On Tuesday 27 May 2008 16:49:49 NotFound wrote: > <[EMAIL PROTECTED]> wrote: > > There's a segfault in t/dynpmc/dynlexpad_2.pir with the gcdebug core. It > > happens at the end of the program, during global destruction, when trying > > to call the destroy() entry on the DynLexPad PMC. > > Is this ticket still valid? I'm unable to reproduce the problem: > > $ ./parrot t/dynpmc/dynlexpad_2.pir > ok 1 > ok 2 > 13013 > > $ ./parrot --runcore=gcdebug t/dynpmc/dynlexpad_2.pir > ok 1 > ok 2 > 13013
I fixed that segfault, but the order-of-destruction bug is still a problem. I think this is the one where we can't destroy Library PMCs until we've destroyed all instances of the PMCs they may define, otherwise their function pointers will have gone away. -- c
