tags 367831 +pending thanks -- hi,
well, i got to reproduce this bug on my 'new' powerpc machine. but, mmh, i still haven't found a way around it though. the problem seems to lie somewhere in topo_sort, a recursive function, when it reaches the end of graph branches, hitting NULL pointer there. fusebirth is only used to generate fused_loop.c, which is then used to build the final freebirth. so shipping a generated fused_loop.c instead of trying to generate it every time would be a solution. i still wonder why this code (although quite ugly) behaves correctly on the other arches, and runs on powerpc with -O0... cheers, Paul On Wed, Oct 03, 2007 at 06:51:55PM +0200, Martin Michlmayr wrote: > * Paul Brossier <[EMAIL PROTECTED]> [2007-09-28 14:51]: > > unfortunately, i don't have access to powerpc at the moment, so Martin, > > any help would be greatly appreciated. > > I just tried to compile the current version of freebirth with the > current gcc (4.2) version in unstable, and it still segfaults. > -- > Martin Michlmayr > http://www.cyrius.com/ > > >
signature.asc
Description: Digital signature