gtk4.git
11 years agoAdd an expander reftest
Matthias Clasen [Tue, 3 Jun 2014 02:33:20 +0000 (22:33 -0400)]
Add an expander reftest

This test check that resizing the window when expanding
the expander yields the same end result as having the
expander expanded to begin with. The test uses the inhibit
mechanism introduced in the previous commit.

11 years agoAdd a way to delay snapshots in reftests
Matthias Clasen [Tue, 3 Jun 2014 02:31:49 +0000 (22:31 -0400)]
Add a way to delay snapshots in reftests

This adds an inhibit api that code from the reftest module
can use to delay the taking of the snapshot. Also refactor
the code in gtk-reftest to use the inhibit mechanism for
its own delaying of the snapshot until after the first
expose.

11 years agoForgotten file
Matthias Clasen [Tue, 3 Jun 2014 01:43:04 +0000 (21:43 -0400)]
Forgotten file

11 years agoAdd a reftest for flipping icons
Matthias Clasen [Tue, 3 Jun 2014 01:41:32 +0000 (21:41 -0400)]
Add a reftest for flipping icons

Taking advantage of our new direction flipping ability.

11 years agowindow: Only check for the event widget if clicked on the "content" region
Carlos Garnacho [Mon, 2 Jun 2014 16:04:36 +0000 (18:04 +0200)]
window: Only check for the event widget if clicked on the "content" region

Every button press/release event reaching the the multipress gesture in GtkWindow
and happening in the "title" region must be handled, regardless of the event widget.
Children there wanting the event(s) for themselves are (and were always) expected
to stop event propagation.

So the only place to check for the event widget's "window-dragging" style property
is the "content" region, which matches the pre-gestures behavior.

This fixes some issues with sequences being mistakenly claimed (and events not
propagated further) on situations it shouldn't.

11 years agowindow: Handle manually bubbled events
Carlos Garnacho [Mon, 2 Jun 2014 15:55:26 +0000 (17:55 +0200)]
window: Handle manually bubbled events

The multipress gesture must react to either direct events on the
GtkWindow (special cased through _gtk_widget_check_handle_wm_event),
or bubbled events from child widgets. Ensure bubbled events go
through the gesture, those are fed manually to make sure events are
only handled once, in either one or other place. The implicit grab
will ensure that doesn't change mid-action.

11 years agowindow: Refactor _gtk_window_check_handle_wm_event()
Carlos Garnacho [Mon, 2 Jun 2014 15:53:53 +0000 (17:53 +0200)]
window: Refactor _gtk_window_check_handle_wm_event()

Separate the event widget check from event handling

11 years agotreeview: Ensure the event is consumed on ::row-activate
Carlos Garnacho [Mon, 2 Jun 2014 15:47:10 +0000 (17:47 +0200)]
treeview: Ensure the event is consumed on ::row-activate

Otherwise the event is possibly handled, but still propagated further anyway.
Ensure the event is consumed by claiming the current sequence on the
GtkGestureMultiPress::pressed handler.

11 years agotreeview: restrict the multipress gesture to GDK_BUTTON_PRIMARY
Carlos Garnacho [Sun, 1 Jun 2014 18:20:08 +0000 (20:20 +0200)]
treeview: restrict the multipress gesture to GDK_BUTTON_PRIMARY

::row-activated only used to be triggered by GDK_BUTTON_PRIMARY, so make
the multipress gesture handling this now to be only triggered by that same
button.

https://bugzilla.gnome.org/show_bug.cgi?id=731020

11 years agoRevert "Use minimum/natural size semantics"
Matthias Clasen [Sun, 1 Jun 2014 14:16:01 +0000 (10:16 -0400)]
Revert "Use minimum/natural size semantics"

This reverts commit 6d53c2339f79baa0b295ecc614f41f9daab2e132.

https://bugzilla.gnome.org/show_bug.cgi?id=731054 showed some
major regressions caused by this commit.

11 years agoRevert "Better resize of expandable columns"
Matthias Clasen [Sun, 1 Jun 2014 14:15:41 +0000 (10:15 -0400)]
Revert "Better resize of expandable columns"

This reverts commit 0050d469b592ec571a5940f1ab1d842a76905e17.

Conflicts:
gtk/gtktreeview.c

11 years agoAvoid another revert conflict
Matthias Clasen [Sun, 1 Jun 2014 14:11:14 +0000 (10:11 -0400)]
Avoid another revert conflict

Moving this bit up a few lines lets the following revert apply
cleanly.

11 years agoRevert "treeview: Remove unneeded checks"
Matthias Clasen [Sun, 1 Jun 2014 14:09:52 +0000 (10:09 -0400)]
Revert "treeview: Remove unneeded checks"

This reverts commit a158a2aa48c8023f99963642cc2657bff207b82d.

The commit conflicts with a later revert, so move it out of the way.

11 years agoAdd more gestures tests
Matthias Clasen [Sat, 31 May 2014 22:46:43 +0000 (18:46 -0400)]
Add more gestures tests

Some of these sequences look wrong/questionable to me.

11 years agogestures test: make more readable
Matthias Clasen [Sat, 31 May 2014 16:54:36 +0000 (12:54 -0400)]
gestures test: make more readable

11 years agogestures: Add a test for grouping
Matthias Clasen [Sat, 31 May 2014 16:48:00 +0000 (12:48 -0400)]
gestures: Add a test for grouping

I'm not entirely convinced the test produces the expected
outcome, here.

11 years agoExpand gesture tests a big more
Matthias Clasen [Sat, 31 May 2014 16:12:10 +0000 (12:12 -0400)]
Expand gesture tests a big more

Print out sequence state changes, and cancellation, and verify that
claiming a gestures causes the expected state changes.

11 years agotextview: Only trigger drag gesture on GDK_BUTTON_PRIMARY
Carlos Garnacho [Sat, 31 May 2014 12:54:32 +0000 (14:54 +0200)]
textview: Only trigger drag gesture on GDK_BUTTON_PRIMARY

This gesture was only meant to react on GDK_BUTTON_PRIMARY (either
through real pointer events, or implicitly assumed from touch events),
as it used to behave before gestures. Otherwise the gtk_drag_begin*()
call assumes being triggered by button 1, and the drag misbehaves
because that button isn't really in the state mask.

https://bugzilla.gnome.org/show_bug.cgi?id=731016

11 years agoUpdated POTFILES.skip
Piotr Drąg [Sat, 31 May 2014 13:08:21 +0000 (15:08 +0200)]
Updated POTFILES.skip

11 years agogtk-reftest: Add GMODULE_CFLAGS/LIBS
Benjamin Otte [Sat, 31 May 2014 12:39:27 +0000 (14:39 +0200)]
gtk-reftest: Add GMODULE_CFLAGS/LIBS

They aren't included in GTK_DEP_CFLAGS/LIBS so they need to be added
manually.

11 years agotestsuite: put installed tests in the right place
Matthias Clasen [Sat, 31 May 2014 04:16:07 +0000 (00:16 -0400)]
testsuite: put installed tests in the right place

Follow the prevailing practice and put the installed tests
into $libdir/installed-tests/$PACKAGE instead of
$pkglibdir/installed-tests.

11 years agoinspector: better type detection for resources
Matthias Clasen [Sat, 31 May 2014 03:45:11 +0000 (23:45 -0400)]
inspector: better type detection for resources

Pass the name to g_content_type_guess. With this, most
CSS gets sniffed as C source code.

11 years agogtk-reftest: Don't warn needlessly
Matthias Clasen [Sat, 31 May 2014 03:30:56 +0000 (23:30 -0400)]
gtk-reftest: Don't warn needlessly

Only warn about the --direction argument if the option was
actually specified.

11 years agoinspector: Go to the right tab when changing objects
Matthias Clasen [Sat, 31 May 2014 03:21:13 +0000 (23:21 -0400)]
inspector: Go to the right tab when changing objects

When going from attribute mapping to model, it makes most sense
to go directly to the data tab, and when going from an action
name to the owner, we want to show the actions tab. Make it so.

11 years agoinspector: Allow jumping to actions
Matthias Clasen [Sat, 31 May 2014 02:56:12 +0000 (22:56 -0400)]
inspector: Allow jumping to actions

The property editor for action-name properties now offers
to go to the object where the action is defined.

11 years agoinspector: Trivial string change
Matthias Clasen [Sat, 31 May 2014 02:20:31 +0000 (22:20 -0400)]
inspector: Trivial string change

'Size Groups' looks better than 'Sizegroups'.

11 years agoreftests: Add reftest for direction change
Benjamin Otte [Sat, 31 May 2014 03:17:51 +0000 (05:17 +0200)]
reftests: Add reftest for direction change

This reftest makes use of the new feature to add signal handlers.

It adds a libreftest.so module containing all the code for the reftests.
When adding a test named reftest.ui, please keep code contained in a
source file names reftest.c and add that file to Makefile.am.

https://bugzilla.gnome.org/show_bug.cgi?id=730833

11 years agoUpdate GtkImage icons when the direction is changed
Debarshi Ray [Tue, 27 May 2014 18:41:50 +0000 (20:41 +0200)]
Update GtkImage icons when the direction is changed

Don't call gtk_widget_set_state_flags(), it filters the direction flags.
Instead, call gtk_widget_update_state_flags() directly.

https://bugzilla.gnome.org/show_bug.cgi?id=730833

11 years agogtk-reftest: Clarify --direction argument parameters
Benjamin Otte [Sat, 31 May 2014 02:52:53 +0000 (04:52 +0200)]
gtk-reftest: Clarify --direction argument parameters

... and print a proper error message when the parameter is not
recognized.

11 years agogtk-reftest: Allow loading code for reftests
Benjamin Otte [Fri, 30 May 2014 20:36:10 +0000 (22:36 +0200)]
gtk-reftest: Allow loading code for reftests

When connecting signal names, gtk-reftest now allows you to use a colon
in the signal handler name like so:
  module:function_name
where module is a module loaded from the same directory (or the .libs
subdirectory for compatibility with uninstalled libtool) as the running
test and the function is resolved in that module. Of course, normal
function names work as before.

11 years agoreftests: Add a reftest with signals
Benjamin Otte [Fri, 30 May 2014 02:33:06 +0000 (04:33 +0200)]
reftests: Add a reftest with signals

11 years agoreftests: Automatically connect signals
Benjamin Otte [Fri, 30 May 2014 02:31:40 +0000 (04:31 +0200)]
reftests: Automatically connect signals

11 years agobuilder: Cleanup
Benjamin Otte [Thu, 29 May 2014 22:21:18 +0000 (00:21 +0200)]
builder: Cleanup

Use the usual way of providing user data to vfuncs:
- Don't allocate it, use the stack
- Use CamelCase for naming the struct

11 years agoinspector: Handle attribute mapping changes better
Matthias Clasen [Sat, 31 May 2014 01:47:40 +0000 (21:47 -0400)]
inspector: Handle attribute mapping changes better

When the mapping is changed, notify the property and
queue a draw on the widget. This makes both the inspector
and the app update their display.

11 years agoinspector: some code reshuffling
Matthias Clasen [Sat, 31 May 2014 01:35:39 +0000 (21:35 -0400)]
inspector: some code reshuffling

11 years agoinspector: Support saving CSS
Matthias Clasen [Sat, 31 May 2014 00:35:43 +0000 (20:35 -0400)]
inspector: Support saving CSS

We have a file chooser, lets use it.

11 years agoGtkPrintUnixDialog: Remove some dead code
Matthias Clasen [Fri, 30 May 2014 22:50:07 +0000 (18:50 -0400)]
GtkPrintUnixDialog: Remove some dead code

Automatic template children get cleaned up and NULLed out in destroy,
the g_clear_pointer() calls in finalize never do anything.

11 years agogtk: port many nullability annotation fixes from Vala bindings
Evan Nemerson [Mon, 26 May 2014 16:13:02 +0000 (09:13 -0700)]
gtk: port many nullability annotation fixes from Vala bindings

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agoMake notebook content redraw during drag
Matthias Clasen [Fri, 30 May 2014 17:38:26 +0000 (13:38 -0400)]
Make notebook content redraw during drag

The extra condition here that caused the current child to
not redraw during reordering was introduced in f383e1f1
during the port to ::draw, but was not explained in the
commit message, and removing it has no obvious negative
effect.

https://bugzilla.gnome.org/show_bug.cgi?id=730767

11 years agoGtkBuilder: Ensure types are fully initialized
Matthias Clasen [Fri, 30 May 2014 16:39:07 +0000 (12:39 -0400)]
GtkBuilder: Ensure types are fully initialized

Just calling get_type() does not ensure that the signals, properties
and everything else gets set up properly. Ensure it is, by calling
g_type_class_ref() before using the type. This fixes the testcase
added in the previous commit.

11 years agoAdd a failing builder testcase
Matthias Clasen [Fri, 30 May 2014 16:09:43 +0000 (12:09 -0400)]
Add a failing builder testcase

This shows <signal> failing because class_init hasn't been run
yet.

11 years agoGtkActionable: Update the docs
Matthias Clasen [Fri, 30 May 2014 11:05:50 +0000 (07:05 -0400)]
GtkActionable: Update the docs

Actionable is no longer restricted to "app" and "win" actions.

11 years agoGtkScrolledWindow: Add some NULL checks back
Matthias Clasen [Fri, 30 May 2014 01:52:40 +0000 (21:52 -0400)]
GtkScrolledWindow: Add some NULL checks back

Turns out these were not quite redundant, gtk_container_forall
may be called very early during construction.

https://bugzilla.gnome.org/show_bug.cgi?id=730924

11 years agoRevert "GtkMisc: Keep gtk_misc_set_alignment working"
Matthias Clasen [Thu, 29 May 2014 14:30:57 +0000 (10:30 -0400)]
Revert "GtkMisc: Keep gtk_misc_set_alignment working"

This reverts commit aabc356b80fab025d460f61b6c08743b2c10e20d.

Labels are back to looking at x/yalign (for now).

11 years agoeventcontroller: Ensure controllers stay alive during event emission
Carlos Garnacho [Wed, 28 May 2014 17:39:07 +0000 (19:39 +0200)]
eventcontroller: Ensure controllers stay alive during event emission

Those might trigger the destruction of some widget that would dispose the
event controller while the event is still being handled, so keep an extra
ref on the controller during event processing.

11 years agodocs: Add some more information in the "input handling model" chapter
Carlos Garnacho [Wed, 28 May 2014 13:50:06 +0000 (15:50 +0200)]
docs: Add some more information in the "input handling model" chapter

Grabs/Touch/Gestures are now fairly well covered. Only keyboard handling
is left.

11 years agogtk: fix several out argument annotations
Evan Nemerson [Mon, 26 May 2014 16:37:13 +0000 (09:37 -0700)]
gtk: fix several out argument annotations

These mostly just switch from allow-none to optional, nullable, or
both, as necessary.

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agogtk: add missing ownership annotations ported from Vala
Evan Nemerson [Mon, 26 May 2014 16:42:23 +0000 (09:42 -0700)]
gtk: add missing ownership annotations ported from Vala

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agogtk: add missing type annotations ported from Vala bindings
Evan Nemerson [Mon, 26 May 2014 16:41:21 +0000 (09:41 -0700)]
gtk: add missing type annotations ported from Vala bindings

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agogtk: fix annotation syntax and missing some missing annotations
Evan Nemerson [Mon, 26 May 2014 16:32:32 +0000 (09:32 -0700)]
gtk: fix annotation syntax and missing some missing annotations

These changes clean up various errors and omissions resulting from
either slightly incorrect G-I/gtk-doc syntax or missing documentation
blocks.

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agogtk: port missing array annotations from Vala bindings
Evan Nemerson [Mon, 26 May 2014 16:29:56 +0000 (09:29 -0700)]
gtk: port missing array annotations from Vala bindings

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agogtk: fix many callback annotations to include closure information
Evan Nemerson [Mon, 26 May 2014 16:24:00 +0000 (09:24 -0700)]
gtk: fix many callback annotations to include closure information

Without this information introspection-based consumers don't realize
they can include context information, but instead think that they
receive an extra gpointer argument (which they don't know how to
handle).

https://bugzilla.gnome.org/show_bug.cgi?id=730745

11 years agoAdd some gesture tests
Matthias Clasen [Tue, 27 May 2014 23:15:40 +0000 (19:15 -0400)]
Add some gesture tests

This test starts to assemble tests related to event propagation,
interaction with traditional event signal handlers, claiming of
events, etc.

11 years agoForgotten file
Matthias Clasen [Tue, 27 May 2014 21:56:39 +0000 (17:56 -0400)]
Forgotten file

Oops.

11 years agoBump version
Matthias Clasen [Tue, 27 May 2014 21:49:22 +0000 (17:49 -0400)]
Bump version

11 years agoMove extract-strings to its own directory
Matthias Clasen [Tue, 27 May 2014 19:54:38 +0000 (15:54 -0400)]
Move extract-strings to its own directory

This will hopefully help resolve the circular dependency between
libgtk linking against inspector/libgtkinspector and inspector/
needing extract-strings from gtk/.

I didn't preserve the EXEEXT decorations in this operation -
automake gave me stern warnings about it, so I just dropped them
all. Somebody who cross-builds GTK+ will have to reconstruct this.

11 years agoFix distcheck
Matthias Clasen [Tue, 27 May 2014 19:14:02 +0000 (15:14 -0400)]
Fix distcheck

11 years ago3.13.2
Matthias Clasen [Tue, 27 May 2014 17:23:07 +0000 (13:23 -0400)]
3.13.2

11 years agoDon't assert in gtk_event_controller_constructed
Matthias Clasen [Tue, 27 May 2014 18:30:47 +0000 (14:30 -0400)]
Don't assert in gtk_event_controller_constructed

This prevents some of our generic object implementation tests
from working with gesture objects. Instead, add g_return_if_fail
checks in all the gesture constructors.

11 years agoDocumentation fixes
Matthias Clasen [Tue, 27 May 2014 17:40:57 +0000 (13:40 -0400)]
Documentation fixes

Add a few missing symbols, and match up parameters in headers
and doc comments.

11 years agowayland: Use event->key.time for setting key event time
Kristian Høgsberg [Mon, 26 May 2014 21:45:21 +0000 (14:45 -0700)]
wayland: Use event->key.time for setting key event time

We were using event->button.time before. That works because it's part of
the common event header, but it's wrong.

11 years agowayland: Remove unused XSERVER_TIME_IS_LATER macro
Kristian Høgsberg [Mon, 26 May 2014 20:43:48 +0000 (13:43 -0700)]
wayland: Remove unused XSERVER_TIME_IS_LATER macro

11 years agoSilence gtk-doc warnings
Matthias Clasen [Tue, 27 May 2014 16:45:04 +0000 (12:45 -0400)]
Silence gtk-doc warnings

11 years agoRevert "GtkImage: Drop use of GtkMisc::x/yalign"
Matthias Clasen [Tue, 27 May 2014 17:08:25 +0000 (13:08 -0400)]
Revert "GtkImage: Drop use of GtkMisc::x/yalign"

This reverts commit 54d5b426fca71dcda22ea5467054fa801556c20a.

Conflicts:
gtk/gtkimage.c

11 years agoRevert "GtkLabel: Drop use of GtkMisc::x/yalign"
Matthias Clasen [Tue, 27 May 2014 17:07:03 +0000 (13:07 -0400)]
Revert "GtkLabel: Drop use of GtkMisc::x/yalign"

This reverts commit b320c198b6d325c35e6f6d32f4f841a913efa956.

11 years agoRevert "Fix up GtkLabel and GtkImage sizing"
Matthias Clasen [Tue, 27 May 2014 17:04:43 +0000 (13:04 -0400)]
Revert "Fix up GtkLabel and GtkImage sizing"

This reverts commit 3b30cfa344ebc035ae7318dda5cfa11b8ab3aee7.

11 years agoRevert "More label and image sizing fixes"
Matthias Clasen [Tue, 27 May 2014 17:04:34 +0000 (13:04 -0400)]
Revert "More label and image sizing fixes"

This reverts commit 68c1e83cf0fd7432dbb151ff351ce225708e3ffb.

11 years agowindow: Do not raise border windows on each size allocation
Carlos Garnacho [Tue, 27 May 2014 09:44:45 +0000 (11:44 +0200)]
window: Do not raise border windows on each size allocation

This prevents these windows from being raised above popovers, which
is about the only child widget that may overlap with those anyway.

https://bugzilla.gnome.org/show_bug.cgi?id=730589

11 years agotreeview: Make the multipress gesture listen only to bin_window
Carlos Garnacho [Tue, 27 May 2014 15:37:07 +0000 (17:37 +0200)]
treeview: Make the multipress gesture listen only to bin_window

This gesture is meant to handle row-activated signaling, so make the
gesture ignore events from any other window.

11 years agobutton: Claim the sequence on multipress::pressed
Carlos Garnacho [Tue, 27 May 2014 15:35:36 +0000 (17:35 +0200)]
button: Claim the sequence on multipress::pressed

In order to avoid the event from bubbling, as it used to (not) do.

11 years agogesture: Fix wrong assert in gtk_gesture_set_window
Carlos Garnacho [Tue, 27 May 2014 15:33:31 +0000 (17:33 +0200)]
gesture: Fix wrong assert in gtk_gesture_set_window

The window must be part of the event controller widget, so
fix the inverted logic.

11 years agoeventcontroller: Hide more details
Carlos Garnacho [Tue, 27 May 2014 10:41:30 +0000 (12:41 +0200)]
eventcontroller: Hide more details

Signals are left hidden, so the only entry point to handle_event/reset
is API.

11 years agogtk: Update GtkGesture users to the GtkPropagationPhase semantics change
Carlos Garnacho [Tue, 27 May 2014 10:38:49 +0000 (12:38 +0200)]
gtk: Update GtkGesture users to the GtkPropagationPhase semantics change

Mainly doing s/TARGET/BUBBLE/ on the fully ported widgets, but GtkTreeView
where the double click handler has moved to GTK_PHASE_TARGET so it runs
parallelly to the still existing event handlers.

11 years agowidget: Change GtkCapturePhase behavior in event controllers' handling
Carlos Garnacho [Tue, 27 May 2014 10:14:52 +0000 (12:14 +0200)]
widget: Change GtkCapturePhase behavior in event controllers' handling

Previously, there would be globally just a capture and a bubble phase,
with the event just going down the hierarchy once, and the up once.
GTK_PHASE_TARGET actually meaning "run within event handlers", so in
a hierarchy of 3 widgets, emission would be:

  Capture(C)
  Capture(B)
  Capture(A)
  Target(A) (if event handlers allow)
  Bubble(A)
  Target(B) (if event handlers allow)
  Bubble(B)
  Target(C) (if event handlers allow)
  Bubble(C)

This commit changes this behavior and uses GTK_PHASE_TARGET in a less
misleading way, running only on the widget that was meant to receive
the event. And GTK_PHASE_BUBBLE has taken over the execution place of
GTK_PHASE_TARGET, so the emission remains:

  Capture(C)
  Capture(B)
  Capture(A)
  Target(A)
  Bubble(A) (if event handlers allow)
  Bubble(B) (...)
  Bubble(C) (...)

As it was, GTK_PHASE_BUBBLE was useful for running event controllers
paralelly to event handlers, without modifying a single line in those.
For those mixed scenarios, Any of the other phases will have to be
used at discretion, or the event handlers eventually changed to chain
up and let the default event handlers in GtkWidget to be run.

11 years agowindow: Avoid double event processing in the multipress event controller
Carlos Garnacho [Tue, 27 May 2014 10:14:17 +0000 (12:14 +0200)]
window: Avoid double event processing in the multipress event controller

The events to those are fed outside the regular event propagation scheme,
through _gtk_window_check_handle_wm_event(), so set the controller to
GTK_PHASE_NONE so events aren't processed first manually, and then
automatically.

11 years agogesturezoom: Just return a double in get_scale_delta()
Carlos Garnacho [Mon, 26 May 2014 12:32:59 +0000 (14:32 +0200)]
gesturezoom: Just return a double in get_scale_delta()

Checking whether the gesture is active is a responsibility of the caller.

11 years agogesturerotate: Just return a double in get_angle_delta()
Carlos Garnacho [Mon, 26 May 2014 12:29:17 +0000 (14:29 +0200)]
gesturerotate: Just return a double in get_angle_delta()

Whether the gesture is active or recognized should be checked before in
callers, or just not used naively.

11 years agogesture: Replace gtk_gesture_attach/detach with event controller API
Carlos Garnacho [Mon, 26 May 2014 12:02:30 +0000 (14:02 +0200)]
gesture: Replace gtk_gesture_attach/detach with event controller API

Event controllers now auto-attach, and the GtkCapturePhase only determines
when are events dispatched, but all controllers are managed by the widget wrt
grabs.

All callers have been updated.

11 years agogtk: Update callers of GtkGesturePan orientation
Carlos Garnacho [Mon, 26 May 2014 10:38:06 +0000 (12:38 +0200)]
gtk: Update callers of GtkGesturePan orientation

11 years agoeventcontroller: Make the event-mask property and methods private.
Carlos Garnacho [Mon, 26 May 2014 10:08:55 +0000 (12:08 +0200)]
eventcontroller: Make the event-mask property and methods private.

This will rarely have any use if no subclassing is allowed, so just make it
private for our own.

11 years agogtk: Replace GtkPanOrientation with GtkOrientation
Carlos Garnacho [Mon, 26 May 2014 09:58:18 +0000 (11:58 +0200)]
gtk: Replace GtkPanOrientation with GtkOrientation

And document GtkOrientation to be more generic. There's little added
value in a separate enum for this.

11 years agoMore label and image sizing fixes
Matthias Clasen [Tue, 27 May 2014 10:00:21 +0000 (06:00 -0400)]
More label and image sizing fixes

Go back to respecting GtkMisc::xpad/ypad. Not doing so breaks
the misc-alignment reftest. As long as we still derive from
GtkMisc, we may as well do this.

11 years agoFix up GtkLabel and GtkImage sizing
Matthias Clasen [Tue, 27 May 2014 04:21:46 +0000 (00:21 -0400)]
Fix up GtkLabel and GtkImage sizing

When GtkMisc was removed, I forgot to apply the style context
padding. This caused the info-bar-message-types reftest to fail.

11 years agoDon't use hicolor as the default icon-theme
Matthias Clasen [Tue, 27 May 2014 03:15:53 +0000 (23:15 -0400)]
Don't use hicolor as the default icon-theme

The icontheme lookup code has a special-case that prefers builtin icons
if the theme name is "hicolor". This is problematic for our reftests,
which run in a barebones environment with not settings.

11 years agodocs: enum cleanup
Matthias Clasen [Mon, 26 May 2014 14:37:34 +0000 (10:37 -0400)]
docs: enum cleanup

Move GtkIMPreeditStyle nad GtkIMStatusStyle to the settings where
they are solely used.

11 years agodocs: enum cleanup
Matthias Clasen [Mon, 26 May 2014 14:24:02 +0000 (10:24 -0400)]
docs: enum cleanup

Move GtkStateType to the GtkWidget apis where it is used, and
mark the enum as deprecated.

11 years agodocs: remove markup
Matthias Clasen [Mon, 26 May 2014 14:17:07 +0000 (10:17 -0400)]
docs: remove markup

The overview chapters don't go through shorthand processing, so
don't use them there.

11 years agoGtkApplication: Fix two muxer-related memory leaks
Kalev Lember [Mon, 19 May 2014 11:19:11 +0000 (13:19 +0200)]
GtkApplication: Fix two muxer-related memory leaks

Drop the ref on the action muxer in finalize, and also make sure
shutdown() tears down the muxer setup done in startup().

When GtkApplication adds itself to a muxer, it causes the muxer to take
a ref on the GtkApplication. This has to be undone in shutdown() to make
sure the GtkApplication doesn't end up holding a ref on itself.

https://bugzilla.gnome.org/show_bug.cgi?id=730383

11 years agoinspector: Allow changing attribute mappings
Matthias Clasen [Sun, 25 May 2014 03:30:01 +0000 (23:30 -0400)]
inspector: Allow changing attribute mappings

11 years agodocs: Move GdkColor off into a deprecated section
Matthias Clasen [Sun, 25 May 2014 00:46:30 +0000 (20:46 -0400)]
docs: Move GdkColor off into a deprecated section

11 years agoBump ATK dependency to 2.12
Kalev Lember [Sat, 24 May 2014 22:17:55 +0000 (00:17 +0200)]
Bump ATK dependency to 2.12

... for the new AtkValue interface that we implement.

11 years agoSome more updates
Matthias Clasen [Sat, 24 May 2014 22:41:16 +0000 (18:41 -0400)]
Some more updates

11 years agoTypo fixes
Matthias Clasen [Sat, 24 May 2014 22:18:36 +0000 (18:18 -0400)]
Typo fixes

11 years agoOrganize the icons
Matthias Clasen [Sat, 24 May 2014 15:16:35 +0000 (11:16 -0400)]
Organize the icons

Sort the icon list in check-icon-theme.c, and add
comments so we have a chance of keeping the order.

11 years agoDrop an unused object
Matthias Clasen [Sat, 24 May 2014 15:06:10 +0000 (11:06 -0400)]
Drop an unused object

A downside of using templates is that we can have 'parasitic'
objects in there that don't get used for anything.

11 years agoGtkIconTheme: Add more debug spew
Matthias Clasen [Sat, 24 May 2014 15:01:27 +0000 (11:01 -0400)]
GtkIconTheme: Add more debug spew

Still trying to track down test problems in gnome-continuous.

11 years agowayland: Add support for show_window_menu
Jasper St. Pierre [Sat, 24 May 2014 03:01:27 +0000 (23:01 -0400)]
wayland: Add support for show_window_menu

11 years agoimage: Draw the image again
Benjamin Otte [Sat, 24 May 2014 17:38:56 +0000 (19:38 +0200)]
image: Draw the image again

An important line of code was removed in
54d5b426fca71dcda22ea5467054fa801556c20a that lead to an undefined
variable. Add that line back.

11 years agocombobox: Implement clip
Benjamin Otte [Sat, 24 May 2014 14:21:02 +0000 (16:21 +0200)]
combobox: Implement clip

11 years agoentry: Implement clip
Benjamin Otte [Sat, 24 May 2014 14:20:48 +0000 (16:20 +0200)]
entry: Implement clip