gtk4.git
15 years agoUpdated Japanese translation
Takayuki KUSANO [Mon, 20 Sep 2010 09:50:58 +0000 (18:50 +0900)]
Updated Japanese translation

15 years agoUpdated Gujarati Translations
Sweta Kothari [Mon, 20 Sep 2010 07:20:44 +0000 (12:50 +0530)]
Updated Gujarati Translations

15 years agoAdd ug to LINGUAS.
Aron Xu [Mon, 20 Sep 2010 05:17:10 +0000 (05:17 +0000)]
Add ug to LINGUAS.

15 years agoAdd new error trap api to gdk.symbols and gdk3-sections.txt
Matthias Clasen [Mon, 20 Sep 2010 02:42:19 +0000 (22:42 -0400)]
Add new error trap api to gdk.symbols and gdk3-sections.txt

15 years agoAdd deprecation guards for gtk_cell_renderer_get_size()
Javier Jardón [Sun, 19 Sep 2010 15:46:13 +0000 (17:46 +0200)]
Add deprecation guards for gtk_cell_renderer_get_size()

15 years agoUpdated Danish translation
Kenneth Nielsen [Sun, 19 Sep 2010 15:40:56 +0000 (17:40 +0200)]
Updated Danish translation

15 years agoUpdated Dutch translation by Vincent van Adrighem
Wouter Bolsterlee [Sun, 19 Sep 2010 14:39:54 +0000 (16:39 +0200)]
Updated Dutch translation by Vincent van Adrighem

15 years agoFix a few dubious return values in return_if_fail calls
Matthias Clasen [Sun, 19 Sep 2010 05:58:46 +0000 (01:58 -0400)]
Fix a few dubious return values in return_if_fail calls

These were pointed out in bug 623958.

15 years agoAdd a rename-to annotation for gtk_tree_path_get_indices_with_depth
Matthias Clasen [Sun, 19 Sep 2010 05:15:17 +0000 (01:15 -0400)]
Add a rename-to annotation for gtk_tree_path_get_indices_with_depth

https://bugzilla.gnome.org/show_bug.cgi?id=621590

15 years agoPoint to the new GtkWidget align and margin properties
Matthias Clasen [Sun, 19 Sep 2010 05:01:09 +0000 (01:01 -0400)]
Point to the new GtkWidget align and margin properties

15 years agoAnother fix for the initial focus heuristic
Matthias Clasen [Sun, 19 Sep 2010 04:54:50 +0000 (00:54 -0400)]
Another fix for the initial focus heuristic

We were trying to avoid selecting a label initially, but the code
was sometimes leaving labels selected when the focus eventually
ended up on a button instead.

15 years agoFix an annotation
Matthias Clasen [Sun, 19 Sep 2010 04:22:37 +0000 (00:22 -0400)]
Fix an annotation

15 years agoUse gdk_error_trap_pop_ignore() where appropriate
Matthias Clasen [Sun, 19 Sep 2010 03:17:31 +0000 (23:17 -0400)]
Use gdk_error_trap_pop_ignore() where appropriate

15 years agoImprove tests for X error traps, fix two bugs
Havoc Pennington [Sun, 19 Sep 2010 03:03:31 +0000 (23:03 -0400)]
Improve tests for X error traps, fix two bugs

* don't lose track of previous X error handler
  if nested traps are pushed
* free any remaining traps when display
  is finalized

Test will fail unless bug 630032 is closed so
gdk_display_close() works.

https://bugzilla.gnome.org/show_bug.cgi?id=630033

15 years agoFix bugs that crashed gdk_display_close() on x11
Havoc Pennington [Sun, 19 Sep 2010 02:57:36 +0000 (22:57 -0400)]
Fix bugs that crashed gdk_display_close() on x11

* _gdk_device_set_associated_device() did not allow NULL device
* GdkDisplay should dispose device manager to avoid devices
  trying to touch the display in finalize
* GdkDeviceManagerXI did not ref devices in id hash
* GdkDisplayX11 did not ref devices in ->input_devices

15 years agoRevamp and modernize X error traps
Havoc Pennington [Sat, 18 Sep 2010 22:19:27 +0000 (18:19 -0400)]
Revamp and modernize X error traps

* add per-display gdk_x11_display_error_trap_push()
  (X11-specific because gdk_error_trap_push() probably
  should have been)
* make gdk_error_trap_push() handle only GDK displays
  not displays opened without a GDK wrapper
* make gdk_error_trap_pop() and gdk_x11_display_error_trap_pop()
  automatically sync only if needed, so manual gdk_flush() is not
  required
* add gdk_error_trap_pop_ignored() which just asynchronously
  ignores errors, so never needs to sync
* add G_GNUC_WARN_UNUSED_RESULT to plain pop(), because
  if you use plain pop() and don't need the return value,
  the async gdk_error_trap_pop_ignored() should be used
  instead. This results in lots of warnings to clean
  up in a later patch.

The main objective here was to avoid the need to sync just
to ignore an error. Now, syncing is automatic, and only
happens when we need to know the error code.

https://bugzilla.gnome.org/show_bug.cgi?id=629608

15 years agoAdd some minimal test for X error traps
Matthias Clasen [Sat, 18 Sep 2010 22:18:36 +0000 (18:18 -0400)]
Add some minimal test for X error traps

15 years agoReplace crufty old code for gdk error traps with GQueue and GSlice
Havoc Pennington [Sat, 18 Sep 2010 20:44:25 +0000 (16:44 -0400)]
Replace crufty old code for gdk error traps with GQueue and GSlice

No need to do a manual free list these days.

https://bugzilla.gnome.org/show_bug.cgi?id=629608

15 years agoRemove trailing whitespace and obsolete doc comments from gdk_error_trap code
Havoc Pennington [Sat, 18 Sep 2010 20:39:56 +0000 (16:39 -0400)]
Remove trailing whitespace and obsolete doc comments from gdk_error_trap code

(there are actual docs in the template file, these were some kind of
pre-gtk-doc comments without useful info)

https://bugzilla.gnome.org/show_bug.cgi?id=629608

15 years agoAdd a couple of missing annotations
Johan Dahlin [Sat, 18 Sep 2010 16:42:15 +0000 (13:42 -0300)]
Add a couple of missing annotations

15 years ago[introspection] Do not parse private sources
Johan Dahlin [Sat, 18 Sep 2010 16:41:30 +0000 (13:41 -0300)]
[introspection] Do not parse private sources

Private sources include private headers which we may not be able
to parse properly since some of them use unsupported syntax such
as objective-c.

15 years ago[introspection] Do not include internal api
Johan Dahlin [Sat, 18 Sep 2010 16:28:52 +0000 (13:28 -0300)]
[introspection] Do not include internal api

GtkTextLayout is internal and unsupported, do not include
it in the gir

15 years agodepend on GLib 2.27.0
Ryan Lortie [Sat, 18 Sep 2010 00:49:30 +0000 (20:49 -0400)]
depend on GLib 2.27.0

15 years agogail: Do not use deprecated gtk_cell_renderer_get_size()
Javier Jardón [Wed, 15 Sep 2010 19:03:34 +0000 (21:03 +0200)]
gail: Do not use deprecated gtk_cell_renderer_get_size()

Use gtk_cell_size_request_get_size() instead

https://bugzilla.gnome.org/show_bug.cgi?id=629785

15 years agoDo not use deprecated gtk_cell_renderer_get_size()
Javier Jardón [Mon, 13 Sep 2010 00:50:04 +0000 (02:50 +0200)]
Do not use deprecated gtk_cell_renderer_get_size()

Use gtk_cell_size_request_get_size() instead

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=629785

15 years agoChange GtkEditable typedef from GtkEditableClass to GtkEditabeInterface
Javier Jardón [Sat, 11 Sep 2010 02:12:42 +0000 (04:12 +0200)]
Change GtkEditable typedef from GtkEditableClass to GtkEditabeInterface

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=323904

15 years agoFix a trivial typo in a doc comment
Matthias Clasen [Fri, 17 Sep 2010 14:30:37 +0000 (10:30 -0400)]
Fix a trivial typo in a doc comment

https://bugzilla.gnome.org/show_bug.cgi?id=629733

15 years agoAdd some element-type annotations to gtk apis
Matthias Clasen [Fri, 17 Sep 2010 13:26:39 +0000 (09:26 -0400)]
Add some element-type annotations to gtk apis

15 years agoFix a doc comment
Matthias Clasen [Fri, 17 Sep 2010 13:25:49 +0000 (09:25 -0400)]
Fix a doc comment

15 years agoScan gdk/x11 sources for Gdk-3.0.gir
Matthias Clasen [Fri, 17 Sep 2010 13:23:30 +0000 (09:23 -0400)]
Scan gdk/x11 sources for Gdk-3.0.gir

The doc comments (with annotations) for some of the gdk api are down
in x11/ sources, so we need to scan those too, when building Gdk-3.0.gir.

15 years agoAdd annotations
Matthias Clasen [Fri, 17 Sep 2010 04:18:20 +0000 (00:18 -0400)]
Add annotations

The goi scanner warns about these nowadays.

15 years agoUpdate Simplified Chinese translation of po-properties.
Aron Xu [Thu, 16 Sep 2010 17:34:27 +0000 (17:34 +0000)]
Update Simplified Chinese translation of po-properties.

15 years agoRename h/v-align to h/valign
Matthias Clasen [Wed, 15 Sep 2010 22:37:20 +0000 (18:37 -0400)]
Rename h/v-align to h/valign

And adjust the getters and setters to match. Also include some
documentation by Havoc Pennington about adjustment of size requests
and allocations.

15 years agointrospection: Only use 0.9.3 API
Colin Walters [Wed, 15 Sep 2010 02:01:00 +0000 (22:01 -0400)]
introspection: Only use 0.9.3 API

15 years agoUpdate Czech translation
Petr Kovar [Wed, 15 Sep 2010 18:31:29 +0000 (20:31 +0200)]
Update Czech translation

15 years agoUpdate Czech translation
Petr Kovar [Wed, 15 Sep 2010 18:01:47 +0000 (20:01 +0200)]
Update Czech translation

15 years agoAdd docs for new API
Matthias Clasen [Wed, 15 Sep 2010 17:52:52 +0000 (13:52 -0400)]
Add docs for new API

15 years agoAdd new api to gtk.symbols
Matthias Clasen [Wed, 15 Sep 2010 17:49:00 +0000 (13:49 -0400)]
Add new api to gtk.symbols

15 years agoUse gint instead of in for local consistency
Matthias Clasen [Wed, 15 Sep 2010 17:22:51 +0000 (13:22 -0400)]
Use gint instead of in for local consistency

15 years agoAdded a default size of a magic number to testwrapbox.c
Tristan Van Berkom [Wed, 15 Sep 2010 16:40:32 +0000 (01:40 +0900)]
Added a default size of a magic number to testwrapbox.c

This was added to the test only for the sake of making it
easier to reproduce a bug with scrolled windows (bug 629778).

Expected behaviour: The vertical scrollbar should dissapear as soon
as the required height for the full allocation width (without any
vertical scrollbar) is small enough to not need a scrollbar.

15 years agogdk: Only use XComposite if it is available
Sebastian Dröge [Wed, 15 Sep 2010 14:41:41 +0000 (16:41 +0200)]
gdk: Only use XComposite if it is available

Fixes build if it isn't available. Bug #629748.

15 years agoFix build on Windows
Tor Lillqvist [Wed, 15 Sep 2010 11:19:40 +0000 (14:19 +0300)]
Fix build on Windows

No idea of the widget embedding crack has worked in recent versions,
even less now. But at least it now compiles.

15 years agospinner: Dont inherit from GtkDrawingArea part 2.
Tristan Van Berkom [Wed, 15 Sep 2010 07:52:29 +0000 (16:52 +0900)]
spinner: Dont inherit from GtkDrawingArea part 2.

Seems commit 7b42d4feda5448cf4f929aeb79f56ba9e524ac39 makes
GtkSpinner a direct subclass of GtkWidget but forgets to
update the header file to include gtkwidget.h instead and
declare the instance and class structures properly (assuming
this was just a missed file in the commit).

15 years agospinner: timeout should add/remove in map/unmap, not realize/unrealize.
Javier Jardón [Tue, 14 Sep 2010 18:09:43 +0000 (20:09 +0200)]
spinner: timeout should add/remove in map/unmap, not realize/unrealize.

Otherwise hiding the spinner won't stop it

15 years agogtkentry: remove unnneded check in map/unmap functions
Javier Jardón [Wed, 15 Sep 2010 01:28:24 +0000 (03:28 +0200)]
gtkentry: remove unnneded check in map/unmap functions

15 years agoAdd deprecation guards for gtk_widget_size_request()
Javier Jardón [Wed, 15 Sep 2010 01:06:47 +0000 (03:06 +0200)]
Add deprecation guards for gtk_widget_size_request()

15 years agodemos: Use gtk_size_request_get_size() instead gtk_widget_size_request()
Javier Jardón [Wed, 15 Sep 2010 01:13:58 +0000 (03:13 +0200)]
demos: Use gtk_size_request_get_size() instead gtk_widget_size_request()

15 years agoUse gtk_size_request_get_size() instead deprecated gtk_widget_size_request()
Javier Jardón [Tue, 14 Sep 2010 01:33:06 +0000 (03:33 +0200)]
Use gtk_size_request_get_size() instead deprecated gtk_widget_size_request()

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=629598

Signed-off-by: Javier Jardón <jjardon@gnome.org>
Signed-off-by: Tristan Van Berkom <tristanvb@openismus.com>
15 years ago[l10n] Updated German translation
Christian Kirbach [Tue, 14 Sep 2010 21:38:07 +0000 (23:38 +0200)]
[l10n] Updated German translation

15 years agoAdjusted new documentation for GtkSizeGroup.
Tristan Van Berkom [Tue, 14 Sep 2010 19:51:36 +0000 (04:51 +0900)]
Adjusted new documentation for GtkSizeGroup.

15 years agoAdded documentation to gtksizegroup.sgml explaining what happens with height-for...
Tristan Van Berkom [Tue, 14 Sep 2010 18:54:12 +0000 (03:54 +0900)]
Added documentation to gtksizegroup.sgml explaining what happens with height-for-width widgets.

When grouping height-for-width trading widgets (wrapping labels for instance)
vertically; the height for the minimum width will always be used for the entire
group... this patch warns about this in the docs.

15 years agoFixed GtkSizeGroups to adjust both minimum and natural requisitions
Tristan Van Berkom [Tue, 14 Sep 2010 05:52:59 +0000 (14:52 +0900)]
Fixed GtkSizeGroups to adjust both minimum and natural requisitions

Originally the GtkSizeRequestIface patches left GtkSizeGroup working
only by bumping the minimum sizes - this commit fixes size groups to take
both minimum and natural requests into account.

15 years agoAdd deprecation guards for gtk_cell_view_get_size_of_row()
Javier Jardón [Tue, 14 Sep 2010 18:40:15 +0000 (20:40 +0200)]
Add deprecation guards for gtk_cell_view_get_size_of_row()

15 years agospinner: Don't inherit from drawing area
Benjamin Otte [Mon, 13 Sep 2010 23:28:01 +0000 (01:28 +0200)]
spinner: Don't inherit from drawing area

A drawing area is not meant to be subclassed from. It also doesn't
provide any feature a spinner would need either.

15 years agospinner: Implement the size request interface
Benjamin Otte [Mon, 13 Sep 2010 23:25:07 +0000 (01:25 +0200)]
spinner: Implement the size request interface

Doing size requests in expose is broken.

15 years agospinner: Move G_DEFINE_TYPE() below function definitions
Benjamin Otte [Mon, 13 Sep 2010 23:19:48 +0000 (01:19 +0200)]
spinner: Move G_DEFINE_TYPE() below function definitions

Also remove redundant double definition.

15 years agowrapbox: Lookie, an unused variable
Benjamin Otte [Mon, 13 Sep 2010 23:17:17 +0000 (01:17 +0200)]
wrapbox: Lookie, an unused variable

15 years agospinner: Define desired size at top of file
Benjamin Otte [Mon, 13 Sep 2010 23:09:53 +0000 (01:09 +0200)]
spinner: Define desired size at top of file

No magic numbers in the code please

15 years agoGtkWrapBox: let GtkContainer handle border width
Matthias Clasen [Mon, 13 Sep 2010 22:46:49 +0000 (18:46 -0400)]
GtkWrapBox: let GtkContainer handle border width

15 years agoRemove fill options from GtkWrapBox
Matthias Clasen [Mon, 13 Sep 2010 22:36:51 +0000 (18:36 -0400)]
Remove fill options from GtkWrapBox

GtkWidget alignment properties make this unnecessary in new containers.

15 years agoRemove padding from GtkWrapBox
Matthias Clasen [Mon, 13 Sep 2010 22:21:21 +0000 (18:21 -0400)]
Remove padding from GtkWrapBox

GtkWidget margins make this unnecessary in new containers.

15 years agoGtkBox: let GtkContainer handle border width
Matthias Clasen [Mon, 13 Sep 2010 21:54:08 +0000 (17:54 -0400)]
GtkBox: let GtkContainer handle border width

15 years agoAdd deprecation guards for gtk_widget_get_child_requisition()
Javier Jardón [Mon, 13 Sep 2010 20:11:42 +0000 (22:11 +0200)]
Add deprecation guards for gtk_widget_get_child_requisition()

15 years agodemos: Do not use deprecated gtk_widget_get_child_requisition()
Javier Jardón [Mon, 13 Sep 2010 20:10:47 +0000 (22:10 +0200)]
demos: Do not use deprecated gtk_widget_get_child_requisition()

Use gtk_size_request_get_size() instead

15 years agobgo#486839 - Fix path bar size group in the file chooser
Milan Bouchet-Valat [Wed, 7 Oct 2009 20:08:41 +0000 (22:08 +0200)]
bgo#486839 - Fix path bar size group in the file chooser

Since the location button is hidden in save mode, we need to add the
path bar to the size group too. The location button still has to be in
the group though, because it's larger than the path bar (when
shown). Instead of using the recent/search icons, add their hboxes so
that themed widget spacings don't introduce variations.

15 years agoUse gtk_size_request_get_size() instead deprecated gtk_widget_get_child_requisition()
Javier Jardón [Thu, 9 Sep 2010 13:35:58 +0000 (15:35 +0200)]
Use gtk_size_request_get_size() instead deprecated gtk_widget_get_child_requisition()

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=629177

15 years agoFixed GtkWrapBox to allocate variable row AND column heights when in ALIGNED mode
Tristan Van Berkom [Mon, 13 Sep 2010 17:28:23 +0000 (02:28 +0900)]
Fixed GtkWrapBox to allocate variable row AND column heights when in ALIGNED mode

With this new approach at request and allocate time, the average child size
is used to determine a good guess at how many columns will fit the box
width; afterwards extra columns are appended and checked to fit.
Then the row heights are calculated based on height-for-width of each
child in the row which now may have individual widths.

15 years agoMake testwrapbox scrolled window scrollbars automatic.
Tristan Van Berkom [Mon, 13 Sep 2010 11:32:13 +0000 (20:32 +0900)]
Make testwrapbox scrolled window scrollbars automatic.

Did this in the hope to reproduce infinite recursion bugs
with height-for-width in scrolled windows (see complex bgo #611740
for reference for now).

15 years agoChanged GtkWrapBox code to avoid a growing GArray of a known size.
Tristan Van Berkom [Mon, 13 Sep 2010 11:31:33 +0000 (20:31 +0900)]
Changed GtkWrapBox code to avoid a growing GArray of a known size.

15 years ago[GdkWindowCache] Don't ignore the CompositeOverlayWindow
Adel Gadllah [Mon, 13 Sep 2010 16:20:22 +0000 (18:20 +0200)]
[GdkWindowCache] Don't ignore the CompositeOverlayWindow

Add the composite overlay window to the cache, as this can be a reasonable Xdnd proxy as well.

This is only done when the screen is composited in order to avoid mapping
the COW. We assume that the CM is using the COW (which is true for pretty
much any CM currently in use).

https://bugzilla.gnome.org/show_bug.cgi?id=601731

15 years agoUpdate gtkstyle.c to account for removal of row-ending-details
Kristian Rietveld [Mon, 13 Sep 2010 15:13:21 +0000 (17:13 +0200)]
Update gtkstyle.c to account for removal of row-ending-details

The detail strings now have more "detail" by default, so gtkstyle.c
needed to be updated to properly handle this.  Tests like testtreeview,
testtreesort now have proper background drawing again.

This strncmp trick was the best I could think of so quickly, if anybody
has an idea to do this in a better way, let me know.

15 years agoMark declaration of symbol _gtk_text_unknown_char_utf8 as extern
Kristian Rietveld [Mon, 13 Sep 2010 14:47:14 +0000 (16:47 +0200)]
Mark declaration of symbol _gtk_text_unknown_char_utf8 as extern

So that the symbol is not re-created each time the header is included.
The symbol is defined in gtktexttypes.c already.

15 years agoUse accessor functions to access GtkWidget
Kristian Rietveld [Mon, 13 Sep 2010 14:39:55 +0000 (16:39 +0200)]
Use accessor functions to access GtkWidget

15 years agoUpdate for GDK key name changes
Kristian Rietveld [Mon, 13 Sep 2010 14:27:08 +0000 (16:27 +0200)]
Update for GDK key name changes

15 years agoUpdate Quartz backend for GDK key name changes
Kristian Rietveld [Mon, 13 Sep 2010 14:26:02 +0000 (16:26 +0200)]
Update Quartz backend for GDK key name changes

15 years agoUpdated Japanese translation.
Takayuki KUSANO [Mon, 13 Sep 2010 13:19:59 +0000 (22:19 +0900)]
Updated Japanese translation.

15 years agoAdd padding and alignment tests to testadjustsize.c
Havoc Pennington [Sun, 5 Sep 2010 16:59:29 +0000 (12:59 -0400)]
Add padding and alignment tests to testadjustsize.c

15 years agoAdd margin and alignment properties to GtkWidget
Havoc Pennington [Sun, 5 Sep 2010 16:21:27 +0000 (12:21 -0400)]
Add margin and alignment properties to GtkWidget

h-align = START,END,CENTER,FILL
v-align = START,END,CENTER,FILL
margin-left,right,top,bottom
margin

These should obsolete all such similar properties on
layout containers, GtkMisc, GtkAlignment, GtkContainer::border-width

Margin is outside the size request.
If margin were not outside the set_size_request() it would not work the
same way as container-supplied (child property) padding.

Conceptually set_size_request() forces the value from the subclass
(the original unadjusted request) and then we go on to adjust
the request further by adding the margin.

15 years agoGtkButton: let GtkContainer handle border width
Havoc Pennington [Sun, 5 Sep 2010 06:22:47 +0000 (02:22 -0400)]
GtkButton: let GtkContainer handle border width

15 years agoadd gtk_container_class_handle_border_width() so subclasses can ignore border_width
Havoc Pennington [Sun, 5 Sep 2010 06:12:10 +0000 (02:12 -0400)]
add gtk_container_class_handle_border_width() so subclasses can ignore border_width

A subclass calls gtk_container_class_handle_border_width()
in its class_init

This marks the subclass as expecting GtkContainer to deal with
border width automatically, which GtkContainer then does.

15 years agoAdd testadjustsize test, to test new adjust size methods and related features
Havoc Pennington [Sun, 5 Sep 2010 03:40:32 +0000 (23:40 -0400)]
Add testadjustsize test, to test new adjust size methods and related features

This will test size adjust, and interactions with other padding and border

15 years agoGtkWidget: add adjust_size_request adjust_size_allocation virtual funcs
Havoc Pennington [Sun, 5 Sep 2010 16:19:14 +0000 (12:19 -0400)]
GtkWidget: add adjust_size_request adjust_size_allocation virtual funcs

Use these new methods to handle set_size_request (aka aux_info)
inside gtkwidget.c, instead of having external code mess with it.

The virtual functions can be used for other purposes in the
future. For example, GtkContainer::border_width could be
automatically implemented for all container subclasses.

15 years agoUse _gtk_widget_get_aux_info_or_defaults() when possible in gtkwidget.c
Havoc Pennington [Sun, 5 Sep 2010 16:14:40 +0000 (12:14 -0400)]
Use _gtk_widget_get_aux_info_or_defaults() when possible in gtkwidget.c

Did not update uses in other files because the plan is to
get rid of those other uses anyhow. So don't want to make
this function available in the header.

15 years agoadd _gtk_widget_get_aux_info_or_defaults()
Havoc Pennington [Sun, 5 Sep 2010 16:11:47 +0000 (12:11 -0400)]
add _gtk_widget_get_aux_info_or_defaults()

This is better than peeking aux info then testing != NULL
in several ways:
- it returns const aux info so if we don't create we can't write
- it ensures that the default we assume if aux_info is NULL is
  the same as the default we set if we've created the aux info
- it avoids typing in != NULL checks

15 years agoFix more SizeRequest implementations to avoid recursive calls to wrapper API
Havoc Pennington [Mon, 6 Sep 2010 02:19:31 +0000 (22:19 -0400)]
Fix more SizeRequest implementations to avoid recursive calls to wrapper API

GtkFrame, GtkComboBox, GtkExpander, GtkMenu, GtkWrapBox

These are all the examples I could find so far.

https://bugzilla.gnome.org/show_bug.cgi?id=628829

15 years agoWarn about recursively calling size req wrappers on the same object and orientation
Havoc Pennington [Sun, 5 Sep 2010 06:48:36 +0000 (02:48 -0400)]
Warn about recursively calling size req wrappers on the same object and orientation

We are not re-entrant and there is no reason for widgets to
do this, most likely they'll just get unexpected bugs because
the wrappers may modify the request.

Computing the request should logically rely only on the
widget itself, not on any adjustments caused by set_size_request,
size groups, and so forth.

https://bugzilla.gnome.org/show_bug.cgi?id=628829

15 years agodefault impls of width_for_height,hfw should chain directly not use wrapper API
Havoc Pennington [Sun, 5 Sep 2010 05:42:14 +0000 (01:42 -0400)]
default impls of width_for_height,hfw should chain directly not use wrapper API

In GtkBin and GtkWidget we tried to provide handy defaults that
call get_width if there's no get_width_for_height and
get_height for get_height_for_width.

However, they used the wrapper API on GtkSizeRequest instead of
chaining directly to the other method implementation.

This could result in all kinds of surprising behavior, for example,
get_width_for_height() would now already include the effects of set_size_request().

If nothing else it's inefficient. But it's just conceptually wrong,
because to chain to another implementation, we should call the other
implementation, not call a wrapper around the other implementation
(when we're already inside a previous invocation of the wrapper,
i.e. compute_size_for_orientation() ends up reinvoking itself
in the same orientation on the same object which it pretty
likely isn't intending to do)

https://bugzilla.gnome.org/show_bug.cgi?id=628829

15 years agoUse gint16 for GtkBorder
Havoc Pennington [Sat, 11 Sep 2010 23:25:57 +0000 (19:25 -0400)]
Use gint16 for GtkBorder

32K of border ought to be enough for any pixel dimensions. At least
until screens are so huge we start using doubles.

This saves a nice 64 bits of space when we have a GtkBorder
stored somewhere.

Signed integers are used to avoid surprising unsigned math
issues. Just search GTK's whole git log from inception
for "unsigned" if you want to find any number of commits
fixing signed/unsigned bugs.

https://bugzilla.gnome.org/show_bug.cgi?id=629387

15 years agoFix some compilation warnings
Javier Jardón [Wed, 25 Aug 2010 15:55:48 +0000 (17:55 +0200)]
Fix some compilation warnings

GScanner token field is declared as GTokenType instead a guint
See Glib bug: https://bugzilla.gnome.org/show_bug.cgi?id=627890

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=627962

15 years agodocs: Fix gdk_display_get_core_pointer() docs
Javier Jardón [Mon, 13 Sep 2010 00:46:46 +0000 (02:46 +0200)]
docs: Fix gdk_display_get_core_pointer() docs

Point to gdk_device_manager_get_client_pointer(),
not gdk_display_get_client_pointer().

15 years agostyle: Fix drawing of insensitive text
Benjamin Otte [Sun, 12 Sep 2010 13:55:57 +0000 (15:55 +0200)]
style: Fix drawing of insensitive text

When writing the original code I erroneously assumed that the current
point of the cairo context would be saved by cairo_save/restore(), but
of course the current point is part of the path and therefor isn't
saved.

Also do a cairo_new_path() before rendering any text so that we are sure
the text ends up at the right spot.

15 years agoDon't use g_idle_add to schedule idles in GDK
Matthias Clasen [Sat, 11 Sep 2010 02:28:55 +0000 (22:28 -0400)]
Don't use g_idle_add to schedule idles in GDK

We need to use gdk_threads_add_idle, in order to keep GDK code
under the GDK lock.

Bug 629277

15 years agogtk/gtkwidget.c: Use accessor functions to access GtkWindowGroup
Javier Jardón [Sat, 11 Sep 2010 01:27:37 +0000 (03:27 +0200)]
gtk/gtkwidget.c: Use accessor functions to access GtkWindowGroup

15 years agoRevert "gtk/gtkmain.c: Use accessor functions to access GtkWindowGroup"
Javier Jardón [Sat, 11 Sep 2010 02:24:54 +0000 (04:24 +0200)]
Revert "gtk/gtkmain.c: Use accessor functions to access GtkWindowGroup"

This reverts commit 65b9fc3e5bbae5b5efb4568c02df60939956f722.

15 years agogtk/gtkmain.c: Use accessor functions to access GtkWindowGroup
Javier Jardón [Sat, 11 Sep 2010 01:28:11 +0000 (03:28 +0200)]
gtk/gtkmain.c: Use accessor functions to access GtkWindowGroup

15 years agoGtkWindow: move public members to private structure
Javier Jardón [Tue, 17 Aug 2010 13:37:57 +0000 (15:37 +0200)]
GtkWindow: move public members to private structure

Signed-off-by: Christian Dywan <christian@twotoasts.de>
15 years agoBump version to 2.90.8
Matthias Clasen [Fri, 10 Sep 2010 16:31:15 +0000 (12:31 -0400)]
Bump version to 2.90.8

15 years agoInclude gdkkeysyms.h in gdk.h
Matthias Clasen [Fri, 10 Sep 2010 15:08:18 +0000 (11:08 -0400)]
Include gdkkeysyms.h in gdk.h

Followup to bug 629093

15 years agoGail: Remove gail_cell_type_add_action_interface, as it no longer does anything
Mike Gorse [Fri, 10 Sep 2010 13:46:34 +0000 (09:46 -0400)]
Gail: Remove gail_cell_type_add_action_interface, as it no longer does anything

15 years agoGail: Take fill/page size into account when returning a range's maximum value
Mike Gorse [Fri, 10 Sep 2010 10:09:41 +0000 (06:09 -0400)]
Gail: Take fill/page size into account when returning a range's maximum value