What you just 'fixed' is an assertion against a memory corruption bug in all Zope versions between 2.4.0 and 2.4.3, inclusive. Zope version 2.4.4 (which hasn't been released, Hey ZC! wake up! stabilizing 2.4 is more important than releasing 2.5 in my book! Now that you've done the latter can you care to do the former, please? :-), should fix your problem, as would migrate to Zope 2.5. Legend has it that the current 2.4 branch is exactly what 2.4.4 will be. Check the archives for this mailing list (link below) for instructions on how to take the current 2.4 branch from CVS. BTW, Chris, what Olivier saw couldn't automatically lead him to correlate with the other Zope crashes, since the other reports don't mention the 'assertion' error message he saw, which he concluded was the cause of his problems. I'm really disappointed with ZC for putting out a new release of Zope instead of a fixed version of the release most everyone is currently using. Regards, Leo On Sun, 2002-01-27 at 14:45, Olivier Deckmyn wrote:
Hi,
I decided to upgrade my Python 2.1.1 to Python 2.1.2 (in order to gently move to 2.5 later).
My Zope instances were working quite fine....but...randomly I got a core dump (!) and the following message in logs :
assertion "STACK_LEVEL() <= f->f_stacksize" failed: file "Python/ceval.c", line 687
Bug seems to appear when playing with ZCatalog (CMF sites could'nt even render a page) and Python scripts ?
After some hours of researches, I have a (dirty, really) fix :
I had to modify Python sources and hack line 687 in ceval.c : I removed it... Then recompile and it seems to work. I know this is dirty, and my Zope my crash later.
PLEASE CHECK and HELP !
This looks like a major serious bug to me.
Cheers,
Olivier.
_______________________________________________ Zope-Dev maillist - Zope-Dev@zope.org http://lists.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope )