Mike Gorse [Mon, 14 Mar 2016 22:21:48 +0000 (17:21 -0500)]
2.19.92
Samuel Thibault [Tue, 8 Mar 2016 18:15:08 +0000 (12:15 -0600)]
Make sure the runtime directory exists
It may happen that the home directory of the user does not contain the
runtime directory yet, when it's a freshly-created home for instance.
In that case, establishing the p2p socket in spi_atk_create_socket
will fail since it only tries to create a socket in that directory
without checking that the directory exists.
https://bugzilla.gnome.org/show_bug.cgi?id=763274
Samuel Thibault [Tue, 8 Mar 2016 18:00:33 +0000 (12:00 -0600)]
Don't drop the main context when p2p socket configuration fails
Otherwise we can re-enter the main loop when sending a key
notification, causing problems.
https://bugzilla.gnome.org/show_bug.cgi?id=763273
Mike Gorse [Tue, 1 Mar 2016 01:43:15 +0000 (19:43 -0600)]
2.19.91
Mike Gorse [Wed, 24 Feb 2016 22:18:11 +0000 (16:18 -0600)]
Don't send children of gtk menus to be cached
This is o(n^2) because of the way gtk's atk implementation implements
get_index_in_parent(). WOrking around it here for now.
https://bugzilla.gnome.org/show_bug.cgi?id=754048
Mike Gorse [Mon, 15 Feb 2016 22:30:56 +0000 (16:30 -0600)]
NEWS: Fix typo
Mike Gorse [Mon, 15 Feb 2016 22:20:13 +0000 (16:20 -0600)]
2.19.90
Mike Gorse [Mon, 7 Dec 2015 17:53:49 +0000 (11:53 -0600)]
Always accept connections as root if unable to find an owning user
When running as root, we look through /proc to try to identify our
parent and check that it matches the user making a connection. However,
this check does not always yield useful information and also will
always fail on anything that isn't Linux. Unsure what security
implications this change has, but then it isn't ideal to run UI as
the super-user in the first place, yet users sometimes need to do so
for various reasons, and a11y needs to be enabled, so adding this
change unless a better solution is found.
Mike Gorse [Tue, 24 Nov 2015 00:54:05 +0000 (18:54 -0600)]
2.19.2
Patryk Kaczmarek [Tue, 29 Sep 2015 13:24:49 +0000 (15:24 +0200)]
New tests for last interfaces
* Tested interfaces:
- Document
- Hyperlink
- Hypertext
- Selectoin
- State Set
- Table
- Table Cell
Patryk Kaczmarek [Tue, 29 Sep 2015 10:33:07 +0000 (12:33 +0200)]
Next tests for accessible interfaces
* New tested interfaces:
- Collection
- EditableText
- Image
- Text
- Value
* Commented tests for depricated function in Component tests deleted
* Unused variables deleted
* Added compilation flag -Wall
Patryk Kaczmarek [Mon, 28 Sep 2015 11:36:31 +0000 (13:36 +0200)]
Text, Value and Table Cell adaptors fixed
https://bugzilla.gnome.org/show_bug.cgi?id=755724
Patryk Kaczmarek [Mon, 21 Sep 2015 18:48:38 +0000 (20:48 +0200)]
Test fixes up to atk-adaptor version and cache changes
Mike Gorse [Sun, 27 Sep 2015 22:10:41 +0000 (17:10 -0500)]
2.18.0
Mike Gorse [Mon, 17 Aug 2015 19:31:41 +0000 (14:31 -0500)]
Release v2.17.90
Mike Gorse [Fri, 19 Jun 2015 01:05:48 +0000 (20:05 -0500)]
Modify for the new caching API
Send an object's index and child count in the cache data rather than a list of
children.
Also bumping the libatspi requirement, since an older libatspi will not
work with this version of at-spi2-atk.
https://bugzilla.gnome.org/show_bug.cgi?id=650090
Mike Gorse [Mon, 20 Jul 2015 23:52:59 +0000 (18:52 -0500)]
Release v2.17.1
Mike Gorse [Wed, 15 Jul 2015 16:16:20 +0000 (11:16 -0500)]
Updated AUTHORS/MAINTAINERS/doap
Mark Doffman and Li Yuan were listed as maintainers, but effectively
neither have been involved for several years.
Also added Li to AUTHORS.
https://bugzilla.gnome.org/show_bug.cgi?id=752231
Patrick Welche [Fri, 19 Jun 2015 09:56:56 +0000 (10:56 +0100)]
build fix: add tests/data/Makefile.am
https://bugzilla.gnome.org/show_bug.cgi?id=751192
Mike Gorse [Fri, 19 Jun 2015 01:32:24 +0000 (20:32 -0500)]
Add some missing files to the distribution
https://bugzilla.gnome.org/show_bug.cgi?id=751138
Alexandre Rostovtsev [Thu, 18 Jun 2015 04:17:04 +0000 (00:17 -0400)]
tests: fix out-of-source and parallel build
https://bugzilla.gnome.org/show_bug.cgi?id=751137
Christian Hergert [Fri, 15 May 2015 20:07:18 +0000 (13:07 -0700)]
atk-adapter: be careful with potential NULL GObjects
I've been seeing warnings in my Gtk application for a while due to this.
Mike Gorse [Mon, 23 Mar 2015 23:22:02 +0000 (18:22 -0500)]
2.16.0
Patryk Kaczmarek [Fri, 6 Mar 2015 09:58:16 +0000 (10:58 +0100)]
Component interface tests
Mike Gorse [Mon, 16 Feb 2015 21:23:44 +0000 (15:23 -0600)]
Release v2.15.90
Patryk Kaczmarek [Mon, 16 Feb 2015 14:15:42 +0000 (15:15 +0100)]
Action tests added and some test framework refactor
* Accessible Action Test addded
* XML loader unificated for different Atk Object types
* Atk Suite now can run single interface tests
* get_root_obj function changed
* Created tests utils for test unification
* Makefile.am files formating changed and unused parts deleted
* MyAtkObject attributes receiving fixed
Joanmarie Diggs [Fri, 23 Jan 2015 15:12:49 +0000 (10:12 -0500)]
Add new roles for fractions, roots, subscripts, and superscripts
Mike Gorse [Mon, 19 Jan 2015 22:57:28 +0000 (16:57 -0600)]
Release v2.15.4
Joanmarie Diggs [Thu, 15 Jan 2015 22:24:34 +0000 (17:24 -0500)]
Fix for bug 690008 - Add support for ATSPI_STATE_READ_ONLY/ATK_STATE_READ_ONLY
Joanmarie Diggs [Tue, 21 Oct 2014 18:03:16 +0000 (14:03 -0400)]
Add support for ATK_STATE_HAS_TOOLTIP
https://bugzilla.gnome.org/show_bug.cgi?id=738953
Benjamin Otte [Thu, 18 Dec 2014 20:45:36 +0000 (21:45 +0100)]
Don't parse non-existing arguments
The definition of the signal declares 0 arguments, so the listener can't
assume there are any.
Found by Michael Stahl via this valgrind snippet:
==4770== Conditional jump or move depends on uninitialised value(s)
==4770== at 0x3104C10812: text_selection_changed_event_listener
(event.c:1036)
==4770== by 0x30EF42160A: signal_emit_unlocked_R (gsignal.c:3519)
==4770== by 0x30EF42A180: g_signal_emit_valist (gsignal.c:3309)
==4770== by 0x30EF42A8F9: g_signal_emit_by_name (gsignal.c:3405)
==4770== by 0x1835EE96:
AtkListener::notifyEvent(com::sun::star::accessibility::AccessibleEventObject
const&) (atklistener.cxx:454)
==4770== by 0x56E46BA:
comphelper::AccessibleEventNotifier::addEvent(unsigned int,
https://bugzilla.gnome.org/show_bug.cgi?id=741734
Mike Gorse [Mon, 15 Dec 2014 23:43:00 +0000 (18:43 -0500)]
Release v2.15.3
Mike Gorse [Mon, 15 Dec 2014 23:42:08 +0000 (18:42 -0500)]
Always generate the tests Makefiles
We always want Makefiles in the tests directory, so that make distcheck
works.
Patryk Kaczmarek [Tue, 2 Dec 2014 11:18:41 +0000 (12:18 +0100)]
Unit test framework
* New test framework for testing Atspi library
* https://bugzilla.gnome.org/show_bug.cgi?id=730505
Joanmarie Diggs [Wed, 19 Nov 2014 13:05:50 +0000 (08:05 -0500)]
Add mapping between ATK_ROLE_STATIC and ATSPI_ROLE_STATIC
https://bugzilla.gnome.org/show_bug.cgi?id=740362
Alejandro Piñeiro [Mon, 6 Oct 2014 12:28:54 +0000 (14:28 +0200)]
Remove tests directory
Reasons:
* They don't belong to at-spi2-atk (no relation to atk at all)
* They are clearly outdated.
* They are not tests but examples
* Real tests are coming (bug 730505)
Mike Gorse [Mon, 13 Oct 2014 22:55:55 +0000 (17:55 -0500)]
Release v2.14.1
Alejandro Piñeiro [Wed, 24 Sep 2014 17:40:51 +0000 (19:40 +0200)]
Aggressively filtering AddChildren with STATE_MANAGES_DESCENDANTS
Mike Gorse [Mon, 22 Sep 2014 16:40:37 +0000 (11:40 -0500)]
Release as 2.14.0
Piotr Drąg [Thu, 31 Jul 2014 15:17:03 +0000 (17:17 +0200)]
doap: add <programming-language>
Olav Vitters [Wed, 30 Jul 2014 13:58:34 +0000 (15:58 +0200)]
doap category core
Mike Gorse [Sun, 20 Jul 2014 21:15:59 +0000 (16:15 -0500)]
Release v2.13.4
Mike Gorse [Sat, 19 Jul 2014 20:01:42 +0000 (15:01 -0500)]
Correctly remove the socket on exit
https://bugzilla.gnome.org/show_bug.cgi?id=684076
Jacobo Aragunde Pérez [Thu, 19 Jun 2014 15:27:51 +0000 (17:27 +0200)]
Do not overwrite text attributes with the default ones
As it was implemented, the default attributes are appended to the end
of the list, and later in time only the last occurence of every
attribute will be preserved.
Just changing the order of the list fixes this bug.
https://bugzilla.gnome.org/show_bug.cgi?id=731980
Samuel Thibault [Sat, 7 Jun 2014 20:39:56 +0000 (15:39 -0500)]
Fix typo in .pc file
Requires.Private should be Requires.private.
https://bugzilla.gnome.org/show_bug.cgi?id=721719
Mike Gorse [Mon, 28 Apr 2014 23:49:14 +0000 (18:49 -0500)]
2.13.1
Mike Gorse [Mon, 21 Apr 2014 23:08:19 +0000 (18:08 -0500)]
Only send cache-related events for transients when explicitly requested
Hopefully this will improve performance when, ie, gtk floods us with
children-changed events when populating a large treeview.
https://bugzilla.gnome.org/show_bug.cgi?id=728319
Mike Gorse [Mon, 14 Apr 2014 20:21:24 +0000 (15:21 -0500)]
Add an atexit handler to remove the D-Bus socket
We should try not to leave stale sockets and their directories when exiting.
They are removed when calling atk_bridge_adaptor_cleanup, but gtk does not
currently call this function when exiting, and there is not a good place
to call it from gtk.
https://bugzilla.gnome.org/show_bug.cgi?id=684076
Mike Gorse [Mon, 24 Mar 2014 21:26:51 +0000 (16:26 -0500)]
2.12.0
Alejandro Piñeiro [Thu, 6 Mar 2014 13:01:47 +0000 (14:01 +0100)]
Bump dbus version needed
This commit:
9eae1f7c2b818b5bd702f99ed97056ca76ea01f6
Started to use some methods not included until 1.5, so bumping the
version needed
Mike Gorse [Mon, 3 Mar 2014 23:56:23 +0000 (17:56 -0600)]
Release v2.11.91
Mike Gorse [Mon, 24 Feb 2014 23:09:52 +0000 (17:09 -0600)]
Allow EventListenerRegistered with old signature
Add_event_from_iter checks for the old form anyway, so might as well
allow it. Improves compatibility with at-spi2-registryd <= 3.10.
Mike Gorse [Thu, 20 Feb 2014 16:37:41 +0000 (10:37 -0600)]
Collection: Fix an infinite loop when encountering a NULL parent
Mike Gorse [Tue, 18 Feb 2014 18:40:08 +0000 (12:40 -0600)]
Release v2.11.90
Mike Gorse [Tue, 18 Feb 2014 18:07:33 +0000 (12:07 -0600)]
Add AtspitableCell, tracking the new atk interface
Mike Gorse [Tue, 4 Feb 2014 00:08:34 +0000 (18:08 -0600)]
Release v2.11.5
Mike Gorse [Fri, 20 Sep 2013 17:16:31 +0000 (12:16 -0500)]
Support sending data with events
Handle requests for additional data from event listeners (ie, registered with
atspi_event_listener_register_full), and allow sending interfaces and screen
extents in this way.
https://bugzilla.gnome.org/show_bug.cgi?id=708695
Mike Gorse [Mon, 16 Dec 2013 21:57:50 +0000 (16:57 -0500)]
Release v2.11.3
Alejandro Piñeiro [Thu, 28 Nov 2013 17:14:20 +0000 (18:14 +0100)]
document: add support for newly added AtkDocument methods and signals
It also bumps the atk dependency
https://bugzilla.gnome.org/show_bug.cgi?id=719508
Mike Gorse [Wed, 20 Nov 2013 20:45:30 +0000 (14:45 -0600)]
Fix GetBoundedRanges
Mike Gorse [Wed, 20 Nov 2013 20:45:23 +0000 (14:45 -0600)]
Post-release version bump
Mike Gorse [Tue, 19 Nov 2013 00:08:09 +0000 (18:08 -0600)]
Release v2.11.2
Alejandro Piñeiro [Mon, 18 Nov 2013 18:17:31 +0000 (19:17 +0100)]
states: add ATSPI_STATE_CHECKABLE and ATSPI_STATE_HAS_POPUP
https://bugzilla.gnome.org/show_bug.cgi?id=712609
Mike Gorse [Mon, 18 Nov 2013 20:49:43 +0000 (14:49 -0600)]
Bump AT-SPI version requirement
New roles are present in 2.11.2 and later.
Alejandro Piñeiro [Mon, 18 Nov 2013 18:14:35 +0000 (19:14 +0100)]
roles: create new AT-SPI2 roles for ARIA roles (and one for rating)
https://bugzilla.gnome.org/show_bug.cgi?id=712608
Alejandro Piñeiro [Tue, 29 Oct 2013 17:51:57 +0000 (18:51 +0100)]
Support for roles TITLE_BAR, BLOCK_QUOTE, AUDIO and VIDEO
https://bugzilla.gnome.org/show_bug.cgi?id=710464
Mike Gorse [Mon, 28 Oct 2013 22:43:36 +0000 (17:43 -0500)]
Release v2.11.1
Mike Gorse [Wed, 23 Oct 2013 20:43:58 +0000 (15:43 -0500)]
Fix atk_component_contains wrapper
The method call should be called Contains, not contains, per the D-Bus and
AT-SPI specs.
https://bugzilla.gnome.org/show_bug.cgi?id=710730
Mike Gorse [Mon, 23 Sep 2013 22:04:39 +0000 (17:04 -0500)]
Release as v2.10.0
Mike Gorse [Mon, 16 Sep 2013 21:39:45 +0000 (16:39 -0500)]
Release v2.9.92
Mike Gorse [Sun, 15 Sep 2013 16:59:03 +0000 (11:59 -0500)]
Allow a :-separated list of attribute values when reading match rules
https://bugzilla.gnome.org/show_bug.cgi?id=700865
Mike Gorse [Thu, 5 Sep 2013 03:29:15 +0000 (22:29 -0500)]
Properly clean up when timing out on a keystroke listener
If we time out waiting for a keystroke listener, then exit the main
loop, and reset the DBus connections' main loop contexts becak to the
default loop.
https://bugzilla.gnome.org/show_big.cgi?id=707218
Mike Gorse [Mon, 19 Aug 2013 21:45:47 +0000 (16:45 -0500)]
Release v2.9.90
Mario Sanchez Prada [Tue, 6 Aug 2013 11:06:36 +0000 (13:06 +0200)]
Implemented GetStringAtOffset in the text adaptor in the ATK bridge
Also, raised the version of the ATK dependency up to 2.9.4.
https://bugzilla.gnome.org/show_bug.cgi?id=705581
Mike Gorse [Tue, 9 Jul 2013 17:45:20 +0000 (12:45 -0500)]
Fix another crash when we're initialized/shut down repeatedly
Mike Gorse [Mon, 8 Jul 2013 21:37:26 +0000 (16:37 -0500)]
Release v2.9.4
Mike Gorse [Mon, 1 Jul 2013 19:35:04 +0000 (14:35 -0500)]
Fix deadlock with key event listeners in our own process
If a key event was sent out and there was also a listener in the same
process, then the listener would never see the event, since the main
loop was being run in a separate context, while libatspi would defer
handling of the event, adding an idle handler into the default context.
We should call atspi_set_main_context (new function) to let libatspi
know which context to use when adding an idle callback. This was
previously hacked around by setting ATSPI_CLIENT in pyatspi and avoiding
switching main loop contexts if it was set, but this was specific to
pyatspi (ie, did not take into account direct usage of libatspi or use
directly through introspection, such as in Javascript), so this check
was removed in 2.9.3.
Mike Gorse [Mon, 17 Jun 2013 21:43:11 +0000 (16:43 -0500)]
Release v2.9.3
Mike Gorse [Sun, 16 Jun 2013 19:04:01 +0000 (14:04 -0500)]
Use atspi_is_initialized to decide whether to expose p2p connection
Previously, direct D-Bus connections were not exposed if AT_SPI_CLIENT
was set in the environment, since they are causing deadlock if the
application and the AT are in the same process. However, AT_SPI_CLIENT
was only being set by pyatspi, meaning that libatspi consumers not using
pyatspi (Javascript embedded in gnome-shell that uses the libatspi
introspection bindings directly, for instance) would still have problems
when attempting to inspect their own accessibles.
Also removed the code in event.c to disable main loop context switching
when AT_SPI_CLIENT was set. Modifying it to also check
atspi_is_initialized would cause deadlock with gnome-shell when Orca is
running. Hopefully that code is not needed anymore.
Mike Gorse [Tue, 28 May 2013 16:06:10 +0000 (11:06 -0500)]
Release v2.9.2
Mike Gorse [Mon, 20 May 2013 21:25:46 +0000 (16:25 -0500)]
Don't warn if an app returns NULL when referencing a child
If atk_object_ref_accessible_child returns NULL when creating the cache,
then don't attempt to process/unref the NULL item, although applications
probably shouldn't be returning NULL in response to a call to
atk_object_ref_accessible_child, so it seems debatable whether we should
be printing a warning here.
Mike Gorse [Fri, 3 May 2013 14:48:50 +0000 (09:48 -0500)]
Add a NULL check into register_reply
Register_reply could access invalid memory if it is called after
atk_bridge_adaptor_cleanup() has been called.
This code should probably be refactored to make the instance act as a
GObject, but committing this for now to fix the crash.
https://bugzilla.gnome.org/show_bug.cgi?id=699554
Mike Gorse [Thu, 2 May 2013 20:24:22 +0000 (15:24 -0500)]
More DBusError leak fixes
Mike Gorse [Tue, 30 Apr 2013 19:12:05 +0000 (14:12 -0500)]
Ensure that DBus errors are freed
Call dbus_error_free in some cases where it is needed. Just pass NULL to
dbus_message_get_args where we aren't using the error.
https://bugzilla.gnome.org/show_bug.cgi?id=698951
Mike Gorse [Wed, 10 Apr 2013 17:40:47 +0000 (12:40 -0500)]
Add NULL check to fix crash when receiving a dbus reply after cleanup
Xan Lopez [Wed, 27 Mar 2013 20:40:12 +0000 (21:40 +0100)]
socket-adaptor: fix memory leak
Set a destructor for the object data. Other than fixing a leak this
also allows to simplify the code a bit.
https://bugzilla.gnome.org/show_bug.cgi?id=696733
Mike Gorse [Mon, 25 Mar 2013 22:23:36 +0000 (17:23 -0500)]
Release as v2.8.0
Mike Gorse [Mon, 4 Mar 2013 21:53:55 +0000 (15:53 -0600)]
Release v2.7.91
Mike Gorse [Tue, 26 Feb 2013 00:39:08 +0000 (18:39 -0600)]
Add Locale property, mapped to atk_object_get_object_locale
https://bugzilla.gnome.org/show_bug.cgi?id=694368
Mike Gorse [Mon, 18 Feb 2013 22:57:03 +0000 (16:57 -0600)]
Release v2.7.90
Dominique Leuenberger [Tue, 5 Feb 2013 09:05:16 +0000 (10:05 +0100)]
configure: bump atk dependency to 2.7.5 in order to enusre ATK_ROLE_LEVEL_BAR is available.
Fixes bug 693189.
Mike Gorse [Mon, 4 Feb 2013 22:35:28 +0000 (16:35 -0600)]
Release v2.7.5
Mike Gorse [Mon, 4 Feb 2013 19:34:48 +0000 (13:34 -0600)]
Update README
The README was very outdated.
Mike Gorse [Mon, 4 Feb 2013 18:36:16 +0000 (12:36 -0600)]
Fix various warnings
Mike Gorse [Thu, 31 Jan 2013 22:31:52 +0000 (16:31 -0600)]
Support ATK_ROLE_LEVEL_BAR
Mike Gorse [Fri, 4 Jan 2013 18:10:55 +0000 (12:10 -0600)]
Simplify autogen.sh
Simply calling autoreconf will do most of what is needed; we don't need
the full gnome-autogen.sh script, and this eliminates incompatibilities
with new versions of automake (ie, 1.13) that the GNOME script did not
check for.
Mike Gorse [Mon, 17 Dec 2012 22:12:54 +0000 (17:12 -0500)]
Release v2.7.3
Jasper St. Pierre [Sat, 8 Dec 2012 08:21:08 +0000 (03:21 -0500)]
droute: Fix another memory few memory leaks
We need to free the pointer array here, along with the path.
https://bugzilla.gnome.org/show_bug.cgi?id=689887
Mike Gorse [Mon, 26 Nov 2012 20:04:05 +0000 (14:04 -0600)]
Stop using deprecated glib functions
GStaticRecMutex is deprecated. Use GRecMutex instead. Also, this means
that we need at least glib 2.32, so specify that in configure.ac.
Mike Gorse [Tue, 20 Nov 2012 22:29:23 +0000 (16:29 -0600)]
Re-register if the registry goes away and later returns