platform/upstream/at-spi2-core.git
12 years agoRelease v2.1.3 AT_SPI2_CORE_2_1_3
Mike Gorse [Mon, 4 Jul 2011 21:47:21 +0000 (16:47 -0500)]
Release v2.1.3

12 years agoAdded Lithuanian translation.
Aurimas Černius [Wed, 29 Jun 2011 10:58:34 +0000 (13:58 +0300)]
Added Lithuanian translation.

12 years agoFix a failure to respond to children-changed events in some cases
Mike Gorse [Tue, 28 Jun 2011 20:00:41 +0000 (15:00 -0500)]
Fix a failure to respond to children-changed events in some cases

When a children-changed event is received, the cache of children should
be updated if present.  However, to test this, the code was checking for
the presence of a non-NULL list of children, which would be the case if
the object initially had 0 children.  The correct thing to do is to
check the accessible's cached_properties.

12 years agoAdd atspi_accessible_get_process_id
Mike Gorse [Mon, 27 Jun 2011 22:18:05 +0000 (17:18 -0500)]
Add atspi_accessible_get_process_id

12 years agoSet the gerror for a dbus error in _atspi_dbus_call
Trevor Saunders [Mon, 27 Jun 2011 14:32:08 +0000 (09:32 -0500)]
Set the gerror for a dbus error in _atspi_dbus_call

12 years agoTry to fix gtk-doc flags/libs
Mike Gorse [Sun, 26 Jun 2011 03:02:04 +0000 (22:02 -0500)]
Try to fix gtk-doc flags/libs

12 years agoAdded Belarusian translation.
Ihar Hrachyshka [Sat, 25 Jun 2011 15:04:40 +0000 (18:04 +0300)]
Added Belarusian translation.

12 years agoBGO#652596: Use DBus property to set value
Mike Gorse [Wed, 22 Jun 2011 16:06:19 +0000 (11:06 -0500)]
BGO#652596: Use DBus property to set value

12 years agoOne check for gio is enough.
Patrick Welche [Tue, 21 Jun 2011 11:02:29 +0000 (12:02 +0100)]
One check for gio is enough.

12 years agoDon't pollute installed header files with config.h.
Patrick Welche [Wed, 15 Jun 2011 22:11:32 +0000 (23:11 +0100)]
Don't pollute installed header files with config.h.

atspi.h -> atspi-gmain.h -> config.h

12 years agoUse stored values for AT-SPI role names
Mike Gorse [Wed, 15 Jun 2011 20:33:33 +0000 (15:33 -0500)]
Use stored values for AT-SPI role names

When atspi_accessible_get_role returns a recognized role that isn't
ATSPI_ROLE_EXTENDED, use a stored role name in order to take advantage
of roles being cached.

12 years agoRelease v2.1.2 AT_SPI2_CORE_2_1_2
Mike Gorse [Mon, 13 Jun 2011 22:31:38 +0000 (17:31 -0500)]
Release v2.1.2

12 years agoMake the bus launcher abort if an instance is already running
Mike Gorse [Thu, 9 Jun 2011 16:50:58 +0000 (11:50 -0500)]
Make the bus launcher abort if an instance is already running

Upon starting up, at-spi-bus-launcher now looks for an X property with
the bus address and tries to connect to it, aborting if successful.
This fixes a bug where, if the launcher was run twice (perhaps by both
an autostart script and a request for the bus address), the second
instance would start its main loop, eventually abort, and delete the X
property, which would cause a new X property to be created later,
overriding the old one, if a different user requested the address,
leading to a11y breakage.

12 years agoRevert "Clean up enums"
Mike Gorse [Sun, 29 May 2011 22:27:02 +0000 (17:27 -0500)]
Revert "Clean up enums"

This reverts commit a12d7458110c57e183916255d701b1e0d7c070e1.

Conflicts:

atspi/atspi-constants.h

Reverting these enum clean-ups for now, since they require pygobject 2.28.0
for pyatspi, at least one user is pulling git while using pygobject 2.26.0,
upgrading pygobject is likely to be a complicated process that may break
things, and these changes are only code clean-ups that will not really benefit
users.  Will reinstate in a separate branch and merge some time in the future.

12 years agoRevert "Install enum header, and try to get role name from enum"
Mike Gorse [Sun, 29 May 2011 22:24:44 +0000 (17:24 -0500)]
Revert "Install enum header, and try to get role name from enum"

This reverts commit 9ddf777a096440afe521c094544942e85e3ee838.

12 years agoRevert "Remove redundant tables of state and role names"
Mike Gorse [Sun, 29 May 2011 22:24:38 +0000 (17:24 -0500)]
Revert "Remove redundant tables of state and role names"

This reverts commit 73b811b78954dfb0f8b407354c146d9fbd47dea3.

12 years agoFix the build
Mike Gorse [Sun, 29 May 2011 13:37:17 +0000 (08:37 -0500)]
Fix the build

Include atspi-gmain.h from dbind rather than atspi.h, since atspi.h includes
atspi-enum-types.h, which has not yet been built.

12 years ago[l10n]Added Catalan (Valencian) translation
Gil Forcada [Sun, 29 May 2011 11:57:30 +0000 (13:57 +0200)]
[l10n]Added Catalan (Valencian) translation

12 years agoAdd Li Yuan to MAINTAINERS and .doap
Mike Gorse [Fri, 27 May 2011 04:36:22 +0000 (23:36 -0500)]
Add Li Yuan to MAINTAINERS and .doap

12 years agoMake deferred_messages a GQueue rather than a GList
Mike Gorse [Fri, 27 May 2011 04:03:15 +0000 (23:03 -0500)]
Make deferred_messages a GQueue rather than a GList

If an application sends a large number of events, then deferred_messages can
become quite large (possibly holding several thousand messages), so appending
to a GList can become a performance hit.  Thus, use a GQueue instead.

12 years agoAllow caching of attributes
Mike Gorse [Thu, 26 May 2011 22:27:28 +0000 (17:27 -0500)]
Allow caching of attributes

Attributes can now be cached, but not enabling by default, since there
is currently no event to notify AT-SPI that attributes have changed (see
BGO#649771), so this is dangerous but may improve performance if we can
reliably assume that attributes will not change.

12 years agoRemove redundant tables of state and role names
Mike Gorse [Wed, 25 May 2011 19:36:52 +0000 (14:36 -0500)]
Remove redundant tables of state and role names

Use the generated enum type to convert state names to numeric values rather
than a string table that needs to be maintained in parallel.  Also, remove
atspi_role_get_name since it was unused, marked deprecated, and relied on a
string table which is now redundant with the enum type.

12 years agoInstall enum header, and try to get role name from enum
Mike Gorse [Wed, 25 May 2011 18:00:24 +0000 (13:00 -0500)]
Install enum header, and try to get role name from enum

Have atspi_accessible_get_role_name try to use the internal string associated
with the accessible's role rather than calling GetRoleName.  This avoids
making a DBus call when the role is cached.

Also, atspi.h now includes atspi-enum-types.h, and the latter is now installed.

12 years agoCatch toolkit name/version and AT-SPI version per-application
Mike Gorse [Wed, 25 May 2011 15:10:10 +0000 (10:10 -0500)]
Catch toolkit name/version and AT-SPI version per-application

12 years agoDisable caching if ATSPI_NO_CACHE is set in the environment and not 0
Mike Gorse [Tue, 24 May 2011 22:15:13 +0000 (17:15 -0500)]
Disable caching if ATSPI_NO_CACHE is set in the environment and not 0

12 years agoAdd atspi_accessible_clear_cache
Mike Gorse [Tue, 24 May 2011 21:56:42 +0000 (16:56 -0500)]
Add atspi_accessible_clear_cache

12 years agoClean up enums
Mike Gorse [Tue, 24 May 2011 21:24:09 +0000 (16:24 -0500)]
Clean up enums

Use glib-enums to generate GTypes for enums, and modify enum definitions so
that they will be marked as bitflags where appropriate.  This helps with
introspection and allows functions that take flags to be prototyped as such
rather than being marked as taking a gint.

12 years agoFix get_toolkit_version, and add get_atspi_version for accessibles
Mike Gorse [Mon, 23 May 2011 23:04:01 +0000 (18:04 -0500)]
Fix get_toolkit_version, and add get_atspi_version for accessibles

12 years agoDisable caching when not in atspi_event_main
Mike Gorse [Mon, 23 May 2011 22:46:46 +0000 (17:46 -0500)]
Disable caching when not in atspi_event_main

If we are not in a GLib main loop, then we will not see signals to update
the cache, so cached data will become invalid.  Assume that, if
atspi-event_main has not been called, then we are, for instance, in a simple
pyatspi script without a main loop, so we should disable caching.

12 years agoUpdated Vietnamese translation
Trương Ứng Minh [Mon, 23 May 2011 12:15:05 +0000 (19:15 +0700)]
Updated Vietnamese translation

12 years agoDynamically allocate reentrant call closures
Mike Gorse [Mon, 23 May 2011 02:53:36 +0000 (21:53 -0500)]
Dynamically allocate reentrant call closures

Putting reentrant call closures on the stack can be dangerous if set_reply
is called after the caller has timed out and returned, for instance, so better
to dynamically allocate them.  This will hopefully fix some crashes and
hopefully not leak memory.

12 years agoFix/rework selective caching
Mike Gorse [Sun, 22 May 2011 23:37:27 +0000 (18:37 -0500)]
Fix/rework selective caching

Previously, atspi_accessible_set_cache_mask would not affect accessibles
which were already cached, but this was not very useful, as caching is done
automatically when a new app is discovered.  Functions that may retrieve
cached data now test the cache setting for the application--not just whether
there is a value stored for the data--whe ndetermining whether to call the app.

Also, annotate atspi_accessible_set_cache_mask as a gint.  This is temporary;
properly generating GTypes for enums/flags is TODO for 2.2.

12 years agoAdded Aragonese translation
Daniel Martinez Cucalon [Thu, 19 May 2011 08:49:10 +0000 (10:49 +0200)]
Added Aragonese translation

12 years agoFix table properties that return objects
Mike Gorse [Tue, 17 May 2011 22:06:39 +0000 (17:06 -0500)]
Fix table properties that return objects

12 years agoAnnotate atspi_table_get_row_column_extents_at_index out parameters
Mike Gorse [Tue, 17 May 2011 21:21:50 +0000 (16:21 -0500)]
Annotate atspi_table_get_row_column_extents_at_index out parameters

12 years agoIn atspi_match_rule_get_attributes, copy the hash rather than ref it
Mike Gorse [Tue, 17 May 2011 19:00:40 +0000 (14:00 -0500)]
In atspi_match_rule_get_attributes, copy the hash rather than ref it

It is not safe to ref a hash marked (transfer none) and expect its elements
to remain, since the caller will assume that it owns the elements and may
deallocate them (this is done by pygobject).  It would also be problematic
to simply mark the hash (transfer full), since we need to either assume that
the caller passed a destroy function when creating the hash (pygobject does
not) or deallocate the values ourselves when the match rule is finalized
(which requires making assumptions about the correct free function needed to
deallocate the values that the caller allocated).  So it seems that the only
safe thing to do is to copy the hash.

12 years agoAdd Esperanto translation
Kristjan SCHMIDT [Sun, 15 May 2011 13:35:30 +0000 (15:35 +0200)]
Add Esperanto translation

12 years agoBuild fixes.
Patrick Welche [Sat, 14 May 2011 16:57:42 +0000 (17:57 +0100)]
Build fixes.

- atspi-application.h makes use of DBusConnection* .
- -Werror-implicit-function-declaration doesn't contain an equal sign in
  my copy of gcc, and is gcc specific.
- The default behaviour of AC_CHECK_LIB glib-2.0 is to add -lglib-2.0 to
  LIBS which can then cause the search for a dlopen function to fail
  because the flags to find libglib aren't included. Solution is to
  tell AC_CHECK_LIB to do nothing(!)

13 years agoRelease v2.1.1 AT_SPI2_CORE_2_1_1
Mike Gorse [Mon, 9 May 2011 13:55:28 +0000 (15:55 +0200)]
Release v2.1.1

13 years agoCall g_error if unable to find the accessibility bus
Mike Gorse [Fri, 6 May 2011 15:59:45 +0000 (10:59 -0500)]
Call g_error if unable to find the accessibility bus

Previously, AT-SPI would wind up passing a NULL connection to libdbus if it
couldn't connect to the bus, causing a cryptic abort.  This change makes it
at least abort with a sensible error message in that case.
Also, if unable to find the accessibility bus address via the session bus,
save and print the DBusError, if any.

13 years agoFix memory leaks in get_toolkit_name and get_toolkit_version
Mike Gorse [Fri, 6 May 2011 00:39:33 +0000 (19:39 -0500)]
Fix memory leaks in get_toolkit_name and get_toolkit_version

_atspi_dbus_get_property calls g_strdup when it returns a string, so we
should not g_strdup the result unless we're caching a copy for later use.

13 years agoRemove some redundant out-of-memory checks
Mike Gorse [Thu, 5 May 2011 22:14:02 +0000 (17:14 -0500)]
Remove some redundant out-of-memory checks

If malloc returns NULL, then g_malloc will call g_error and thus abort, so
it is redundant to check the return value of g_new0 and the like, since glib
will already have aborted the program rather than returning a NULL value.
"Cleaning up" the code by removing some of these checks.

13 years agoDon't translate g_warning messages
Mike Gorse [Thu, 5 May 2011 21:57:25 +0000 (16:57 -0500)]
Don't translate g_warning messages

The consensus seems to be that messages written out as a result of calls to
g_warning/g_log should ot be translated; ie:
http://mail.gnome.org/archives/desktop-devel-list/2011-March/msg00070.html

So removing translations from these calls.

13 years agoFix a possible crash when an application is short-lived
Mike Gorse [Wed, 4 May 2011 19:34:48 +0000 (14:34 -0500)]
Fix a possible crash when an application is short-lived

When GetAPplicationBusAddress is called, at-spi sets a callback to handle the
reply.  It is possible that the application will go away and the
AtspiApplication will be disposed before this callback is reached, in which
case the reference to the accessibility bus connection will have gone away.
In this case, we should not call GetItems, since app->bus will be NULL and
libdbus will throw a SIGABRT.

13 years agoOnly flush a display if we have one
Matthias Clasen [Fri, 29 Apr 2011 14:31:02 +0000 (10:31 -0400)]
Only flush a display if we have one

This was causing occasional segfaults in the gdm session.
https://bugzilla.gnome.org/show_bug.cgi?id=648942

13 years agoRelease v2.0.1 AT_SPI2_CORE_2_0_1
Mike Gorse [Mon, 25 Apr 2011 23:41:51 +0000 (18:41 -0500)]
Release v2.0.1

13 years agoRemove doc about non-existent ATSPI_KEYSET_ALL_KEYS.
Daiki Ueno [Mon, 25 Apr 2011 02:02:27 +0000 (11:02 +0900)]
Remove doc about non-existent ATSPI_KEYSET_ALL_KEYS.

13 years agoUpdated asturian translation
Xandru Armesto [Mon, 25 Apr 2011 06:23:15 +0000 (08:23 +0200)]
Updated asturian translation

13 years agoAdded asturian language
Xandru Armesto [Mon, 25 Apr 2011 06:22:59 +0000 (08:22 +0200)]
Added asturian language

13 years agoAdded Turkish translation
Muhammet Kara [Fri, 22 Apr 2011 21:13:55 +0000 (00:13 +0300)]
Added Turkish translation

13 years agoUpdated Persian translation
Arash Mousavi [Fri, 22 Apr 2011 20:03:52 +0000 (00:33 +0430)]
Updated Persian translation

13 years agoUpdated Persian translation
Arash Mousavi [Fri, 22 Apr 2011 20:01:25 +0000 (00:31 +0430)]
Updated Persian translation

13 years agoUpdated Persian translation
Arash Mousavi [Fri, 22 Apr 2011 19:57:36 +0000 (00:27 +0430)]
Updated Persian translation

13 years agoRemove po/Makefile.in.in & add to .gitignore
Craig Keogh [Mon, 18 Apr 2011 11:55:19 +0000 (21:25 +0930)]
Remove po/Makefile.in.in & add to .gitignore

13 years agoAdd atspi_accessible_get_id
Mike Gorse [Mon, 18 Apr 2011 20:38:32 +0000 (15:38 -0500)]
Add atspi_accessible_get_id

13 years agoFix for BGO#648130: event host_application should return NULL for the desktop
Mike Gorse [Mon, 18 Apr 2011 16:44:28 +0000 (11:44 -0500)]
Fix for BGO#648130: event host_application should return NULL for the desktop

13 years agoFix a crash introduced in last commit
Mike Gorse [Mon, 18 Apr 2011 16:33:51 +0000 (11:33 -0500)]
Fix a crash introduced in last commit

13 years agoAdd intltoolize to autogen.sh (GNOME bug 648014)
Craig Keogh [Sun, 17 Apr 2011 12:27:46 +0000 (21:57 +0930)]
Add intltoolize to autogen.sh (GNOME bug 648014)

13 years agoHonor aclocal flags (GNOME bug 648013)
Craig Keogh [Sun, 17 Apr 2011 12:07:18 +0000 (21:37 +0930)]
Honor aclocal flags (GNOME bug 648013)

13 years agoFix for BGO#648059: get_application should return something w/ROLE_APPLICATION
Mike Gorse [Mon, 18 Apr 2011 10:07:14 +0000 (05:07 -0500)]
Fix for BGO#648059: get_application should return something w/ROLE_APPLICATION

atspi_accessible_get_application will walk up the tree until it finds a child
of the desktop, but this causes it to return the wrong accessible if an
object's parent is set to NULL for whatever reason, so check for this and
return the object's root accessible in that case.

13 years agofix atk_text_get_text leak
Michael Meeks [Fri, 8 Apr 2011 18:27:13 +0000 (19:27 +0100)]
fix atk_text_get_text leak

13 years agoUpdating Persian Translation
Mahyar Moghimi [Tue, 5 Apr 2011 14:16:33 +0000 (18:46 +0430)]
Updating Persian Translation

13 years agoRelease v2.0.0 AT_SPI2_CORE_2_0_0
Mike Gorse [Mon, 4 Apr 2011 22:34:18 +0000 (17:34 -0500)]
Release v2.0.0

13 years ago[l10n] Updated German translation, umlauts fixed
Wolfgang Stöggl [Sun, 3 Apr 2011 17:59:21 +0000 (19:59 +0200)]
[l10n] Updated German translation, umlauts fixed

13 years agoAdded Basque language
Inaki Larranaga Murgoitio [Sun, 3 Apr 2011 15:09:33 +0000 (17:09 +0200)]
Added Basque language

13 years ago[l10n]Fixes on Catalan translation
Jordi Serratosa [Sat, 2 Apr 2011 18:16:14 +0000 (20:16 +0200)]
[l10n]Fixes on Catalan translation

13 years agoregistry: Use correct signature for EndSessionResponse
Matthias Clasen [Fri, 1 Apr 2011 15:41:17 +0000 (11:41 -0400)]
registry: Use correct signature for EndSessionResponse

It turns out that using the wrong signature here adds a 30 second
wait to login/logout, when accessibility is turned on.
Thanks to Panu Matilainen for tracking this down.
https://bugzilla.gnome.org/show_bug.cgi?id=646403

13 years agoUpdated Danish translation
Kris Thomsen [Tue, 29 Mar 2011 21:50:30 +0000 (23:50 +0200)]
Updated Danish translation

13 years agoAdded da to list of languages
Kenneth Nielsen [Tue, 29 Mar 2011 21:50:29 +0000 (23:50 +0200)]
Added da to list of languages

13 years agodbus-glib-1 -> dbus-1 in AT-SPI GIR flags
Mike Gorse [Tue, 29 Mar 2011 14:54:49 +0000 (09:54 -0500)]
dbus-glib-1 -> dbus-1 in AT-SPI GIR flags

13 years agoUpdated Polish translation
Piotr Drąg [Mon, 28 Mar 2011 18:51:59 +0000 (20:51 +0200)]
Updated Polish translation

13 years agohindi added and translated and lingua upodated
Rajesh Ranjan [Mon, 28 Mar 2011 08:57:47 +0000 (14:27 +0530)]
hindi added and translated and lingua upodated

13 years agoAdded Brazilian Portuguese translation.
Og B. Maciel [Sun, 27 Mar 2011 22:13:10 +0000 (18:13 -0400)]
Added Brazilian Portuguese translation.

13 years agoUpdated Brazilian Portuguese translation. Reviewed by Flamarion Jorge and Antonio...
Rodrigo Padula de Oliveira [Sun, 27 Mar 2011 21:04:36 +0000 (17:04 -0400)]
Updated Brazilian Portuguese translation. Reviewed by Flamarion Jorge and Antonio Fernandes C. Neto.

13 years agoRelease v1.91.93
Mike Gorse [Fri, 25 Mar 2011 16:28:17 +0000 (09:28 -0700)]
Release v1.91.93

13 years agoAdded UG translation
Abduxukur Abdurixit [Sun, 27 Mar 2011 17:12:23 +0000 (19:12 +0200)]
Added UG translation

13 years agoupdate tranlation for Punjabi by A S Alam
A S Alam [Sun, 27 Mar 2011 04:17:56 +0000 (09:47 +0530)]
update tranlation for Punjabi by A S Alam

13 years agoAdded Vietnamese translation
Nguyễn Thái Ngọc Duy [Sat, 26 Mar 2011 09:40:49 +0000 (16:40 +0700)]
Added Vietnamese translation

13 years agoClear AT_SPI_BUS property on launcher exit
Colin Walters [Thu, 24 Mar 2011 17:30:51 +0000 (13:30 -0400)]
Clear AT_SPI_BUS property on launcher exit

Clear the X property if our bus is gone; in the case where e.g.
GDM is launching a login on an X server it was using before,
we don't want early login processes to pick up the stale address.

13 years agoUpdated Italian translation
Luca Ferretti [Wed, 23 Mar 2011 18:19:35 +0000 (19:19 +0100)]
Updated Italian translation

13 years agoAdded Marathi Translations
Sandeep Shedmake [Wed, 23 Mar 2011 10:47:28 +0000 (16:17 +0530)]
Added Marathi Translations

13 years agoAdded mr to LINGUAS
Sandeep Shedmake [Wed, 23 Mar 2011 10:22:37 +0000 (15:52 +0530)]
Added mr to LINGUAS

13 years agoRemoving not required dbus-glib related includes
Alejandro Piñeiro [Tue, 22 Mar 2011 11:18:53 +0000 (12:18 +0100)]
Removing not required dbus-glib related includes

13 years agoRelease v1.91.92 and fix some build errors AT_SPI2_CORE_1_91_92
Mike Gorse [Mon, 21 Mar 2011 23:31:23 +0000 (16:31 -0700)]
Release v1.91.92 and fix some build errors

13 years agoAdd some missing prototypes
Mike Gorse [Mon, 21 Mar 2011 22:40:50 +0000 (15:40 -0700)]
Add some missing prototypes

13 years agoPass the right parameter to _atspi_dbus_call() from atspi_value_set_current_value().
Mario Sanchez Prada [Mon, 21 Mar 2011 22:22:57 +0000 (23:22 +0100)]
Pass the right parameter to _atspi_dbus_call() from atspi_value_set_current_value().

Bug #35523

13 years agoBump version, remove dbus-glib dependency, and add pc file
Mike Gorse [Mon, 21 Mar 2011 18:44:46 +0000 (11:44 -0700)]
Bump version, remove dbus-glib dependency, and add pc file

Add in a modified version of dbus-gmain.c so that we can stop dependeing on
dbus-glib and include a currently unreviewed fix for FDO#35115.
Also add a pc file and bump version to 1.91.92.

13 years agoMove atspi_get_a11y_bus back into libatspi and publicly export
Mike Gorse [Mon, 21 Mar 2011 02:15:47 +0000 (19:15 -0700)]
Move atspi_get_a11y_bus back into libatspi and publicly export

13 years agoPort registryd session management code from dbus-glib to gdbus
Mike Gorse [Mon, 21 Mar 2011 01:51:24 +0000 (18:51 -0700)]
Port registryd session management code from dbus-glib to gdbus

13 years agoAdding as to list of languages
Amitakhya Phukan [Mon, 21 Mar 2011 07:10:29 +0000 (12:40 +0530)]
Adding as to list of languages

13 years agoUpdated Assamese translations
Amitakhya Phukan [Mon, 21 Mar 2011 07:08:24 +0000 (12:38 +0530)]
Updated Assamese translations

13 years agoUpdated Assamese translations
Amitakhya Phukan [Mon, 21 Mar 2011 07:07:08 +0000 (12:37 +0530)]
Updated Assamese translations

13 years agoUpdated Dutch translation by Wouter Bolsterlee
Wouter Bolsterlee [Sat, 19 Mar 2011 15:41:10 +0000 (16:41 +0100)]
Updated Dutch translation by Wouter Bolsterlee

13 years agoDo not free a key listener's bus name and path twice
Mike Gorse [Fri, 18 Mar 2011 22:30:49 +0000 (15:30 -0700)]
Do not free a key listener's bus name and path twice

13 years agobus: Rewrite a11y bus management, don't fall back to session bus
Colin Walters [Tue, 15 Mar 2011 21:00:35 +0000 (17:00 -0400)]
bus: Rewrite a11y bus management, don't fall back to session bus

First of all, there should *always* be an a11y bus; if I enable
toolkit-accessibility, and apps pick that up but then libat-spi falls
back to the session bus, that's broken.

Fix this by rewriting the a11y bus launcher in C, making it a persistent
session service, and giving it a DBus API (bus name: org.a11y.Bus).  It
will start the bus process as a child dynamically; however if
toolkit-accessibility is enabled, we start it on startup.  For more
information, see the new file bus/README.

Details:

* Create a .service file so the session bus autostarts us if
  necessary
* The .desktop file is changed to use --start-immediately
* Remove the kill_accessibility_bus() hack in registryd; instead
  we chain at-spi-bus-launcher to the session bus lifecycle.
* Change at-spi bus lookup to try both the new session bus API
  and the X11 root window property.
* Create libregistry-internals.la which encapsulates a11y bus
  lookup logic, de-duplicating from atspi/ and registryd/.

https://bugzilla.gnome.org/show_bug.cgi?id=644851

13 years agoAdded Serbian Translation
Мирослав Николић [Fri, 18 Mar 2011 15:56:10 +0000 (16:56 +0100)]
Added Serbian Translation

13 years agoUpdated Traditional Chinese translation(Hong Kong and Taiwan)
Chao-Hsiung Liao [Fri, 18 Mar 2011 07:07:28 +0000 (15:07 +0800)]
Updated Traditional Chinese translation(Hong Kong and Taiwan)

13 years agoUpdated Oriya Translation
Manoj Kumar Giri [Thu, 17 Mar 2011 10:00:12 +0000 (15:30 +0530)]
Updated Oriya Translation

13 years agoAdded Oriya (or) into the List
Manoj Kumar Giri [Thu, 17 Mar 2011 10:00:04 +0000 (15:30 +0530)]
Added Oriya (or) into the List

13 years agoFix atspi_table_get_column_header
Mike Gorse [Tue, 15 Mar 2011 19:31:53 +0000 (12:31 -0700)]
Fix atspi_table_get_column_header