gtk4.git
12 years agoUpdate keyname tables
Matthias Clasen [Fri, 19 Apr 2013 23:35:49 +0000 (19:35 -0400)]
Update keyname tables

This commit is very similar to 8c8853a1f5d29d447c16ba5d92bdb7bb442ee721

We update the keynames.txt file from gdkkeynames.h, and we update
keynames-translate.txt to include all the keysym names that we want
to have translations for. Also strip the XF86 from the translatable
keysym names, since we are returning those names now from
gdk_keyval_name().

keyname-table.h is regenerated from these updated files.

12 years agoKeep XF86 keysym names working
Matthias Clasen [Fri, 19 Apr 2013 23:29:13 +0000 (19:29 -0400)]
Keep XF86 keysym names working

These names are unfortunately stored in gsettings around
the world, so we can't really stop supporting them.

12 years agoAdd some tests for keysyms
Matthias Clasen [Fri, 19 Apr 2013 23:28:09 +0000 (19:28 -0400)]
Add some tests for keysyms

Not very extensive, but we do test that XF86 keysym names
keep working.

12 years agoUpdate gdkkeysyms.h header
Matthias Clasen [Fri, 19 Apr 2013 21:28:35 +0000 (17:28 -0400)]
Update gdkkeysyms.h header

Regenerate these files from the X11 keysymdef.h file, using
gdkkeysyms-update.pl. The only change is the addition of
GDK_KEY_AudioMicMute.

12 years agoMerge branch 'places-sidebar'
Federico Mena Quintero [Fri, 19 Apr 2013 22:26:39 +0000 (17:26 -0500)]
Merge branch 'places-sidebar'

12 years agotreeview: Use equal areas in the computation of drag dest row positions
Federico Mena Quintero [Fri, 19 Apr 2013 22:20:12 +0000 (17:20 -0500)]
treeview: Use equal areas in the computation of drag dest row positions

We used to divide the row in thirds vertically, and use the outer thirds for GTK_TREE_VIEW_DROP_BEFORE and AFTER, respectively.

Now we use *fourths*.  This is so that we get equal areas for these:

  GTK_TREE_VIEW_DROP_BEFORE
  GTK_TREE_VIEW_DROP_INTO_OR_BEFORE
  GTK_TREE_VIEW_DROP_INTO_OR_AFTER
  GTK_TREE_VIEW_DROP_AFTER

This makes hovering tree rows much more positive.

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agoAdjust the drag destination row when it is above the currently highlighted row
Federico Mena Quintero [Fri, 19 Apr 2013 22:22:39 +0000 (17:22 -0500)]
Adjust the drag destination row when it is above the currently highlighted row

This makes the feedback accurate and without hysteresis.  Haven't I written this code ten times before?

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agoDropping on the feedback row is always possible
Federico Mena Quintero [Fri, 19 Apr 2013 20:49:54 +0000 (15:49 -0500)]
Dropping on the feedback row is always possible

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agoAdd a temporary row for feedback when inserting a bookmark is possible
Federico Mena Quintero [Fri, 19 Apr 2013 20:03:17 +0000 (15:03 -0500)]
Add a temporary row for feedback when inserting a bookmark is possible

We'll prettify this gradually.

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years ago[l10n] Add Tajik translation
Victor Ibragimov [Fri, 19 Apr 2013 22:10:36 +0000 (00:10 +0200)]
[l10n] Add Tajik translation

12 years agoRevert "gdk: Update keynames list from gdkkeysyms.h"
Matthias Clasen [Fri, 19 Apr 2013 21:10:39 +0000 (17:10 -0400)]
Revert "gdk: Update keynames list from gdkkeysyms.h"

This reverts commit 8c8853a1f5d29d447c16ba5d92bdb7bb442ee721.

12 years agogdk: Unvfuncify gdk_display_manager_open_display()
Benjamin Otte [Fri, 19 Apr 2013 20:17:17 +0000 (16:17 -0400)]
gdk: Unvfuncify gdk_display_manager_open_display()

This looks like a pretty stupid patch, but it's only a step towards the
ultimate end goal: Get rid of all the displaymanagers.

12 years agoRevert "wayland: Cache the discovery connection and reuse if possible"
Benjamin Otte [Fri, 19 Apr 2013 20:22:32 +0000 (16:22 -0400)]
Revert "wayland: Cache the discovery connection and reuse if possible"

This reverts commit 524ce7f88e26a498c43729f99fda564f5e222582.

12 years agowayland: Get rid of generic macros
Benjamin Otte [Fri, 19 Apr 2013 19:51:25 +0000 (15:51 -0400)]
wayland: Get rid of generic macros

They're unused.

12 years agobroadway: Get rid of generic macros
Benjamin Otte [Fri, 19 Apr 2013 19:50:38 +0000 (15:50 -0400)]
broadway: Get rid of generic macros

Use existing API instead.

12 years agogdk: Update keynames list from gdkkeysyms.h
Benjamin Otte [Fri, 19 Apr 2013 19:36:10 +0000 (15:36 -0400)]
gdk: Update keynames list from gdkkeysyms.h

grep \#define gdkkeysyms.h | sed "s/#define GDK_KEY_\(.*\) \(.*\)/\2
\1/" | grep -v \#define | sort > keynames.txt
./gen-keyname-table.pl keynames.txt keynames-translate.txt

https://bugzilla.gnome.org/show_bug.cgi?id=698385

12 years agox11: Simplify function
Benjamin Otte [Wed, 17 Apr 2013 22:04:54 +0000 (23:04 +0100)]
x11: Simplify function

Displays get made default displays automatically, so there's no need to
do it in the vfunc.

12 years agodisplaymanager: Emit display-opened directly
Benjamin Otte [Wed, 17 Apr 2013 22:04:05 +0000 (23:04 +0100)]
displaymanager: Emit display-opened directly

Instead of letting every backend do it manually.

12 years agoquartz: Emit GdkDisplay::opened signal
Benjamin Otte [Wed, 17 Apr 2013 22:00:00 +0000 (23:00 +0100)]
quartz: Emit GdkDisplay::opened signal

12 years agowin32: Emit GdkDisplay::opened signal
Benjamin Otte [Wed, 17 Apr 2013 21:59:21 +0000 (22:59 +0100)]
win32: Emit GdkDisplay::opened signal

12 years agodisplay: Add display to displaymanager later
Benjamin Otte [Wed, 17 Apr 2013 21:56:22 +0000 (22:56 +0100)]
display: Add display to displaymanager later

Instead of GdkDisplay::init, only add the display to the display manager
in GdkDisplay::opened. This avoids spurious changes of the default
display in gtk_init() when we're trying to find the one that works and
try to open lots of different ones.

12 years agodisplay: Make opened signal have a vfunc
Benjamin Otte [Wed, 17 Apr 2013 21:53:42 +0000 (22:53 +0100)]
display: Make opened signal have a vfunc

... instead of g_signal_connect()ing in every init function.

12 years agogdk: Refactor function
Benjamin Otte [Tue, 16 Apr 2013 15:29:05 +0000 (17:29 +0200)]
gdk: Refactor function

Split out a GdkBackend type and iterate over it.

12 years agoextract-strings: Deal with multiline strings
Matthias Clasen [Fri, 19 Apr 2013 19:22:03 +0000 (15:22 -0400)]
extract-strings: Deal with multiline strings

When putting out N_(...) macros, we must split multiline strings,
otherwise xgettext complains.

12 years agoRevert more POTFILES.in breakage
Matthias Clasen [Fri, 19 Apr 2013 18:39:23 +0000 (14:39 -0400)]
Revert more POTFILES.in breakage

12 years agoRevert adding .ui files to POTFILES.in
Piotr Drąg [Fri, 19 Apr 2013 18:27:40 +0000 (20:27 +0200)]
Revert adding .ui files to POTFILES.in

This breaks the build unfortunately. Developers will need to resolve that one.

12 years agoUpdated POTFILES.in
Piotr Drąg [Fri, 19 Apr 2013 17:40:56 +0000 (19:40 +0200)]
Updated POTFILES.in

Oops.

12 years agoRemove "tg" from po/LINGUAS until po-properties is fixed
Colin Walters [Fri, 19 Apr 2013 17:27:22 +0000 (13:27 -0400)]
Remove "tg" from po/LINGUAS until po-properties is fixed

Commit dcad786ba040413dd10e416d03d0cf7f3df226b6 introduced "tg" but
didn't add po-properties/tg.po, which is necessary in order for the
build to not break.

Therefore I'm removing tg until the file is added.

12 years agoUpdated POTFILES.in
Piotr Drąg [Fri, 19 Apr 2013 17:18:52 +0000 (19:18 +0200)]
Updated POTFILES.in

12 years agoMove the drop feedback code to stub functions
Federico Mena Quintero [Fri, 19 Apr 2013 16:25:03 +0000 (11:25 -0500)]
Move the drop feedback code to stub functions

For now it is the old, simple call to gtk_tree_view_set_drag_dest_row().  We'll do something
fancier in the following commits.

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agoAdd a DropState enum to the internal state of the sidebar
Federico Mena Quintero [Fri, 19 Apr 2013 16:00:36 +0000 (11:00 -0500)]
Add a DropState enum to the internal state of the sidebar

We want to show a little animation when a DnD operation is happening and the user hovers
between existing bookmarks.  The animation should indicate the user that he's about to
create a new bookmark from the file right there.

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agodoc: Documented icon-shadow CSS property
P. F. Chimento [Wed, 17 Apr 2013 08:33:43 +0000 (10:33 +0200)]
doc: Documented icon-shadow CSS property

The docs say that text-shadow applies a drop shadow to both text
and icons, but icons need to use the icon-shadow property.

https://bugzilla.gnome.org/show_bug.cgi?id=698181

12 years agoFix mouse interaction in combo cell renderer
Matthias Clasen [Fri, 19 Apr 2013 14:07:44 +0000 (10:07 -0400)]
Fix mouse interaction in combo cell renderer

Pass the event on when calling gtk_cell_editable_start_editing.
https://bugzilla.gnome.org/show_bug.cgi?id=504901

12 years agowayland: Cache the discovery connection and reuse if possible
Rob Bradford [Fri, 19 Apr 2013 12:59:58 +0000 (13:59 +0100)]
wayland: Cache the discovery connection and reuse if possible

This will prevent us opening, closing and reopening the same display
connection in the majority of cases.

https://bugzilla.gnome.org/show_bug.cgi?id=694465

12 years agoTajik Translation updated
Victor Ibragimov [Fri, 19 Apr 2013 12:27:17 +0000 (17:27 +0500)]
Tajik Translation updated

12 years agoTajik translation added (translation is still in progress)
Victor Ibragimov [Fri, 19 Apr 2013 12:23:35 +0000 (17:23 +0500)]
Tajik translation added (translation is still in progress)

12 years agomenu binding docs: Link to insert_action_group()
Ryan Lortie [Thu, 18 Apr 2013 11:11:15 +0000 (07:11 -0400)]
menu binding docs: Link to insert_action_group()

It's pretty useless to make a menu without actions behind it and people
who are using gtk_menu_shell_bind_model() directly are probably not
interested in doing it the GtkApplicationWindow way (so they won't get
the "app" and "win" groups for free).  People are going to need to call
gtk_widget_insert_action_group(), so mention this in the docs to help
them along.

12 years agobindings: Fix gtk_binding_entry_skip
Christian Persch [Fri, 15 Mar 2013 17:48:44 +0000 (18:48 +0100)]
bindings: Fix gtk_binding_entry_skip

This fixes a regression from commit 9a3466832cc73ae29b704547747b8699132c1250.

https://bugzilla.gnome.org/show_bug.cgi?id=695380

12 years agogdk: Unvfuncify generic key functions
Benjamin Otte [Tue, 16 Apr 2013 13:27:15 +0000 (15:27 +0200)]
gdk: Unvfuncify generic key functions

This makes Wayland and X11 no longer call into XKB and libX11 for these
functions but use GDK's own copy of these functions, just like the
win32, quartz and broadway backends.

12 years agogdk: Refactor default key vfuncs
Benjamin Otte [Tue, 16 Apr 2013 13:24:02 +0000 (15:24 +0200)]
gdk: Refactor default key vfuncs

Instead of copying them all over the place, keep a default
implementation around.

12 years agogdk: Refactor gdk_keyval_convert_case()
Benjamin Otte [Tue, 16 Apr 2013 11:02:23 +0000 (13:02 +0200)]
gdk: Refactor gdk_keyval_convert_case()

... so it doesn't export the function anymore.

12 years agobroadway: Delete 2 files doing nothing
Benjamin Otte [Mon, 15 Apr 2013 13:54:50 +0000 (15:54 +0200)]
broadway: Delete 2 files doing nothing

A function was doing nothing but calling a function that was in its own
source file doing nothing but calling a function in its own source file
that did nothing.

12 years agodocs: Move property section docs to correct file
Benjamin Otte [Mon, 15 Apr 2013 13:52:10 +0000 (15:52 +0200)]
docs: Move property section docs to correct file

12 years agoGtkSwitch: moved focus drawing from outside border inside the handle.
Juan Pablo Ugarte [Mon, 15 Apr 2013 20:24:48 +0000 (17:24 -0300)]
GtkSwitch: moved focus drawing from outside border inside the handle.

12 years agoMerge branch 'merge-places-sidebar'
Federico Mena Quintero [Tue, 16 Apr 2013 00:29:37 +0000 (19:29 -0500)]
Merge branch 'merge-places-sidebar'

This is the new GtkPlacesSidebar widget, which is the sidebar used in
GtkFileChooser to display a list of places/folders that the user may
want to use.  File managers are encouraged to use this widget for
their own global list of folders.

12 years agoLoad the file chooser's widgets from the .ui file
Tristan Van Berkom [Fri, 12 Apr 2013 01:37:01 +0000 (20:37 -0500)]
Load the file chooser's widgets from the .ui file

This is Tristan's *excellent* work, minus the old code for the shortcuts bar - that is all done
in GtkPlacesSidebar now.

The UI gets loaded from a Glade resource; most of the old code to create the UI by hand is gone.

There is still code for save_widgets_create(); this needs to be moved into the UI file, but it
is not a big deal.

gtk_file_chooser_default_init() calls a new post_process_ui() that takes care of all the things
that cannot be done directly in Glade.

12 years agoFor now, add GtkPlacesSidebar to the Glade catalog for private widgets
Federico Mena Quintero [Sat, 13 Apr 2013 00:57:35 +0000 (19:57 -0500)]
For now, add GtkPlacesSidebar to the Glade catalog for private widgets

Although GtkPlacesSidebar is a public widget, we can't have that in Glade's general catalog
as this branch is not merged into GTK+ proper yet.  We'll move the widget to Glade's
catalog when this is done.

12 years agoAdd .ui file for GtkFileChooserDefault
Federico Mena Quintero [Sat, 13 Apr 2013 00:58:16 +0000 (19:58 -0500)]
Add .ui file for GtkFileChooserDefault

12 years agoUpdated Interlingua translation
Nik Kalach [Mon, 15 Apr 2013 19:42:57 +0000 (21:42 +0200)]
Updated Interlingua translation

12 years agogdk: Make atoms handled generically
Benjamin Otte [Mon, 15 Apr 2013 13:40:55 +0000 (15:40 +0200)]
gdk: Make atoms handled generically

This is another step towards making GdkDisplayManager backend-agnostic.

Most of the backends profit from this as their atom implementations
where generic anyway - x11 needed that to allow multiple X displays and
broadway, quartz and wayland don't have the concept of displays.

The X11 backend still did things, so I only #if 0'd some code but did
not actually update anything.

12 years agogdk: Remove unused function
Benjamin Otte [Wed, 10 Apr 2013 21:27:52 +0000 (23:27 +0200)]
gdk: Remove unused function

12 years agodisplaymanager: Handle list of displays in base class
Benjamin Otte [Wed, 10 Apr 2013 21:11:52 +0000 (23:11 +0200)]
displaymanager: Handle list of displays in base class

This moves the add/remove_display() functions from the subclasses to
GdkDisplay and GdkDisplayManager. It also gets rid of the list_displays
vfunc.

12 years agodoc: Clarify gtk_widget_queue_draw_region() doc
Simon Kågedal Reimer [Tue, 9 Apr 2013 21:10:38 +0000 (23:10 +0200)]
doc: Clarify gtk_widget_queue_draw_region() doc

12 years agodisplaymanager: Move set_default_display vfunc
Benjamin Otte [Tue, 9 Apr 2013 11:32:03 +0000 (13:32 +0200)]
displaymanager: Move set_default_display vfunc

... to GdkDisplayClass.make_default. It's only implemented by X11
anyway.

12 years agodisplaymanager: Handle the default display
Benjamin Otte [Tue, 9 Apr 2013 11:25:31 +0000 (13:25 +0200)]
displaymanager: Handle the default display

... instead of having every backend do it on their own.

12 years agoprint: disconnect source_changed_handler on printeroptionwidget finalize
Alban Browaeys [Thu, 21 Mar 2013 08:45:32 +0000 (09:45 +0100)]
print: disconnect source_changed_handler on printeroptionwidget finalize

Printing a web page without window.print (which still segfault),
that is Ctrl+P is fixed (does not attempt to execute the
source_changed_handler on a printeroptionwidget that has been
destroyed) by disconnecting this handler in the printeroptionwidget
finalize.

https://bugzilla.gnome.org/show_bug.cgi?id=696622

12 years agogtk-demo: Fix typo
Volker Sobek [Tue, 8 May 2012 01:32:28 +0000 (03:32 +0200)]
gtk-demo: Fix typo

https://bugzilla.gnome.org/show_bug.cgi?id=675649

12 years agogtkprintjob: release the surface and backend before the output io channel
Carlos Garcia Campos [Wed, 3 Oct 2012 17:24:52 +0000 (19:24 +0200)]
gtkprintjob: release the surface and backend before the output io channel

If the GtkPrintJob is freed too early when the surface has been created
but the job hasn't been sent to the printer, it's possible that the
file print backend tries to write to the io chaneel when it is already
closed. This produces runtime critical warnings:

GLib-CRITICAL **: g_io_channel_write_chars: assertion `channel->is_writeable' failed

https://bugzilla.gnome.org/show_bug.cgi?id=685420

12 years agogtkprintbackendfile: Fix infinite loop in _cairo_write()
Carlos Garcia Campos [Wed, 3 Oct 2012 17:34:09 +0000 (19:34 +0200)]
gtkprintbackendfile: Fix infinite loop in _cairo_write()

It can happen if the io channel has been closed. In that case
g_io_channel_write_chars() returns early because of a g_return macro
that checks if the io channel is writable. When returning from g_return
macros, the bytes written output parameter is not updated and the error
is not filled, so the error is not detected and the written variable is
used uninitialized. We should check the return value of
g_io_channel_write_chars() to break the loop.

https://bugzilla.gnome.org/show_bug.cgi?id=685419

12 years agofix mem leak in gtk_drag_set_icon_surface, destroy pattern
ammonkey [Wed, 8 Aug 2012 17:07:17 +0000 (19:07 +0200)]
fix mem leak in gtk_drag_set_icon_surface, destroy pattern

https://bugzilla.gnome.org/show_bug.cgi?id=681446

12 years agoPrecache more atoms
Matthias Clasen [Sat, 13 Apr 2013 23:48:38 +0000 (19:48 -0400)]
Precache more atoms

Add a few more entries to the list of precached atoms.

12 years agowayland: Don't get the shell surface twice
Matthias Clasen [Sat, 13 Apr 2013 21:20:05 +0000 (17:20 -0400)]
wayland: Don't get the shell surface twice

Noticed by Pu Xingyu,
https://bugzilla.gnome.org/show_bug.cgi?id=697947

12 years agoRevert "Use natural size to set adjustment ranges"
Matthias Clasen [Sat, 13 Apr 2013 20:17:24 +0000 (16:17 -0400)]
Revert "Use natural size to set adjustment ranges"

This reverts commit 666d10ec7676a15b6861b785397563cbd4ef21e6.

This change severely broke any treeviews without horizontal
scrollbars. Basically, ellipsization never kicks in, and instead
the treeview content just extends outside the visible area,
rendering it inaccessible. This broke e.g. the control-center
keyboard shortcuts panel, the gnome-disks device list, etc etc.

12 years agocsd: Don't use a content_window for o-r windows
Matthias Clasen [Sat, 13 Apr 2013 02:45:13 +0000 (22:45 -0400)]
csd: Don't use a content_window for o-r windows

This fixes drawing of dnd icons.
https://bugzilla.gnome.org/show_bug.cgi?id=697886

12 years agowayland: Don't "probe" for a Wayland compositor if started from compositor
Rob Bradford [Fri, 12 Apr 2013 17:39:20 +0000 (18:39 +0100)]
wayland: Don't "probe" for a Wayland compositor if started from compositor

In the case that the client is started directly by the compositor the
WAYLAND_SOCKET environment variable is set containing the fd to use that was
created by a socketpair.

This environment variable is consumed by a call to wl_display_connect so a
second call will not take advantage of it.

https://bugzilla.gnome.org/show_bug.cgi?id=697673

12 years agoMove all of GtkFileChooserDefault's fields into a private structure
Federico Mena Quintero [Fri, 12 Apr 2013 03:44:19 +0000 (22:44 -0500)]
Move all of GtkFileChooserDefault's fields into a private structure

I love Emacs macros, by the way.

12 years agoMove GtkFileChooserDefault to gtkfilechooserdefault.c
Federico Mena Quintero [Fri, 12 Apr 2013 02:44:21 +0000 (21:44 -0500)]
Move GtkFileChooserDefault to gtkfilechooserdefault.c

We will move all the fields to a private structure next.

12 years agoMove GtkFileChooserWidgetPrivate to gtkfilechooserwidget.c
Federico Mena Quintero [Fri, 12 Apr 2013 01:50:57 +0000 (20:50 -0500)]
Move GtkFileChooserWidgetPrivate to gtkfilechooserwidget.c

In addition, remove the unused "file_system" field from that
structure.

12 years agoMove GtkFileChooserDialogPrivate into gtkfilechooserdialog.c
Federico Mena Quintero [Fri, 12 Apr 2013 01:46:30 +0000 (20:46 -0500)]
Move GtkFileChooserDialogPrivate into gtkfilechooserdialog.c

We'll take private structures out of gtkfilechooserprivate.h and move
them to the appropriate place.

12 years agoRemove unused #define
Federico Mena Quintero [Fri, 12 Apr 2013 01:42:43 +0000 (20:42 -0500)]
Remove unused #define

12 years agoDon't include gtkfilechooserprivate.h from the file chooser's tests
Federico Mena Quintero [Fri, 12 Apr 2013 01:40:05 +0000 (20:40 -0500)]
Don't include gtkfilechooserprivate.h from the file chooser's tests

We will make that file a fully private header; the tests have no
business poking inside the file chooser's internals at this point.

12 years agoGtkPlacesSidebar: Hold a reference to the sidebar while a D-Bus proxy is being created
Federico Mena Quintero [Fri, 12 Apr 2013 01:04:53 +0000 (20:04 -0500)]
GtkPlacesSidebar: Hold a reference to the sidebar while a D-Bus proxy is being created

The creation is async, and the sidebar could be destroyed while the D-Bus proxy is being created.

Signed-off-by: Federico Mena Quintero <federico@gnome.org>
12 years agoMerge branch 'places-sidebar' into master
Federico Mena Quintero [Fri, 12 Apr 2013 00:45:12 +0000 (19:45 -0500)]
Merge branch 'places-sidebar' into master

This lands the GtkPlacesSidebar widget.  It is used in
GtkFileChooserDefault, and it can also be used by third-party
applications.

12 years agoRevert "GtkFileChooserDefault: Define children with a GtkBuilder template"
Federico Mena Quintero [Thu, 11 Apr 2013 21:47:38 +0000 (16:47 -0500)]
Revert "GtkFileChooserDefault: Define children with a GtkBuilder template"

Having the changes for composite widget templates makes it impossible
to merge the places-sidebar branch.  So, we will merge that branch,
and *then* apply the changes for composite templates.

This reverts commit bf909f56157311174d5d9089bdec954e7def8873.

12 years agoxi2: Improve pointer emulation debug reporting
David King [Thu, 11 Apr 2013 13:29:11 +0000 (14:29 +0100)]
xi2: Improve pointer emulation debug reporting

Reporting "true" or "false" is nicer than the value of the flag.

https://bugzilla.gnome.org/show_bug.cgi?id=697795

12 years agogdkwindow: Handle updates created by outstanding moves in same update
Alexander Larsson [Thu, 11 Apr 2013 10:15:30 +0000 (12:15 +0200)]
gdkwindow: Handle updates created by outstanding moves in same update

If gdk_window_flush_outstanding_moves() creates new update area
we handle this directly in the same draw to avoid flashing.

This mainly affects win32 as X11 does its exposes from moves async.
However, its important for win32 since ScrollDC seems to sometimes
invalidate (and not copy) unexected regions.

http://bugzilla.gnome.org/show_bug.cgi?674051

12 years agowin32: Report ScrollDC update region directly
Alexander Larsson [Thu, 11 Apr 2013 10:12:35 +0000 (12:12 +0200)]
win32: Report ScrollDC update region directly

Rather than set the window update region and repaint this region
when we get a WM_PAINT we just directly add it to the update
region. No need to roundtrip via win32.

This lets us also make sure we do this drawing in the same update
cycle. This seems especially important on Win7, because ScrollDC
seems to act kind of weird there, not using bitblt in areas where
it seemingly could, which makes scrolling look really flashy.

http://bugzilla.gnome.org/show_bug-cgi?id=674051

12 years agoMake gdk_flush_outstanding_moves "reentrant"
Alexander Larsson [Thu, 11 Apr 2013 09:35:17 +0000 (11:35 +0200)]
Make gdk_flush_outstanding_moves "reentrant"

If do_move_region_bits_on_impl causes updates they will propely
be moved by the outstanding moves we have not yet flushed.

12 years agoUpdate gtk.symbols
Alexander Larsson [Thu, 11 Apr 2013 09:23:15 +0000 (11:23 +0200)]
Update gtk.symbols

This fixes win32 build

12 years agowin32: Fix build
Alexander Larsson [Thu, 11 Apr 2013 09:22:27 +0000 (11:22 +0200)]
win32: Fix build

Need to take ".exe" into account for file dependencies on win32.

12 years agowin32: Make build with latest gdkdisplay changes
Alexander Larsson [Thu, 11 Apr 2013 09:22:04 +0000 (11:22 +0200)]
win32: Make build with latest gdkdisplay changes

12 years agoxi2: Reset scroll valuators on synthesized crossing events
Carlos Garnacho [Thu, 28 Mar 2013 13:57:00 +0000 (14:57 +0100)]
xi2: Reset scroll valuators on synthesized crossing events

On crossing events resulting from moving windows (eg. workspace switch),
deviceid equals sourceid, so make those reset scroll valuators on all
slave devices to avoid misleading jumps in scroll events

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=690275

12 years agocsd: Use symbolic icons for window buttons
Matthias Clasen [Wed, 10 Apr 2013 10:13:26 +0000 (06:13 -0400)]
csd: Use symbolic icons for window buttons

With symbolic icons, we can recolor the icons as needed, and we
can use icon shadows, etc.

12 years agoDon't free a list that has not been initialized
Matthias Clasen [Wed, 10 Apr 2013 01:18:05 +0000 (21:18 -0400)]
Don't free a list that has not been initialized

12 years agowayland: Use more explicit out parameter names
Chris Cummins [Thu, 28 Mar 2013 14:47:03 +0000 (14:47 +0000)]
wayland: Use more explicit out parameter names

Prevent confusion between positional values and hotspot values.

Signed-off-by: Rob Bradford <rob@linux.intel.com>
12 years agoUpdate GtkCssProvider box-shadow documentation
Chris Cummins [Tue, 9 Apr 2013 10:56:17 +0000 (11:56 +0100)]
Update GtkCssProvider box-shadow documentation

Non-inset box-shadows are no longer ignored by the GTK+ theming engine,
outset shadows are supported.

Signed-off-by: Rob Bradford <rob@linux.intel.com>
12 years agocds: enable outset shadows
Chris Cummins [Fri, 5 Apr 2013 09:54:03 +0000 (10:54 +0100)]
cds: enable outset shadows

Adds conditional code paths to GdkCssShadowValue for painting outset
shadows, and allows shadows to be applied in two passes (first outset
then inset). This can be used to draw csd shadows in outer window
borders.

https://bugzilla.gnome.org/show_bug.cgi?id=695998
Signed-off-by: Rob Bradford <rob@linux.intel.com>
12 years agowayland: don't leave root window values uninitialised
Rob Bradford [Tue, 9 Apr 2013 10:49:42 +0000 (11:49 +0100)]
wayland: don't leave root window values uninitialised

Under Wayland we don't know the absolute position of the device but there are
some API calls that expect to get an root window position. Previously we were
not assigning any value to these out parameters potentially leaving the values
undefined.

This change returns the current surface relative position of the device.

12 years agoquartz: Fix call to missing symbol
Carlos Garnacho [Mon, 8 Apr 2013 15:25:19 +0000 (17:25 +0200)]
quartz: Fix call to missing symbol

This patch can be considered a leftover of commit 6f607fc8b4cd ,
replace the no longer existing function with the GDK function
meant to replace it.

This was noticeable on regular clipboard operations in quartz

12 years agowindow: Remember current size
Benjamin Otte [Tue, 9 Apr 2013 10:08:08 +0000 (12:08 +0200)]
window: Remember current size

Don't just look at previously remembered sizes, also look at the current
size.
This is useful for cases where the window was resized by the user or WM
and not by the application itself.

https://bugzilla.gnome.org/show_bug.cgi?id=696882

12 years agowindow: Split out a function
Benjamin Otte [Tue, 9 Apr 2013 10:06:51 +0000 (12:06 +0200)]
window: Split out a function

We don't want to remember sizes in the not resizable case. Also a
function named "guess_default_size" should not look at previous sizes,
it should guess.

12 years agowindow: Remove an unused function.
Benjamin Otte [Tue, 9 Apr 2013 09:20:49 +0000 (11:20 +0200)]
window: Remove an unused function.

Luckily, we don't need to support an implementation of a minimal WM for
linux-fb anymore (which is what this function was added for).

12 years agodocs: Fix Since tag for frame clock functions
Benjamin Otte [Mon, 8 Apr 2013 16:05:53 +0000 (18:05 +0200)]
docs: Fix Since tag for frame clock functions

12 years agowayland: Clean up gdkwayland.h
Matthias Clasen [Tue, 9 Apr 2013 02:28:43 +0000 (22:28 -0400)]
wayland: Clean up gdkwayland.h

Make this an include-only header

13 years agowindow: restore size after hide/show properly
Benjamin Otte [Mon, 8 Apr 2013 14:25:29 +0000 (16:25 +0200)]
window: restore size after hide/show properly

Old code assumed the size was stored in widget.allocation. This is no
longer true as the allocation is cleared upon hide. However, we store
the last configure request, and that one tracks the last size, so we can
just use that number.

Sometimes things are so easy - once you figure them out...

https://bugzilla.gnome.org/show_bug.cgi?id=696882

13 years agoplug: Don't unregister windows that were never registered
Benjamin Otte [Mon, 8 Apr 2013 13:28:22 +0000 (15:28 +0200)]
plug: Don't unregister windows that were never registered

Fixes warnings when trying to unregister the socket window. The socket
window is foreign, so we shouldn't register it in the first place.

https://bugzilla.gnome.org/show_bug.cgi?id=697427

13 years agobuild: Fix srcdir != builddir
Colin Walters [Mon, 8 Apr 2013 14:16:58 +0000 (10:16 -0400)]
build: Fix srcdir != builddir

13 years agoExtract strings from .ui files without intltool
Matthias Clasen [Mon, 8 Apr 2013 03:42:55 +0000 (23:42 -0400)]
Extract strings from .ui files without intltool

13 years agoGtkVolumeButton: Define children with a GtkBuilder template
Tristan Van Berkom [Sat, 30 Mar 2013 09:51:58 +0000 (18:51 +0900)]
GtkVolumeButton: Define children with a GtkBuilder template

13 years agoGtkFontButton: Define children with a GtkBuilder template
Tristan Van Berkom [Sat, 30 Mar 2013 08:35:54 +0000 (17:35 +0900)]
GtkFontButton: Define children with a GtkBuilder template