Liviu Andronic wrote: > On Sat, Dec 4, 2010 at 4:55 PM, Pavel Sanda <sa...@lyx.org> wrote: > > most of them are not "ours"... anyway, we are havent moved a single > > step as far as crashes is concerned. > > > Is it normal that I cannot replicate any crashes under valgrind, even > though I can easily (within max 2 mins) replicate under gdb or simple > execution? What would valgrind execution change for this to happen?
it may be dependent on some timing synchrony which will change with different speed under valgrind... i guess the crash happens during some export? we switched to threads, you know :) iirc we have in our sources still some #ifdef for using threads. if you switch it off and the crash disappears at least we know the area of origin. pavel
