This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Current CVS'd snapshot build dies on m68k-next-nextstep3compiling cccp.c for stage 3
- To: law at cygnus dot com
- Subject: Re: Current CVS'd snapshot build dies on m68k-next-nextstep3compiling cccp.c for stage 3
- From: Toon Moene <toon at moene dot indiv dot nluug dot nl>
- Date: Fri, 2 Oct 98 15:03:14 +0200
- Cc: egcs at cygnus dot com
- Organization: Moene Computational Physics, Maartensdijk, The Netherlands
- References: <9748.907316911@hurl.cygnus.com>
I wrote:
> Yep, I'd had my hopes centered on that one (if for nothing else
> than the amount of work David put into tracking it down).
> Unfortunately, the snapshot I CVS'd yesterday (199809191000UTC),
> which contains his updates, still failed the same way.
and Jeff replied:
> Could also be the reload1.c bug that Joern fixed today.
> It was causing codegen problems on my PAs. Or Richard's
> function.c change which is causing codegen problems on
> big endian targets.
Close but no cigar (oops, that ruins my PCness ;-)
The 19980919 problem was solved by the following patch from Andreas
Schwab:
* final.c (final_scan_insn): Disable CC tracking across branches,
it fails to preserve all CC flags if the output routine needs to
modify them.
The current snapshot (CVS'd 19981002730 UTC) also doesn't build -
this probably is due to Richard's function.c change.
Cheers,
Toon.