Re: gcc-3_0 march=athlon && (-Os || -ggdb)

[Jan Hubicka <jh at suse dot cz>]

> > Hmm, I marked it critical/high severity/priority.
> > 
> > Well, here is the preprocessed file bzip2'd that I submitted to gcc
> > gnats. I really hope this helps.
> Thanks.
> I can reproduce the crash with -march=pentium as well, so it definitly
> classify as regression.
> I am trying to hunt it down right now. I am not dwarf2 expert so if I will
> fail to do so, I will set priority back to high and add comment that it
> actually is regression.
Hi,
I am puzzled by the bug - it seems to be caused by using nested function C
extension. I don't understand much to the dwarf code, so I've updated the
bugreport (set it to high priority, changed category from c++ to 'debug' and
marked as regression over 2.95) in hope that someone else will get enlightened
sooner than I do.

Honza
> 
> Honza
> > 
> > -- 
> > Gordon Sadler
> > 
>