>>>> "Gunnar" == Gunnar Evermann
<ge204(a)eng.cam.ac.uk> writes:
Gunnar> Hrvoje Niksic <hniksic(a)arsdigita.com> writes:
> Raymond Toy <toy(a)rtp.ericsson.se> writes:
>
> > o xemacs crashes for me quite often in either gc or
> > make_lisp_buffer_stream. (Already reported). I have a sneaking
> > suspicion that it's a gcc 2.95.2 problem. After recompiling with
> > Sun CC 5.0, I haven't run into it today in the last few hours,
> > whereas it was happening as often as every 30 minutes with gcc
> > 2.95.2. I'll have to try with gcc 2.7.2.2 too. Perhaps a note in
> > the PROBLEMS file?
>
> Maybe, maybe not. Remember that it could also be a memory corruption
> bug, and those can go away when you change *anything* -- and come back
> to haunt you a year later.
>
> Is gcc 2.95.2 a stable release of Gcc?
Gunnar> It's supposed to be. Ray, could you try compiling with gcc but with
Gunnar> less aggressive optimisation (say -O1 instead of -O3)?
I reconfigured with --cflags=-O1 and rebuilt. I get this:
Dumping under the name xemacs
unexec(): dldump(/apps/public/XEmacs/src/xemacs-21.2.36/src/xemacs): ld.so.1: ./temacs:
fatal: /apps/public/XEmacs/src/xemacs-21.2.36/src/xemacs: unknown dynamic entry:
1879048176
So gcc 2.95.2 on Solaris with -O1 optimization can't even produce a
working temacs. It does, however, produce a working temacs with -O3
optimization, but that seems to cause crashes while using xemacs.
I think I'll stick to gcc 2.7.2.2 or Sun C compiler for compiling
xemacs. I've been using the Sun C compiled version for about a day
and a half, and it's been quite stable, unlike the crash every 30-60
minutes scenario I was getting last week with 2.95.2.
Ray