Le 2001.04.03 01:45:12 +0200, Not Zed a écrit :
>
> > > Does it still crash at all (during mail retrieval)? That bug should
> be
> > > fixed.
> >
> > Well, it still hasn't crashed (I tried several times with various mail
> > loads). I think I can safely say this annoying bug is fixed.
>
>
> I reckon you're just lucky, since it crashed in libibex and that hasn't
> changed for months.
Well, this is not the same bug ! The libibex bug was caused by a corrupted
mbox.ibex and crashed the mailer systematically when syncronising mail just
after retreival. But the bug I was talking about in my previous mail is
caused by getting mail from multiple POP accounts simultaneously. And it
was quite frequent (nearly each time there was mail on both accounts) and
occured on the 2nd or 3rd mail.
> I've been running some regression tests (the last one ran all night on
> an athlon 1200 :(, and I am not getting anywhere. Under very rare
> circumstances I can get it to crash (actually i have to turn on mcheck),
> but if i run it under any sort of debugger it never fails (and we're
> talking a lot of heavy testing, millions of flat out updates while
> performing flat out lookups from another thread).
>
> I've never seen any corrupted file errors either, just mcheck can bomb
> out, but nothing else (i save/reload the file often).
>
> If I use mcheck i sometimes get clobbered memory but the result doesn't
> make a lot of sense, and i have no idea what's causing it. It never
> crashes (so far) when I do single-threaded operation. But that doesn't
> make sense either, the locking is so simple, etc.
>
> So, either mcheck is getting confused, or its not, or ibex memory is
> being corrupted by other things going on, or well I dont know. I need
> to do some purify.
>
> !Z
OK, so there's still something waiting in the dark to bite us ...
Xav
_______________________________________________
evolution maillist - [EMAIL PROTECTED]
http://lists.helixcode.com/mailman/listinfo/evolution