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 622048 - Orca locks up when accessing Pidgin
Orca locks up when accessing Pidgin
Status: RESOLVED NOTGNOME
Product: orca
Classification: Applications
Component: general
2.31.x
Other All
: Normal normal
: ---
Assigned To: Orca Maintainers
Orca Maintainers
Depends on:
Blocks:
 
 
Reported: 2010-06-19 03:00 UTC by Nolan Darilek
Modified: 2010-07-05 16:41 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
debug.out (94.18 KB, application/octet-stream)
2010-06-19 03:00 UTC, Nolan Darilek
Details
Newer debug.out (966.14 KB, application/octet-stream)
2010-06-19 04:10 UTC, Nolan Darilek
Details

Description Nolan Darilek 2010-06-19 03:00:35 UTC
Created attachment 164055 [details]
debug.out

I applied a bunch of OS updates today, and now Orca seems to lock up when I load Pidgin.

Orca version is Git master, though I've also tried with 2.30.

I've tried with both Pidgins 2.6.6 and 2.7.1.

Sometimes Orca locks up completely. If I kill Pidgin from another terminal, it seems like a bunch of events are being queued and processed at once. This seems pretty similar to some of the old desktop hanging bugs.

I don't see this behavior with any other apps as of yet, making me wonder if perhaps some of the new Pidgin bugfixing/refactoring may be at fault here.

When I look at the attached debug.out, I see at least one Pidgin-related traceback which increases this likelihood. :) There are other tracebacks as well.

I've also removed and reinstalled my local build of Orca, but this didn't change anything either. Help? :) I miss my Pidgin.
Comment 1 Joanmarie Diggs (IRC: joanie) 2010-06-19 03:12:05 UTC
Nolan, looking at your traceback, I see an error related to script_settings in a nolan directory. In particular:

Traceback (most recent call last):
  • File "/usr/local/stow/orca/lib/python2.6/site-packages/orca/focus_tracking_presenter.py", line 709 in _processObjectEvent
    self.loadAppSettings(orca_state.activeScript)
  • File "/usr/local/stow/orca/lib/python2.6/site-packages/orca/focus_tracking_presenter.py", line 414 in loadAppSettings
    __import__(name, globals(), locals(), [''])
  • File "/home/nolan/.orca/app-settings/pidgin.py", line 26 in <module>
    orca.scripts.apps.pidgin.script_settings.prefixChatMessage = True
AttributeError: 'module' object has no attribute 'script_settings'

So, yes, you're right about your theory. :-) This is why, in my email to the Orca list [1], I had advised that two things be done:

[snip]

Please note:

1. It probably wouldn't hurt to do a 'make uninstall' before building
Orca to remove the old script_settings.py file from the pidgin module.

2. If you have custom Pidgin settings in your $HOME/.orca/app-settings,
you should remove/reconfigure them.

[snip]

You're suffering from problem number 2. Please remove $HOME/.orca/app-settings and try again.

[1] http://mail.gnome.org/archives/orca-list/2010-June/msg00294.html
Comment 2 Joanmarie Diggs (IRC: joanie) 2010-06-19 03:27:20 UTC
Nolan, one more thing: The other tracebacks are indicative of whatever application you were accessing crashing. Any idea what app the string 'prosody' might be associated with?
Comment 3 Nolan Darilek 2010-06-19 04:10:49 UTC
Created attachment 164058 [details]
Newer debug.out

Apologies for missing the note about removing my old preferences. That did in fact resolve the first traceback, but the second still appears.

"Prosody" refers to a chatroom I have Pidgin configured to enter on startup. As such, the event you see corresponds with the main chat window gaining focus.

Note that, though I thought I kept this short, the original debug.out was just over 4 megs. I trimmed it down to a single meg, which is how far I had to search in for the first traceback. If you need any more than that, Gods help us, then let me know. :)
Comment 4 Joanmarie Diggs (IRC: joanie) 2010-06-19 04:30:03 UTC
Thanks.

Tracebacks like this one are, again, indicative of the application having crashed:

Traceback (most recent call last):
  • File "/usr/local/stow/orca/lib/python2.6/site-packages/orca/focus_tracking_presenter.py", line 618 in _processObjectEvent
    state = event.source.getState()
  • File "/usr/lib/python2.6/dist-packages/pyatspi/accessible.py", line 238 in _inner
    raise LookupError(e) LookupError

There used to be an Atk bug which, when combined with a Pidgin text attribute, caused Pidgin to crash if Orca (or Accerciser) was used to obtain the default text attributes. If you have enabled the 'underlining' in braille of text attributes, you could be hitting this bug. Any chances that would be the case?
Comment 5 Nolan Darilek 2010-06-19 04:42:58 UTC
I just ran Pidgin from a console. It seemed to run perfectly fine, except for the whole me not being able to use it bit. :) I didn't get any errors or assertion failures printed to the console, and Pidgin only ended when I sent it a SIGKILL.

I did have the underline text in Braille option checked on the text formatting tab. Unchecking it made no difference. :(
Comment 6 Joanmarie Diggs (IRC: joanie) 2010-06-19 04:57:25 UTC
Dang, I'm honestly not sure Nolan.

I've been using a number of different versions of Pidgin in a number of different versions of GNOME in both OpenSolaris and Linux. After I put out the call for testing, I received a couple of private emails confirming that things were working as expected. So I'm wondering if it's something unique to your system.

Would you mind posting your issue on the list and seeing if anyone else is experiencing a similar problem?
Comment 7 Nolan Darilek 2010-06-19 14:07:07 UTC
Good news of a sort: I wiped out and recreated my Pidgin configuration, and now it doesn't hang on login. So perhaps it was Pidgin hanging somehow, not sure.

I'm slowly adding back elements to my configuration to see if something causes this, but having just gotten it working after a day of failed upgrades on multiple fronts, I'm not quite ready to put it all back and have things start breaking again. :) If I recreate a configuration similar to what I had and the crash goes away, then I'll close this bug in a few days. If I do duplicate it then it's by all likelihood a Pidgin issue, but we can figure that out here first I'd guess.

Thanks a bunch.
Comment 8 Joanmarie Diggs (IRC: joanie) 2010-07-05 02:45:34 UTC
(3.0 Planning Spam-o-rama. Sorry!)
Comment 9 Joanmarie Diggs (IRC: joanie) 2010-07-05 16:41:49 UTC
Nolan, since it sounds like you've got it worked out, I'm going to close this out. Please feel free to re-open if the problem re-appears. Thanks!