]> Pileus Git - ~andy/gtk/log
~andy/gtk
11 years agostylecontext: Pass a bitmask to the changed values
Benjamin Otte [Fri, 30 Nov 2012 12:58:11 +0000 (13:58 +0100)]
stylecontext: Pass a bitmask to the changed values

11 years agoreftests: Fix a reftest
Benjamin Otte [Fri, 30 Nov 2012 13:57:02 +0000 (14:57 +0100)]
reftests: Fix a reftest

Recent Adwaita changes caused diffs. We don't want that, so we run this
test against the default settings.

11 years agoicon cache: Fix double free
Alexander Larsson [Fri, 30 Nov 2012 15:18:14 +0000 (16:18 +0100)]
icon cache: Fix double free

The symbolic icon cache code free crashed with a double free
due to a missing ->next in the loop.

11 years agogtk: convert keyvals to unicode before committing to the imcontext
Michael Natterer [Fri, 30 Nov 2012 14:06:48 +0000 (15:06 +0100)]
gtk: convert keyvals to unicode before committing to the imcontext
(cherry picked from commit 1eb0d98d734d96a6612e320d32a2aded423d2746)

11 years agogtk: add more OSX special casing for the deadacute and deaddoubleacute keys
Michael Natterer [Fri, 30 Nov 2012 13:39:42 +0000 (14:39 +0100)]
gtk: add more OSX special casing for the deadacute and deaddoubleacute keys

So " plus foo prduces foo-with-diaereses and ' plus c produces
c-with-cedilla.
(cherry picked from commit c0102b30894401a9a1d6d93a972bcc6c37f1b5ac)

11 years agocss: Make test pass
Alexander Larsson [Fri, 30 Nov 2012 13:18:39 +0000 (14:18 +0100)]
css: Make test pass

The new css tree may change the order of selectors (keeping the
same semantics). This affects how the selectors are printed later,
which causes some css parsing tests to not match the references.

Fortunately the order is consistent between runs given the same
css, so we just have to switch around the order in some of the
.ref.css files.

11 years agocss: Allocate the css tree in a single chunk
Alexander Larsson [Thu, 29 Nov 2012 19:52:46 +0000 (20:52 +0100)]
css: Allocate the css tree in a single chunk

This gives us several advantages:

* Smaller struct on 64bit (32bit indexes vs 64bit pointers)
* Less overhead for allocation
* Less fragmentation

11 years agocss: Fix leak of lists while building tree
Alexander Larsson [Thu, 29 Nov 2012 18:10:57 +0000 (19:10 +0100)]
css: Fix leak of lists while building tree

11 years agocss: Add accessor functions for traversing css tree
Alexander Larsson [Thu, 29 Nov 2012 16:56:00 +0000 (17:56 +0100)]
css: Add accessor functions for traversing css tree

This will let us later change how the tree is stored

11 years agocss: Add const to _gtk_css_selector_tree_match_all arg
Alexander Larsson [Thu, 29 Nov 2012 16:55:26 +0000 (17:55 +0100)]
css: Add const to _gtk_css_selector_tree_match_all arg

11 years agocss: Don't keep around linear selectors
Alexander Larsson [Thu, 29 Nov 2012 16:28:51 +0000 (17:28 +0100)]
css: Don't keep around linear selectors

Now we use the selector tree everywhere, so there is no need to
keep around the linear selectors unless we're using them to
verify the tree correctness, so free them.

11 years agocss: Remove gtk_css_ruleset_matches
Alexander Larsson [Thu, 29 Nov 2012 16:23:36 +0000 (17:23 +0100)]
css: Remove gtk_css_ruleset_matches

This is only not needed anymore, and only the VERIFY_TREE code
should access ->selector.

11 years agocss: Use tree for gtk_css_provider_get_style_property
Alexander Larsson [Thu, 29 Nov 2012 16:13:59 +0000 (17:13 +0100)]
css: Use tree for gtk_css_provider_get_style_property

11 years agocss: Implement ruleset_get_change() with the tree
Alexander Larsson [Thu, 29 Nov 2012 15:43:02 +0000 (16:43 +0100)]
css: Implement ruleset_get_change() with the tree

We traverse the tree on the matches instead of using
the linear selectors.

11 years agocss: Ensure the tree built is always the same
Alexander Larsson [Thu, 29 Nov 2012 15:20:14 +0000 (16:20 +0100)]
css: Ensure the tree built is always the same

We add some "artificial" ordering to the otherwise unordered
tree nodes. This means the tree will be the same every time for the
same input. This is good because e.g. tree order affects the
reordering of the simple selectors, which may affect how
css providers are printed, which need to be consistent for
the css tests to work.

11 years agocss: Use the tree to print css selectors
Alexander Larsson [Thu, 29 Nov 2012 15:00:01 +0000 (16:00 +0100)]
css: Use the tree to print css selectors

11 years agocss: Track the tree selector matches
Alexander Larsson [Thu, 29 Nov 2012 14:29:14 +0000 (15:29 +0100)]
css: Track the tree selector matches

11 years agocss: Fix type of GtkCssSelectorRuleSetInfo match
Alexander Larsson [Thu, 29 Nov 2012 14:01:47 +0000 (15:01 +0100)]
css: Fix type of GtkCssSelectorRuleSetInfo match

The old type was a leftover from a previous version.

11 years agocss: Track parent in the css tree nodes
Alexander Larsson [Thu, 29 Nov 2012 13:58:08 +0000 (14:58 +0100)]
css: Track parent in the css tree nodes

This way we can reconstruct matched css rules

11 years agocss: Better tree match verification
Alexander Larsson [Thu, 29 Nov 2012 13:40:10 +0000 (14:40 +0100)]
css: Better tree match verification

11 years agocss: Fix up position with region tree matching
Alexander Larsson [Thu, 29 Nov 2012 13:25:53 +0000 (14:25 +0100)]
css: Fix up position with region tree matching

This was using the wrong result in case of a match (results from
the position, not the region. Also, the descendant checks were
wrong.

11 years agocss: Don't reorder some selectors when building selector tree
Alexander Larsson [Thu, 29 Nov 2012 13:21:20 +0000 (14:21 +0100)]
css: Don't reorder some selectors when building selector tree

When building the tree we generally reorder the selectors inside
the same simple selector in order to pick a good first selector
to balance the tree better. However, some kinds of selectors
can't really be reordered, even thought they are simple.

This is since the matching code for some types handle
the existance of a directly preceeding selector differently:

 REGION and ANY selectors look for a DESCENDANT previous
 POSITION selector look for a REGION previous

11 years agocss: Fixed typo in PRINT_TREE debug code
Alexander Larsson [Wed, 28 Nov 2012 11:13:09 +0000 (12:13 +0100)]
css: Fixed typo in PRINT_TREE debug code

11 years agocss: Create and use a tree for css selector matching
Alexander Larsson [Wed, 28 Nov 2012 10:21:06 +0000 (11:21 +0100)]
css: Create and use a tree for css selector matching

11 years agocss: Add GtkCssSelectorTree creation and matching
Alexander Larsson [Wed, 28 Nov 2012 10:15:53 +0000 (11:15 +0100)]
css: Add GtkCssSelectorTree creation and matching

From a set of GtkCssSelectors and the rulesets they match to
we create a large decision tree that lets us efficitently match
against all the rules and return the set of matched rulesets.

The tree is created such that at each level we pick the initial rule[1]
in all the considered selectors for that level and use put the
one that is in most selectors in the node. All selectors matching that
are put in the previous part of the tree.

11 years agocss: Add _gtk_css_matcher_matches_any()
Alexander Larsson [Wed, 28 Nov 2012 10:07:52 +0000 (11:07 +0100)]
css: Add _gtk_css_matcher_matches_any()

This returns true if the matcher matches *anything*. We need
to check this later, because such matchers are dangerous in loops
that iterate over all parents/siblings since such loops would not
terminate.

11 years agocss: Track which selectors are "simple"
Alexander Larsson [Tue, 27 Nov 2012 09:50:59 +0000 (10:50 +0100)]
css: Track which selectors are "simple"

11 years agoReuse rendered symbolic icons
Alexander Larsson [Mon, 26 Nov 2012 12:52:05 +0000 (13:52 +0100)]
Reuse rendered symbolic icons

With the previous commit all loads of the same icon will share a single
GtkIconInfo, which typicallty means the pixbuf is shared via Info->pixbuf.

However, atm we don't share symbolic icons, which causes these to be re-read
and re-parsed every time. This is especially bad if the icon is used many times
in some form of list. So, we cache the pixbufs and reuse them.

https://bugzilla.gnome.org/show_bug.cgi?id=689081

11 years agoCache GtkIconInfo
Alexander Larsson [Mon, 26 Nov 2012 12:49:49 +0000 (13:49 +0100)]
Cache GtkIconInfo

In order to avoid loading and keeping around the same icon multiple times
we keep a cache of all outstanding GtkIconInfo objects for a given theme.

Additionally we return to the app not the normal pixbuf from the info,
but rather a proxy copy of it sharing the same data, but no extra
reference. This allows us to track when the app is no longer using
the pixbuf, and we can thus ensure that the GtkIconInfo in the cache
stays around for at least as long as the pixbuf is alive.

When the app unrefs the pixbuf we put the Info on a short LRU list
to keep it alive a bit longer, in case the app needs it in a short
while.

https://bugzilla.gnome.org/show_bug.cgi?id=689081

11 years agoAdd a missing va_end call
Matthias Clasen [Fri, 30 Nov 2012 04:18:49 +0000 (23:18 -0500)]
Add a missing va_end call

va_copy must be matched with va_end.

11 years agoAvoid unreachable code
Matthias Clasen [Fri, 30 Nov 2012 04:08:24 +0000 (23:08 -0500)]
Avoid unreachable code

Reestablish an else that went missing in commit 5ff328d2. Without
it, the return is reached unconditionally, and the code behind
it is dead.

11 years agoDon't return prematurely
Matthias Clasen [Fri, 30 Nov 2012 04:05:07 +0000 (23:05 -0500)]
Don't return prematurely

This code was clearly meant to return _after_ the loop.
Don't put the return inside the loop body, then.

11 years agoGtkAboutDialog: Initialize a variable
Matthias Clasen [Fri, 30 Nov 2012 04:01:10 +0000 (23:01 -0500)]
GtkAboutDialog: Initialize a variable

Coverity complains about it being uninitialized, otherwise.

11 years agoGtkNotebook: Clarify a switch statement
Matthias Clasen [Fri, 30 Nov 2012 03:48:27 +0000 (22:48 -0500)]
GtkNotebook: Clarify a switch statement

We were intentionally falling through here, this commit
just adds comments to make that explicit.

11 years agoGtkCellAccessible: Add missing break statements
Matthias Clasen [Fri, 30 Nov 2012 03:47:02 +0000 (22:47 -0500)]
GtkCellAccessible: Add missing break statements

This switch was entirely breakless. Amazing accessibility code,
found by Coverity.

11 years agoGtkExpander: Clarify a nested switch
Matthias Clasen [Fri, 30 Nov 2012 03:45:24 +0000 (22:45 -0500)]
GtkExpander: Clarify a nested switch

This code confused Coverity into thinking we were falling through,
when we were not. Add a few explicit breaks to clear that up.

11 years agoAdd a forgotten break in render_border
Matthias Clasen [Fri, 30 Nov 2012 03:42:36 +0000 (22:42 -0500)]
Add a forgotten break in render_border

The GTK_BORDER_STYLE_DOUBLE case was clearly not meant to
fall through to the subsequent cases, yet it did.
Found by Coverity.

11 years agoGtkIconViewAccessible: Fix a copy-paste error
Matthias Clasen [Fri, 30 Nov 2012 03:14:12 +0000 (22:14 -0500)]
GtkIconViewAccessible: Fix a copy-paste error

Don't access vadjustment after checking hadjustment for
not being NULL. Found by Coverity.

11 years agoGtkActionHelper: Fix a copy-paste error
Matthias Clasen [Fri, 30 Nov 2012 03:12:58 +0000 (22:12 -0500)]
GtkActionHelper: Fix a copy-paste error

The code clearly meant to reset action->active here.

This was found by Coverity.

11 years agogtk_css_value_bg_size_equal: actually compare contain values
Matthias Clasen [Fri, 30 Nov 2012 03:09:31 +0000 (22:09 -0500)]
gtk_css_value_bg_size_equal: actually compare contain values

Due to a copy-paste error, the code failed to compare the
contain members of value1 and value2.

This was found by Coverity.

11 years agoDon't check the same crossing mode twice
Matthias Clasen [Fri, 30 Nov 2012 03:06:18 +0000 (22:06 -0500)]
Don't check the same crossing mode twice

This was broken since commit b2aaa94 in 2008. Its commit message
clearly states that the intention was to check for GTK_GRAB,
GTK_UNGRAB and STATE_CHANGED. Lets do that, then.

This was found by Coverity.

11 years agoGtkLabel: Fix a pointless expression
Matthias Clasen [Fri, 30 Nov 2012 03:02:16 +0000 (22:02 -0500)]
GtkLabel: Fix a pointless expression

Clearly, what was meant here was that we create a new attribute
list if either of the input attribute lists are non-NULL.
This was found by Coverity.

11 years agoObtain the recent files max age setting from xsettings
Matthias Clasen [Sun, 25 Nov 2012 20:45:26 +0000 (15:45 -0500)]
Obtain the recent files max age setting from xsettings

This will help with implementing desktop-wide policy for
retaining of history.

https://bugzilla.gnome.org/show_bug.cgi?id=689047

11 years agoIn the file chooser, use a proper apostrophe rather than a prime mark (tick mark)
Timothy Arceri [Sun, 25 Nov 2012 06:05:45 +0000 (17:05 +1100)]
In the file chooser, use a proper apostrophe rather than a prime mark (tick mark)

https://bugzilla.gnome.org/show_bug.cgi?id=689012

11 years agoFix broken function pointer declarations on windows
Alan McGovern [Fri, 23 Nov 2012 20:38:34 +0000 (15:38 -0500)]
Fix broken function pointer declarations on windows

Both flashing a window and setting the window opacity were using
incorrect declarations for function pointers. They were missing the
WINAPI annotation as defined in windows.h. As a result, the stack
could be corrupted when these functions were invoked.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=689235
(cherry picked from commit 5637ef1f97ee46666c97707ed7f6bae459007163)

11 years agoquartz: add mountain lion as version 8 to enum GdkOSXVersion
Michael Natterer [Thu, 29 Nov 2012 13:51:59 +0000 (14:51 +0100)]
quartz: add mountain lion as version 8 to enum GdkOSXVersion

Also use GDK_OSX_UNSUPPORTED instead of 0 in gdk_quartz_osx_version().
(cherry picked from commit 9644e910a85f97aedf3f5f4fee974229f29766c3)

11 years agoquartz: call Gestalt() only once per session in gdk_quartz_osx_version()
Michael Natterer [Thu, 29 Nov 2012 13:44:04 +0000 (14:44 +0100)]
quartz: call Gestalt() only once per session in gdk_quartz_osx_version()

Thanks to Paul Davis for pointing this out.
(cherry picked from commit d6533ffc44d77b42cce7987f249fa481581b7c50)

11 years agoUse the ppd groups "name" instead of "text" (bug #687065)
Benjamin Berg [Sun, 28 Oct 2012 21:19:35 +0000 (22:19 +0100)]
Use the ppd groups "name" instead of "text" (bug #687065)

This commit fixes a regression caused by a patch to remove Cups 1.2
ifdefs. This resulted in the "installable options" to appear in the
print dialog.

11 years agomenuitem: Remove unused variables
Benjamin Otte [Wed, 28 Nov 2012 20:15:05 +0000 (21:15 +0100)]
menuitem: Remove unused variables

11 years agotests: Add CSS test for comment detection
Benjamin Otte [Wed, 28 Nov 2012 20:11:22 +0000 (21:11 +0100)]
tests: Add CSS test for comment detection

Tests recent fix.

11 years agocssparser: Don't detect /*/ as a comment
Marcus Kraßmann [Wed, 28 Nov 2012 20:09:37 +0000 (21:09 +0100)]
cssparser: Don't detect /*/ as a comment

11 years agomenubar: deprecate internal-padding style property
Cosimo Cecchi [Tue, 27 Nov 2012 14:54:42 +0000 (09:54 -0500)]
menubar: deprecate internal-padding style property

https://bugzilla.gnome.org/show_bug.cgi?id=689168

11 years agomenubar: add support for CSS padding
Cosimo Cecchi [Tue, 27 Nov 2012 14:14:20 +0000 (09:14 -0500)]
menubar: add support for CSS padding

In preparation for the removal of the internal-padding style property.

https://bugzilla.gnome.org/show_bug.cgi?id=689168

11 years agomenu: deprecate horizontal-padding and vertical-padding
Cosimo Cecchi [Tue, 27 Nov 2012 14:04:12 +0000 (09:04 -0500)]
menu: deprecate horizontal-padding and vertical-padding

https://bugzilla.gnome.org/show_bug.cgi?id=689168

11 years agomenuitem: deprecate horizontal-padding style property
Cosimo Cecchi [Tue, 27 Nov 2012 13:57:23 +0000 (08:57 -0500)]
menuitem: deprecate horizontal-padding style property

https://bugzilla.gnome.org/show_bug.cgi?id=689168

11 years agoAssamese translation updated
Nilamdyuti Goswami [Wed, 28 Nov 2012 08:19:44 +0000 (13:49 +0530)]
Assamese translation updated

11 years agoAssamese translation updated
Nilamdyuti Goswami [Wed, 28 Nov 2012 08:17:00 +0000 (13:47 +0530)]
Assamese translation updated

11 years agoUpdated Slovenian translation
Matej Urbančič [Mon, 26 Nov 2012 20:12:00 +0000 (21:12 +0100)]
Updated Slovenian translation

11 years agoUpdated Slovenian translation
Matej Urbančič [Mon, 26 Nov 2012 20:07:25 +0000 (21:07 +0100)]
Updated Slovenian translation

11 years agodocs: Add gtk-doc comment for GtkTreeRowReference
David King [Thu, 22 Nov 2012 18:38:21 +0000 (18:38 +0000)]
docs: Add gtk-doc comment for GtkTreeRowReference

https://bugzilla.gnome.org/show_bug.cgi?id=93381

11 years agoSanitize memory handling in cups_request_printer_list_cb
Christophe Fergeau [Thu, 22 Nov 2012 18:04:48 +0000 (19:04 +0100)]
Sanitize memory handling in cups_request_printer_list_cb

gtk+ was trying to display already freed strings, leaking memory,
...I noticed this because I was getting weird blinking characters
as the status of my cups printers, and valgrind confirmed something
was wrong.

https://bugzilla.gnome.org/show_bug.cgi?id=683072

11 years agoi18n: Update British English translation
Philip Withnall [Mon, 26 Nov 2012 09:00:32 +0000 (09:00 +0000)]
i18n: Update British English translation

11 years agoBug 595615 — Use proper ellipses
Philip Withnall [Mon, 26 Nov 2012 08:59:27 +0000 (08:59 +0000)]
Bug 595615 — Use proper ellipses

Use ‘…’ instead of ‘...’ in translatable strings.

Closes: https://bugzilla.gnome.org/show_bug.cgi?id=595615
11 years agoUpdated POTFILES.skip
Piotr Drąg [Sun, 25 Nov 2012 23:46:05 +0000 (00:46 +0100)]
Updated POTFILES.skip

11 years agoreftests: Add reftest for broken fixed height mode
Benjamin Otte [Sun, 25 Nov 2012 22:50:27 +0000 (23:50 +0100)]
reftests: Add reftest for broken fixed height mode

https://bugzilla.gnome.org/show_bug.cgi?id=687816

11 years agotreeview: Delay computing fixed height
Benjamin Otte [Sun, 25 Nov 2012 23:06:40 +0000 (00:06 +0100)]
treeview: Delay computing fixed height

In the setter, we only set fixed height mode and queue a revalidation of
the row heights.

https://bugzilla.gnome.org/show_bug.cgi?id=687816

11 years agosizegroup: Use g_hash_table_add()
Emmanuele Bassi [Sun, 25 Nov 2012 21:57:41 +0000 (22:57 +0100)]
sizegroup: Use g_hash_table_add()

This is the function to use when treating hash tables as a set.

11 years agostyleproperty: Use _gtk_css_initial_value_get()
Benjamin Otte [Sun, 25 Nov 2012 14:35:52 +0000 (15:35 +0100)]
styleproperty: Use _gtk_css_initial_value_get()

See previous commit(s).

11 years agocssvalue: Use _gtk_css_initial_value_get() for the inherit value
Benjamin Otte [Sun, 25 Nov 2012 14:35:27 +0000 (15:35 +0100)]
cssvalue: Use _gtk_css_initial_value_get() for the inherit value

See previous commit(s).

11 years agocssvalue: Add _gtk_css_initial_value_get()
Benjamin Otte [Sun, 25 Nov 2012 14:31:59 +0000 (15:31 +0100)]
cssvalue: Add _gtk_css_initial_value_get()

... so we don't bump a refcount whenever we get the initial singleton.

We want to use this function instead of
_gtk_css_style_property_get_initial_value() everywhere where we compute
values, because some initial values may depend on settings soon.

11 years agoizegroup: Improve performance
Benjamin Otte [Sun, 25 Nov 2012 12:32:21 +0000 (13:32 +0100)]
izegroup: Improve performance

Keep a list of all groups and avoid groups we already handled. Speeds up
GtkToolpalette a lot.

11 years agowidget: Remove an unneeded queue_resize()
Benjamin Otte [Sun, 25 Nov 2012 03:41:49 +0000 (04:41 +0100)]
widget: Remove an unneeded queue_resize()

Resizes are queued via
   gtk_widget_propagate_state()
=> gtk_style_context_set_state()
=> gtk_style_context_queue_invalidate()
=> gtk_style_context_validate()
=> _gtk_widget_style_context_invalidated()
so there's no need to queue an extra one.

11 years agodeprecations: Move files into deprecated/ dir
Benjamin Otte [Sun, 25 Nov 2012 03:16:43 +0000 (04:16 +0100)]
deprecations: Move files into deprecated/ dir

11 years agogtkgradient: Deprecate
Benjamin Otte [Sun, 25 Nov 2012 02:47:51 +0000 (03:47 +0100)]
gtkgradient: Deprecate

11 years agosymboliccolor: Deprecate
Benjamin Otte [Sun, 25 Nov 2012 02:34:08 +0000 (03:34 +0100)]
symboliccolor: Deprecate

Symbolic colors are an implementation detail of the CSS engine and have
been superceded by GtkCssColorValue. We don't want them clobbering the
public API. In particular because the only use I could find in the
public API is people using it to shade colors.

11 years agothemingengine: Do shading with GtkHSLA
Benjamin Otte [Sun, 25 Nov 2012 02:33:33 +0000 (03:33 +0100)]
themingengine: Do shading with GtkHSLA

... instead of with symbolic colors.

11 years agohsla: Add _gtk_hsla_shade()
Benjamin Otte [Sun, 25 Nov 2012 02:24:47 +0000 (03:24 +0100)]
hsla: Add _gtk_hsla_shade()

We use it in multiple places, so better split it out.

11 years agosymboliccolor: Remove extra includes
Benjamin Otte [Sun, 25 Nov 2012 01:59:15 +0000 (02:59 +0100)]
symboliccolor: Remove extra includes

That header was included in way too many places.

11 years agostyleprovider: Change function prototype
Benjamin Otte [Sun, 25 Nov 2012 01:28:59 +0000 (02:28 +0100)]
styleprovider: Change function prototype

Make _gtk_style_provider_private_get_color() return a GtkCssValue (a
GtkCssColorValue to be exact) instead of GtkSymbolicColor.

With this, the symbolic color usage inside GTK is minimized.

11 years agosymboliccolor: Implement using GtkCssValue
Benjamin Otte [Sun, 25 Nov 2012 00:55:53 +0000 (01:55 +0100)]
symboliccolor: Implement using GtkCssValue

11 years agosymboliccolor: symbolic color is no longer a CssValue
Benjamin Otte [Sat, 24 Nov 2012 21:04:24 +0000 (22:04 +0100)]
symboliccolor: symbolic color is no longer a CssValue

This is not needed anymore, because we have GtkCssColorValue for that
now.

11 years agosymboliccolor: Change prototype of function
Benjamin Otte [Sat, 24 Nov 2012 20:56:31 +0000 (21:56 +0100)]
symboliccolor: Change prototype of function

Makes it easier to use the function in the places where it's still used.

11 years agosettings: Remove support for color schemes
Benjamin Otte [Sat, 24 Nov 2012 20:27:33 +0000 (21:27 +0100)]
settings: Remove support for color schemes

Color schemes were unused and their interaction with CSS3 themes was
undefined. So we decided to remove support for them.

This commit does that.

11 years agostyleprovider: Remove (broken) implementations of get_style
Benjamin Otte [Sat, 24 Nov 2012 20:08:11 +0000 (21:08 +0100)]
styleprovider: Remove (broken) implementations of get_style

The implementations were broken and get_style() is deprecated and
documented to return NULL now.

11 years agostylecascade: Remove get_icon_factory() implementation
Benjamin Otte [Sat, 24 Nov 2012 19:59:37 +0000 (20:59 +0100)]
stylecascade: Remove get_icon_factory() implementation

It was redundant.

11 years agostyleprovider: Deprecate two non-working functions
Benjamin Otte [Sat, 24 Nov 2012 19:57:12 +0000 (20:57 +0100)]
styleprovider: Deprecate two non-working functions

Both of them weren't used inside GTK. And apparently they weren't used
outside of GTK either, as alex recently mentioned them being severly
broken.

11 years agodocs: gtk_file_chooser_get_filename() returns absolute paths
David King [Thu, 22 Nov 2012 21:15:06 +0000 (21:15 +0000)]
docs: gtk_file_chooser_get_filename() returns absolute paths

The documentation for gtk_file_chooser_get_filenames() states that the
returned filenames are absolute paths, and uses g_file_get_path() to
construct the filename. The same function is used to construct the
filename in gtk_file_chooser_get_filename(), so it should also return
absolute paths.

https://bugzilla.gnome.org/show_bug.cgi?id=371034

11 years agoAssamese translation updated
Nilamdyuti Goswami [Fri, 23 Nov 2012 14:48:36 +0000 (20:18 +0530)]
Assamese translation updated

11 years agoAssamese translation updated
Nilamdyuti Goswami [Fri, 23 Nov 2012 14:47:24 +0000 (20:17 +0530)]
Assamese translation updated

11 years agodocs: Fix typo in gtk_tree_view_set_tooltip_column comment
David King [Thu, 22 Nov 2012 18:57:15 +0000 (18:57 +0000)]
docs: Fix typo in gtk_tree_view_set_tooltip_column comment

https://bugzilla.gnome.org/show_bug.cgi?id=688884

11 years agodocs: Improve GtkTreeModel iteration pattern
David King [Thu, 22 Nov 2012 17:38:41 +0000 (17:38 +0000)]
docs: Improve GtkTreeModel iteration pattern

Iterating over the model in this way means that use of continue is less
error-prone, as the increment is part of the loop construct.

https://bugzilla.gnome.org/show_bug.cgi?id=548793

11 years agodocs: Correct GtkWidget::scroll-event mask documentation
David King [Thu, 22 Nov 2012 17:15:56 +0000 (17:15 +0000)]
docs: Correct GtkWidget::scroll-event mask documentation

Since GTK+ 3.3.18, GDK_SCROLL_MASK has been used as the mask for scroll
events. Update the documentation to reflect this.

https://bugzilla.gnome.org/show_bug.cgi?id=677339

11 years agoaboutdialog: Fix add_credit_section() annotations
Martin Pitt [Thu, 22 Nov 2012 13:51:52 +0000 (14:51 +0100)]
aboutdialog: Fix add_credit_section() annotations

The people argument is an array.

https://mail.gnome.org/archives/python-hackers-list/2012-November/msg00010.html

11 years agogtk: don't let insensitive children eat scroll events when bubbling down
Michael Natterer [Thu, 22 Nov 2012 12:37:58 +0000 (13:37 +0100)]
gtk: don't let insensitive children eat scroll events when bubbling down

When event capturing is enabled, stop propagating scroll events
at insensitive widgets, but don't handle them (don't return TRUE),
so they can bubble up again and reach their handling widgets.

11 years agoUpdated Dutch translation
Wouter Bolsterlee [Thu, 22 Nov 2012 11:02:09 +0000 (12:02 +0100)]
Updated Dutch translation

11 years agoBug 688710 - splashscreen shouldn't be always on top on OS X
Camillo Lugaresi [Tue, 20 Nov 2012 08:52:39 +0000 (02:52 -0600)]
Bug 688710 - splashscreen shouldn't be always on top on OS X

Activate the "hides on deactivate" behavior for splashscreens,
torn-off menus, utility windows, tooltips and notifications: when
another application is brought to the front, these windows are hidden
so as not to obscure it. This is the expected behavior for
application-specific floating windows on OS X.
(cherry picked from commit 0596f5591f36f57b99103914ba03abbed5765e8b)

11 years agotree-view: don't use gtk_style_context_set_background()
Cosimo Cecchi [Tue, 20 Nov 2012 17:43:37 +0000 (12:43 -0500)]
tree-view: don't use gtk_style_context_set_background()

Render a background with gtk_render_background() in draw() instead.
Note that we still use gtk_style_context_set_background() for the header
window.

https://bugzilla.gnome.org/show_bug.cgi?id=688744

11 years agoiconview: don't use gtk_style_context_set_background()
Cosimo Cecchi [Tue, 20 Nov 2012 17:28:14 +0000 (12:28 -0500)]
iconview: don't use gtk_style_context_set_background()

Render a background with gtk_render_background() during draw instead.

https://bugzilla.gnome.org/show_bug.cgi?id=688744

11 years agoquartz: fix the mapping of GDK_KP_Enter in known_numeric_keys[]
Michael Natterer [Tue, 20 Nov 2012 18:10:06 +0000 (19:10 +0100)]
quartz: fix the mapping of GDK_KP_Enter in known_numeric_keys[]

so it will actually be used, instead of delivering GDK_Return
when it should be GDK_KP_Enter.
(cherry picked from commit 27f3fcf12c39d551532872cf26823993c93eb04d)

11 years agoPost-release version bump
Matthias Clasen [Tue, 20 Nov 2012 12:00:53 +0000 (07:00 -0500)]
Post-release version bump