This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: The current egcs is broken on x86
- To: mark at markmitchell dot com
- Subject: Re: The current egcs is broken on x86
- From: Manfred Hollstein <manfred at s-direktnet dot de>
- Date: Sun, 14 Jun 1998 14:15:54 +0200 (MEST)
- Cc: wilson at cygnus dot com, hjl at lucon dot org, egcs at cygnus dot com, meissner at cygnus dot com
- References: <199806122210.PAA04254@rtl.cygnus.com><199806131649.JAA25046@mail.earthlink.net>
- Reply-To: manfred at s-direktnet dot de, Manfred dot Hollstein at ks dot sel dot alcatel dot de
On Sat, 13 June 1998, 09:49:19, mark@markmitchell.com wrote:
> >>>>> "Jim" == Jim Wilson <wilson@cygnus.com> writes:
>
> Jim> My patched rebuild failed in stage2 while running the stage2
> Jim> compiler to build the SYSCALLS.c file. I suspect another bug
> Jim> in the recent regmove changes.
>
> I can confirm this on x86. I was able to bootstrap only by completely
> removing the regmove change, not simply by applying your patch.
>
> Since the repository is available to the public, and since problems
> like this really do hinder development by others (I was working on
> some optimizations of my own, and not sure whether I'd broken things,
> or whether it was the regmove patch), I would like people to refrain
> from checking in any significant changes that could affect code
> generation without bootstrapping the compiler on at least one
> platform. I don't know whether that was done in this case or not, of
> course.
>
> Do people think that's a reasonable guideline, in general?
Yep, I can only support this advice, though I must admit I'd thought
everybody was working this way already/always... ;-)
manfred