Comment 4 for bug 238593

It seems like there might be some relationship between some of the conditions around this bug and bug 430449 submitted to the OpenSUSE team that maintains their branch of OpenOffice (see https://bugzilla.novell.com/show_bug.cgi?id=430449). There is a stack trace attached to that bug report which might be useful to compare with the base bug report.

The OpenSUSE bug report was also seen when using Vista, but was not seen when using Windows XP. Thorsten Behrens described the issue as a 3d transition apparently not being initialized properly, and the "Random Transition" crashed because of it.

The Vista failure is also reported to the OpenOffice.org team as bug 94389 *http://www.openoffice.org/issues/show_bug.cgi?id=94389).

Thorsten and Radek committed a change to the OpenSUSE code which is intended to fix issue 430449 on their system. They didn't say if their fix was being pushed back to the OpenOffice team, or if it would remain forever on only the OpenSUSE code line.

My failure case was using an OpenSUSE 11.0 machine with the open source video drivers over an NVidia GEForce4 440 Go. The computer has 512 MB of memory and is relatively slow.

When I wanted to test dual head mode, I had to install the NVidia binary driver. Once I installed that binary driver, my crash stopped happening. I ran in multiple different configurations with multiple tests (multi-head, and Xinerama) and saw no failures in all my test runs.

I also saw the same failure on an OpenSUSE 11.0 computer at work with a different video card and 1 GB of memory. It might be worth your while to check with Thorsten to see if he can offer insights about the likely use of this case.