]> Pileus Git - ~andy/gtk/log
~andy/gtk
12 years agofile printbackend: Protect a callback by GDK_THREADS_ENTER
Matthias Clasen [Thu, 17 May 2012 04:11:15 +0000 (00:11 -0400)]
file printbackend: Protect a callback by GDK_THREADS_ENTER

https://bugzilla.gnome.org/show_bug.cgi?id=649569

12 years agoMake certain gtk3 types conditional (GST-style)
Руслан Ижбулатов [Sat, 12 May 2012 16:37:56 +0000 (20:37 +0400)]
Make certain gtk3 types conditional (GST-style)

https://bugzilla.gnome.org/show_bug.cgi?id=675895

12 years agodoc: Add missing CDATA around a XML programlisting
Volker Sobek [Fri, 4 May 2012 00:08:21 +0000 (02:08 +0200)]
doc: Add missing CDATA around a XML programlisting

https://bugzilla.gnome.org/show_bug.cgi?id=675423

12 years agoFix gdkx include path
Руслан Ижбулатов [Sat, 12 May 2012 16:27:41 +0000 (20:27 +0400)]
Fix gdkx include path

Fixes #675951

12 years agoBug 676087-configure.ac: isnan() and isinf() are macros, not functions
Natanael Copa [Tue, 15 May 2012 11:09:24 +0000 (13:09 +0200)]
Bug 676087-configure.ac: isnan() and isinf() are macros, not functions

The isnan() and isinf() are C99 macros not functions so use
AC_CHECK_DECL instead of AC_CHECK_FUNCS for those.

12 years agoImprove GtkColorButton behaviour
Hiroyuki Ito [Thu, 17 May 2012 03:45:48 +0000 (23:45 -0400)]
Improve GtkColorButton behaviour

Select the current color of a GtkColorButton should be selected in the
GtkColoroChooserDialog which is shown when the color button is clicked.

https://bugzilla.gnome.org/show_bug.cgi?id=676198

12 years agothemingengine: don't use gtk_theming_engine_get()
Cosimo Cecchi [Wed, 16 May 2012 22:27:21 +0000 (18:27 -0400)]
themingengine: don't use gtk_theming_engine_get()

This also avoids some useless GdkRGBA heap allocations when drawing a
border.

12 years agothemingengine: add missing deprecation annotation
Cosimo Cecchi [Wed, 16 May 2012 22:13:31 +0000 (18:13 -0400)]
themingengine: add missing deprecation annotation

gtk_theming_engine_state_is_running() is deprecated for 3.6.

12 years agothemingengine: don't use gtk_theming_engine_state_is_running()
Cosimo Cecchi [Wed, 16 May 2012 22:13:02 +0000 (18:13 -0400)]
themingengine: don't use gtk_theming_engine_state_is_running()

It's deprecated, and always returns FALSE. This allows us to simplify a
bunch of code.

12 years agoFix a case of 'stuck grab'
Matthias Clasen [Wed, 16 May 2012 21:18:51 +0000 (17:18 -0400)]
Fix a case of 'stuck grab'

This was showing up when using a combo box in list mode. After popping
up the list, the keyboard grab appeared stuck. What was stuck here is
only the client-side grab, since we forgot to clean up our grabs
when receiving an UnmapNotify.

This bug was introduced in 1c970036648bffd5d58499bb7c8e2439b84d4397.

12 years agogrid: add missing transfer annotation
Martin Pitt [Wed, 16 May 2012 16:00:04 +0000 (18:00 +0200)]
grid: add missing transfer annotation

gtk_grid_get_child_at() is missing a transfer annotation for the returned
widget, making it not introspectable.

12 years agotestscale: add a scale with a fill level to the test
Cosimo Cecchi [Wed, 16 May 2012 14:29:05 +0000 (10:29 -0400)]
testscale: add a scale with a fill level to the test

12 years agoBe more careful when freeing ui manager nodes
Matthias Clasen [Wed, 16 May 2012 01:49:24 +0000 (21:49 -0400)]
Be more careful when freeing ui manager nodes

https://bugzilla.redhat.com//show_bug.cgi?id=812035 has
a stacktrace that shows recursion via
free_node -> unref -> dispose -> ui manager api
which ends in a crash, since we run free_node over the entire
tree and it leaves lots of dangling pointers behind.
So, better be careful by setting all pointers to NULL after
freeing them.

12 years agoUpdated Dutch translation
Wouter Bolsterlee [Tue, 15 May 2012 20:20:12 +0000 (22:20 +0200)]
Updated Dutch translation

12 years agoUpdated Dutch translation
Rachid BM [Tue, 15 May 2012 20:18:56 +0000 (22:18 +0200)]
Updated Dutch translation

12 years agogdk: fix the build
Cosimo Cecchi [Tue, 15 May 2012 17:27:21 +0000 (13:27 -0400)]
gdk: fix the build

Regression from commit 861c57e3043229e769a45d7349b6e71d99c752de

12 years agothemingengine: render pane-separator grip only if we have no bg-image
Cosimo Cecchi [Mon, 14 May 2012 13:20:56 +0000 (09:20 -0400)]
themingengine: render pane-separator grip only if we have no bg-image

So that it's possible to implement this with an SVG asset if we want.

12 years agotests: add background-shorthand parser test for multiple backgrounds
Cosimo Cecchi [Sat, 12 May 2012 00:43:09 +0000 (20:43 -0400)]
tests: add background-shorthand parser test for multiple backgrounds

These CSS files test the background shorthand property when it's used
with multiple background layers.

12 years agotests: move background-shorthand test to background-shorthand-single
Cosimo Cecchi [Tue, 15 May 2012 16:45:39 +0000 (12:45 -0400)]
tests: move background-shorthand test to background-shorthand-single

And update/fix it for the new shorthand parsing, since the old test did
a couple of things wrong:
- it erroneously flipped background-clip and background-origin when
  testing
- it was missing default initial values for background-size and
  background-position from the generated test output

12 years agoshorthand: implement multiple backgrounds for the background shorthand
Cosimo Cecchi [Sat, 12 May 2012 00:38:36 +0000 (20:38 -0400)]
shorthand: implement multiple backgrounds for the background shorthand

12 years agocsspositionvalue: make empty error value non fatal
Cosimo Cecchi [Sat, 12 May 2012 00:46:18 +0000 (20:46 -0400)]
csspositionvalue: make empty error value non fatal

If there's just no value to parse (e.g. because we're parsing a
shorthand property), don't error out while trying to parse a number.

12 years agoshorthand: add background-position and background-size to bg shorthand
Cosimo Cecchi [Fri, 11 May 2012 20:33:13 +0000 (16:33 -0400)]
shorthand: add background-position and background-size to bg shorthand

12 years agoshorthand: use array values for shorthand properties parsing
Cosimo Cecchi [Fri, 11 May 2012 18:30:07 +0000 (14:30 -0400)]
shorthand: use array values for shorthand properties parsing

Right now, just parse the first value in the array.

12 years agotests: add trailing commas to background-shorthand.css test file
Cosimo Cecchi [Fri, 11 May 2012 18:29:50 +0000 (14:29 -0400)]
tests: add trailing commas to background-shorthand.css test file

12 years agoreftests: add a reftest for multiple background layers
Cosimo Cecchi [Fri, 11 May 2012 14:30:46 +0000 (10:30 -0400)]
reftests: add a reftest for multiple background layers

12 years agothemingbackgorund: render all background layers
Cosimo Cecchi [Fri, 11 May 2012 14:25:55 +0000 (10:25 -0400)]
themingbackgorund: render all background layers

Now that background properties are arrays, use them in
GtkThemingBackground when rendering.

12 years agocolorswatch: don't use the clip box directly
Cosimo Cecchi [Fri, 11 May 2012 14:22:24 +0000 (10:22 -0400)]
colorswatch: don't use the clip box directly

We can render our checkboard pattern on the padding box instead. This is
in preparation for GtkThemingBackground supporting multiple layers.

12 years agostyleproperty: make background-image an array value
Cosimo Cecchi [Thu, 10 May 2012 15:18:29 +0000 (11:18 -0400)]
styleproperty: make background-image an array value

This also requires having the GtkCssArrayValue parse func forward the
base GFile from the style property parse implementation.

12 years agostyleproperty: make background-size an array property
Cosimo Cecchi [Thu, 10 May 2012 14:43:43 +0000 (10:43 -0400)]
styleproperty: make background-size an array property

12 years agostyleproperty: make background-position an array property
Cosimo Cecchi [Thu, 10 May 2012 14:39:25 +0000 (10:39 -0400)]
styleproperty: make background-position an array property

12 years agostyleproperty: make background-repeat an array property
Cosimo Cecchi [Thu, 10 May 2012 14:31:06 +0000 (10:31 -0400)]
styleproperty: make background-repeat an array property

12 years agostyleproperty: make background-clip and background-area array properties
Cosimo Cecchi [Thu, 10 May 2012 14:25:33 +0000 (10:25 -0400)]
styleproperty: make background-clip and background-area array properties

Right now just make them arrays, but don't consider values after the
first.

12 years agogdkoffscreenwindow: Don't crash when resetting cursor
Bastien Nocera [Thu, 10 May 2012 14:23:57 +0000 (15:23 +0100)]
gdkoffscreenwindow: Don't crash when resetting cursor

When the toplevel is a GdkOffscreenWindow which doesn't
implement the set_device_cursor() vfunc, we would have
crashed. Implement a dummy ->set_device_cursor vfunc.

https://bugzilla.gnome.org/show_bug.cgi?id=675809

12 years agoRevert "gdk: Don't crash when resetting cursor"
Bastien Nocera [Tue, 15 May 2012 16:57:33 +0000 (17:57 +0100)]
Revert "gdk: Don't crash when resetting cursor"

Patch didn't get reviewed yet.

This reverts commit 779591b62fbc5ee4ebf9912a4416b655cdcf4e32.

12 years agostatusicon: Fix implicit declaration warning
Bastien Nocera [Tue, 15 May 2012 16:55:36 +0000 (17:55 +0100)]
statusicon: Fix implicit declaration warning

gtkstatusicon.c:1333:4: warning: implicit declaration of function 'abs' [-Wimplicit-function-declaration]

Added in commit c8378099600e4279cb5c608b546bbe778c451233

12 years agogdk: Don't crash when resetting cursor
Bastien Nocera [Thu, 10 May 2012 14:23:57 +0000 (15:23 +0100)]
gdk: Don't crash when resetting cursor

When the toplevel is a GdkOffscreenWindow which doesn't
implement the set_device_cursor() vfunc.

https://bugzilla.gnome.org/show_bug.cgi?id=675809

12 years agostatusicon: use absolute delta to find the nearest icon size
Fernando Lemos [Tue, 15 May 2012 14:28:37 +0000 (10:28 -0400)]
statusicon: use absolute delta to find the nearest icon size

Otherwise, if we only look at (pixel_size - stock_size), we miss the
case when the pixel size is smaller than any provided stock size.

https://bugzilla.gnome.org/show_bug.cgi?id=673749

12 years agoUpdated Norwegian bokmål translation
Kjartan Maraas [Tue, 15 May 2012 12:17:31 +0000 (14:17 +0200)]
Updated Norwegian bokmål translation

12 years agoiconview: Bring back rtl flipping
Matthias Clasen [Tue, 15 May 2012 05:12:28 +0000 (01:12 -0400)]
iconview: Bring back rtl flipping

This was lost in the recent layout rewrite.

12 years agoGDK X11 DND: Fix infinite loop
Edward Sheldrake [Mon, 14 May 2012 12:06:43 +0000 (13:06 +0100)]
GDK X11 DND: Fix infinite loop

Related: https://bugzilla.gnome.org/show_bug.cgi?id=664646

12 years agoUpdated Spanish translation
Daniel Mustieles [Mon, 14 May 2012 12:52:24 +0000 (14:52 +0200)]
Updated Spanish translation

12 years agoGTK PrintToFile settings
Timothy Arceri [Tue, 6 Mar 2012 10:15:01 +0000 (21:15 +1100)]
GTK PrintToFile settings

To make setting output directory and filename simpler in the PrintToFile
dialog two gtkprintsettings have been added GTK_PRINT_SETTINGS_OUTPUT_DIR
 and GTK_PRINT_SETTINGS_OUTPUT_BASENAME.

This will reduce the code needed to implement a better name than "output.pdf"
and actually makes more sense than the existing setting
GTK_PRINT_SETTINGS_OUTPUT_URI which doesn't work seamlessly with
GTK_PRINT_SETTINGS_OUTPUT_FILE_FORMAT like the new settings do.

https://bugzilla.gnome.org/show_bug.cgi?id=657322

12 years agoFix links to developer.gnome.org
Frédéric Péters [Fri, 11 May 2012 12:17:19 +0000 (14:17 +0200)]
Fix links to developer.gnome.org

12 years agoquartz: Ignore events from all mouse buttons past the resize boundary
Kristian Rietveld [Sun, 13 May 2012 15:57:59 +0000 (17:57 +0200)]
quartz: Ignore events from all mouse buttons past the resize boundary

Before, right click events were still let through into GDK. In this
case, also middle/right button events with x-coordinates in the range
[-3, 0] are processed, resulting in failures/crashes in the window
finding code because no GdkWindows are present in this range.

12 years agoRemove lies
Matthias Clasen [Fri, 11 May 2012 21:42:47 +0000 (17:42 -0400)]
Remove lies

We forgot the documentation when we simplified the GtkApplication
session api.

12 years agoUpdated POTFILES.skip
Piotr Drąg [Fri, 11 May 2012 20:01:19 +0000 (22:01 +0200)]
Updated POTFILES.skip

12 years agovisuals: add some visuals UI files for h/v linked button boxes
Cosimo Cecchi [Fri, 11 May 2012 15:37:56 +0000 (11:37 -0400)]
visuals: add some visuals UI files for h/v linked button boxes

12 years agobox: set orientation style classes when setting orientation
Cosimo Cecchi [Fri, 11 May 2012 15:36:05 +0000 (11:36 -0400)]
box: set orientation style classes when setting orientation

Since we override the property, we have to call the internal helper
manually.

12 years agocssprovider: Remove unused members from scanner
Benjamin Otte [Wed, 18 Apr 2012 20:19:30 +0000 (22:19 +0200)]
cssprovider: Remove unused members from scanner

Now that the parser takes care of the file, we don't need to keep track
of it ourselves anymore.

12 years agocsssection: Remove file argument from constructor
Benjamin Otte [Wed, 18 Apr 2012 20:15:32 +0000 (22:15 +0200)]
csssection: Remove file argument from constructor

The parser has the file available.

12 years agostyleproperty: Remove base argument
Benjamin Otte [Wed, 18 Apr 2012 20:04:44 +0000 (22:04 +0200)]
styleproperty: Remove base argument

from _gtk_style_property_parse_value(). The GtkCssParser takes care of
that now.

12 years agocssstyleproperty: Remove base argument
Benjamin Otte [Wed, 18 Apr 2012 19:58:22 +0000 (21:58 +0200)]
cssstyleproperty: Remove base argument

from parse functions.

12 years agocssstylefuncs: Remove base argument
Benjamin Otte [Wed, 18 Apr 2012 19:52:36 +0000 (21:52 +0200)]
cssstylefuncs: Remove base argument

The GtkCssParser keeps track of the base file now.

12 years agocssshorthand: Don't pass base file anymore
Benjamin Otte [Wed, 18 Apr 2012 19:48:05 +0000 (21:48 +0200)]
cssshorthand: Don't pass base file anymore

It's part of the parser now.

12 years agocssimage: No need to pass base file anymore
Benjamin Otte [Wed, 18 Apr 2012 19:44:08 +0000 (21:44 +0200)]
cssimage: No need to pass base file anymore

The file is now part of the GtkCssParser

12 years agocssparser: Change APIs that work with files
Benjamin Otte [Wed, 18 Apr 2012 17:43:24 +0000 (19:43 +0200)]
cssparser: Change APIs that work with files

Now that we keep track of the parser's file, we don't require it to be
passed anymore.

12 years agocssparser: Keep track of the file
Benjamin Otte [Wed, 18 Apr 2012 14:54:04 +0000 (16:54 +0200)]
cssparser: Keep track of the file

12 years agoRevert "cssprovider: Remove unused members from scanner"
Benjamin Otte [Fri, 11 May 2012 14:41:25 +0000 (16:41 +0200)]
Revert "cssprovider: Remove unused members from scanner"

This reverts commit b501e64a8a78e78c49b0abd1f34fc554ccacfd65.

12 years agocssprovider: Remove unused members from scanner
Benjamin Otte [Wed, 18 Apr 2012 20:19:30 +0000 (22:19 +0200)]
cssprovider: Remove unused members from scanner

Now that the parser takes care of the file, we don't need to keep track
of it ourselves anymore.

12 years ago[broadway] Fixup build
Alexander Larsson [Fri, 11 May 2012 12:49:15 +0000 (14:49 +0200)]
[broadway] Fixup build

Seems the return value of device_query_state was removed, it was
TRUE always anyway.

12 years agogtk: Fix warnings for some uses of GtkLinkButton
Bastien Nocera [Thu, 10 May 2012 14:15:43 +0000 (15:15 +0100)]
gtk: Fix warnings for some uses of GtkLinkButton

Not setting a URI but catching the activate-link signal is a
valid use of GtkLinkButton, but we shouldn't allow showing a
popup menu which offers to copy the URI if there's none.

12 years agoquartz: add a special case to GtkIMContextSimply for entering '"'
Michael Natterer [Thu, 10 May 2012 08:57:16 +0000 (10:57 +0200)]
quartz: add a special case to GtkIMContextSimply for entering '"'

Turn dead_doubleacute plus space into '"' and not into a double
acute because that's the way to enter double quotes on the
US-International keyboard layout.
(cherry picked from commit 71164e57b9b999f07a613806058ee87b9cbf882d)

12 years agoiconview: Don't draw unallocated items
Benjamin Otte [Wed, 9 May 2012 12:23:10 +0000 (14:23 +0200)]
iconview: Don't draw unallocated items

This is more a workaround for Clutter-GTK than anything else. Most of
all, it makes gnome-boxes start again.

12 years agol10n: updated Italian translation
Luca Ferretti [Wed, 9 May 2012 09:48:36 +0000 (11:48 +0200)]
l10n: updated Italian translation

12 years agoquartz: handle yet another dead tilde variant
Michael Natterer [Tue, 8 May 2012 14:07:31 +0000 (16:07 +0200)]
quartz: handle yet another dead tilde variant

which is delivered by the US-International keyboard layout.

12 years agogtk: Add private headers to EXTRA_DIST
Rico Tzschichholz [Tue, 8 May 2012 07:16:51 +0000 (09:16 +0200)]
gtk: Add private headers to EXTRA_DIST

This got lost with 985881ffcc315e20012a42dfba6cf5ab0046a92c

12 years agoiconview: Only compute item sizes in one place
Benjamin Otte [Tue, 8 May 2012 01:57:07 +0000 (03:57 +0200)]
iconview: Only compute item sizes in one place

Code was using different places to compute the icon size when layouting
and when doing size requests. This resulted in non-matching behavior in
obscure cases. And that lead to iconviews that were too small.

12 years agoiconview: Bring back per-row contexts
Benjamin Otte [Mon, 7 May 2012 23:53:48 +0000 (01:53 +0200)]
iconview: Bring back per-row contexts

Apparently GtkCellAreaContext is broken enough to not allow multiple
allocations...

12 years agoiconview: Move wrap width adjustment to size request code
Benjamin Otte [Mon, 7 May 2012 23:51:51 +0000 (01:51 +0200)]
iconview: Move wrap width adjustment to size request code

12 years agostylecontext: Make queue_invalidate() always work
Benjamin Otte [Mon, 7 May 2012 23:35:00 +0000 (01:35 +0200)]
stylecontext: Make queue_invalidate() always work

Handle both the case where a widget is set and also the case where a
widget path is set.

12 years agoimmulticontext: Always check if the global context id changed
Rui Matos [Sun, 29 Apr 2012 00:44:23 +0000 (02:44 +0200)]
immulticontext: Always check if the global context id changed

This makes sure that if the gtk-im-module setting changes we update
our internal state immediately on the next event whichever it is.

In particular this fixes the case of the gtk-im-module setting
changing while the user is typing and the slave context remaining the
same, effectively ignoring the setting change.

https://bugzilla.gnome.org/show_bug.cgi?id=675365

12 years agoinfobar: set default border/spacing on action/content areas
Cosimo Cecchi [Mon, 7 May 2012 21:34:46 +0000 (17:34 -0400)]
infobar: set default border/spacing on action/content areas

Commit cdf473ec100d23ea6a9a8750ad301307a116b674 deprecated the style
properties associated with the container border width and box spacing of
the action and content areas, in favour of using
gtk_container_set_border_width() and gtk_box_set_spacing() on the
widgets themselves, but failed to initialize those values to the
defaults.

12 years ago[Bug 675501] gtkquartz.h is not in the gtk+-3.5.2.tar.xz archive
John Ralls [Mon, 7 May 2012 21:58:10 +0000 (14:58 -0700)]
[Bug 675501] gtkquartz.h is not in the gtk+-3.5.2.tar.xz archive

Rearrange the gtk_private_h_sources so that they're all added to EXTRA_DIST instead of only the ones included in the current build.

12 years agoRename gtkquartz-menu to gtkmodelmenu-quartz
John Ralls [Sat, 5 May 2012 20:11:10 +0000 (13:11 -0700)]
Rename gtkquartz-menu to gtkmodelmenu-quartz

Makes name consistent with other quartz-only modules and makes it clear that this works with the GMenuModel system rather than the older GtkMenu system.

12 years agobgo#675421 - Set the wm_role on GtkFileChooserDialog
Federico Mena Quintero [Mon, 7 May 2012 20:42:32 +0000 (15:42 -0500)]
bgo#675421 - Set the wm_role on GtkFileChooserDialog

The window's role is 'GtkFileChooserDialog', so that window managers can match it
for positioning.

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agosizerequest: Improve warning message
Benjamin Otte [Mon, 7 May 2012 15:37:36 +0000 (17:37 +0200)]
sizerequest: Improve warning message

Actually print out the function we're warning about

12 years agoiconview: Fix negative natural size request on empty iconview
Benjamin Otte [Mon, 7 May 2012 15:12:36 +0000 (17:12 +0200)]
iconview: Fix negative natural size request on empty iconview

12 years agobroadway: Fix vfunc prototype
Benjamin Otte [Mon, 7 May 2012 02:12:29 +0000 (04:12 +0200)]
broadway: Fix vfunc prototype

12 years agobuild: Don't invoke pkg-config directly
Benjamin Otte [Mon, 7 May 2012 01:14:35 +0000 (03:14 +0200)]
build: Don't invoke pkg-config directly

Instead, always use PKG_CHECK_MODULES(). That macro actually gets it
right. In particular the erroring out part when you miss xkbcommon or
wayland-client.

12 years agoiconview: Don't expand items to more than natural size
Benjamin Otte [Sun, 6 May 2012 23:35:00 +0000 (01:35 +0200)]
iconview: Don't expand items to more than natural size

This ensures that items stay left-aligned instead of slowly expanding into
empty space when widening the iconview. It's also what the iconview did
pre-refactoring.

Note that for cases where natural width != minimum width, the cells
might still expand and shrink back.

12 years agotests: Remove unused include
Benjamin Otte [Sun, 6 May 2012 22:16:45 +0000 (00:16 +0200)]
tests: Remove unused include

Fixes build

12 years agoiconview: Redo layouting
Benjamin Otte [Sun, 6 May 2012 03:39:58 +0000 (05:39 +0200)]
iconview: Redo layouting

12 years agoiconview: Respect item width in size requestr code
Benjamin Otte [Sun, 6 May 2012 01:21:18 +0000 (03:21 +0200)]
iconview: Respect item width in size requestr code

12 years agoiconview: Reduce complexity of sizing code
Benjamin Otte [Sat, 5 May 2012 22:51:14 +0000 (00:51 +0200)]
iconview: Reduce complexity of sizing code

Always assume max-columns and min-rows. The old approach was kinda
insane.

As an example, try to write an algorithm that optimizes the minimum size
for infinite (take a reasonably large number like 2520) word-wrapped
Monospace text cells containing the text "XXXXX XXX XXX XXXXX" (keep in
mind that this is the easy problem, because it's assuming equal cell
renderers). There's 4 ways to reasonably lay out this text:
19 glyphs (19x1):
  XXXXX XXX XXX XXXXX
18 glyphs (9x2):
  XXXXX XXX
  XXX XXXXX
21 glyphs (7x3):
  XXXXX
  XXX XXX
  XXXXX
20 glyphs (5x4):
  XXXXX
  XXX
  XXX
  XXXXX
The best thing to do usually is using the 9x2 approach, but that's
neither the one using the natural nor the one using the minimum size.

As a side note, this does not include spacing and padding, which might
also influence the decision. Nor does it include height-for-width
considerations. Look at this table (numbers given in glyphs, not pixels,
as for pixel-sizes it gets even more interesting):
  given  best solution
  width  columns  sizing  glyphs per cell
   6      1       6x4           20
   7      1       7x3           21
   8      1       7x3           24
   9      1       9x2           18
  10      1/2     9x2/5x4       20
  11      1/2     9x2/5x4       22
  12      1/2     9x2/5x4       24
  13      1/2     9x2/5x4       26
  14      2       7x3           21
  15      3       5x4           20
  16      3       5x4           21.3
  17      3       5x4           22.7
  18      2       9x2           18
  19      1/2    19x1/8x2       19
  20      1/2/4  19x1/8x2/5x4   20
  21      1-4     any           21
  22      1-4     any           22
  23      1-4     any           23
  24      1-4     any           24
  25      5       5x4           20
  26      5       5x4           20.8
  27      3       9x2           18
  28      3       9x2           18.7
  29      3       9x2           19.3
  30      3/6     9x2/5x4       20

Now of course, nobody wants the number of columns to randomly change in
inexplicable ways while they enlarge or shrink an iconview, so we not
only have to optimize for smallest or other size measurements, but we
also have to optimize for "most pleasing to the eye".

And last but not least, I'd like to once again remind you - if you kept
up until now - that this discussion was for identically-sized cells
only.

12 years agotests: Make testcellarea use an iconview
Benjamin Otte [Sat, 5 May 2012 03:39:58 +0000 (05:39 +0200)]
tests: Make testcellarea use an iconview

12 years agoiconview: Redo size requests
Benjamin Otte [Sat, 5 May 2012 02:33:47 +0000 (04:33 +0200)]
iconview: Redo size requests

Instead of just returning the last allocated numbers, we now compute the
proper sizes from scratch. This is a bit less trivial, but it results in
proper height-for-width handling.

12 years agoiconview: Don't cache iters
Benjamin Otte [Fri, 4 May 2012 16:44:39 +0000 (18:44 +0200)]
iconview: Don't cache iters

This is a huge quest to remove all caching from GtkIconview to simplify
the code. As it turns out, iconview performance is a joke, so the caches
are kinda unnecessary.

If we need caching, we can add it in a useful way later.

12 years agoiconview: Remove needless requirement
Benjamin Otte [Thu, 3 May 2012 22:31:14 +0000 (00:31 +0200)]
iconview: Remove needless requirement

If we use
  &item->cell_area
instead of
  (GdkRectangle *) item
there is no need anymore to keep the cell_area as the first member of
the the item. And we cget compile-time checks for changes to the item
struct.

12 years agogtk_entry_[get|set]_inner_border(): Clarify the deprecation docs.
José Alburquerque [Sun, 6 May 2012 19:05:27 +0000 (15:05 -0400)]
gtk_entry_[get|set]_inner_border(): Clarify the deprecation docs.

12 years ago[socket]: fix gtk-doc warnings.
Paolo Borelli [Sun, 6 May 2012 10:37:59 +0000 (12:37 +0200)]
[socket]: fix gtk-doc warnings.

12 years agogdk: do not send leave event on touchpad button release
Marc-André Lureau [Thu, 3 May 2012 16:15:48 +0000 (18:15 +0200)]
gdk: do not send leave event on touchpad button release

This is along the changes introduced in
bd55519f7e6bb5da2fdf6f5e99bd79213059f81b,
A TOUCHPAD device behaves like a mouse"

https://bugzilla.gnome.org/show_bug.cgi?id=674549

12 years agoUpdated Hebrew translation.
Yaron Shahrabani [Sat, 5 May 2012 15:22:56 +0000 (18:22 +0300)]
Updated Hebrew translation.

12 years agoUpdated Galician translations
Fran Diéguez [Sat, 5 May 2012 13:18:26 +0000 (15:18 +0200)]
Updated Galician translations

12 years agothemingengine: don't hardcode the slider lines in the theming engine
Cosimo Cecchi [Fri, 10 Jun 2011 17:44:51 +0000 (13:44 -0400)]
themingengine: don't hardcode the slider lines in the theming engine

It's wrong to hardcode the slider lines here; those should be rendered
as an additional background layer using background-image if the theme
specifies so.

https://bugzilla.gnome.org/show_bug.cgi?id=652304

12 years agoentry-completion: set size request on the toplevel window
Cosimo Cecchi [Tue, 27 Mar 2012 18:23:48 +0000 (14:23 -0400)]
entry-completion: set size request on the toplevel window

Instead of setting it on the child scrolled window. This is needed
because the whole window's allocation must be equal to the one of the
entry (in case the popup-set-width property is TRUE); if we set the size
request on a children of the window, there might be other children with
borders/paddings in between the toplevel and the child we set the size
request too, which will break alignment.

https://bugzilla.gnome.org/show_bug.cgi?id=672947

12 years agoentry-completion: don't subtract entry borders from popup allocation
Cosimo Cecchi [Tue, 27 Mar 2012 18:22:56 +0000 (14:22 -0400)]
entry-completion: don't subtract entry borders from popup allocation

The completion popup should be aligned with the GtkEntry allocation, not
less than that.

https://bugzilla.gnome.org/show_bug.cgi?id=672947

12 years agoUpdated Spanish translation
Daniel Mustieles [Fri, 4 May 2012 12:16:45 +0000 (14:16 +0200)]
Updated Spanish translation

12 years agocellrenderertext: Add placeholder-text property
Carlos Garcia Campos [Fri, 4 May 2012 03:25:08 +0000 (23:25 -0400)]
cellrenderertext: Add placeholder-text property

It works similar to placeholder text in a GtkEntry, if the cell renderer
is editable and it's empty, the placeholder text is displayed.
https://bugzilla.gnome.org/show_bug.cgi?id=645514

12 years agodocs: Improve GtkMessageDialog description
David King [Mon, 20 Feb 2012 18:05:12 +0000 (18:05 +0000)]
docs: Improve GtkMessageDialog description

Add mention of default setting of GtkWindow:skip-taskbar-hint.

https://bugzilla.gnome.org/show_bug.cgi?id=587873

12 years agodocs: Move GtkSelectionData to gtk3-Selections
David King [Mon, 20 Feb 2012 17:05:08 +0000 (17:05 +0000)]
docs: Move GtkSelectionData to gtk3-Selections

Put the (albeit empty) documentation for the GtkSelectionData struct
with the gtk_selection_data_*() methods.

https://bugzilla.gnome.org/show_bug.cgi?id=663072