GNOME Bugzilla – Bug 462239
OpenOffice OOo-dev 2.3.0 Presentation application crashes when trying to open an existing presentation.
Last modified: 2007-08-07 18:40:01 UTC
This is a tracking bug for OOo issue #80228 http://www.openoffice.org/issues/show_bug.cgi?id=80228 Using the latest OOo-dev 2.3.0 distribution on Ubuntu gutsy (with all the latest updates). Steps to reproduce: 1/ Start simpress 2/ Trying to open an existing presentation. 3/ Watch simpress <strike>crash</strike> go into error recovery mode. I tried it with: http://www.netbeans.org/files/documents/4/948/NetBeansIDE.odp and http://www.presentationhelper.co.uk/impress-templates/subtlewaves.odp Note that Orca doesn't need to be running for simpress to crash. Having said that, we aren't going to be able to test with Orca until this problem is fixed though, so monitoring it here as well as in the OOo issues bug database.
Just added another scenerio to the OOo issue, that cause simpress to crash: The following steps also crash simpress too: 1/ Start simpress. The startup wizard appears with the "Create new presentation" radio button selected and focus on the Next button. 2/ Press Space. The second screen of the startup wizard is displayed, with focus still on the Next button. 3/ Press Space. The third screen of the startup wizard appears and focus in now on the Create button. 4/ Press Space. The main simpress window appears. After a few seconds, it disappears and simpress goes into error recovery mode. This is without Orca running. In short, we are not going to be able to do much Orca testing with simpress until this issue is fixed.
Rich: w.r.t. the first issue, I cannot reproduce it. w.r.t the 2nd issue, I can only reproduce it if Orca is running. So I'm wondering: 1. Since you can reproduce the problems without Orca running, can you reproduce them without assistive technologies enabled in GNOME? If not: 2. If you grabbed the latest at-spi from SVN trunk could you reproduce the problem without Orca running? (Li just checked in a patch)
Great ideas. I'm going to work on something else first, but then I'll try out what you suggested (probably tomorrow). Thanks.
Another data point: if in atspi.getRelationString(), I comment out relations = self.relations and make relations an empty list instead, simpress doesn't hang. If instead I wrap relations = self.relations in a try/except, it does hang. It seems that we're getting a defunct object, trying to print details, calling toString(), and making simpress really unhappy when we look at its relations.
I'm confused. I don't see hangs w.r.t. this bug, but I do with bug #462547. Did you add the comment to the wrong bug?
So NOW he wants me to add the comments to the bug they actually apply to.... Sheesh.... ;-) ;-) ;-) ;-) Yeah, my bad. Sorry!!
> 1. Since you can reproduce the problems without Orca running, > can you reproduce them without assistive technologies enabled in GNOME? Yes. simpress successfully crashes with the GNOME gconf setting for "enable accessibility" set to false.I've updated the OOo issue to that effect. Seems like quite a serious issue to me. I wonder when somebody will take a look at it.
Well I've upgrade to OOo-dev 2.3.0 (build #223), plus latest Ubuntu Gutsy packages (as of today), plus latest Orca plus deleted all my GNOME "dot" files, and the problem has gone away. I've closed out the OOo issue and am closing this bug out as FIXED.