Sorry to reply on my own mail. However it is indeed the same ICE still with the workaround -O1. I assumed this ICE was fixed as I did not find any bug on the gcc bugs page in the debian bug tracker.
I have an updated package ready that re-enables -O1 for m68k, however I'm not 100% sure that is the correct way to proceed. Joost On Saturday 04 February 2006 14:47, Joost Yervante Damad wrote: > On Wednesday 01 February 2006 05:14, Henrique de Moraes Holschuh wrote: > > OpenMSX requires quite a lot of CPU, and especially on slow machines like > > most m68k, it really benefits from -O2. -O1 would have to be enabled > > selectively for m68k, but would OpenMSX still be usable in m68k under > > -O1? > > To be honest, I don't know if openMSX is usable on m68k at all. The same > might be the case for other architectures. > > It's unclear to me what is the prefered way here: only enable the archs > that are known to work, or enable as many as possible, and preferably all? > > > About optimizations,the default openmsx flavour is "opt" which uses -O3. > See also build/flavour-opt.mk > > Greetings, Joost Damad -- The planet Andete is famous for it's killer edible poets.
pgpF09TZ7iSYG.pgp
Description: PGP signature