On Thu, 13 Nov 2008, Frank Niessink wrote:

> 2008/11/13 Jerome Laheurte <[EMAIL PROTECTED]>:

>> There's another problem though; on my PowerBook G4 running MacPython
>> 2.5.1 + wxPython 2.8.9.1, all unit tests pass. On my Mac Mini running
>> stock Python 2.5.1 + wxPython 2.8.9.1, I get a bunch of
>> warnings/errors/failures/whatever (see the Buildbot output for
>> details). The only (but big) difference between the two is that the
>> PowerBook is running Panther whereas the Mini is running Leopard.
>>
>> This began with a commit that couldn't possibly have this outcome
>> (replacing "xxx ?" with "xxx?"). It looks like the SIGSEGV problems
>> under Linux.
>>
>> I'm a bit at loss here... What version of MacOS do you use on your own
>> Mini, and do the unit tests pass on it ?
>
> They did the other day when I released 0.71.3. I'll check again tomorrow.

Okay, I've tried all combinations of stock Python 2.5.1/MacPython 
2.5.2 with wxPython 2.8.4.0/2.8.7.1/2.8.9.1 on my Mini and still got 
the problem. Then I rebooted the machine, and the unit tests passed. I 
first thought it was a problem with Spaces since I had a pass/doesn't 
pass occurence with the only difference between the two was that I 
switched on another Space and launched Safari there, but I can't 
reproduce this after another reboot.

There's something strange about this. When the unit tests fail, 
Task Coach itself crashes with a "bus error", not unlike the SIGSEGV I 
regularly get under Linux. It's obviously a wxWidgets bug, but it may 
take some time until I can isolate it.

Cheers
Jérôme

Reply via email to