After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 738463 - [a11y] Fix various problems with fetching children
[a11y] Fix various problems with fetching children
Status: RESOLVED FIXED
Product: evolution
Classification: Applications
Component: Calendar
3.14.x (obsolete)
Other Linux
: Normal major
: ---
Assigned To: evolution-calendar-maintainers
Evolution QA team
Depends on:
Blocks:
 
 
Reported: 2014-10-13 13:55 UTC by Vadim Rutkovsky
Modified: 2014-10-20 13:31 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Reproducer (179 bytes, text/x-python)
2014-10-13 13:55 UTC, Vadim Rutkovsky
  Details
proposed evo patch (4.40 KB, patch)
2014-10-13 18:31 UTC, Milan Crha
none Details | Review
Alternate patch. (4.19 KB, patch)
2014-10-14 01:29 UTC, Mike Gorse
committed Details | Review

Description Vadim Rutkovsky 2014-10-13 13:55:51 UTC
Created attachment 288389 [details]
Reproducer

See attached reproducer - it finds all the searchbars in the app. When calendar/tasks/memo view is opened the scripts hangs up with lots of warnings:
 (evolution:27576): GLib-GObject-CRITICAL **: g_object_unref: assertion 'G_IS_OBJECT (object)' failed

(evolution:27576): evolution-util-CRITICAL **: eti_get_accessible: assertion 'eti' failed

This blocks tests from checking calendar/tasks/memo views
Comment 1 Milan Crha 2014-10-13 17:13:48 UTC
This seems to be a regression after bu #722938, but not completely, because I noticed some parts which were different even before. In any case, I asked Mike to join this bug.

Vadim, could you state the version of gtk3 you have installed, please?
Comment 2 Milan Crha 2014-10-13 18:31:35 UTC
Created attachment 288413 [details] [review]
proposed evo patch

for evolution;

This is a patch I used to get rid of the runtime warnings from the run of the script.py. It reverts some parts of the patch from bug #722938, but I do not know how much correct it is.

I took current git master of evolution (at commit 4f02ea569ba8328a83c1108ba6) and applied on top of it this proposed patch and built evolution for Fedora's rawhide at [1] for easier testing to Vadim. Please give it a try.

[1] http://koji.fedoraproject.org/koji/taskinfo?taskID=7852959
Comment 3 Vadim Rutkovsky 2014-10-13 19:07:15 UTC
Yes, that build did fix the hangup for me, thanks!
Comment 4 Mike Gorse 2014-10-14 01:29:16 UTC
Created attachment 288428 [details] [review]
Alternate patch.

I think that your patch disables some of the a11y classes. Anyway, I've modified your patch to keep the calls to gtk_widget_class-set_accessible_type and fix some cases in gal-a11y-e-table.c where et_get_n_children and et_ref_child didn't behave consistently. I no longer see the hang or warnings with this applied, but it would be good if someone else wanted to test it.
Comment 5 Milan Crha 2014-10-14 09:38:15 UTC
Thanks Mike, your patch doesn't claim any errors on console on my machine too.

Vadim, here is a test build with Mike's patch:
http://koji.fedoraproject.org/koji/taskinfo?taskID=7859001
Comment 6 Milan Crha 2014-10-20 13:31:20 UTC
I've got a confirmation from Vadim about Mike's patch working properly, thus I'm committing this to sources:

Created commit 7bcec80 in evo master (3.13.7+) [1]
Created commit 710943b in evo evolution-3-12 (3.12.8+)

[1] https://git.gnome.org/browse/evolution/commit/?id=7bcec80