This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: EGCS 1.1.1 on BSDI 3.1
- To: jonas dot bulow at ac dot salcom dot se, egcs at cygnus dot com
- Subject: Re: EGCS 1.1.1 on BSDI 3.1
- From: N8TM at aol dot com
- Date: Fri, 8 Jan 1999 00:30:06 EST
In a message dated 1/7/99 11:22:05 AM Pacific Standard Time,
jonas.bulow@ac.salcom.se writes:
<< _udivsi3
/tmp/cc011487.s: Assembler messages:
/tmp/cc011487.s:11: Error: Rest of line ignored. >>
This resembles the problem which the snapshots have exhibited for 2 months now
on targets including linux-gnulibc1 and cygwin, where options -Os
-march=pentiumpro produce a cc1 which fails just before this point (when
attempting to compile _divdi3) at the end of the stage1 compile. On those
platforms, egcs-1.1.1 does not exhibit this failure. I wonder if it's some
kind of piping configuration problem; you might try preprocessing to a file
and feeding that to cc1 to see if the bug still occurs. Not that I know what
to make of that. Apparently the experts are stumped by this thing not doing
the same on all x86 targets and not occurring for the most common combinations
of targets and compiler flags.