gtk4.git
11 years agoscrolledwindow: Be smarter at invalidating overshoot areas
Carlos Garnacho [Sun, 27 Jul 2014 11:39:51 +0000 (13:39 +0200)]
scrolledwindow: Be smarter at invalidating overshoot areas

The previous way to invalidate was meant to work on the overshoot window
so it wouldn't be as taxing. Since the overshoot window is gone, this
would invalidate way more than intended. So constrain invalidated areas
to the sides where overshoot is happenning at that moment.

https://bugzilla.gnome.org/show_bug.cgi?id=735223

11 years agotooltip: Use subsurfaces on wayland
Carlos Garnacho [Fri, 22 Aug 2014 11:13:50 +0000 (13:13 +0200)]
tooltip: Use subsurfaces on wayland

This places tooltips in the right place, and avoids having those attempt
to steal the focus.

https://bugzilla.gnome.org/show_bug.cgi?id=735531

11 years agoAdwaita: border rationalization and cleanup
Lapo Calamandrei [Thu, 28 Aug 2014 13:42:17 +0000 (15:42 +0200)]
Adwaita: border rationalization and cleanup

Rely only on border-style for border, take out border-width and
border-style from drawing functions.

11 years agocss: Implement font-stretch
Emmanuele Bassi [Thu, 28 Aug 2014 11:50:49 +0000 (12:50 +0100)]
css: Implement font-stretch

The font-stretch CSS property is defined in the Level 3 CSS Fonts
module, available at:

  http://dev.w3.org/csswg/css-fonts/#propdef-font-stretch

It allows defining a normal, condensed, or expanded face to the font
description. Pango already supports it, so this is literally just the
CSS parser machinery needed to bridge our CSS to the FontDescription
API.

https://bugzilla.gnome.org/show_bug.cgi?id=735593

11 years agoAdwaita: take named colors out of _colors.scss
Lapo Calamandrei [Thu, 28 Aug 2014 11:21:14 +0000 (13:21 +0200)]
Adwaita: take named colors out of _colors.scss

To be able to use sass for shipped css, importing _colors.scss
and _drawings.scss.

11 years agoAdwaita: rationalize font sizes.
Lapo Calamandrei [Thu, 28 Aug 2014 10:48:32 +0000 (12:48 +0200)]
Adwaita: rationalize font sizes.

...making GtkSwitch font-size relative in the process.

11 years agotheme: Specify the weight, not the font, for header bar titles
Emmanuele Bassi [Thu, 28 Aug 2014 10:17:10 +0000 (11:17 +0100)]
theme: Specify the weight, not the font, for header bar titles

Using `font: Bold` will override the font with whatever the system
resolves a `Bold` font name to, including the font's own default size.

We want the font weight to be bold, but we want the default font name
and size to come from the GtkSettings:font-name property, like we do for
the rest of the system.

11 years agoAdwaita: Make expanders turn again
Matthias Clasen [Thu, 28 Aug 2014 02:17:29 +0000 (22:17 -0400)]
Adwaita: Make expanders turn again

The theme needed some small adjustments for the recent expander
state changes ACTIVE to CHECKED.

11 years agoUse CHECKED instead of ACTIVE for expanded expanders
Benjamin Otte [Wed, 27 Aug 2014 19:25:53 +0000 (21:25 +0200)]
Use CHECKED instead of ACTIVE for expanded expanders

https://bugzilla.gnome.org/show_bug.cgi?id=733967

11 years agoconfigure: Move AC_USE_SYSTEM_EXTENSIONS up
Benjamin Otte [Wed, 27 Aug 2014 19:24:27 +0000 (21:24 +0200)]
configure: Move AC_USE_SYSTEM_EXTENSIONS up

We don't want to get warnings from autoconf about
AC_USE_SYSTEM_EXTENSIONS coming after AC_COMPILE_IFELSE.

11 years agoquartz-dnd: make sure to set source window
Jesse van den Kieboom [Wed, 27 Aug 2014 12:08:49 +0000 (14:08 +0200)]
quartz-dnd: make sure to set source window

https://bugzilla.gnome.org/show_bug.cgi?id=658722

11 years agoAdwaita: no edge hilight for popover checks/radios
Lapo Calamandrei [Wed, 27 Aug 2014 16:29:32 +0000 (18:29 +0200)]
Adwaita: no edge hilight for popover checks/radios

11 years agoreftests: Fix Makefile
Benjamin Otte [Wed, 27 Aug 2014 15:51:09 +0000 (17:51 +0200)]
reftests: Fix Makefile

with non-installed tests the build would get an empty $(reftestdir)
which would screw up the LDFLAGS.

An rpath seems to be required to make libtool build a shared object.
Without an rpath line, it only builds a static object.

https://bugzilla.gnome.org/show_bug.cgi?id=735401

11 years agogtk-demo: Don't use mad mime type guessing
Benjamin Otte [Tue, 26 Aug 2014 02:18:52 +0000 (04:18 +0200)]
gtk-demo: Don't use mad mime type guessing

... for displaying resources. Instead use the proven and way more
reliable method of trial and error.

It's less code and more portable for a start.
But most of all it displays PNM files as text if you fail to compile
the gdk-pixbuf loader for it.

11 years agomodifierstyle: Remove the changed signal
Benjamin Otte [Mon, 25 Aug 2014 14:35:11 +0000 (16:35 +0200)]
modifierstyle: Remove the changed signal

It's been unused since last commit

11 years agowidget: Don't connect to GtkModifierStyle::changed
Benjamin Otte [Mon, 25 Aug 2014 14:33:33 +0000 (16:33 +0200)]
widget: Don't connect to GtkModifierStyle::changed

Changes to the modifier style will be picked up by the
GtkStyleProviderPrivate::changed signal. And that signal will actually
invalidate properly.

11 years agowayland: Ensure the parent is committed after repositioning subsurfaces
Carlos Garnacho [Wed, 27 Aug 2014 10:03:28 +0000 (12:03 +0200)]
wayland: Ensure the parent is committed after repositioning subsurfaces

Subsurface position is deemed part of the state of the parent surface, so
ensure wl_surface_commit() happens on the parent surface if none is
scheduled, so the repositioning takes place.

11 years agotreeview: Reset multipress controller after 2nd click
Carlos Garnacho [Wed, 27 Aug 2014 12:36:53 +0000 (14:36 +0200)]
treeview: Reset multipress controller after 2nd click

The pre-gestures code used to compare the current button press with the
previous one on !activate_on_single_click, and unset the previous event
data so ::row-activated would be emitted every 2 clicks.

So do the same with the multipress gesture and reset it after every 2nd
click to have ::row-activated emitted multiple times while manic clicking.

https://bugzilla.gnome.org/show_bug.cgi?id=735353

11 years agothreads: Do not release the GDK lock if it hasn't been acquired yet
Emmanuele Bassi [Tue, 26 Aug 2014 11:15:06 +0000 (12:15 +0100)]
threads: Do not release the GDK lock if it hasn't been acquired yet

Since GLib ≥ 2.41, attempting to release an unlocked mutex will abort(),
as it happens on most systems already.

Given the lack of proper documentation on how to use GDK with threads,
there is code in the wild that does:

    gdk_threads_init ();
    gdk_init ();

    ...

    gtk_main ();

instead of the idiomatically correct:

    gdk_threads_init ();
    gdk_threads_enter ();

    gtk_init ();

    ...

    gtk_main ();

    ...

    gdk_threads_leave ();

Which means that gtk_main() will try to release the GDK lock, and thus
trigger an error from GLib.

we cannot really fix all the wrong code everywhere, and since it does
not cost us anything, we can work around the issue inside GDK itself, by
trying to acquire the GDK lock inside gdk_threads_leave() with
trylock().

https://bugzilla.gnome.org/show_bug.cgi?id=735428

11 years agobuild: make reftests work without installed tests
Benjamin Otte [Mon, 25 Aug 2014 18:39:10 +0000 (20:39 +0200)]
build: make reftests work without installed tests

The libreftest.so module needs to be built in all cases.
Without installed tests, it needs to not be installed though.

https://bugzilla.gnome.org/show_bug.cgi?id=735401

11 years agoCleanup invalid comments (duplication of words)
Ikey Doherty [Sun, 24 Aug 2014 00:26:36 +0000 (01:26 +0100)]
Cleanup invalid comments (duplication of words)

https://bugzilla.gnome.org/show_bug.cgi?id=735301

Signed-off-by: Ikey Doherty <michael.i.doherty@intel.com>
11 years agobutton: Track gesture point to maintain priv->in_button
Carlos Garnacho [Wed, 20 Aug 2014 18:12:35 +0000 (20:12 +0200)]
button: Track gesture point to maintain priv->in_button

This makes the active state work invariably with both mouse/touch, and
regardless of X11 pointer emulation being friendly and sending crossing
events for the emulated pointer events in the latter.

This makes GtkButtons' active state look correct when pressing on
touchscreens on wayland.

https://bugzilla.gnome.org/show_bug.cgi?id=731380

11 years agoentrycompletion: Mark popup window as subsurface on wayland
Carlos Garnacho [Thu, 21 Aug 2014 15:36:45 +0000 (17:36 +0200)]
entrycompletion: Mark popup window as subsurface on wayland

This will make the completion window use GDK_WINDOW_SUBSURFACE and be
shown relative to the parent window without involving compositor grabs.

https://bugzilla.gnome.org/show_bug.cgi?id=695504

11 years agogtkwindow: Add private API to mark as subsurface
Carlos Garnacho [Thu, 21 Aug 2014 15:34:09 +0000 (17:34 +0200)]
gtkwindow: Add private API to mark as subsurface

This is a rather hackish way to let GTK+ widgets declare popup windows
as subsurfaces, so they may work on wayland without the need of xdg_popup,
and without many changes yet on the GTK+ side.

https://bugzilla.gnome.org/show_bug.cgi?id=695504

11 years agoAdwaita: menu checks and radios making sense
Lapo Calamandrei [Tue, 26 Aug 2014 18:32:36 +0000 (20:32 +0200)]
Adwaita: menu checks and radios making sense

11 years agoAdwaita: set a background for GtkTextView
Lapo Calamandrei [Tue, 26 Aug 2014 17:54:41 +0000 (19:54 +0200)]
Adwaita: set a background for GtkTextView

Needed for gedit line numbers, overridden by .view {} normally

11 years agoAdwaita: minor fixes to GtkLabel
Lapo Calamandrei [Tue, 26 Aug 2014 17:24:44 +0000 (19:24 +0200)]
Adwaita: minor fixes to GtkLabel

11 years agoAdwaita: take rtl into consideration for GtkAssistant
Lapo Calamandrei [Tue, 26 Aug 2014 16:02:28 +0000 (18:02 +0200)]
Adwaita: take rtl into consideration for GtkAssistant

11 years agoAdwaita: GtkAssistant tweakings
Lapo Calamandrei [Tue, 26 Aug 2014 15:51:45 +0000 (17:51 +0200)]
Adwaita: GtkAssistant tweakings

11 years agoAdwaita: treeview checks/radios minor fix
Lapo Calamandrei [Tue, 26 Aug 2014 14:56:22 +0000 (16:56 +0200)]
Adwaita: treeview checks/radios minor fix

Superfluous icon shadow in checked hover state removed.

11 years agoAdwaita: in app notification fixes
Lapo Calamandrei [Tue, 26 Aug 2014 14:54:19 +0000 (16:54 +0200)]
Adwaita: in app notification fixes

11 years agoscalebutton: Handle smooth scroll events
Carlos Garnacho [Fri, 22 Aug 2014 17:15:57 +0000 (19:15 +0200)]
scalebutton: Handle smooth scroll events

This has the double effect to make scale button truly smooth, and make
scrolling work on wayland, since all scrolling is "smooth" there.

https://bugzilla.gnome.org/show_bug.cgi?id=720269

11 years agowayland: Fix x/y coordinate arguments on wl_data_device events
Carlos Garnacho [Thu, 21 Aug 2014 17:01:32 +0000 (19:01 +0200)]
wayland: Fix x/y coordinate arguments on wl_data_device events

Those are wl_fixed_t, not int.

https://bugzilla.gnome.org/show_bug.cgi?id=697855

11 years agowayland: ensure gdk_device_manager_get_client_pointer() returns a master pointer
Carlos Garnacho [Wed, 13 Aug 2014 19:16:39 +0000 (21:16 +0200)]
wayland: ensure gdk_device_manager_get_client_pointer() returns a master pointer

The internal ordering of devices isn't quite guaranteed, so ensure we're returning
the right one.

https://bugzilla.gnome.org/show_bug.cgi?id=697855

11 years agoRevert "Adwaita: notebook background"
Jakub Steiner [Tue, 26 Aug 2014 11:46:54 +0000 (13:46 +0200)]
Revert "Adwaita: notebook background"

Not a good solution to the problem at hand (gedit line numbering)

This reverts commit fde03c1f8174a16120264eb689415b6f7535fc81.

11 years agowayland: unset sequence if the xdg surface is moved/resized through touch
Carlos Garnacho [Wed, 20 Aug 2014 18:23:58 +0000 (20:23 +0200)]
wayland: unset sequence if the xdg surface is moved/resized through touch

The latest implicit grab serial is used in order to start the compositor
grab, If it belongs to a touch event, remove that touch sequence, as the
rest of the sequence will be gone for good.

This avoids stale sequences (and implicit grab info) after a window is
moved/resized.

https://bugzilla.gnome.org/show_bug.cgi?id=731380

11 years agowayland: Add internal API to unset a touch implicit grab
Carlos Garnacho [Wed, 20 Aug 2014 18:22:42 +0000 (20:22 +0200)]
wayland: Add internal API to unset a touch implicit grab

This removes both the wayland specific accounting, and the Gdk implicit
grab tracking.

https://bugzilla.gnome.org/show_bug.cgi?id=731380

11 years agowayland: Update the display serial on touch down/up events
Carlos Garnacho [Wed, 20 Aug 2014 18:52:39 +0000 (20:52 +0200)]
wayland: Update the display serial on touch down/up events

https://bugzilla.gnome.org/show_bug.cgi?id=734374

11 years agowayland: Improve implicit grab serial retrieval
Carlos Garnacho [Wed, 13 Aug 2014 19:05:58 +0000 (21:05 +0200)]
wayland: Improve implicit grab serial retrieval

_gdk_wayland_device_get_button_press_serial() has been replaced by
_gdk_wayland_device_get_implicit_grab_serial(), which takes a touch/pointer
event and figures out the relevant serial, and
_gdk_wayland_device_get_last_implicit_grab_serial() which returns
the most recent serial.

The button press serial was currently used when operating popping up
xdg_shell/surface popups and window menus, so this is now touch aware, of
some sort.

https://bugzilla.gnome.org/show_bug.cgi?id=734374

11 years agowayland: Store per-touch touch_down serial
Carlos Garnacho [Wed, 13 Aug 2014 19:05:12 +0000 (21:05 +0200)]
wayland: Store per-touch touch_down serial

https://bugzilla.gnome.org/show_bug.cgi?id=734374

11 years agowayland: rename internal touch management functions
Carlos Garnacho [Wed, 13 Aug 2014 18:59:20 +0000 (20:59 +0200)]
wayland: rename internal touch management functions

These functions arent quite related to the device manager, so rename these
more conveniently.

https://bugzilla.gnome.org/show_bug.cgi?id=734374

11 years agowayland: Protect against invalid keymaps gotten from the compositor
Carlos Garnacho [Mon, 25 Aug 2014 13:55:22 +0000 (15:55 +0200)]
wayland: Protect against invalid keymaps gotten from the compositor

If the compositor sends a keymap that fails on "compilation",
xkb_keymap_new_from_string() returns NULL, which makes xkb_state_new()
crash when assuming there is a keymap.

In these cases, gdk must remain with a xkb_state to handle modifiers/keys
properly, so warn about the invalid keymap string, and keep the previous
keymap (currently initialized to "us")

https://bugzilla.gnome.org/show_bug.cgi?id=735389

11 years agowayland: create a wl_subsurface interface for GDK_WINDOW_SUBSURFACE windows
Carlos Garnacho [Thu, 21 Aug 2014 15:31:50 +0000 (17:31 +0200)]
wayland: create a wl_subsurface interface for GDK_WINDOW_SUBSURFACE windows

This subsurface is currently dependent on the transient_for parent, so the
subsurface is repositioned relative to it.

https://bugzilla.gnome.org/show_bug.cgi?id=729215

11 years agowayland: Acquire wl_subcompositor interface
Carlos Garnacho [Thu, 21 Aug 2014 15:27:26 +0000 (17:27 +0200)]
wayland: Acquire wl_subcompositor interface

This will be needed for GDK_WINDOW_SUBSURFACE windows.

https://bugzilla.gnome.org/show_bug.cgi?id=729215

11 years agogdk: Add GDK_WINDOW_SUBSURFACE window type
Carlos Garnacho [Thu, 21 Aug 2014 15:26:18 +0000 (17:26 +0200)]
gdk: Add GDK_WINDOW_SUBSURFACE window type

This window type can only be used on wayland so far, so NULL is returned
if it's attempted to be used on any other windowing backend.

https://bugzilla.gnome.org/show_bug.cgi?id=729215

11 years agowayland: Make toplevels' X/Y coordinates be 0
Carlos Garnacho [Thu, 21 Aug 2014 15:39:09 +0000 (17:39 +0200)]
wayland: Make toplevels' X/Y coordinates be 0

To all effects each window has its own "root" coordinates system, so set
toplevels at 0,0 in that coordinate system, so root coordinate calculations
are locally right.

https://bugzilla.gnome.org/show_bug.cgi?id=729215

11 years agoAdwaita: notebook background
Jakub Steiner [Tue, 26 Aug 2014 10:32:05 +0000 (12:32 +0200)]
Adwaita: notebook background

11 years agoTextView: clear the background using the default bg color
Paolo Borelli [Sun, 24 Aug 2014 15:58:36 +0000 (17:58 +0200)]
TextView: clear the background using the default bg color

Try to tidy up how the background is set on the textview:
 *) the .view class should be applied only to the text window, not
    to the margins
 *) when setting the background on the margins we must use .left etc
 *) use context_set_bg instead of manually setting the color

https://bugzilla.gnome.org/show_bug.cgi?id=735368

11 years agoGtkTreeView: Avoid a crash with empty trees
Matthias Clasen [Tue, 26 Aug 2014 04:05:14 +0000 (00:05 -0400)]
GtkTreeView: Avoid a crash with empty trees

The gesture code was assuming that the rb tree exists, which
will not be the case if the treeview is empty.

https://bugzilla.gnome.org/show_bug.cgi?id=735300

11 years agoAdwaita: define styling for location-bar toolbars
Jakub Steiner [Mon, 25 Aug 2014 17:20:49 +0000 (19:20 +0200)]
Adwaita: define styling for location-bar toolbars

- same inset style as searchbars

11 years agoAdwaita: legible labels in headerbars in selection mode
Jakub Steiner [Mon, 25 Aug 2014 13:24:44 +0000 (15:24 +0200)]
Adwaita: legible labels in headerbars in selection mode

11 years agoAdwaita: selection-mode checked togglebuttons
Jakub Steiner [Mon, 25 Aug 2014 13:18:58 +0000 (15:18 +0200)]
Adwaita: selection-mode checked togglebuttons

- exposed in gnome-contacts for example.

11 years agoAdwaita: selected text in GtkLabel
Jakub Steiner [Mon, 25 Aug 2014 12:15:32 +0000 (14:15 +0200)]
Adwaita: selected text in GtkLabel

https://bugzilla.gnome.org/show_bug.cgi?id=735310

11 years agoinspector: fix a crash in the CSS editor
Sébastien Wilmet [Sun, 24 Aug 2014 18:27:13 +0000 (20:27 +0200)]
inspector: fix a crash in the CSS editor

gtk_css_section_get_end_position() can return a position one byte beyond
the end of the last line.

gtk_text_buffer_get_iter_at_line_index() accepts only valid
line_number/byte_index pairs. Another solution is to make the
GtkTextBuffer function less strict, by returning a boolean if the exact
position was found.

The CSS parser should also be fixed to always return valid positions.
But it's better to have a safety net in the CSS editor, just in case (a
warning could be print).

https://bugzilla.gnome.org/show_bug.cgi?id=735337

11 years agogdkwindow-quartz: Support native fullscreen mode
Jesse van den Kieboom [Sat, 23 Aug 2014 14:04:46 +0000 (16:04 +0200)]
gdkwindow-quartz: Support native fullscreen mode

This implements native fullscreen mode on OS X >= 10.7. This patch
adds tracking of the fullscreen mode if externally modified, as well
as toggling the native fullscreen mode of the window as needed.

11 years agoRef the popover before activating the item
Garrett Regier [Sat, 23 Aug 2014 13:55:07 +0000 (06:55 -0700)]
Ref the popover before activating the item

Otherwise the item could cause the popover to be destroyed,
for example a Quit item, and cause a warning to be printed.

11 years agoGtkAboutDialog: Drop visibility-notify-event handler
Matthias Clasen [Sat, 23 Aug 2014 05:25:35 +0000 (01:25 -0400)]
GtkAboutDialog: Drop visibility-notify-event handler

The signal is deprecated, and the handler is only helping
for a really tiny corner case anyway.

https://bugzilla.gnome.org/show_bug.cgi?id=734961

11 years agoChange gdk-pixbuf minimum required version to 2.30.0
Ikey Doherty [Fri, 22 Aug 2014 22:31:17 +0000 (23:31 +0100)]
Change gdk-pixbuf minimum required version to 2.30.0

https://bugzilla.gnome.org/show_bug.cgi?id=735253

Signed-off-by: Ikey Doherty <michael.i.doherty@intel.com>
11 years agoSome release notes for 3.14
Matthias Clasen [Sat, 23 Aug 2014 03:11:11 +0000 (23:11 -0400)]
Some release notes for 3.14

11 years agoAdwaita: use osd style for in-app notifications
Jakub Steiner [Fri, 22 Aug 2014 16:16:23 +0000 (18:16 +0200)]
Adwaita: use osd style for in-app notifications

11 years agoxdg-shell: Bump version
Jasper St. Pierre [Fri, 22 Aug 2014 16:05:21 +0000 (12:05 -0400)]
xdg-shell: Bump version

11 years agoAdwaita: added a workaround to make sass behave
Lapo Calamandrei [Fri, 22 Aug 2014 15:25:25 +0000 (17:25 +0200)]
Adwaita: added a workaround to make sass behave

11 years agoGtkColorChooser: Make color plane react to mouse again
Matthias Clasen [Fri, 22 Aug 2014 15:21:52 +0000 (11:21 -0400)]
GtkColorChooser: Make color plane react to mouse again

Fallout from changing the defaults for gesture properties.

11 years agoAdwaita: dark active buttons for osd
Lapo Calamandrei [Fri, 22 Aug 2014 14:55:23 +0000 (16:55 +0200)]
Adwaita: dark active buttons for osd

11 years agoAdwaita: @define-color "compatible" with sass 3.4.0
Lapo Calamandrei [Fri, 22 Aug 2014 12:53:39 +0000 (14:53 +0200)]
Adwaita: @define-color "compatible" with sass 3.4.0

11 years agowayland: Add defines to convert between event sequences and touch slots
Carlos Garnacho [Fri, 22 Aug 2014 11:33:14 +0000 (13:33 +0200)]
wayland: Add defines to convert between event sequences and touch slots

This makes it clearer in code what's going on there.
GDK_SLOT_TO_EVENT_SEQUENCE isn't used yet, but will be useful in the
future.

11 years agotextview: Unset bubble timeout ID right after requesting the clipboard
Carlos Garnacho [Thu, 21 Aug 2014 15:43:10 +0000 (17:43 +0200)]
textview: Unset bubble timeout ID right after requesting the clipboard

The timeout ID used to be unset after we got the targets from the
clipboard, but there's still a moment between the clipboard request and
the GDK_SELECTION_NOTIFY event that the ID points to an already gone
timeout.

11 years agox11: Fix possible leak of GdkEvents
Carlos Garnacho [Wed, 20 Aug 2014 18:53:22 +0000 (20:53 +0200)]
x11: Fix possible leak of GdkEvents

If the filter didn't return GDK_EVENT_TRANSLATE, the event struct would be
left unused and leaked.

11 years agoAdwaita: kill a bogus background-color
Lapo Calamandrei [Fri, 22 Aug 2014 11:07:44 +0000 (13:07 +0200)]
Adwaita: kill a bogus background-color

11 years agoAdwaita: first round at osd revamp
Lapo Calamandrei [Fri, 22 Aug 2014 10:55:30 +0000 (12:55 +0200)]
Adwaita: first round at osd revamp

...buttons, gtkscale and toolbars, more to follow...

11 years agoGtkFileChooserWidget: Use ratio in formatted times
Matthias Clasen [Fri, 22 Aug 2014 05:47:21 +0000 (01:47 -0400)]
GtkFileChooserWidget: Use ratio in formatted times

This matches what we do in other places now.

11 years agoAvoid picking up long description as translator comment
Matthias Clasen [Fri, 22 Aug 2014 05:18:53 +0000 (01:18 -0400)]
Avoid picking up long description as translator comment

Rearrange a few lines to help xgettext avoid picking up the
long description of GtkFileChooserButton as translator comment.

11 years agoAvoid picking up long description as translator comment
Matthias Clasen [Fri, 22 Aug 2014 05:10:24 +0000 (01:10 -0400)]
Avoid picking up long description as translator comment

Rearrange a few lines to help xgettext avoid picking up the
long description of GtkAboutDialog as translator comment.

11 years agoRemove an extraneous translatable string
Matthias Clasen [Fri, 22 Aug 2014 05:01:19 +0000 (01:01 -0400)]
Remove an extraneous translatable string

11 years agoAdwaita: Insensitive labels
Matthias Clasen [Fri, 22 Aug 2014 04:42:52 +0000 (00:42 -0400)]
Adwaita: Insensitive labels

There is an insensitive label in widget-factory which was
not getting any different from the sensitive label next
to it. With this patch, it does, again.

11 years agoAdwaita: Add assistant sidebar theming
Matthias Clasen [Fri, 22 Aug 2014 03:07:57 +0000 (23:07 -0400)]
Adwaita: Add assistant sidebar theming

This is just enough to draw the sidebar background and
markt he current page as active.

11 years agoGtkAssistant: Don't use a draw cb for the sidebar
Matthias Clasen [Fri, 22 Aug 2014 03:07:04 +0000 (23:07 -0400)]
GtkAssistant: Don't use a draw cb for the sidebar

We can do this with css nowadays.

11 years agogtk: Fix return-annotation of gtk_get_default_language to "transfer none"
Rico Tzschichholz [Thu, 21 Aug 2014 21:10:32 +0000 (23:10 +0200)]
gtk: Fix return-annotation of gtk_get_default_language to "transfer none"

11 years agoGtkPageSetup: Improve formatting of size
Matthias Clasen [Thu, 21 Aug 2014 21:01:54 +0000 (17:01 -0400)]
GtkPageSetup: Improve formatting of size

Use U+00D7 MULTIPLICATION SIGN and U+200A HAIR SPACE instead of plain
old 'x' and ' ', following https://wiki.gnome.org/Design/OS/Typography.

Sadly, many fonts don't have space variations, so Pango/harfbuzz fall
back to using the regular space glyph anyway.

11 years agoAdd tests for parsing int64/uin64 in GtkBuilder
Matthias Clasen [Thu, 21 Aug 2014 18:21:34 +0000 (14:21 -0400)]
Add tests for parsing int64/uin64 in GtkBuilder

This tests the fix in the previous commit.

11 years agoGtkBuilder: Support parsing int64/uin64 values
Matthias Clasen [Thu, 21 Aug 2014 18:20:28 +0000 (14:20 -0400)]
GtkBuilder: Support parsing int64/uin64 values

This was missing; the code to parse these numbers was there,
but the switch didn't have cases for these types.

https://bugzilla.gnome.org/show_bug.cgi?id=602443

11 years agoInclude config.h first
Matthias Clasen [Thu, 21 Aug 2014 17:19:26 +0000 (13:19 -0400)]
Include config.h first

Otherwise, we miss the _GNU_SOURCE #define that configure
puts there.

11 years agoSimplify _gtk_text_buffer_get_line_log_attrs()
Sébastien Wilmet [Wed, 20 Aug 2014 17:06:05 +0000 (19:06 +0200)]
Simplify _gtk_text_buffer_get_line_log_attrs()

NULL was returned in case of an empty last line. Every users needed to
special-case this. Now it will return the expected result: char_len of 0
with one PangoLogAttr.

In compute_log_attrs(), 'paragraph' will be the empty string "" with
'char_len' == 0.
pango_get_log_attrs() works fine with an empty string, it will return
one correct PangoLogAttr (because there is one text position for the
empty string).

It fixes the unit tests for gtk_text_iter_is_cursor_position().

https://bugzilla.gnome.org/show_bug.cgi?id=156164

11 years agotextbuffer: unit tests for the empty last line
Sébastien Wilmet [Wed, 20 Aug 2014 14:57:05 +0000 (16:57 +0200)]
textbuffer: unit tests for the empty last line

For functions using _gtk_text_buffer_get_line_log_attrs():
- gtk_text_buffer_backspace()
- some gtk_text_iter functions (word/sentence/cursor boundaries)

As the FIXME comments show, there is a bug with
gtk_text_iter_is_cursor_position() for an empty last line.

https://bugzilla.gnome.org/show_bug.cgi?id=156164

11 years agoGtkPopover: Fix rtl positioning
Matthias Clasen [Thu, 21 Aug 2014 15:57:19 +0000 (11:57 -0400)]
GtkPopover: Fix rtl positioning

There was some confusion between unflipped and flipped positions.
Both final_position and current_position are meant to be unflipped,
and get_effective_position() needs to be applied to them to get
a flipped position. _gtk_window_set_popover_position() also expects
an unflipped position.

https://bugzilla.gnome.org/show_bug.cgi?id=735014

11 years agoAdwaita: try to be nicer on performances with entry shadows
Lapo Calamandrei [Thu, 21 Aug 2014 12:41:29 +0000 (14:41 +0200)]
Adwaita: try to be nicer on performances with entry shadows

11 years agoUpdated POTFILES.skip
Piotr Drąg [Thu, 21 Aug 2014 13:33:36 +0000 (15:33 +0200)]
Updated POTFILES.skip

11 years agoAdwaita: no edge highlight for backdrop check/radios
Jakub Steiner [Thu, 21 Aug 2014 13:03:32 +0000 (15:03 +0200)]
Adwaita: no edge highlight for backdrop check/radios

11 years agoAdwaita: remove edge highlight form the assets
Jakub Steiner [Thu, 21 Aug 2014 12:32:41 +0000 (14:32 +0200)]
Adwaita: remove edge highlight form the assets

- using icon-shadow now, allows to treat special cases
  without introducing assets

11 years agoUpdated Traditional Chinese translation(Hong Kong and Taiwan)
Chao-Hsiung Liao [Thu, 21 Aug 2014 10:50:35 +0000 (18:50 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)

11 years agoAdwaita: some separators simplification
Lapo Calamandrei [Thu, 21 Aug 2014 10:24:01 +0000 (12:24 +0200)]
Adwaita: some separators simplification

Accidentally fixing the backdrop separators in sidebars turning
black

11 years agoAdwaita: inline toolbar buttons fixes
Lapo Calamandrei [Thu, 21 Aug 2014 09:52:20 +0000 (11:52 +0200)]
Adwaita: inline toolbar buttons fixes

Properly styled single buttons in inline toolbars and fixed
a superflous border chiming in in backdrop.

11 years agoDeprecate GTK_CHECK_BACKEND in favour of PKG_CHECK_MODULES
Simon McVittie [Wed, 6 Aug 2014 10:23:53 +0000 (11:23 +0100)]
Deprecate GTK_CHECK_BACKEND in favour of PKG_CHECK_MODULES

This is easy to do with pkg-config too.

Bug: https://bugzilla.gnome.org/show_bug.cgi?id=641638
Bug-Debian: https://bugs.debian.org/756476
Reviewed-by: Matthias Clasen
11 years agoDeprecate AM_PATH_GTK_3_0 in favour of PKG_CHECK_MODULES
Simon McVittie [Wed, 6 Aug 2014 10:23:23 +0000 (11:23 +0100)]
Deprecate AM_PATH_GTK_3_0 in favour of PKG_CHECK_MODULES

That's what GNOME applications use.

Bug: https://bugzilla.gnome.org/show_bug.cgi?id=641638
Bug-Debian: https://bugs.debian.org/756476
Reviewed-by: Matthias Clasen
11 years agoUse AC_USE_SYSTEM_EXTENSIONS to get _GNU_SOURCE, _XOPEN_SOURCE etc.
Simon McVittie [Sun, 17 Aug 2014 14:00:54 +0000 (15:00 +0100)]
Use AC_USE_SYSTEM_EXTENSIONS to get _GNU_SOURCE, _XOPEN_SOURCE etc.

Similar to Bug #684123 in GLib.

Bug: https://bugzilla.gnome.org/show_bug.cgi?id=641638
Bug-Debian: https://bugs.debian.org/756476
Reviewed-by: Matthias Clasen
11 years agocss: Round shadow extents properly
Benjamin Otte [Wed, 20 Aug 2014 21:19:59 +0000 (23:19 +0200)]
css: Round shadow extents properly

Otherwise drawing will be clipped.

Testcase included

11 years agoroundedbox: Only grow border-radius if there is one
Benjamin Otte [Wed, 20 Aug 2014 22:10:43 +0000 (00:10 +0200)]
roundedbox: Only grow border-radius if there is one

This is relevant for shadow spread and goes in line with the web.

Testcase included.

11 years agolabel: Respect box-shadow when computing clip
Benjamin Otte [Wed, 20 Aug 2014 22:02:29 +0000 (00:02 +0200)]
label: Respect box-shadow when computing clip

Testclase included.

Also reorder headers to be alphabetic.

11 years agowidget: Make _gtk_set_simple_clip() take an optional content clip
Benjamin Otte [Wed, 20 Aug 2014 21:49:00 +0000 (23:49 +0200)]
widget: Make _gtk_set_simple_clip() take an optional content clip

11 years agostylecontext: Remove GTK_CSS_CHANGE_FORCE_INVALIDATE
Benjamin Otte [Wed, 20 Aug 2014 02:39:54 +0000 (04:39 +0200)]
stylecontext: Remove GTK_CSS_CHANGE_FORCE_INVALIDATE

This was introduced as a hackish way in 3.6 to make font updates
propagate properly. But since then, font handling has been changed and
this flag is no longer necessary.