gtk4.git
15 years agolink-button: Add ::activate-link signal
Emmanuele Bassi [Thu, 14 Oct 2010 14:22:19 +0000 (15:22 +0100)]
link-button: Add ::activate-link signal

We cannot use the GtkButton::clicked signal to override the default
behavior of GtkLinkButton (i.e. call gtk_show_uri()), because
GtkButton::clicked is registered as a RUN_FIRST signal, which obviously
prevents any other signal handler connected to it to stop the
propagation before the class handler has a chance to run.

For this reason we can add a GtkLinkButton::activate-link signal, which
will be emitted by the default GtkButton::clicked signal handler; the
::activate-link signal has a boolean return value, which allows simpler
code for stopping the propagation to the next signal handler.

https://bugzilla.gnome.org/show_bug.cgi?id=632150

15 years agoRemoved screwy GtkScrolledWindow logic
Tristan Van Berkom [Fri, 15 Oct 2010 05:14:14 +0000 (14:14 +0900)]
Removed screwy GtkScrolledWindow logic

Removed portion of scrolled-window that observes user-set
size request data (aux_info->width/height) on the child directly
in order to derive it's minimum possible size... if the scrolled
window has (auto/always) scrollbars in a said orientation; only
request enough space for the scrollbars (bug 631976).

15 years agoImplement get_preferred_width/height in GtkEventBox
Matthias Clasen [Fri, 15 Oct 2010 03:03:39 +0000 (23:03 -0400)]
Implement get_preferred_width/height in GtkEventBox

And drop size_request.

15 years agoUse gtk_container_handle_border_width() in GtkEventBox
Matthias Clasen [Fri, 15 Oct 2010 02:34:39 +0000 (22:34 -0400)]
Use gtk_container_handle_border_width() in GtkEventBox

15 years agoInclusion cleanups in sources
Matthias Clasen [Fri, 15 Oct 2010 02:05:51 +0000 (22:05 -0400)]
Inclusion cleanups in sources

Try to do inclusions in the same sequence, more or less.

15 years agoInclusion cleanups in headers
Matthias Clasen [Tue, 5 Oct 2010 20:11:49 +0000 (16:11 -0400)]
Inclusion cleanups in headers

15 years agoMake gdk.h an include-only header
Matthias Clasen [Sat, 25 Sep 2010 03:46:32 +0000 (23:46 -0400)]
Make gdk.h an include-only header

Move declarations to other headers, and introduce some new ones:
gdkmain.h, gdkrectangle.h, gdkthreads.h.

15 years agoDrop dead gxid_proto.h header
Matthias Clasen [Tue, 5 Oct 2010 20:44:55 +0000 (16:44 -0400)]
Drop dead gxid_proto.h header

This header is no longer used.

15 years agoSome GDK documentation fixes
Matthias Clasen [Fri, 15 Oct 2010 01:40:33 +0000 (21:40 -0400)]
Some GDK documentation fixes

15 years agoEnable the build of MS Windows engine again
Fridrich Štrba [Thu, 14 Oct 2010 18:56:17 +0000 (20:56 +0200)]
Enable the build of MS Windows engine again

The MS Windows engine compiles anew, so enable it so that one can fix remaining issues

15 years agoPart of bgo#613728 - [GtkTreeView] Focus the list, not the headers, when Tab-focusing
Federico Mena Quintero [Thu, 14 Oct 2010 18:17:46 +0000 (13:17 -0500)]
Part of bgo#613728 - [GtkTreeView] Focus the list, not the headers, when Tab-focusing

The idea is that it is way more common to want to manipulate the
actual list, rather than the headers.  Once you Tab into the treeview
(and the list part gets the focus), you can use Shift-Tab to focus
the headers.

This means that some hysteresis is added to the focus chain, but
it makes the treeview a lot more convenient to focus with the keyboard.

Signed-off-by: Federico Mena Quintero <federico@novell.com>
15 years agoInitialize device_cursor hash table before creating impl window
Kristian Rietveld [Thu, 14 Oct 2010 16:01:51 +0000 (18:01 +0200)]
Initialize device_cursor hash table before creating impl window

15 years agoDo not call validate_visible_area() from the draw method
Kristian Rietveld [Thu, 14 Oct 2010 15:26:17 +0000 (17:26 +0200)]
Do not call validate_visible_area() from the draw method

Because validate_visible_area() can modify the window size (and thus
the underlying surface), it should not be called from within the draw
method.  Given that the presize handler is run with a higher priority
than redraw, and the presize handler will validate the visible area,
there should not be cases wherein the draw method is called and
validate_visible_area() has not been run yet.

However, one such a case was gdk_window_process_updates(), which would
trigger the draw method at some point.  We now work around this
by factoring this in a new gtk_tree_view_bin_process_updates() function
that will run the presize handler first if needed.

Note: for other platforms, it might still be the case that the draw
method is called and validate_visible_area() has not been run yet.
(For example the Mac backend calls gdk_window_process_updates() from
the drawRect method, and the redraw-in-idle handling thus works
differently).  This does not seem to be a problem now, if
it will be in the future we need to take care of that then.

15 years agoAdd test for bug 163214
Kristian Rietveld [Thu, 14 Oct 2010 13:07:06 +0000 (15:07 +0200)]
Add test for bug 163214

15 years agoAdd test for bug 111500
Kristian Rietveld [Thu, 14 Oct 2010 12:11:30 +0000 (14:11 +0200)]
Add test for bug 111500

This test is constructed in such a way that it breaks when you
(currently!) do not trigger a size-request from validate_visible_area.
Especially row expansion appears to have a need for this currently.

15 years agoAdd unit test for bug 93584
Kristian Rietveld [Thu, 14 Oct 2010 09:24:46 +0000 (11:24 +0200)]
Add unit test for bug 93584

15 years agoAlso test scrolling to a new row in a mixed height model
Kristian Rietveld [Thu, 14 Oct 2010 08:57:20 +0000 (10:57 +0200)]
Also test scrolling to a new row in a mixed height model

15 years agoTrying to fix MS Windows engine
Martin Schlemmer [Thu, 14 Oct 2010 12:15:06 +0000 (14:15 +0200)]
Trying to fix MS Windows engine

15 years agoInitialize variables
Matthias Clasen [Thu, 14 Oct 2010 12:12:11 +0000 (08:12 -0400)]
Initialize variables

valgrind was complaining about uninitialized values here, and
sure enough, gdkdevice-xi.c doesn't set device position.

15 years agobuild: Require cairo-gobject for GDK
Benjamin Otte [Thu, 14 Oct 2010 11:53:23 +0000 (13:53 +0200)]
build: Require cairo-gobject for GDK

With the new signal added, GDK now also requires cairo-gobject.

15 years agogdk: add slots for all GdkWindowObject signals
Michael Natterer [Thu, 14 Oct 2010 11:43:08 +0000 (13:43 +0200)]
gdk: add slots for all GdkWindowObject signals

Also add padding for future extension.

15 years agoBug 631599 - Allow to use arbitrary surfaces for offscreen windows
Michael Natterer [Thu, 14 Oct 2010 11:25:23 +0000 (13:25 +0200)]
Bug 631599 - Allow to use arbitrary surfaces for offscreen windows

Add signal GdkWindow::create-surface which allows to use any
surface type as storage for offscreen windows.

Test the new signal in tests/gdkoffscreenbox.c

15 years agogtk: add cast to fix warning in my last commit
Michael Natterer [Thu, 14 Oct 2010 09:32:44 +0000 (11:32 +0200)]
gtk: add cast to fix warning in my last commit

15 years agoImplement style_set() for GtkWindow to resize window grips when the style changes.
Cody Russell [Thu, 14 Oct 2010 04:04:11 +0000 (23:04 -0500)]
Implement style_set() for GtkWindow to resize window grips when the style changes.

15 years agoDeprecate GtkComboBoxEntry in favor of added properties to GtkComboBox
Tristan Van Berkom [Thu, 14 Oct 2010 01:52:27 +0000 (21:52 -0400)]
Deprecate GtkComboBoxEntry in favor of added properties to GtkComboBox

GtkComboBox now sports a construct-only "has-entry" property which
decides if it uses a GtkEntry to allow additional user input. Also
it has a new "entry-text-column" to fetch strings for the entry
from the model.

This patch deprecates the GtkComboBoxEntry and updates the rest of GTK+
to use the new semantics on GtkComboBox instead.

GtkComboBoxEntry will be removed altogether before GTK+ 3, in a
later commit.

15 years agoAdd new API to docs and symbol list
Matthias Clasen [Wed, 13 Oct 2010 15:24:56 +0000 (11:24 -0400)]
Add new API to docs and symbol list

15 years agoSupport GtkWidget expand properties in GtkNotebook
Matthias Clasen [Mon, 11 Oct 2010 19:23:03 +0000 (15:23 -0400)]
Support GtkWidget expand properties in GtkNotebook

We expand a tab if either tab-expand is set, or the generic widget
expand property in the correct direction is set. And we do not
propagate expand flags from tab labels to the notebook, only
the expand flags from the pages.

15 years agoSupport GtkWidget expand properties in GtkTable
Matthias Clasen [Mon, 11 Oct 2010 17:56:52 +0000 (13:56 -0400)]
Support GtkWidget expand properties in GtkTable

We expand a child when it either has the table-specific expand flag
or the generic expand property set. Override compute_expand so that
it also takes the table-specific expand flags of children into
account.

https://bugzilla.gnome.org/show_bug.cgi?id=628902

15 years agoAdd a GtkTable testcase
Matthias Clasen [Mon, 11 Oct 2010 17:26:11 +0000 (13:26 -0400)]
Add a GtkTable testcase

15 years agoSupport GtkWidget expand properties in GtkBox
Havoc Pennington [Mon, 6 Sep 2010 16:30:40 +0000 (12:30 -0400)]
Support GtkWidget expand properties in GtkBox

This consists of:
* expand a child if either child->expand || gtk_widget_get_expand(child)
* override compute_expand so that child->expand will cause us to
  return TRUE for gtk_widget_get_expand()

https://bugzilla.gnome.org/show_bug.cgi?id=628902

15 years agoadd tests/testexpand.c used to test the expand props on GtkWidget
Havoc Pennington [Mon, 6 Sep 2010 16:31:54 +0000 (12:31 -0400)]
add tests/testexpand.c used to test the expand props on GtkWidget

There are two colored boxes with toggle buttons nested
inside several GtkBox. Toggling these to expand mode
should automatically propagate expansion up through
the several GtkBox such that resizing the window
results in resizing the colored boxes.

https://bugzilla.gnome.org/show_bug.cgi?id=628902

15 years agoAdd horizontal and vertical expand flags, compute_expand() to GtkWidget
Havoc Pennington [Mon, 6 Sep 2010 16:26:32 +0000 (12:26 -0400)]
Add horizontal and vertical expand flags, compute_expand() to GtkWidget

GtkWidget now has flags for horizontal and vertical expand, and
a compute_expand() method. compute_expand() is used by containers
to set a default expand flag. (If a widget has expand set explicitly,
it always overrides the results of compute_expand.)

GtkContainer has a default implementation of compute_expand which
simply walks over all child widgets and sets expand=TRUE
if any child is expanding.

The overall effect is that you only have to set expand on
leaf nodes in the widget tree, while previously you had to
set expand all the way up the tree as you packed every
container. Also, now containers need not have their own child
properties for expand.

For old containers which do have "expand" child properties,
they should override compute_expand and include the child
properties in whether the container is expand=TRUE.
Also, the old container should use
"child_prop_expand || gtk_widget_compute_expand()" everywhere
it previously used simply "child_prop_expand"

https://bugzilla.gnome.org/show_bug.cgi?id=628902

15 years agoSome more resize grip cleanups
Matthias Clasen [Wed, 13 Oct 2010 14:39:09 +0000 (10:39 -0400)]
Some more resize grip cleanups

Don't recompute hints whenever we determine the drag edge, and
take the drag edge into account when deciding whether to show
the grip.

15 years agoUnify handling of GtkWindow::resizable property
Matthias Clasen [Mon, 11 Oct 2010 21:23:01 +0000 (17:23 -0400)]
Unify handling of GtkWindow::resizable property

15 years agoRemove GtkWrapBox from the docs, too
Matthias Clasen [Wed, 13 Oct 2010 14:21:04 +0000 (10:21 -0400)]
Remove GtkWrapBox from the docs, too

15 years agoMake GtkRecentManager use standard icon names for themed icons
Matthias Clasen [Wed, 13 Oct 2010 14:18:03 +0000 (10:18 -0400)]
Make GtkRecentManager use standard icon names for themed icons

15 years agoSome resize grip cleanups
Matthias Clasen [Wed, 13 Oct 2010 13:51:54 +0000 (09:51 -0400)]
Some resize grip cleanups

Safe space in the private struct, always set the cursor when we
update visibility, since the cursor depends on resizability conditions
too. And don't set the shape needlessly.

15 years agoRemoved all traces of GtkWrapBox from GTK+.
Tristan Van Berkom [Wed, 13 Oct 2010 13:26:08 +0000 (22:26 +0900)]
Removed all traces of GtkWrapBox from GTK+.

15 years agoDon't start a drag if we are not resizable
Matthias Clasen [Wed, 13 Oct 2010 12:28:21 +0000 (08:28 -0400)]
Don't start a drag if we are not resizable

This was caused by get_drag_edge() returning -1 to mean 'no drag',
but callers didn't really pay attention. Change get_drag_edge()
to return a boolean instead.

15 years agoUse GSettings for the filechooser settings
Christian Persch [Tue, 28 Sep 2010 20:56:02 +0000 (22:56 +0200)]
Use GSettings for the filechooser settings

Bug #630850.

15 years agoUpdated Spanish translation
Jorge González [Wed, 13 Oct 2010 12:26:30 +0000 (14:26 +0200)]
Updated Spanish translation

15 years agoUpdated Spanish translation
Jorge González [Tue, 12 Oct 2010 23:25:28 +0000 (01:25 +0200)]
Updated Spanish translation

15 years agoUpdated Spanish translation
Jorge González [Tue, 12 Oct 2010 19:56:42 +0000 (21:56 +0200)]
Updated Spanish translation

15 years agoBug 563002 - Doesn't call 'update-preview' on set_filename
Michael Natterer [Tue, 12 Oct 2010 19:20:16 +0000 (21:20 +0200)]
Bug 563002 - Doesn't call 'update-preview' on set_filename

Call gtk_tree_view_set_cursor() in addition to
gtk_tree_selection_select_iter() when selecting the file in
show_and_select_files() so the preview update machinery gets
triggered.
(cherry picked from commit 795c8070db6c97ace040bfd220794bae3e2721c2)

15 years agoUpdated Slovenian translation
Matej Urbančič [Tue, 12 Oct 2010 18:09:27 +0000 (20:09 +0200)]
Updated Slovenian translation

15 years agoDon't check resize-grip-visible in default values unit test
Kristian Rietveld [Tue, 12 Oct 2010 17:08:34 +0000 (19:08 +0200)]
Don't check resize-grip-visible in default values unit test

The value for this property is determined at runtime.

15 years agoDisable resize grip on spin button test
Kristian Rietveld [Tue, 12 Oct 2010 16:53:44 +0000 (18:53 +0200)]
Disable resize grip on spin button test

Otherwise, the resize grip obscures the decrement button causing the
test to fail.

15 years agoFix scrolled window policy so the size allocation does not "flap"
Kristian Rietveld [Tue, 12 Oct 2010 16:25:22 +0000 (18:25 +0200)]
Fix scrolled window policy so the size allocation does not "flap"

15 years agoHandle off-by-one errors due to rounding
Kristian Rietveld [Tue, 12 Oct 2010 15:39:26 +0000 (17:39 +0200)]
Handle off-by-one errors due to rounding

15 years agoUse double iso. float since that is what GtkAdjustment uses
Kristian Rietveld [Tue, 12 Oct 2010 15:25:21 +0000 (17:25 +0200)]
Use double iso. float since that is what GtkAdjustment uses

15 years agoUse GtkAdjustment accessors to properly compress changed signals
Kristian Rietveld [Tue, 12 Oct 2010 15:20:44 +0000 (17:20 +0200)]
Use GtkAdjustment accessors to properly compress changed signals

15 years agoRework GdkPangoRenderer example to use existing api
Matthias Clasen [Tue, 12 Oct 2010 15:28:21 +0000 (11:28 -0400)]
Rework GdkPangoRenderer example to use existing api

15 years agodocs: add a longdesc for X-specifics
Matthias Clasen [Tue, 12 Oct 2010 15:01:10 +0000 (11:01 -0400)]
docs: add a longdesc for X-specifics

15 years agodocs: Update cairo-related docs
Matthias Clasen [Tue, 12 Oct 2010 14:53:44 +0000 (10:53 -0400)]
docs: Update cairo-related docs

15 years agodocs: Remove mentions of GdkPixmap from cursor docs
Matthias Clasen [Tue, 12 Oct 2010 13:59:05 +0000 (09:59 -0400)]
docs: Remove mentions of GdkPixmap from cursor docs

15 years agoMove GdkVisualType docs inline
Matthias Clasen [Tue, 12 Oct 2010 13:45:11 +0000 (09:45 -0400)]
Move GdkVisualType docs inline

15 years agodocs: Correct some link targets
Matthias Clasen [Tue, 12 Oct 2010 12:59:46 +0000 (08:59 -0400)]
docs: Correct some link targets

15 years agoUpdated Spanish translation
Jorge González [Tue, 12 Oct 2010 15:16:12 +0000 (17:16 +0200)]
Updated Spanish translation

15 years agoAdd two new functions to gtk.symbols
Tor Lillqvist [Tue, 12 Oct 2010 11:39:30 +0000 (14:39 +0300)]
Add two new functions to gtk.symbols

15 years agoBug 631599 - Allow to use arbitrary surfaces for offscreen windows
Michael Natterer [Tue, 12 Oct 2010 09:32:42 +0000 (11:32 +0200)]
Bug 631599 - Allow to use arbitrary surfaces for offscreen windows

As a first step, create surfaces lazily and factor surface creation
out to a single function.

15 years agoAdded logic to GtkScrolledWindow when allocating height-for-width children.
Tristan Van Berkom [Tue, 12 Oct 2010 08:16:32 +0000 (17:16 +0900)]
Added logic to GtkScrolledWindow when allocating height-for-width children.

This patch makes the scrolled window reconsider allocating the child
the full width or height (depending on the child's request mode) without
a scrollbar. For instance when the child is height-for-width; the child
will first be tested if the content's height for full allocated width
(without a vscrollbar) will allow the contents height for that width
to fit the allocated height.

Patch is a simplified version of code inspected in st-scroll-view.c.
Note that this patch assumes children will begin to scroll only after
reaching their minimum size; adding a property to the future
GtkScrollableIface to decide whether to scroll-to-minimum or scroll-to-natural
will effect this code (it should then reconsider whether the child
will scroll below the natural size instead of the minimum).

Patch addresses bug 629778.

15 years agoMake GtkScrolledWindow offload border-width calculations to GtkContainerClass.
Tristan Van Berkom [Tue, 12 Oct 2010 08:10:55 +0000 (17:10 +0900)]
Make GtkScrolledWindow offload border-width calculations to GtkContainerClass.

15 years agoDon't link to old stuff
Matthias Clasen [Tue, 12 Oct 2010 05:10:32 +0000 (01:10 -0400)]
Don't link to old stuff

15 years agoAdd a missing close tag
Matthias Clasen [Tue, 12 Oct 2010 05:05:39 +0000 (01:05 -0400)]
Add a missing close tag

15 years agoRemove a note about GTK 1.2 api
Matthias Clasen [Tue, 12 Oct 2010 04:57:13 +0000 (00:57 -0400)]
Remove a note about GTK 1.2 api

15 years agoFix a trivial typo
Matthias Clasen [Tue, 12 Oct 2010 04:50:18 +0000 (00:50 -0400)]
Fix a trivial typo

15 years agoReorder some sections
Matthias Clasen [Tue, 12 Oct 2010 04:46:59 +0000 (00:46 -0400)]
Reorder some sections

15 years agoRemove mentions of dialog separator from the docs
Matthias Clasen [Tue, 12 Oct 2010 04:38:04 +0000 (00:38 -0400)]
Remove mentions of dialog separator from the docs

15 years agoSome documentation updates
Matthias Clasen [Tue, 12 Oct 2010 04:19:35 +0000 (00:19 -0400)]
Some documentation updates

15 years agoInspect geometry when setting resize grip cursors
Matthias Clasen [Tue, 12 Oct 2010 03:26:58 +0000 (23:26 -0400)]
Inspect geometry when setting resize grip cursors

We want to use the same cursor as metacity when the window is only
resizable in one dimension.

15 years agoGtkWindow: Allow setting size from geometry
Owen W. Taylor [Sun, 10 Oct 2010 02:38:54 +0000 (22:38 -0400)]
GtkWindow: Allow setting size from geometry

If you set a geometry widget via gtk_window_set_geometry_hints() it
becomes very hard to compute appropriate toplevel sizes in pixels
to make the window a particular size. Synthesizing strings and passing
them to gtk_window_parse_geometry() is possible, but to avoid
avoid such ugliness, add functions:

 gtk_window_set_default_geometry()
 gtk_window_resize_to_geometry()

That act like gtk_window_set_default_size() and
gtk_window_resize() but are in terms of the resize increments of the
geometry widget.

https://bugzilla.gnome.org/show_bug.cgi?id=631796

15 years agoWarn when calling gtk_window_parse_geometry() on an empty window
Owen W. Taylor [Sun, 10 Oct 2010 02:29:20 +0000 (22:29 -0400)]
Warn when calling gtk_window_parse_geometry() on an empty window

gtk_window_parse_geometry() gets the size of the window in order
to interpret the position of the window; calling it before, say,
calling gtk_widget_show_all() on a window is a subtle trap, so
add a warning in the case we can easily detect.

https://bugzilla.gnome.org/show_bug.cgi?id=631794

15 years agoFix handling of the geometry widget
Owen W. Taylor [Sun, 10 Oct 2010 02:15:34 +0000 (22:15 -0400)]
Fix handling of the geometry widget

The geometry widget feature of gtk_window_set_geometry_hints() has
never really worked right because the calculation that GTK+ did to
compute the base size of the window only worked when the geometry
widget had a larger minimum size than anything else in the window.

Setup:
* Move the GtkSizeGroup private functions to a new private header
  gtksizegroup-private.h
* Add the possibilty to pass flags to _gtk_size_group_queue_resize(),
  with the flag GTK_QUEUE_RESIZE_INVALIDATE_ONLY to suppress adding
  the widget's toplevel to the resize queue.
* _gtk_container_resize_invalidate() is added to implement that feature
* _gtk_widget_override_size_request()/_gtk_widget_restore_size_request()
  allow temporarily forcing a large minimum size on the geometry
  widget without creating resize loops.

GtkWindow:
* Compute the extra width/height around the geometry widget
  correctly; print a warning if the computation fails.
* Always make the minimum size at least the natural minimum
  size of the toplevel; GTK+ now fails badly with underallocation.
* Always set the base size hint; we were failing to set it
  properly when the specified minimum size was overriden, but
  it's harmless to always set it.

Tests:
* New test 'testgeometry' that replaces the 'gridded geometry' test
  from testgtk. The new test is roughly similar but creates a bunch
  of windows showing different possibilities.
* The testgtk test is removed. No need to have both.

https://bugzilla.gnome.org/show_bug.cgi?id=68668

15 years agoAdded documentation to GtkContainer about implementing containers using height-for...
Tristan Van Berkom [Mon, 11 Oct 2010 15:47:43 +0000 (00:47 +0900)]
Added documentation to GtkContainer about implementing containers using height-for-width apis.

15 years agobuild: Distribute README.multipress
Vincent Untz [Mon, 11 Oct 2010 11:59:06 +0000 (13:59 +0200)]
build: Distribute README.multipress

https://bugzilla.gnome.org/show_bug.cgi?id=631872

15 years agoUpdated Galician translations
Fran Diéguez [Mon, 11 Oct 2010 08:22:04 +0000 (10:22 +0200)]
Updated Galician translations

15 years ago[l10n] Updated Estonian translation
Ivar Smolin [Mon, 11 Oct 2010 06:40:35 +0000 (09:40 +0300)]
[l10n] Updated Estonian translation

15 years agodocs: Move documentation do inline comments: GtkContainer
Javier Jardón [Mon, 11 Oct 2010 02:21:36 +0000 (04:21 +0200)]
docs: Move documentation do inline comments: GtkContainer

15 years agodocs: Improve reference crossing in GtkWidget docs
Javier Jardón [Mon, 11 Oct 2010 02:59:19 +0000 (04:59 +0200)]
docs: Improve reference crossing in GtkWidget docs

15 years agoPlug memory leak in GtkWindow
Xan Lopez [Sun, 10 Oct 2010 09:18:27 +0000 (18:18 +0900)]
Plug memory leak in GtkWindow

1,936 (112 direct, 1,824 indirect) bytes in 4 blocks are definitely lost in loss record 13,453 of 13,673
    at 0x4005BDC: malloc (vg_replace_malloc.c:195)
    by 0x69615A1: cairo_region_create (cairo-region.c:196)
    by 0x683BFBF: gdk_cairo_region_create_from_surface (gdkcairo.c:455)
    by 0x670C882: set_grip_shape (gtkwindow.c:5020)
    by 0x670CFBF: resize_grip_create_window (gtkwindow.c:5271)
    by 0x670C466: gtk_window_realize (gtkwindow.c:4902)

Signed-off-by: Benjamin Otte <otte@redhat.com>
15 years agoPlug leak in gdkcairo.c
Xan Lopez [Sun, 10 Oct 2010 15:16:52 +0000 (00:16 +0900)]
Plug leak in gdkcairo.c

1,968 (1,236 direct, 732 indirect) bytes in 1 blocks are definitely lost in loss record 11,816 of 11,947
   at 0x4025BDC: malloc (vg_replace_malloc.c:195)
   by 0x6950676: _context_get (cairo.c:250)
   by 0x6950940: cairo_create (cairo.c:370)
   by 0x685CD1E: _gdk_cairo_surface_extents (gdkcairo.c:381)
   by 0x685CEDA: gdk_cairo_region_create_from_surface (gdkcairo.c:433)

Signed-off-by: Benjamin Otte <otte@redhat.com>
15 years agodocs: Move documentation to inline comments: GtkScrolledWindow
Javier Jardón [Mon, 11 Oct 2010 00:01:14 +0000 (02:01 +0200)]
docs: Move documentation to inline comments: GtkScrolledWindow

15 years agodocs: Region returned by gdk_cairo_region_create_from_surface() should be freed
Javier Jardón [Mon, 11 Oct 2010 00:10:24 +0000 (02:10 +0200)]
docs: Region returned by gdk_cairo_region_create_from_surface() should be freed

15 years agoUpdated Catalan translation
Gil Forcada [Sun, 10 Oct 2010 21:22:40 +0000 (23:22 +0200)]
Updated Catalan translation

15 years agol10n: Updated Greek translation for gtk+ properties
Simos Xenitellis [Sun, 10 Oct 2010 18:18:39 +0000 (21:18 +0300)]
l10n: Updated Greek translation for gtk+ properties

15 years agoUpdated Spanish translation
Jorge González [Sat, 9 Oct 2010 12:26:20 +0000 (14:26 +0200)]
Updated Spanish translation

15 years agoupdating translation for Punjabi
A S Alam [Sat, 9 Oct 2010 02:57:32 +0000 (08:27 +0530)]
updating translation for Punjabi

15 years agoFix a problem with 'resize grip avoidance' in scrollbars
Matthias Clasen [Sat, 9 Oct 2010 02:04:33 +0000 (22:04 -0400)]
Fix a problem with 'resize grip avoidance' in scrollbars

We need to be a little more careful when determining the overlap
between the new allocation and the grip area. This was causing
vertical scrollbars in evince to overlap with the grip.

15 years agoCheck for grip_window != NULL in gtk_window_draw
Matthias Clasen [Sat, 9 Oct 2010 00:17:51 +0000 (20:17 -0400)]
Check for grip_window != NULL in gtk_window_draw

This is safer, since GtkWindow subclasses may override realize
and end up with has_resize_grip == TRUE, but grip_window still
being NULL - as is the case with GtkPlug.

15 years agogail: GtkTreeViewColumn doesn't have a destroy signal anymore
William Jon McCann [Fri, 8 Oct 2010 23:57:06 +0000 (19:57 -0400)]
gail: GtkTreeViewColumn doesn't have a destroy signal anymore

Replace with g_object_set_qdata_full usage.

15 years agoDon't assume widgets are containers when walking the focus tree
William Jon McCann [Fri, 8 Oct 2010 23:26:01 +0000 (19:26 -0400)]
Don't assume widgets are containers when walking the focus tree

This causes problems for links in labels.

15 years agoautogen: Explicitly allow libtool > 2.2
Colin Walters [Fri, 8 Oct 2010 20:43:38 +0000 (16:43 -0400)]
autogen: Explicitly allow libtool > 2.2

Apparently 2.4 is out, I'm not aware of a reason we shouldn't
allow it.

15 years agoFix grip window positioning
Matthias Clasen [Fri, 8 Oct 2010 18:21:11 +0000 (14:21 -0400)]
Fix grip window positioning

We need to position the grip window in size_allocate after all,
otherwise the grip ends up in the wrong place when the window
is realized early.

Also, avoid shadowing a variable.

15 years agoDon't show resize grips on override-redirect window
Matthias Clasen [Fri, 8 Oct 2010 18:01:46 +0000 (14:01 -0400)]
Don't show resize grips on override-redirect window

15 years agoMake GtkPaned use gtk_container_class_handle_border_width
Matthias Clasen [Fri, 8 Oct 2010 15:11:33 +0000 (11:11 -0400)]
Make GtkPaned use gtk_container_class_handle_border_width

15 years agoMake GtkPaned use gtk_container_class_handle_border_width
Matthias Clasen [Fri, 8 Oct 2010 14:45:24 +0000 (10:45 -0400)]
Make GtkPaned use gtk_container_class_handle_border_width

15 years agoMake GtkPathBar use gtk_container_class_handle_border_width
Matthias Clasen [Fri, 8 Oct 2010 14:25:27 +0000 (10:25 -0400)]
Make GtkPathBar use gtk_container_class_handle_border_width

15 years agoMigrated old GtkSizeRequestIface documentation to GtkWidget and added notes to gtk_wi...
Tristan Van Berkom [Fri, 8 Oct 2010 14:10:42 +0000 (23:10 +0900)]
Migrated old GtkSizeRequestIface documentation to GtkWidget and added notes to gtk_widget_get_preferred_size().

15 years agoUse and print note for key only if there are credentials
Christian Dywan [Fri, 8 Oct 2010 14:01:07 +0000 (16:01 +0200)]
Use and print note for key only if there are credentials

Otherwise key might be printed uninitialized.

15 years agoRemoved call to get_preferred_size (widget, NULL, NULL);
Tristan Van Berkom [Fri, 8 Oct 2010 13:27:02 +0000 (22:27 +0900)]
Removed call to get_preferred_size (widget, NULL, NULL);

This was another case of making sure size-request was called
before other code was doing get_child_requisition(), now caching
is done via the same api so no need for that line.