profile/ivi/clutter.git
13 years agoUpdated Indonesian translation
Andika Triwidada [Wed, 7 Sep 2011 08:46:43 +0000 (15:46 +0700)]
Updated Indonesian translation

13 years agomingw-fetch-dependencies: Build ATK from source
Neil Roberts [Tue, 6 Sep 2011 08:41:06 +0000 (09:41 +0100)]
mingw-fetch-dependencies: Build ATK from source

There is no prebuilt package for ATK 2.1.5 which Clutter now depends
on so let's build it from source instead. The ATK packages don't have
a .gz version so this patch also changes it to download the .bz2
version of json-glib to avoid having to accept both formats.

13 years agobuild: Fix up the release rules
Emmanuele Bassi [Tue, 6 Sep 2011 07:32:07 +0000 (08:32 +0100)]
build: Fix up the release rules

Highlight actions and errors, and update the release message.

13 years agoPost-release version bump to 1.7.15
Emmanuele Bassi [Mon, 5 Sep 2011 20:39:51 +0000 (21:39 +0100)]
Post-release version bump to 1.7.15

13 years agoRelease Clutter 1.7.14
Emmanuele Bassi [Mon, 5 Sep 2011 19:54:35 +0000 (20:54 +0100)]
Release Clutter 1.7.14

13 years agobuild: Fix distcheck in cookbook/examples
Emmanuele Bassi [Mon, 5 Sep 2011 20:15:37 +0000 (21:15 +0100)]
build: Fix distcheck in cookbook/examples

13 years agodocs: Update test-easing and xinclude it into ClutterAnimation
Emmanuele Bassi [Mon, 5 Sep 2011 14:26:27 +0000 (15:26 +0100)]
docs: Update test-easing and xinclude it into ClutterAnimation

The easing test is a nice example of what ClutterAnimation and
clutter_actor_animate() can do. The "tween ball to the pointer
event coordinates" is a bit of a staple in animation libraries
and their documentation.

13 years agostatic-colors: Simplify the defines
Emmanuele Bassi [Mon, 5 Sep 2011 14:25:46 +0000 (15:25 +0100)]
static-colors: Simplify the defines

Use a "private" macro to simplify the defined static colors.

13 years agotext: Make :line-wrap actually work
Emmanuele Bassi [Sun, 4 Sep 2011 16:57:50 +0000 (17:57 +0100)]
text: Make :line-wrap actually work

When we paint a ClutterText we ask the actor for a PangoLayout that fits
inside the actor's allocation - both width and height.

Sadly, whenever a height is set on a PangoLayout, Pango will wrap its
contents - regardless of whether the layout should actually wrap or not.

This means that in certain easy to exploit cases, Clutter will paint a
Text actor with its contents wrapping even if the :wrap property is set
to FALSE.

In order to fix this we need to encode some more cases inside the
::paint implementation of ClutterText, and ask the cache for a layout
that is sized as the allocation's width, but not as its height; we also
need to perform a clip if we detect that the PangoLayout's logical size
is going to overflow the allocated size. This clip might cause some
performance issue, given that clipping breaks batching in the Cogl
journal; hopefully all clips for text are going to be screen-aligned, so
at the end of the batch it'll just scissor them out.

http://bugzilla.clutter-project.org/show_bug.cgi?id=2339

13 years agoclutter-flow-layout: fix non-0,0 offset allocation
Dan Winship [Fri, 2 Sep 2011 16:51:46 +0000 (12:51 -0400)]
clutter-flow-layout: fix non-0,0 offset allocation

https://bugzilla.gnome.org/show_bug.cgi?id=649631

13 years agoupdated Kannada Translation
Shankar Prasad [Mon, 5 Sep 2011 12:33:48 +0000 (18:03 +0530)]
updated Kannada Translation

13 years agoAdded and updated Kannada Translation
Shankar Prasad [Mon, 5 Sep 2011 08:34:50 +0000 (14:04 +0530)]
Added and updated Kannada Translation

13 years agoUpdated Czech translation
Marek Černocký [Sun, 4 Sep 2011 17:35:49 +0000 (19:35 +0200)]
Updated Czech translation

13 years agoUpdated galician translations
Fran Dieguez [Sun, 4 Sep 2011 16:18:55 +0000 (18:18 +0200)]
Updated galician translations

13 years agoUpdated Traditional Chinese translation(Hong Kong and Taiwan)
Chao-Hsiung Liao [Sat, 3 Sep 2011 22:34:25 +0000 (06:34 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)

13 years agoUpdated Swedish translation
Daniel Nylander [Sat, 3 Sep 2011 11:02:29 +0000 (13:02 +0200)]
Updated Swedish translation

13 years agodocs: Fix up deprecation notice for BehaviourScale
Emmanuele Bassi [Fri, 2 Sep 2011 14:51:56 +0000 (15:51 +0100)]
docs: Fix up deprecation notice for BehaviourScale

13 years agodocs: Fix the ClutterTimeline:auto-reverse link
Emmanuele Bassi [Fri, 2 Sep 2011 14:49:05 +0000 (15:49 +0100)]
docs: Fix the ClutterTimeline:auto-reverse link

13 years agodocs: Inline test-constraints.c inside BindConstraint's description
Emmanuele Bassi [Fri, 2 Sep 2011 14:48:27 +0000 (15:48 +0100)]
docs: Inline test-constraints.c inside BindConstraint's description

13 years agotest-constraints: Clean up the example
Emmanuele Bassi [Fri, 2 Sep 2011 14:36:40 +0000 (15:36 +0100)]
test-constraints: Clean up the example

Add some comments to detail what are we doing and why.

13 years agodocs: Inline test-drop.c into DropAction's description
Emmanuele Bassi [Fri, 2 Sep 2011 14:22:08 +0000 (15:22 +0100)]
docs: Inline test-drop.c into DropAction's description

Also, mention that DropAction only works with DragAction.

13 years agodocs: Include test-drag as the DragAction example
Emmanuele Bassi [Fri, 2 Sep 2011 13:50:07 +0000 (14:50 +0100)]
docs: Include test-drag as the DragAction example

13 years agodocs: Remove the 1.8 index for Cally
Emmanuele Bassi [Fri, 2 Sep 2011 13:28:16 +0000 (14:28 +0100)]
docs: Remove the 1.8 index for Cally

No new API has been added in during the 1.7 development cycle.

13 years agodocs: Add missing clutter_backend_get_cogl_context symbol
Emmanuele Bassi [Fri, 2 Sep 2011 13:18:54 +0000 (14:18 +0100)]
docs: Add missing clutter_backend_get_cogl_context symbol

13 years agobind: Check that the source is not contained inside the actor
Emmanuele Bassi [Fri, 2 Sep 2011 12:58:34 +0000 (13:58 +0100)]
bind: Check that the source is not contained inside the actor

Like commit d0439cfb586ca14282c89035119a4acbc0295df7 for
AlignConstraint, let's check that the BindConstraint source is not
a child or a grandchild of the actor attached to the Constraint.

13 years agoalign: Check that source is not contained inside the actor
Emmanuele Bassi [Fri, 2 Sep 2011 12:56:25 +0000 (13:56 +0100)]
align: Check that source is not contained inside the actor

AlignConstraint won't work if the source is a child or a grandchild of
the ClutterActorMeta:actor to which it has been attached to: the
allocation flows from the parent to its children, not the other way
around; in order to avoid weirdness, we better document and check
that when we set the actor and when we set the source.

13 years agodocs: Remove mentions of "frames" in Timeline
Emmanuele Bassi [Fri, 2 Sep 2011 12:54:57 +0000 (13:54 +0100)]
docs: Remove mentions of "frames" in Timeline

We still have some really old naming lying around.

13 years agocookbooks/examples: Use g_signal_handlers_disconnect_by_func()
Emmanuele Bassi [Fri, 2 Sep 2011 10:01:36 +0000 (11:01 +0100)]
cookbooks/examples: Use g_signal_handlers_disconnect_by_func()

The disconnect_matched() function is a bit too complicated, and its
simpler wrapper disconnect_by_func() is functionally equivalent in the
cases used by the cookbook examples.

13 years agoLock the main context when modifying the repaint functions list
Emmanuele Bassi [Thu, 1 Sep 2011 16:12:46 +0000 (17:12 +0100)]
Lock the main context when modifying the repaint functions list

The repaint functions list can (and should) be manipulated from
different threads, but it currently doesn't prevent multiple threads
from accessing it concurrently. We should have a simple lock and take it
when adding and removing elements from the list; the invocation is still
performed under the Big Clutter Lock™, so it doesn't require special
handling.

13 years agodocs/cookbook: Update the introduction
Emmanuele Bassi [Thu, 1 Sep 2011 15:06:04 +0000 (16:06 +0100)]
docs/cookbook: Update the introduction

Clean up some wording.

13 years agobuild: Install the cookbook examples code
Emmanuele Bassi [Thu, 1 Sep 2011 12:51:20 +0000 (13:51 +0100)]
build: Install the cookbook examples code

Use $(datadir)/clutter-1.0/cookbook/examples as the target directory.

13 years agodocs/cookbook: Update copyright
Emmanuele Bassi [Wed, 31 Aug 2011 11:27:30 +0000 (12:27 +0100)]
docs/cookbook: Update copyright

13 years agotests: Go back to the hover state on button-release
Emmanuele Bassi [Wed, 31 Aug 2011 11:23:01 +0000 (12:23 +0100)]
tests: Go back to the hover state on button-release

If we get a button release, going back to the hover state is the most
logical choice.

13 years agodocs/cookbook: Add script->state connection recipe
Emmanuele Bassi [Wed, 31 Aug 2011 11:22:07 +0000 (12:22 +0100)]
docs/cookbook: Add script->state connection recipe

Script definitions can connect a signal to a State transition; this is a
useful thing to document with a recipe in the cookbook.

13 years agoUpdated Spanish translation
Daniel Mustieles [Tue, 30 Aug 2011 19:24:26 +0000 (21:24 +0200)]
Updated Spanish translation

13 years agoactor: make offscreen_redirect prop take flags + default off
Robert Bragg [Tue, 12 Jul 2011 22:03:41 +0000 (23:03 +0100)]
actor: make offscreen_redirect prop take flags + default off

Because we have had several reports about significant performance
regressions since we enabled offscreen redirection by default for
handling correct opacity we are now turning this feature off by default.

We feel that clutter should prioritize performance over correctness in
this case. Correct opacity is still possible if required but the
overhead of the numerous offscreen allocations as well as the cost of
many render target switches per-frame seems too high relative the
improvement in quality for many cases.

On reviewing the offscreen_redirect property so we have a way to
disable redirection by default we realized that it makes more sense for
it to take a set of flags instead of an enum so we can potentially
extend the number of things that might result in offscreen redirection.

We removed the ability to say REDIRECT_ALWAYS_FOR_OPACITY, since it
seems that implies you don't trust the implementation of an actor's
has_overlaps() vfunc which doesn't seem right.

The default value if actor::redirect_offscreen is now 0 which
effectively means don't ever redirect the actor offscreen.

13 years agoactions/drag: Be more reliable when destroying the drag handle
Emmanuele Bassi [Tue, 30 Aug 2011 09:47:12 +0000 (10:47 +0100)]
actions/drag: Be more reliable when destroying the drag handle

Whenever the drag handle gets destroyed mid-drag we need to cancel any
current drag operation and reset the state of the DragAction.

https://bugzilla.gnome.org/show_bug.cgi?id=657681

13 years agoMerge branch 'atkwindow'
Alejandro Piñeiro [Tue, 30 Aug 2011 10:26:35 +0000 (12:26 +0200)]
Merge branch 'atkwindow'

13 years agoPost-release version bump to 1.7.13
Emmanuele Bassi [Mon, 29 Aug 2011 22:29:07 +0000 (23:29 +0100)]
Post-release version bump to 1.7.13

13 years agoRelease Clutter 1.7.12 (snapshot)
Emmanuele Bassi [Mon, 29 Aug 2011 22:05:03 +0000 (23:05 +0100)]
Release Clutter 1.7.12 (snapshot)

13 years agolayout-managers: Take into account the allocations's origin
Emmanuele Bassi [Wed, 24 Aug 2011 17:38:17 +0000 (18:38 +0100)]
layout-managers: Take into account the allocations's origin

If an actor using a LayoutManager has attributes like margin or padding
then it'll have to shave them from the available allocation before
passing it to the LayoutManager::allocate() implementation. Layout
managers should, thus, not assume that the origin of the allocation is
in (0, 0), but take into account that the passed ActorBox might have a
different origin.

https://bugzilla.gnome.org/show_bug.cgi?id=649631

13 years agobox: Munge the allocation passed to the layout manager
Emmanuele Bassi [Wed, 24 Aug 2011 17:36:19 +0000 (18:36 +0100)]
box: Munge the allocation passed to the layout manager

The actor is in charge of providing to the LayoutManager the available
allocation. ClutterBox should not just pass the box it got from its
parent: it should, instead, provide a normalized box, with an origin in
(0, 0) and the available size.

https://bugzilla.gnome.org/show_bug.cgi?id=649631

13 years agobuild: Add -lm to clutter's LIBADD rule
Emmanuele Bassi [Mon, 29 Aug 2011 21:17:53 +0000 (22:17 +0100)]
build: Add -lm to clutter's LIBADD rule

We need to explicitly link against libm when building Clutter, since we
use libm API.

https://bugzilla.gnome.org/show_bug.cgi?id=657529

13 years agoUpdated galician translations
Fran Dieguez [Sun, 28 Aug 2011 15:36:57 +0000 (17:36 +0200)]
Updated galician translations

13 years agobuild: Add -lm to the tests linker flags
Emmanuele Bassi [Sat, 27 Aug 2011 22:58:26 +0000 (23:58 +0100)]
build: Add -lm to the tests linker flags

https://bugzilla.gnome.org/show_bug.cgi?id=657529

13 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Fri, 26 Aug 2011 11:26:49 +0000 (13:26 +0200)]
Updated Norwegian bokmål translation

13 years agoAdds experimental clutter_backend_get_cogl_context API
Robert Bragg [Fri, 12 Aug 2011 15:25:53 +0000 (16:25 +0100)]
Adds experimental clutter_backend_get_cogl_context API

This adds experimental API to be able to get the CoglContext associated
with the ClutterBackend. The CoglContext is required to use some of the
experimental 2.0 Cogl API.

Note: Since CoglContext is itself experimental API this API should
considered experimental too. This patch introduces a
CLUTTER_ENABLE_EXPERIMENTAL_API #ifdef guard which anyone wanting to use
this API must define so it's explicitly clear to developers that they
are playing with experimental API.

Note: This API is not yet supported on OSX because OSX still uses the
stub Cogl winsys and the Clutter backend doesn't explicitly create a
CoglContext.

Note: even though this is experimental API we still promise that it
wont be changed during a stable release cycle. This means for example
that you can depend on this for the lifetime of the clutter-1.8 stable
release cycle.

13 years agoa11y: Using proper way to register to window events
Alejandro Piñeiro [Wed, 10 Aug 2011 17:19:33 +0000 (19:19 +0200)]
a11y: Using proper way to register to window events

13 years agoa11y: We don't need to explicitly manage window events
Alejandro Piñeiro [Wed, 10 Aug 2011 11:06:14 +0000 (13:06 +0200)]
a11y: We don't need to explicitly manage window events

13 years agoa11y: Using AtkWindow on cally-stage
Alejandro Piñeiro [Wed, 10 Aug 2011 11:02:26 +0000 (13:02 +0200)]
a11y: Using AtkWindow on cally-stage

13 years agoactor: Add a comment on the unmapping order
Emmanuele Bassi [Tue, 23 Aug 2011 09:58:09 +0000 (10:58 +0100)]
actor: Add a comment on the unmapping order

Let's add a comment to avoid reshuffling the unmap() implementation and
end up triggering these bugs:

  http://bugzilla.clutter-project.org/show_bug.cgi?id=2621
  https://bugzilla.gnome.org/show_bug.cgi?id=652036

13 years agodocs/cookbook: Point out that Cogl does not do AA drawing
Emmanuele Bassi [Mon, 22 Aug 2011 12:15:42 +0000 (13:15 +0100)]
docs/cookbook: Point out that Cogl does not do AA drawing

13 years agodocs/cookbook: Update the link to the GObject reference
Emmanuele Bassi [Mon, 22 Aug 2011 11:33:20 +0000 (12:33 +0100)]
docs/cookbook: Update the link to the GObject reference

13 years agoUpdated Spanish translation
Daniel Mustieles [Fri, 19 Aug 2011 11:51:32 +0000 (13:51 +0200)]
Updated Spanish translation

13 years agox11/settings: Add Fontconfig/Timestamp key
Emmanuele Bassi [Thu, 18 Aug 2011 14:36:20 +0000 (15:36 +0100)]
x11/settings: Add Fontconfig/Timestamp key

The XSETTINGS key that matches the :fontconfig-timestamp property on
ClutterSettings.

13 years agosettings: Add ClutterSettings:fontconfig-timestamp
Emmanuele Bassi [Thu, 18 Aug 2011 14:33:39 +0000 (15:33 +0100)]
settings: Add ClutterSettings:fontconfig-timestamp

The :fontconfig-timestamp is a write-only property that will get updated
by the underlying platform whenever the fontconfig configuration has
been changed — i.e. when the fontconfig caches should be rebuilt after
the user has installed a new font.

13 years agobuild: Depend on pangoft2 on X11
Emmanuele Bassi [Thu, 18 Aug 2011 14:32:29 +0000 (15:32 +0100)]
build: Depend on pangoft2 on X11

We need the fontconfig API to reinitialize the fontmap caches whenever
the system tells us that the user fonts have changed.

13 years agotext: Implement Scriptable interface for font-description
Robert Bragg [Tue, 16 Aug 2011 16:31:42 +0000 (17:31 +0100)]
text: Implement Scriptable interface for font-description

This makes ClutterText implement the Scriptable interface so that we can
have a custom property parser and setter for the font-description
property. This works by simply passing the string description through
to clutter_text_set_font_name.

Reviewed-by: Emmanuele Bassi <ebassi@linux.intel.com>
13 years agoPost-release version bump to 1.7.11
Emmanuele Bassi [Mon, 15 Aug 2011 23:19:34 +0000 (00:19 +0100)]
Post-release version bump to 1.7.11

13 years agoRelease Clutter 1.7.10 (snapshot)
Emmanuele Bassi [Mon, 15 Aug 2011 22:50:03 +0000 (23:50 +0100)]
Release Clutter 1.7.10 (snapshot)

13 years agobuild: Do not define LDADD twice
Emmanuele Bassi [Mon, 15 Aug 2011 22:52:45 +0000 (23:52 +0100)]
build: Do not define LDADD twice

13 years agoactor: Don't update last-paint-volume during picking
Robert Bragg [Mon, 15 Aug 2011 17:24:32 +0000 (18:24 +0100)]
actor: Don't update last-paint-volume during picking

Actually this change has two notable effects; firstly we no longer
perform culling during picking and secondly we avoid updating the
last-paint-volume of an actor when picking.

We shouldn't perform culling during picking until clutter-stage.c is
updated to setup the clipping planes appropriately.

Since the last-paint-volume is intended to represent the visible region
of the actor the last time it was painted on screen it doesn't make
sense to update this during off screen pick renders since we are liable
to end up with a last-paint-volume that maps to an actors new position
when we next come to paint for real.

This fixes a bug in gnome-shell with dragging dash icons leaving a
messy trail on the screen.

Reviewed-by: Emmanuele Bassi <ebassi@linux.intel.com>
13 years agoUpdate the COPYING file
Emmanuele Bassi [Mon, 15 Aug 2011 16:16:54 +0000 (17:16 +0100)]
Update the COPYING file

The FSF changed address, so we need to update the license accordingly to
avoid triggering issues with automated license checkers used by build
systems, like OBS.

https://bugzilla.gnome.org/show_bug.cgi?id=656588

13 years agobuild: list correct dependencies
Alejandro Piñeiro [Wed, 10 Aug 2011 18:25:43 +0000 (20:25 +0200)]
build: list correct dependencies

Related to bug https://bugzilla.gnome.org/show_bug.cgi?id=656306#c4

13 years agointeractive/test-actors: Constraint the hands group
Emmanuele Bassi [Mon, 15 Aug 2011 14:53:58 +0000 (15:53 +0100)]
interactive/test-actors: Constraint the hands group

Use constraints to align and size the ClutterGroup containing the
spinning hands so that resizing the stage doesn't look like arse.

13 years agobuild: Require the latest stable version of Cogl
Emmanuele Bassi [Mon, 15 Aug 2011 14:35:41 +0000 (15:35 +0100)]
build: Require the latest stable version of Cogl

13 years agoPost-release version bump to 1.7.9
Emmanuele Bassi [Mon, 15 Aug 2011 14:16:24 +0000 (15:16 +0100)]
Post-release version bump to 1.7.9

13 years agoRelease Clutter 1.7.8 (snapshot)
Emmanuele Bassi [Mon, 15 Aug 2011 13:32:37 +0000 (14:32 +0100)]
Release Clutter 1.7.8 (snapshot)

13 years agoconform: Temporarily disable Cally's conformance test
Emmanuele Bassi [Mon, 15 Aug 2011 13:09:24 +0000 (14:09 +0100)]
conform: Temporarily disable Cally's conformance test

See bug:

  https://bugzilla.gnome.org/show_bug.cgi?id=655588

for the failure.

13 years agodocs: Fix up the unused symbols warning
Emmanuele Bassi [Mon, 15 Aug 2011 13:08:57 +0000 (14:08 +0100)]
docs: Fix up the unused symbols warning

13 years agodocs: Fix up warnings in Cally's reference API
Emmanuele Bassi [Mon, 15 Aug 2011 13:08:28 +0000 (14:08 +0100)]
docs: Fix up warnings in Cally's reference API

13 years agostage-cogl: pass top-left relative swap_region coords
Robert Bragg [Fri, 22 Jul 2011 11:28:29 +0000 (12:28 +0100)]
stage-cogl: pass top-left relative swap_region coords

All 2D coordinate spaces in Cogl have their origin at the top-left so we
shouldn't be flipping the coordinates we pass to
cogl_framebuffer_swap_region to be relative to the bottom of the
framebuffer.

This bumps the Cogl version requirement to 1.7.5 since we've had to fix
a bug in the semantics of cogl_framebuffer_swap_region.

Reviewed-by: Neil Roberts <neil@linux.intel.com>
13 years agotest-cogl-blend-strings: Don't disable the TEXTURE_N test
Neil Roberts [Thu, 4 Aug 2011 18:17:23 +0000 (19:17 +0100)]
test-cogl-blend-strings: Don't disable the TEXTURE_N test

This is now fixed in Cogl so there's no need to disable it.

13 years agoa11y: Notify ATK_STATE_SHOWING state change
Alejo Pacin [Thu, 4 Aug 2011 10:02:52 +0000 (12:02 +0200)]
a11y: Notify ATK_STATE_SHOWING state change

https://bugzilla.gnome.org/show_bug.cgi?id=648633

13 years agoclutter-stage-win32: Fix name cogl_win32_onscreen_set_foreign_window
Neil Roberts [Sun, 31 Jul 2011 17:00:43 +0000 (18:00 +0100)]
clutter-stage-win32: Fix name cogl_win32_onscreen_set_foreign_window

Since commit 38b67e2884f of Cogl the naming scheme for winsys-specific
API has changed to be cogl_win32_onscreen_* instead of
cogl_onscreen_win32_* so it wouldn't build on Windows.

13 years agoosx: Avoid leaks in Stage::realize
Kirk A. Baker [Mon, 25 Jul 2011 22:00:30 +0000 (15:00 -0700)]
osx: Avoid leaks in Stage::realize

clutter_stage_osx_realize() can now be called multiple times without
leaking views and windows.

https://bugzilla.gnome.org/show_bug.cgi?id=655307

13 years agoosx: Implement windowDidResignKey
Kirk A. Baker [Tue, 12 Jul 2011 04:56:04 +0000 (21:56 -0700)]
osx: Implement windowDidResignKey

Added isHiding field to _ClutterStageOSX to allow windowDidResignKey
delegate to not order full screen window back whilst the full screen
window was being hidden.  This caused other application windows to be
hidden.  Also added code to keep hidden stage windows from being listed
in the application's Windows menu.

https://bugzilla.gnome.org/show_bug.cgi?id=655311

13 years agoosx: Add windowDidChangeScreen delegate
Kirk A. Baker [Mon, 25 Jul 2011 22:02:46 +0000 (15:02 -0700)]
osx: Add windowDidChangeScreen delegate

Added -windowDidChangeScreen: delegate to handle condition where moving
host window to a different screen would cause pick errors to be output.
The delegate just causes the stage to be redrawn which re-creates the
pick buffer.

https://bugzilla.gnome.org/show_bug.cgi?id=655306

13 years agomodel: Remove an unused variable
Emmanuele Bassi [Fri, 29 Jul 2011 12:17:11 +0000 (13:17 +0100)]
model: Remove an unused variable

13 years agointrospection: add missing introspection data from Vala bindings
Evan Nemerson [Wed, 27 Jul 2011 07:14:10 +0000 (00:14 -0700)]
introspection: add missing introspection data from Vala bindings

https://bugzilla.gnome.org/show_bug.cgi?id=655387

13 years agodocs: Update with the newly added Animatable symbol
Emmanuele Bassi [Fri, 29 Jul 2011 10:48:13 +0000 (11:48 +0100)]
docs: Update with the newly added Animatable symbol

13 years agostate: Use the Animatable interface
Emmanuele Bassi [Fri, 15 Jul 2011 14:06:31 +0000 (15:06 +0100)]
state: Use the Animatable interface

The Animatable interface allows object classes to provide and animate
properties outside of the usual GObject property introspection API.

This change allows ClutterState to defer to the animatable objects the
property introspection and animation, just like ClutterAnimation does.

13 years agoactor: There is no need to implement animate_property()
Emmanuele Bassi [Fri, 15 Jul 2011 13:31:12 +0000 (14:31 +0100)]
actor: There is no need to implement animate_property()

It's a deprecated virtual function, and its replacement does the right
thing by default.

13 years agoanimation: Use Animatable.interpolate_value()
Emmanuele Bassi [Fri, 15 Jul 2011 13:30:39 +0000 (14:30 +0100)]
animation: Use Animatable.interpolate_value()

And stop using the deprecated animate_property() method.

13 years agoanimatable: Deprecate and replace animate_property()
Emmanuele Bassi [Fri, 15 Jul 2011 13:27:26 +0000 (14:27 +0100)]
animatable: Deprecate and replace animate_property()

The animate_property() method of the Animatable interface is far less
than optimal:

  • it has a direct reference to ClutterAnimation;
  • it has an interval decomposed as two values.

These issues tie the Animatable interface with the Animation object,
even though it's neither necessary nor future-proof.

Let's introduce a new method, interpolate_value(), which does not
reference ClutterAnimation and uses a ClutterInterval to express the
initial and final states.

13 years agomodel: Clean up ModelIter vfunc wrappers
Emmanuele Bassi [Fri, 29 Jul 2011 10:29:51 +0000 (11:29 +0100)]
model: Clean up ModelIter vfunc wrappers

All ClutterModelIter virtual functions have a default implementation,
and G_TYPE_INSTANCE_GET_CLASS cannot return NULL unless in case of a
catastrophic event in the type system - which will most likely blow up
any application code way before you could call a ModelIter method.

Thus, the idiom:

  klass = CLUTTER_MODEL_ITER_GET_CLASS (instance);
  if (klass && klass->vfunc)
    klass->vfunc (instance);

is utterly useless complication, and it can be perfectly replaced by:

  CLUTTER_MODEL_ITER_GET_CLASS (instance)->vfunc (instance);

without any loss of safety.

13 years agomodel: Make sure to emit ::row-changed
Emmanuele Bassi [Thu, 28 Jul 2011 14:00:18 +0000 (15:00 +0100)]
model: Make sure to emit ::row-changed

Currently, only clutter_model_iter_set_valist() is in charge of emitting
the ClutterModel::row-changed signal. Both the set() and the
set_valist() functions can be called with multiple columns, so we
coalesce the signal emission at the end of the set_valist(), to have a
single ::row-changed emission per change.

The clutter_model_iter_set_value() function is just a thin wrapper
around the set_value() virtual function, but since it's called
internally we cannot add the signal emission there as well, as we'd
break the signal coalescing.

For this reason, we need some code refactoring inside the various set()
variants of ClutterModelIter:

  - we only use the internal va_arg variant for both the set() and
    set_valist() public functions, to avoid multiple type checks;
  - the internal set_valist() calls an internal set_value() method
    which calls the virtual function from the iterator vtable;
  - a new internal emit_row_changed() method is needed to retrieve
    the ClutterModel from the iterator, and emit the signal;

Now, all three variants of the value setter will call an internal
ClutterModelIter::set_value() wrapper, and emit the ::row-changed
signal.

To check that the intended behaviour has been implemented, and it's not
going to be broken, the test suite has grown a new unit which populates
a model and changes a random row.

13 years agoUpdated Czech translation
Marek Černocký [Thu, 28 Jul 2011 07:31:07 +0000 (09:31 +0200)]
Updated Czech translation

13 years agoUpdated Spanish translation
Daniel Mustieles [Wed, 27 Jul 2011 18:34:39 +0000 (20:34 +0200)]
Updated Spanish translation

13 years agotest/cairo-clock: Remove a double source color set
Emmanuele Bassi [Wed, 27 Jul 2011 10:48:07 +0000 (11:48 +0100)]
test/cairo-clock: Remove a double source color set

Do not call cairo_set_source_rgba() right after calling
clutter_cairo_set_source_color().

13 years agoa11y: Use existing clutter_actor_get_paint_visibility
Alejandro Piñeiro [Tue, 26 Jul 2011 13:56:18 +0000 (15:56 +0200)]
a11y: Use existing clutter_actor_get_paint_visibility

Cally was initially created with Clutter 0.6 in mind. To check
recursively the visibility of a actor a custom method was added.
Since 0.8.4 clutter_actor_get_pain_visibility provides
the same functionality.

Also removed a dummy method. Lets add methods that provide a real
functionality.

13 years agocairo-texture: Add the :auto-resize property
Emmanuele Bassi [Tue, 26 Jul 2011 13:53:26 +0000 (14:53 +0100)]
cairo-texture: Add the :auto-resize property

Keeping the backing Cairo surface of a CairoTexture canvas in sync with
the actor's allocation is tedious and prone to mistakes. We can
definitely do better by simply exposing a property that does the surface
resize and invalidation automagically on ::allocate.

13 years agotests: initial support for cally-text conform tests
Philippe Normand [Mon, 7 Feb 2011 10:59:19 +0000 (11:59 +0100)]
tests: initial support for cally-text conform tests

13 years agoPost-release version bump to 1.7.7
Emmanuele Bassi [Tue, 26 Jul 2011 13:12:08 +0000 (14:12 +0100)]
Post-release version bump to 1.7.7

13 years agoRelease Clutter 1.7.6 (snapshot)
Emmanuele Bassi [Tue, 26 Jul 2011 12:18:43 +0000 (13:18 +0100)]
Release Clutter 1.7.6 (snapshot)

13 years agobin-layout: Fix a leak
Emmanuele Bassi [Tue, 26 Jul 2011 12:55:01 +0000 (13:55 +0100)]
bin-layout: Fix a leak

https://bugzilla.gnome.org/show_bug.cgi?id=655342

13 years agodocs: Fixes for building the API reference
Emmanuele Bassi [Tue, 26 Jul 2011 12:43:37 +0000 (13:43 +0100)]
docs: Fixes for building the API reference

13 years agocairo-texture: Deprecate create()/create_region()
Emmanuele Bassi [Tue, 26 Jul 2011 11:53:22 +0000 (12:53 +0100)]
cairo-texture: Deprecate create()/create_region()

The recommended way of drawing on a ClutterCairoTexture is the ::draw
signal.

13 years agocairo-texture: Use signal-based drawing
Emmanuele Bassi [Tue, 26 Jul 2011 11:40:52 +0000 (12:40 +0100)]
cairo-texture: Use signal-based drawing

The current "create context/draw/destroy context" pattern presents
various problems. The first issue is that it defers memory management to
the caller of the create() or create_region() methods, which makes
bookkeeping of the cairo_t* harder for language bindings and third party
libraries. The second issue is that, while it's easier for
draw-and-forget texturs, this API is needlessly complicated for contents
that have to change programmatically - and it introduces constraints
like calling the drawing code explicitly after a surface resize (e.g.
inside an allocate() implementation).

By using a signal-based approach we can make the CairoTexture actor
behave like other actors, and like other libraries using Cairo as their
2D drawing API.

The semantics of the newly-introduced ::draw signal are the same as the
one used by GTK+:

  - the signal is emitted on invalidation;
  - the cairo_t* context is owned by the actor;
  - it is safe to have multiple callbacks attached to the same
    signal, to allow composition;
  - the cairo_t* is already clipped to the invalidated area, so
    that Cairo can discard geometry immediately before we upload
    the texture data.

There are possible future improvements, like coalescing multiple
invalidations inside regions, and performing clipped draws during
the paint cycle; we could even perform clipped redraws if we know the
extent of the invalidated area.

13 years agoconform/cogl-materials: Fix a compiler warning
Emmanuele Bassi [Mon, 25 Jul 2011 10:07:18 +0000 (11:07 +0100)]
conform/cogl-materials: Fix a compiler warning

The function checking for the presence and use of the GLES2 support in
Cogl should be protected by #ifdef guards, to avoid a compiler warning.