More crash-proof KCrash (was Re: Konq crash backtraces)

[Lubos Lunak]

On Thursday 12 January 2006 00:03, Olivier Goffart wrote:
> Le Mercredi 11 Janvier 2006 17:40, Lubos Lunak a écrit :
> >  This way KCrash should be more reliable and we should (hopefully) always
> > get DrKonqi now.
>
> Even with infinite recursions ?

 Weeeell. No, most probably not, I don't see a way how to handle that in 
KCrash. Infinite recursion is either stopped by running out of memory and the 
hell getting loose as the OOM-killer tries its (poor) luck when trying to 
kill the offender, so in this case we don't need to bother anyway, or it's 
stopped by ulimit for stack. In which case there's no way to go any deeper, 
and so even KCrash cannot call any functions - if KCrash itself would be 
called at all. I just don't see how we could handle this on our own, that'd 
need at least some system support.

> I've just got such crash previous twice this week, and no dr konqui, just
> segfault in the console.
>
> (but i guess dr konqui should also strip the backtrace, like gdb does with
> --- press enter to continue or q to quit ---)

 It could do that, there's some code in drkonqi/backtrace.cpp that 
post-processes the backtrace, but given what's said above I guess there's no 
point right now.

-- 
Lubos Lunak
KDE developer
---------------------------------------------------------------------
SuSE CR, s.r.o.  e-mail: l.lunak at suse.cz , l.lunak at kde.org
Drahobejlova 27  tel: +420 2 9654 2373
190 00 Praha 9   fax: +420 2 9654 2374
Czech Republic   http://www.suse.cz/