summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Release v0.4.1at-spi2-atk-0-4Mike Gorse2010-11-151-1/+1
|
* Remove g_atexit handler.Mike Gorse2010-11-151-2/+0
| | | | | | | BGO#563546: at-spi should not use atexit. Dbus sends a NameOwnerChanged when an application goes away, so at-spi2-registryd should be able to do any necessary handling from that.
* Fire all events until we receive a reply from GetRegisteredEventsMike Gorse2010-11-153-0/+5
|
* Release as v0.4.0AT_SPI2_ATK_0_4_0Mike Gorse2010-09-271-1/+1
|
* Release v0.3.92AT_SPI2_ATK_0_3_92Mike Gorse2010-09-132-1/+5
|
* Have value methods return 0 on failure rather than fail, as in original pyatspiMike Gorse2010-09-011-18/+6
|
* Release v0.3.91.1AT_SPI2_ATK_0_3_91_1Mike Gorse2010-08-312-1/+5
|
* Fix the buildMike Gorse2010-08-311-0/+1
|
* Release v0.3.91Mike Gorse2010-08-302-1/+7
|
* Make the module resident.MatthiAS Clasen2010-08-301-0/+9
| | | | FDO#29880; fix crash when reloading modules.
* Default to --enable-relocate for nowMike Gorse2010-08-301-1/+1
|
* Fix the buildKjartan Maraas2010-08-201-5/+5
|
* Release v0.3.90AT_SPI2_ATK_0_3_90Mike Gorse2010-08-162-1/+10
|
* Only send events when something is listening for themMike Gorse2010-08-164-13/+265
|
* Ensure that an object is registered before adding it to the cacheMike Gorse2010-08-161-0/+5
| | | | | | | We need to ensure that an object is registered before putting it in the cache, since registering the object adds the weak_ref that we use to know when the object should be removed from the cache. Prevents a crash in GetItems.
* Fix droute_add_one / caching (0.3.6 regression)Mike Gorse2010-08-151-1/+1
|
* Have adaptor_init() return -1 on error and not consider itself initializedMike Gorse2010-08-051-3/+9
|
* Change remaining g_error calls to g_warningMike Gorse2010-08-031-4/+14
|
* Define -DG_DISABLE_DEPRECATED, and remove deprecated glib symbolMike Gorse2010-08-024-5/+9
| | | | FDO#29365
* Release v0.3.6AT_SPI2_ATK_0_3_6Mike Gorse2010-08-022-1/+7
|
* Make initial Embed call asynchronousMike Gorse2010-07-301-24/+46
| | | | | | | | The initial Embed does not need to be synchronous, so set a reply callback instead. This prevents a long delay if the registry daemon does not respond. Also unref some related messages that we were leaking. Also fix a warning.
* Improve introspectionMike Gorse2010-07-203-5/+42
|
* Release v0.3.5AT_SPI2_ATK_0_3_5Mike Gorse2010-07-122-1/+8
|
* Have a socket retrieve its state set from its embedded plugMike Gorse2010-07-121-0/+63
|
* AT_SPI_CLIENT -> AT_SPI_REENTER_G_MAIN_LOOPMike Gorse2010-07-091-1/+2
| | | | | | | Rename AT_SPI_CLIENT since we may actually need to reenter the main loop for more than just pyatspi (ie, UiaAtkBridge, since it sends key notifications on a thread separate from the GLib main loop thread) Continue to support AT_SPI_CLIENT for now, for backwards compatibility.
* Release v0.3.4AT_SPI2_ATK_0_3_4Mike Gorse2010-06-282-1/+22
|
* Set /desktop/gnome/interface/at-spi-dbus to false by defaultMike Gorse2010-06-281-1/+1
| | | | | | The at-spi-dbus key is only used if --enable-relocate is specified (which it is not by default). Specifying --enable-relocate would imply that at-spi-corba was still intended to be the default, so setting this key to false.
* Fix a warning and an incorrect return valueMike Gorse2010-06-233-3/+3
|
* Removed the extra boolean parameter added to know when to unref theMario Sanchez Prada2010-06-238-17/+27
| | | | | | | | | AtkObject passed as parameter. Instead, just do g_object_unref() when needed out of that function, right after calling it. Updated all callers Bug #28659
* Added a new boolean parameter to allow specifying when a call toMario Sanchez Prada2010-06-228-12/+13
| | | | | | | | | | g_object_unref should be done by the end of that function, since there are many functions in the ATK API that do not return an extra reference, and therefore do not need this extra unref. Updated all callers accordingly. Bug #622309
* Remove extra call from impl_GetColumnHeaderMike Gorse2010-06-211-1/+0
|
* Have ChildCount and GetChildAtIndex take sockets into accountMike Gorse2010-06-181-3/+30
|
* Fix NSelectedRows/Columns in introspectionMike Gorse2010-06-181-3/+3
|
* Fix NSelectedRows and NSelectedColumnsMike Gorse2010-06-171-2/+2
|
* Add GetChildren to Accessible introspectionMike Gorse2010-06-161-1/+4
|
* ref an object while adding it to the cacheMike Gorse2010-06-161-0/+3
|
* Add const to some dbind prototypes to fix warningsMike Gorse2010-06-083-28/+29
|
* Release v0.3.3AT_SPI2_ATK_0_3_3Mike Gorse2010-06-042-1/+12
|
* Lease objects that send events if they are not cachedMike Gorse2010-06-033-4/+10
| | | | | | | When sending an event, lease it if it is not in the cache, except for children-changed events, which should update the cache. This helps prevent errors with Firefox, which can send property-changed events on a progress bar that was never otherwise exposed.
* Treat children-changed::add:system as equivalent to children-changed::addMike Gorse2010-06-031-1/+2
|
* Use Makefile.include for gtk-3.0 MakefileMike Gorse2010-06-011-33/+1
|
* Install to gtk-3.0 as well as gtk-2.0Mike Gorse2010-05-285-43/+87
|
* Remove unneeded gtk check from configure.acMike Gorse2010-05-281-4/+0
|
* Update news and bump versionAT_SPI2_ATK_0_3_2Mike Gorse2010-05-242-1/+11
|
* Fix cache object path and interface namesMike Gorse2010-05-232-2/+2
|
* Move cache updates to an idleMike Gorse2010-05-182-13/+37
| | | | | | | | When receiving a children-changed signal, we add a whole subtree to the cache, but not all of the objects have necessarily been initialized yet, so we may, eg, indicate in the cache that an object's role is UNKNOWN (see BNC#600024). This also fixes a crash when starting Firefox 3.6 with at-spi2.
* Hyperlink fixesMike Gorse2010-05-173-6/+65
|
* Add NEWS, disable relocate, and bump version to 0.3.1AT_SPI2_ATK_0_3_1Mike Gorse2010-05-032-3/+29
|
* Fix handling of MATCH_ANY for empty statesets in a collection match ruleMike Gorse2010-05-031-3/+3
|
* Fix CLEANFILES to remove schemaMike Gorse2010-05-031-1/+1
|