Mike Gorse [Fri, 7 Dec 2012 16:27:58 +0000 (10:27 -0600)]
Fix another crash when removing hung processes
There was another instance where we were freeing the wrong data.
Mike Gorse [Thu, 6 Dec 2012 15:50:18 +0000 (09:50 -0600)]
Remove the correct datum when clearing a hung process
When clearing hung processes, we were passing the wrong variable to
g_list_remove, meaning that the datum would never be removed from the
list, even though it had been freed, resulting in a crash.
Mike Gorse [Wed, 5 Dec 2012 05:54:55 +0000 (23:54 -0600)]
Fix a crash when parsing events
If an event listener specifies a detail for an event, but an event
arrives with an empty detail, then libatspi might crash.
Rui Matos [Mon, 26 Nov 2012 21:45:25 +0000 (22:45 +0100)]
build: Use gobject-introspection's Makefile instead of rolling our own
Invoking the GI compiler through libtool would cause us to embed an
RPATH in our binaries.
https://bugzilla.gnome.org/show_bug.cgi?id=640303
Mike Gorse [Mon, 26 Nov 2012 19:51:28 +0000 (13:51 -0600)]
Fix atspi_device_listener_new_simple and some other warnings
Mike Gorse [Tue, 20 Nov 2012 22:29:50 +0000 (16:29 -0600)]
Re-register keystroke listeners if the registry goes away and returns
Mike Gorse [Mon, 19 Nov 2012 21:34:26 +0000 (15:34 -0600)]
Fix atspi_accessible_get_index_in_parent when the parent is not cached
Atspi_accessible_get_index_in_parent was returning -1 if the object's
parent is NULL, regardless of whether the accessible_parent field is
currently valid.
https://bugzilla.gnome.org/show_bug.cgi?id=688057
Mike Gorse [Mon, 19 Nov 2012 16:57:56 +0000 (10:57 -0600)]
Release v2.7.2
Mike Gorse [Thu, 15 Nov 2012 19:20:22 +0000 (13:20 -0600)]
[registry] Fix crash handling DeregisterKeystrokeListener
Colin Walters [Wed, 14 Nov 2012 04:47:56 +0000 (23:47 -0500)]
atspi: Readd dbus-1 to introspection generation
It is presently used in AtspiApplication.
https://bugzilla.gnome.org/show_bug.cgi?id=688223
Daiki Ueno [Tue, 13 Nov 2012 08:43:42 +0000 (17:43 +0900)]
Some introspection fixes to make vapigen happy
https://bugzilla.gnome.org/show_bug.cgi?id=688223
Daiki Ueno [Tue, 13 Nov 2012 09:12:36 +0000 (18:12 +0900)]
Fix gtk-doc for AtspiRegistry.
https://bugzilla.gnome.org/show_bug.cgi?id=688224
Mike Gorse [Mon, 12 Nov 2012 16:47:55 +0000 (10:47 -0600)]
Don't distribute dbind-config.h
Dbind-config.h is generated by configure and should not be part of the
release.
https://bugzilla.gnome.org/show_bug.cgi?id=688083
Mike Gorse [Fri, 26 Oct 2012 21:01:35 +0000 (16:01 -0500)]
Only call an event listener once per event
If an event listener is registered for an event more than once (ie,
registering both the event and its superset), then only notify it once.
Mike Gorse [Mon, 22 Oct 2012 16:21:32 +0000 (11:21 -0500)]
Release v2.7.1
Peter Mráz [Sat, 20 Oct 2012 15:35:25 +0000 (16:35 +0100)]
Updated Slovak translation
Mike Gorse [Wed, 17 Oct 2012 14:38:11 +0000 (09:38 -0500)]
Don't output a warning when a p2p connection isn't available
Mike Gorse [Fri, 5 Oct 2012 21:40:55 +0000 (17:40 -0400)]
Update cache in response to role change notifications
We update the cache in response to name, description, and parent
property-change signals, but we were missing code to handle signals for
role changes.
https://bugzilla.gnome.org/show_bug.cgi?id=685469
Mike Gorse [Mon, 24 Sep 2012 20:49:47 +0000 (15:49 -0500)]
Release as v2.6.0
Mike Gorse [Tue, 18 Sep 2012 03:06:39 +0000 (22:06 -0500)]
Release v2.5.92
William Jon McCann [Fri, 14 Sep 2012 16:50:18 +0000 (12:50 -0400)]
Don't leak dbus connections
By using a private connection and closing it.
https://bugzilla.gnome.org/show_bug.cgi?id=684033
Mike Gorse [Mon, 17 Sep 2012 16:16:08 +0000 (11:16 -0500)]
Don't check for NULL when calling _atspi_application_new
g_object_new will error out rather than return NULL when out of memory,
so no need to check its return value.
William Jon McCann [Fri, 14 Sep 2012 17:17:35 +0000 (13:17 -0400)]
Don't leak the display string
https://bugzilla.gnome.org/show_bug.cgi?id=684033
William Jon McCann [Fri, 14 Sep 2012 15:26:14 +0000 (11:26 -0400)]
Make build silent
So we can see warnings better.
https://bugzilla.gnome.org/show_bug.cgi?id=684033
William Jon McCann [Fri, 14 Sep 2012 15:18:53 +0000 (11:18 -0400)]
Don't leak errors
https://bugzilla.gnome.org/show_bug.cgi?id=684033
Mike Gorse [Mon, 10 Sep 2012 17:00:19 +0000 (12:00 -0500)]
Exit without a SIGABRT when unable to open X display
At-spi2-registryd was calling g_error when unable to open the default
display, which led to a SIGABRT and a core dump. This behavior is
inconsistent with what is generally done by GNOME apps, so let's call
g_warning() and exit(1) instead.
https://bugzilla.gnome.org/show_bug.cgi?id=660407
Javier Hernández [Tue, 4 Sep 2012 01:58:18 +0000 (03:58 +0200)]
Fixing 'atspi_accessible_get_id' annotation
Javier Hernández [Tue, 4 Sep 2012 01:58:34 +0000 (03:58 +0200)]
Fix some annotations compilation warnings
Mike Gorse [Mon, 3 Sep 2012 23:10:57 +0000 (18:10 -0500)]
Release v2.5.91
Piotr Drąg [Sat, 1 Sep 2012 18:47:11 +0000 (20:47 +0200)]
Updated Polish translation
Samuel Thibault [Fri, 24 Aug 2012 21:48:03 +0000 (16:48 -0500)]
Add glib-2.0 to requires in at-spi2.pc.in
https://bugzilla.gnome.org/show_bug.cgi?id=682595
Behdad Esfahbod [Tue, 21 Aug 2012 18:15:19 +0000 (14:15 -0400)]
Fix build with -fno-common
Mike Gorse [Wed, 15 Aug 2012 13:41:39 +0000 (08:41 -0500)]
Add license to atspi-constants.h
Mike Gorse [Mon, 6 Aug 2012 20:53:38 +0000 (15:53 -0500)]
Release v2.5.5
Daniel Narvaez [Thu, 19 Jul 2012 20:24:27 +0000 (22:24 +0200)]
Allow to build out of source directory
https://bugzilla.gnome.org/show_bug.cgi?id=680281
Mike Gorse [Mon, 30 Jul 2012 17:02:05 +0000 (12:02 -0500)]
Add atspi_action_get_localized_name
https://bugzilla.gnome.org/show_bug.cgi?id=680598
Mike Gorse [Wed, 18 Jul 2012 17:29:24 +0000 (12:29 -0500)]
Code clean-ups
There isn't really any point in checking for GLib functions returning NULL
due to lack of memory, since they just abort instead.
Mike Gorse [Mon, 16 Jul 2012 20:52:05 +0000 (15:52 -0500)]
Release v2.5.4
Mike Gorse [Mon, 9 Jul 2012 00:35:36 +0000 (19:35 -0500)]
Fix abort on an error while removing a match rule
Patrick Welche [Fri, 29 Jun 2012 17:07:23 +0000 (18:07 +0100)]
Consistently return a value - from joerg@netbsd.org
Mike Gorse [Mon, 25 Jun 2012 22:55:05 +0000 (17:55 -0500)]
Release v2.5.3
Jasper St. Pierre [Sun, 24 Jun 2012 20:32:31 +0000 (16:32 -0400)]
atspi-misc: Prevent leak
We need to free the data that XGetWindowProperty returns back to us.
https://bugzilla.gnome.org/show_bug.cgi?id=678738
Colin Walters [Thu, 14 Jun 2012 20:38:27 +0000 (16:38 -0400)]
autogen.sh: Honor NOCONFIGURE=1
See http://people.gnome.org/~walters/docs/build-api.txt
Mike Gorse [Mon, 4 Jun 2012 21:55:07 +0000 (16:55 -0500)]
Release v2.52
Marcus Habermehl [Mon, 4 Jun 2012 21:31:46 +0000 (16:31 -0500)]
Fix crash in atspi_deregister_keystroke_listener
Atspi_registry_keystroke_listener has code to copy a keyset before passing
to dbind, in order to handle NULL keystrings, but this code was not present
in atspi_deregister_keystroke_listener.
https://bugs.freedesktop.org/show_bug.cgi?id=50578
Mike Gorse [Wed, 23 May 2012 19:19:26 +0000 (14:19 -0500)]
_ATSPI_DBUS_CHECK_SIG: remove warning when no message received
Mike Gorse [Mon, 21 May 2012 22:40:11 +0000 (17:40 -0500)]
Add G_BEGIN_DECLS and G_END_DECLS macros for prototypes to fix c++ linkage
Mike Gorse [Mon, 7 May 2012 16:28:05 +0000 (11:28 -0500)]
Add a check for a null message when returning a hyperlink
Mike Gorse [Mon, 30 Apr 2012 21:22:41 +0000 (16:22 -0500)]
Release v2.5.1
Joanmarie Diggs [Sat, 28 Apr 2012 00:32:41 +0000 (20:32 -0400)]
Fix for bug 675004 - object:state-changed:defunct events are emitted as object:state-change:defunct
O. Atsushi [Mon, 23 Apr 2012 21:28:03 +0000 (16:28 -0500)]
Fix type in atspi_document_get_attribute_value
atspi_document_get_attribute_value was failing because of a typo.
https://bugzilla.gnome.org/show_bug.cgi?id=674515
Mike Gorse [Thu, 12 Apr 2012 16:45:58 +0000 (11:45 -0500)]
Increase start-up timeout to 15 seconds
It seems that a five-second time-out is often not sufficient.
Mike Gorse [Tue, 10 Apr 2012 20:09:32 +0000 (15:09 -0500)]
Fix GError sent when receiving a reply of the wrong type
When a method reply of the wrong type was received, the parameters were
still sent backwards when formatting the GError.
Mike Gorse [Tue, 10 Apr 2012 15:22:51 +0000 (10:22 -0500)]
Set the application start-up timeout to 5 seconds by default
The application timeout is now implemented in 2.4 but is not settable.
Instead, there is a default of 5 seconds, so the default should be
consistent in 2.5.
Also, document the default values.
Mike Gorse [Fri, 6 Apr 2012 21:06:04 +0000 (16:06 -0500)]
Add startup_time parameter to atspi_set_timeout
Add a parameter to allow timeouts not to be enforced until an
application has been running for a given amount of time. Might help
prevent timeout exceptions for applications that get bogged down
performing tasks at initialization.
See https://bugzilla.gnome.org/show_bug.cgi?id=672784
Mike Gorse [Fri, 6 Apr 2012 15:59:53 +0000 (10:59 -0500)]
Add atspi_set_timeout
Mike Gorse [Mon, 26 Mar 2012 20:54:13 +0000 (15:54 -0500)]
Release v2.4.0
Rajesh Ranjan [Mon, 26 Mar 2012 11:12:25 +0000 (16:42 +0530)]
hindi translation by Chandan Kumar
Daniel Mustieles [Sun, 25 Mar 2012 17:58:32 +0000 (19:58 +0200)]
Fixed typo. Closes bug #605662
Mike Gorse [Mon, 19 Mar 2012 21:41:26 +0000 (16:41 -0500)]
Release v2.3.92
Mike Gorse [Sat, 17 Mar 2012 17:12:37 +0000 (12:12 -0500)]
Fix crash fetching error when retrieving a property
Mike Gorse [Fri, 16 Mar 2012 21:10:29 +0000 (16:10 -0500)]
Increase the keystroke event timeout
It seems that it is currently possible for at-spi2-atk to get bogged
down when sending a keystroke notification, so 1 second is a short time
to wait for a reply. It may be possible to mitigate this in the future,
in which case we could reduce the timeout again, but not going to do
this for 3.4. See comment on
https://bugzilla.gnome.org/show_bug.cgi?id=671976.
Mike Gorse [Wed, 14 Mar 2012 20:55:46 +0000 (15:55 -0500)]
[registry] Send a ping to a process when marking as hung
Sometimes keys that should be captured by event listeners are instead
passed through to the application. It seems that at-spi2-registryd might
not always be removing the process from the list of hung processes. I
have not reproduced this recently, and I'm not sure where the root
problem is. Hoping that it will help to send the process a ping and use
the reply as a queue that it is un-hung (previously we were relying on
receiving a reply to the original key event notification).
https://bugzilla.gnome.org/show_bug.cgi?id=671976
Praveen Arimbrathodiyil [Mon, 12 Mar 2012 16:11:58 +0000 (21:41 +0530)]
Malayalam translation by Anish
Timo Jyrinki [Mon, 12 Mar 2012 08:44:10 +0000 (10:44 +0200)]
Finnish translation update from l10n.laxstrom.name/wiki/Gnome_3.4 translation sprint
Khoem Sokhem [Sat, 10 Mar 2012 12:51:31 +0000 (13:51 +0100)]
Add initial Khmer translation
Mike Gorse [Tue, 6 Mar 2012 00:35:36 +0000 (18:35 -0600)]
Fix a couple of NEWS typos
Mike Gorse [Tue, 6 Mar 2012 00:16:01 +0000 (18:16 -0600)]
Release v2.3.91
Mike Gorse [Mon, 5 Mar 2012 23:32:32 +0000 (17:32 -0600)]
Enable caching on a call to atspi_accessible_set_cache_mask
Previously, caching was always disabled unless atspi_event_main() was
called to start a main loop, since we need to handle the case of a
simple application/script that performs an action and then sleeps rather
than entering a main loop. In such a case, libatspi will not be able to
detect that an event has been fired to update the cache, so caching
should be disabled. However, an application may want to start a main
loop through some other mechanism than by calling atspi-event_main(),
and caching should be done in such a case. So we'll assume that, if
atspi_accessible_set_cache_mask is called, then the application expects
caching to work and will run a main loop one way or another. Note that
it is possible for a script to run a main loop in some instances but not
others, and, in such a case, if it is not using atspi_event_main(), then
it should set caching to ATSPI_CACHE_NONE when not in a main loop and
back to ATSPI_CACHE_DEFAULT when running a main loop.
Priit Laes [Thu, 1 Mar 2012 17:40:51 +0000 (19:40 +0200)]
[l10n] Added Estonian (et) to LINGUAS
Mattias Põldaru [Thu, 1 Mar 2012 17:38:43 +0000 (19:38 +0200)]
[l10n] Updated Estonian translation
Mike Gorse [Tue, 28 Feb 2012 20:26:57 +0000 (14:26 -0600)]
Don't cache data for transient objects
If an object is known to be transient, then ignore its cache, since it
seems better not to rely on a toolkit to send updates for transient
objects.
Note that this isn't really a complete solution, since it only disables
the cache if STATE_TRANSIENT is known to be present. If AddAccessible
was not sent for the object and atspi_accessible_get_state_set has not
been called, then the state set is not known, so STATE_TRANSIENT is not
"known to be present..." It may be a good idea to fetch the state set if
it is not known, or check the parent for STATE_MANAGES_DESCENDANTS,
although in the latter case it is also possible that the parent's state
set is not known, and it may be needed to check the parent's parent and
so on.
Bahodir Mansurov [Sun, 26 Feb 2012 05:41:50 +0000 (00:41 -0500)]
Adding Uzbek@cyrillic translation.
Mike Gorse [Fri, 24 Feb 2012 22:25:41 +0000 (16:25 -0600)]
Reduce the method call timeout slightly
It is likely a good idea for the method call timeout to be lower than
the registry daemon's keystroke listener timeout. This way, if an AT
makes a method call that times out while in a keystroke event listener,
it will still have time to exit the listener before the registry daemon
presumes it hung.
Mike Gorse [Fri, 24 Feb 2012 22:22:50 +0000 (16:22 -0600)]
Send key events to listeners assumed to be hung, but don't block
Orca would sometimes fail to notice that its modifier key was released
since it was presumed to be hung and would not be sent the event. So,
rather than ignore event listeners assumed to be hung, we'll send them
the notifications for the keys that they're listening to but not wait
for a reply, instead assuming that they are not consumed.
Mike Gorse [Thu, 23 Feb 2012 22:51:31 +0000 (16:51 -0600)]
Expect a signed int for GetIndexInParent, per the spec
Mike Gorse [Thu, 23 Feb 2012 20:51:51 +0000 (14:51 -0600)]
Fix atspi_table_get_row_column_extents_at_index
atspi_table_get_row_column_extents_at_index was trying to read
parameters from the dbus message in the wrong order and also not
returning the row.
Mike Gorse [Thu, 23 Feb 2012 20:49:16 +0000 (14:49 -0600)]
Fix message displayed when a call returns the wrong type
The message displayed by dbind for a type mismatch had its parameters
inverted.
Mike Gorse [Tue, 21 Feb 2012 16:58:02 +0000 (10:58 -0600)]
Invalidate states when receiving a focus event
Sometimes we get "focus" events from gtk, but the cache still does not
indicate that the object has STATE_FOCUSED. Really should get to the
bottom of this but don't have time to investigate more right now. Might
be some kind of issue with the accessible being added to the cache
before it is fully initialized--perhaps AddAccessible really doesn't
make sense.
https://bugzilla.gnome.org/show_bug.cgi?id=663992
Mike Gorse [Mon, 20 Feb 2012 23:00:48 +0000 (17:00 -0600)]
Release v2.3.90
Mike Gorse [Sun, 19 Feb 2012 22:23:42 +0000 (16:23 -0600)]
Don't g_object_unref a NULL object for event variant
If we get an event with what appears to be an accessible as the variant
but we do not succeed in marshalling the accessible (possibly because it
is the null path), do not unref the null object. Stops a glib warning.
Mike Gorse [Mon, 13 Feb 2012 21:22:23 +0000 (15:22 -0600)]
If a keystroke listener does not respond, disable until it responds
If an AT is hung or is suspended while having a keystroke listener
registered, then this should not lock up the desktop, as would happen if
we wait indefinitely for the keystroke listener to respond.
Mike Gorse [Mon, 13 Feb 2012 17:32:47 +0000 (11:32 -0600)]
At-spi-bus-launcher signal emission fixes
Emit PropertiesChanged from the correct interface. Also, create
GVariantBuilders on the stack; fixes memory leaks.
Mike Gorse [Mon, 6 Feb 2012 22:17:05 +0000 (16:17 -0600)]
Release v2.3.5
Mike Gorse [Mon, 6 Feb 2012 20:50:46 +0000 (14:50 -0600)]
For IsScreenReaderEnabled, listen for the correct gsettings change signal
Mike Gorse [Mon, 6 Feb 2012 15:55:12 +0000 (09:55 -0600)]
Fix atspi_collection_get_matches_to prototype
Mike Gorse [Fri, 3 Feb 2012 23:21:33 +0000 (17:21 -0600)]
For atspi_collection_get_matches_to, rename restrict to limit_scope
Restrict can be a keyword. Don't use it as a variable name.
Also update D-Bus introspection.
https://bugzilla.gnome.org/show_bug.cgi?id=669344
Kjartan Maraas [Tue, 24 Jan 2012 13:57:07 +0000 (14:57 +0100)]
Updated Norwegian bokmål translation
Mike Gorse [Sun, 22 Jan 2012 18:30:48 +0000 (12:30 -0600)]
Add ScreenReaderEnabled d-bus property and proxy for GNOME setting
Add a ScreenReaderEnabled property to org.a11y.Bus, and proxy the
org.gnome.desktop.a11y.applications screen-reader-enabled gsettings key.
This will allow Orca to monitor the key without depending on gsettings
and a GNOME-specific key.
Mike Gorse [Sat, 21 Jan 2012 18:26:51 +0000 (12:26 -0600)]
Call g_settings_sync after updating IsEnabled
Ensure that the GSettings key is updated immediately upon receiving a
D-Bus message to set the property.
Mike Gorse [Fri, 20 Jan 2012 18:55:40 +0000 (12:55 -0600)]
Have dbind_method_call_reentrant_va check the signature of replies
When making a method call, we should check that the signature of the
reply matches what we expect and throw an error if not. Otherwise we may
crash.
José Millán Soto [Fri, 20 Jan 2012 12:13:38 +0000 (13:13 +0100)]
Set correct end offset in atspi_text_get_attributes
Fixes bug 668334
Mike Gorse [Mon, 16 Jan 2012 03:54:21 +0000 (21:54 -0600)]
Release v2.3.4
Mike Gorse [Sun, 15 Jan 2012 18:29:56 +0000 (12:29 -0600)]
Avoid making calls on processes thought to be hung
If a method call times out, record the process as being hung, and send a
ping. Consider the process to be no longer hung when a response to the
ping is received.
Hopefully this will prevent long delays if an AT makes consecutive calls
on an application that is not responding.
José Millán Soto [Wed, 4 Jan 2012 12:48:48 +0000 (13:48 +0100)]
Using value instead of memory address when calling _atspi_dbus_call
The following functions are changed: atspi_selection_select_child,
atspi_selection_deselect_child and atspi_selection_is_child_selected
Fixes bug 667254
Mike Gorse [Thu, 5 Jan 2012 20:36:18 +0000 (14:36 -0600)]
Allow applications to query keyboard and mouse listeners
Add methods to list the currently-active keyboard and mouse listeners
and signals to be fired when listeners are registered or deregistered.
Part of the fix for http://bugzilla.gnome.org/show_bug.cgi?id=666870
Mike Gorse [Thu, 29 Dec 2011 14:23:10 +0000 (08:23 -0600)]
Warn if a device listener returns a non-bool, rather than crash
Mike Gorse [Wed, 28 Dec 2011 03:23:41 +0000 (22:23 -0500)]
Revert part of last commit
Modifiers and hw_code are still sent as int16 for now, so it was wrong
to remove the code to parse them as int16, at least for the time being.
Mike Gorse [Tue, 27 Dec 2011 23:05:16 +0000 (18:05 -0500)]
Add some type checks for dbus messages, and remove some old code
Added a couple of message signature checks where it was still TODO (and
might have led to a crash if the wrong signature was sent). Also removed
some cases where two different signature types were accepted for API
compatibility reasons--the old API is no longer used.
Hideki Yamane [Mon, 26 Dec 2011 00:46:34 +0000 (09:46 +0900)]
Add new Japanese translation