcloudprovider.

7 years agoIgnore deprecations for gdk_flush()
Emmanuele Bassi [Sun, 11 Feb 2018 15:13:06 +0000 (15:13 +0000)]
Ignore deprecations for gdk_flush()

When terminating the main loop, we're really trying to flush all
GdkDisplay connections, so it's actually a legitimate internal use
case.

7 years agodocs: Properly deprecate gdk_keymap_get_default()
Emmanuele Bassi [Sun, 11 Feb 2018 15:07:24 +0000 (15:07 +0000)]
docs: Properly deprecate gdk_keymap_get_default()

We're using a compiler annotation, but there's no deprecation notice in
the gtk-doc stanza.

7 years agoaccelgroup: Do not use deprecated API
Emmanuele Bassi [Sun, 11 Feb 2018 15:06:04 +0000 (15:06 +0000)]
accelgroup: Do not use deprecated API

Use gdk_keymap_get_for_display() with the default display, instead of
the deprecated gdk_keymap_get_default().

7 years agodeprecated: Cast g_object_ref() to the right type
Emmanuele Bassi [Sun, 11 Feb 2018 15:02:30 +0000 (15:02 +0000)]
deprecated: Cast g_object_ref() to the right type

In GLib 2.56, g_object_ref() will check that you're assigning the return
value to a variable of the same type you're passing in.

7 years agoa11y: Cast g_object_ref() to the right type
Emmanuele Bassi [Sun, 11 Feb 2018 15:01:43 +0000 (15:01 +0000)]
a11y: Cast g_object_ref() to the right type

In GLib 2.56, g_object_ref() will check that you're assigning the return
value to a variable of the same type you're passing in.

7 years agoRemove stray format character
Emmanuele Bassi [Sun, 11 Feb 2018 14:58:16 +0000 (14:58 +0000)]
Remove stray format character

We are not using any string in the debug message.

7 years agobroadway: Cast g_object_ref() to the right type
Emmanuele Bassi [Sun, 11 Feb 2018 14:52:22 +0000 (14:52 +0000)]
broadway: Cast g_object_ref() to the right type

In GLib 2.56, g_object_ref() will check that you're assigning the return
value to a variable of the same type you're passing in.

7 years agox11: Balance an error trap
Emmanuele Bassi [Sun, 11 Feb 2018 14:50:27 +0000 (14:50 +0000)]
x11: Balance an error trap

We are pushing an error trap, and never popping it from the stack.

7 years agoFix overzalous method annotations for drag and drop
Alban Browaeys [Sun, 20 Jan 2013 22:45:17 +0000 (23:45 +0100)]
Fix overzalous method annotations for drag and drop

Functional revert of commit 9c4892f2915.

Fixes introspection scanner warnings like:

  Warning: Gtk: gtk_drag_finish: Methods must belong to the same
  namespace as the class they belong to

That is, the gtk_drag_* functions cannot be methods as they have a
"GdkDragContext" as the instance parameter, and that is not a valid
type for the Gtk namespace.

This is not an introspected ABI change, as the generated introspection
data ignores the annotation.

Signed-off-by: Emmanuele Bassi <ebassi@gnome.org>
https://bugzilla.gnome.org/show_bug.cgi?id=692152

7 years agoUpdate Swedish translation
Anders Jonsson [Sun, 11 Feb 2018 19:17:29 +0000 (19:17 +0000)]
Update Swedish translation

7 years agoUpdate Polish translation
Piotr Drąg [Sun, 11 Feb 2018 15:25:35 +0000 (16:25 +0100)]
Update Polish translation

7 years agoAdd AGPL3-only licence to GtkAboutDialog
Emmanuele Bassi [Sun, 28 Jan 2018 14:40:58 +0000 (14:40 +0000)]
Add AGPL3-only licence to GtkAboutDialog

https://bugzilla.gnome.org/show_bug.cgi?id=792793

7 years agoUpdate Friulian translation
Fabio Tomat [Sun, 11 Feb 2018 09:43:48 +0000 (09:43 +0000)]
Update Friulian translation

7 years agoUpdate Friulian translation
Fabio Tomat [Sun, 11 Feb 2018 08:29:56 +0000 (08:29 +0000)]
Update Friulian translation

7 years agoUpdate Galician translation
Fran Dieguez [Sat, 10 Feb 2018 18:43:45 +0000 (18:43 +0000)]
Update Galician translation

7 years agoUpdate Galician translation
Fran Dieguez [Sat, 10 Feb 2018 18:41:38 +0000 (18:41 +0000)]
Update Galician translation

7 years agoUpdate Chinese (Taiwan) translation
Cheng-Chia Tseng [Sat, 10 Feb 2018 07:08:45 +0000 (07:08 +0000)]
Update Chinese (Taiwan) translation

7 years agoUpdate Chinese (Taiwan) translation
Cheng-Chia Tseng [Sat, 10 Feb 2018 06:54:24 +0000 (06:54 +0000)]
Update Chinese (Taiwan) translation

7 years agoMerge branch 'wip/carlosg/fix-issue-1' into 'gtk-3-22'
Carlos Garnacho [Fri, 9 Feb 2018 10:23:31 +0000 (10:23 +0000)]
Merge branch 'wip/carlosg/fix-issue-1' into 'gtk-3-22'

wayland: Improve EOF detection when reading selections

See merge request GNOME/gtk!1

7 years agowayland: Improve EOF detection when reading selections
Carlos Garnacho [Wed, 7 Feb 2018 17:30:18 +0000 (18:30 +0100)]
wayland: Improve EOF detection when reading selections

g_input_stream_read_bytes() roughly provides the same guarantees
than g_input_stream_read() wrt the number of bytes being possibly
read (i.e. it being a best effort, but no real guarantees).

Instead, rely on the 0-len read that we'd get at the end of the
transfer.

Fixes clipboard/DnD transfers possibly being cut short, resulting
on "Broken pipe" errors on the other side.

https://gitlab.gnome.org/GNOME/gtk/issues/1

Closes: #1
8 years agocssprovider: Only free bytes when they exist
Benjamin Otte [Mon, 5 Feb 2018 17:58:54 +0000 (18:58 +0100)]
cssprovider: Only free bytes when they exist

8 years agocss: Add a workaround for gtk_widget_override_font()
Benjamin Otte [Tue, 14 Nov 2017 03:03:57 +0000 (04:03 +0100)]
css: Add a workaround for gtk_widget_override_font()

The problem here is that the CSS machinery expects font sizes to be in
pixels, but gtk_widget_override_font() provides a value in point and the
CSS machinery has no ability to query the DPI and convert.

This patch changes the dconversion DPI we use from a hardcoded 96 to the
default screen's DPI, which should work better than before.
This will of course not listen to changes in the default screen's DPI,
but that shouldn't be a problem.

People who want to workaround this should use gtk_widget_override_font()
with a font that has an absolute size set via
pango_font_description_set_absolute_size (size * PANGO_SCALE *
                                          gdk_screen_get_resolution (screen));

https://bugzilla.gnome.org/show_bug.cgi?id=774248

8 years agoUpdate Finnish translation
Jiri Grönroos [Sun, 4 Feb 2018 09:14:54 +0000 (09:14 +0000)]
Update Finnish translation

8 years agoUpdate Catalan translation
Jordi Mas [Sat, 3 Feb 2018 20:31:58 +0000 (21:31 +0100)]
Update Catalan translation

8 years agoUpdate Catalan translation
Jordi Mas [Sat, 3 Feb 2018 20:31:08 +0000 (21:31 +0100)]
Update Catalan translation

8 years agoiconview: Fix a problem with the previous change
Matthias Clasen [Wed, 31 Jan 2018 16:44:26 +0000 (17:44 +0100)]
iconview: Fix a problem with the previous change

After commit ffef28a7e8d5ffef7de6a3baccb30b0021b6b0ff,
gtk-icon-browser was spewing critical warnings when
changing sections. Avoid that by respecting the return
value of gtk_tree_model_get_iter.

8 years agowayland: Add support for BTN_STYLUS3
Jason Gerecke [Tue, 10 Oct 2017 16:17:43 +0000 (09:17 -0700)]
wayland: Add support for BTN_STYLUS3

BTN_STYLUS3 is defined by the Linux 4.15 kernel and is sent when the
third button on a stylus is pressed. At the moment, only Wacom's "Pro
Pen 3D" has three stylus buttons. Pressing this button triggers a button
8 event to be sent under X11, so we use the same mapping here.

https://bugzilla.gnome.org/show_bug.cgi?id=790033

8 years agotests: Don't add testswitch twice in Makefile.am
Kalev Lember [Sat, 27 Jan 2018 13:49:34 +0000 (14:49 +0100)]
tests: Don't add testswitch twice in Makefile.am

8 years agomenu: Guard against NULL toplevel
Timm Bäder [Tue, 26 Sep 2017 05:10:39 +0000 (07:10 +0200)]
menu: Guard against NULL toplevel

This can happen, as indicated by GtkMenu explicitly connecting to
::destroy of its toplevel window. Do the same thing in GtkComboBox.

8 years agoUpdate Afrikaans translation
Pieter Schalk Schoeman [Wed, 24 Jan 2018 11:40:40 +0000 (11:40 +0000)]
Update Afrikaans translation

8 years agoUpdate Afrikaans translation
Pieter Schalk Schoeman [Wed, 24 Jan 2018 11:36:50 +0000 (11:36 +0000)]
Update Afrikaans translation

8 years agoWidget: Don’t call reset() on NULL EventController
Daniel Boles [Thu, 18 Jan 2018 00:51:24 +0000 (00:51 +0000)]
Widget: Don’t call reset() on NULL EventController

GtkGesture is a GtkEventController. gtk_event_controller_dispose() calls
_gtk_widget_remove_controller(). That NULLs the pointer-to-Controller in
our EventControllerData but does not delete said ECData from our GList.

Subsequently, if that same Widget gets unparent()ed, that method calls
unset_state_flags(), which leads to doing reset_controllers() if we are
insensitive. Now, unlike most most other loops over the GList of ECData,
reset_controllers() does not skip nodes whose pointer-to-Controller is
NULL. So, we call gtk_event_controller_reset(NULL) and get a CRITICAL.

This surfaced in a gtkmm program. The Gesture is destroyed before the
Widget. The Widget then gets dispose()d, which calls unparent()… boom.
I didn’t find an MCVE yet but would hope this logic is correct anyway:

The simplest fix is to make the loop in gtk_widget_reset_controllers()
skip GList nodes with a NULL Controller pointer, like most other such
loops, so we avoid passing the NULL to gtk_event_controller_reset().

In other, live cases, _gtk_widget_run_controllers() loops over the GList
and removes/frees nodes having NULL Controllers, so that should suffice.
But this clearly was not getting a chance to happen in the failing case.

https://bugzilla.gnome.org/show_bug.cgi?id=792624

8 years agoUpdate Esperanto translation
GNOME Translation Robot [Mon, 22 Jan 2018 18:34:48 +0000 (18:34 +0000)]
Update Esperanto translation

8 years agoGtkGestureMultiPress: check event state before emiting released signal
Juan Pablo Ugarte [Mon, 22 Jan 2018 14:20:14 +0000 (11:20 -0300)]
GtkGestureMultiPress: check event state before emiting released signal

Fix bug 771986 "Inconsistent 'row-activated' signal emission before \
drag'n'drop, 'activate-on-single-click'=TRUE, 'reorderable'=TRUE"

8 years agoUpdate Hungarian translation
Balázs Úr [Sun, 21 Jan 2018 22:53:01 +0000 (22:53 +0000)]
Update Hungarian translation

8 years agoUpdate Hungarian translation
Balázs Úr [Sun, 21 Jan 2018 22:49:00 +0000 (22:49 +0000)]
Update Hungarian translation

8 years agoiconview: Fix updates with filter models
Matthias Clasen [Fri, 19 Jan 2018 15:04:18 +0000 (16:04 +0100)]
iconview: Fix updates with filter models

Filter models rely on views taking a ref on every node
they care about. GtkIconView was not doing that. Amazingly,
this has never shown up in a bug so far, until I spotted
the fallout in gnome-font-viewer.

8 years agoAdd a test for ::row-changed vs node refs
Matthias Clasen [Fri, 19 Jan 2018 15:02:42 +0000 (16:02 +0100)]
Add a test for ::row-changed vs node refs

Test that filter models propagate ::row-changed if there is
an external reference on the node, and not otherwise. This
is showing up in buggy icon view behaviour, where the icon
view is not redrawing if the content changes in a model that
is below a filter model.

8 years agofont chooser: Fix property notification
Matthias Clasen [Fri, 19 Jan 2018 00:58:27 +0000 (19:58 -0500)]
font chooser: Fix property notification

We must notify the font and font-desc properties when the
list selection changes, and return NULL values for them
when there is no selection in the list.

8 years agofont chooser dialog: Improve sensitivity handling
Matthias Clasen [Fri, 19 Jan 2018 00:48:12 +0000 (19:48 -0500)]
font chooser dialog: Improve sensitivity handling

Enable the select button when a font is selected, as it should be.

8 years agofont chooser: Small revision of the UI
Matthias Clasen [Fri, 19 Jan 2018 00:36:56 +0000 (19:36 -0500)]
font chooser: Small revision of the UI

Show only the font names in the list, in their own
font, in order to make the list less noisy.

8 years agogtk/queryimmodules.c: Make the output deterministic.
Chris Lamb [Fri, 25 Aug 2017 15:24:38 +0000 (16:24 +0100)]
gtk/queryimmodules.c: Make the output deterministic.

Whilst working on the Reproducible Builds effort [0], we noticed that
queryimmodules generates non-reproducible output as it iterates over the
filesystem without sorting.

Patch attached.

 [0] https://reproducible-builds.org/

Signed-off-by: Chris Lamb <lamby@debian.org>
https://bugzilla.gnome.org/show_bug.cgi?id=786528

8 years agotree model sort: Fix set_sort_column
Matthias Clasen [Tue, 16 Jan 2018 23:16:21 +0000 (18:16 -0500)]
tree model sort: Fix set_sort_column

We were failing to change the sort order for the
default sort column in some cases. Fix that, and
add a testcase for this issue.

https://bugzilla.gnome.org/show_bug.cgi?id=792459

Add a testcase for the previous fix

8 years agotree model sort: Fix initial default sort func
Matthias Clasen [Tue, 16 Jan 2018 23:15:13 +0000 (18:15 -0500)]
tree model sort: Fix initial default sort func

gtk_tree_sortable_has_default_sort_func should return
FALSE initially.

8 years agoUpdate Polish translation
Piotr Drąg [Mon, 15 Jan 2018 21:16:51 +0000 (22:16 +0100)]
Update Polish translation

8 years agoemoji: Skip overly wide fallback rendering
Matthias Clasen [Mon, 18 Dec 2017 03:09:08 +0000 (22:09 -0500)]
emoji: Skip overly wide fallback rendering

Some emoji fonts (such as Emoji One), render Emoji sequences
such as some of the family variations using multiple individual
glyphs. This rendering is too wide and breaks our grid layout.
Therefore, we will just skip any sequence whose rendering is
more than twice as wide as a simple smiley.

8 years agoGdkMonitor: Fix link to nonexistent Display method
Daniel Boles [Sat, 13 Jan 2018 20:03:37 +0000 (20:03 +0000)]
GdkMonitor: Fix link to nonexistent Display method

There is no gdk_display_get_monitors(). Instead, we have to use
gdk_display_get_n_monitors() and gdk_display_get_monitor(int).

8 years agoHeaderBar: Explain use in conjunction w/ GtkWindow
Daniel Boles [Fri, 12 Jan 2018 22:56:11 +0000 (22:56 +0000)]
HeaderBar: Explain use in conjunction w/ GtkWindow

After hinting how good GtkHeaderBar is for GtkWindow, let’s link to
the latter and indicate how users can make the two work together.

8 years agoWindow: Mention GtkHeaderBar in set_titlebar() doc
Daniel Boles [Fri, 12 Jan 2018 21:26:36 +0000 (21:26 +0000)]
Window: Mention GtkHeaderBar in set_titlebar() doc

This is the typical thing passed here and what most users want, so we
should mention it here, rather than requiring users to figure it out.

8 years agoWindow: Clarify resize() doc about titlebar widget
Daniel Boles [Fri, 12 Jan 2018 21:22:11 +0000 (21:22 +0000)]
Window: Clarify resize() doc about titlebar widget

Clarify the reference to HeaderBar, as it applies to any custom title
widget; HeaderBar is only the most common one used. Also, fix a typo.

8 years agowayland: Use portable sed syntax in gdk/wayland/Makefile.am
Ting-Wei Lan [Sat, 6 Jan 2018 18:57:38 +0000 (02:57 +0800)]
wayland: Use portable sed syntax in gdk/wayland/Makefile.am

Replace non-portable \<, \>, \+ with \(, \), \{, \} to avoid build
failure on systems not using GNU sed.

https://bugzilla.gnome.org/show_bug.cgi?id=792279

8 years agoGtkCssImageSurface: add cache for the last drawn size.
Juan Pablo Ugarte [Fri, 29 Dec 2017 18:21:52 +0000 (15:21 -0300)]
GtkCssImageSurface: add cache for the last drawn size.

Keep a copy of the scaled image to speed up rendering multiple times
the image at the same size.

8 years agoAdwaita: Fix typo lenght => length in SASS
Daniel Boles [Tue, 9 Jan 2018 09:46:11 +0000 (09:46 +0000)]
Adwaita: Fix typo lenght => length in SASS

8 years agoUpdate Icelandic translation
Sveinn í Felli [Mon, 8 Jan 2018 17:47:11 +0000 (17:47 +0000)]
Update Icelandic translation

8 years agoUpdate Icelandic translation
Sveinn í Felli [Mon, 8 Jan 2018 17:39:46 +0000 (17:39 +0000)]
Update Icelandic translation

8 years agoScrolledWindow: add() before remove() in snippet
Daniel Boles [Sun, 7 Jan 2018 16:47:24 +0000 (16:47 +0000)]
ScrolledWindow: add() before remove() in snippet

We can't remove() a child widget that wasn't already add()ed, of course.

8 years agofile button: Don't leak rows
Matthias Clasen [Fri, 5 Jan 2018 16:09:04 +0000 (11:09 -0500)]
file button: Don't leak rows

The file chooser button manually manages the memory of
data in its model, so it needs to explicitly free the
rows.

8 years agoRemove a dead macro
Matthias Clasen [Thu, 4 Jan 2018 23:23:36 +0000 (18:23 -0500)]
Remove a dead macro

GCs have been dead for a long time.

8 years agoMake a few more code samples compile
Timm Bäder [Wed, 3 Jan 2018 16:10:21 +0000 (17:10 +0100)]
Make a few more code samples compile

8 years agoentry: Remove some questionable code examples
Timm Bäder [Wed, 3 Jan 2018 15:39:31 +0000 (16:39 +0100)]
entry: Remove some questionable code examples

8 years agowidget: Make all code snippets compile
Timm Bäder [Wed, 3 Jan 2018 15:34:49 +0000 (16:34 +0100)]
widget: Make all code snippets compile

8 years agoliststore: Make a code snippet compile
Timm Bäder [Wed, 3 Jan 2018 14:01:12 +0000 (15:01 +0100)]
liststore: Make a code snippet compile