This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: Workaround for current gcc problem
- To: Andreas Jaeger <aj at suse dot de>
- Subject: Re: Workaround for current gcc problem
- From: Jakub Jelinek <jakub at redhat dot com>
- Date: Mon, 20 Nov 2000 10:01:03 +0100
- Cc: libc-alpha at sources dot redhat dot com, Stanislav Brabec <utx at penguin dot cz>, Neil Booth <NeilB at earthling dot net>
- References: <ho4s13owtp.fsf@gee.suse.de>
- Reply-To: Jakub Jelinek <jakub at redhat dot com>
On Mon, Nov 20, 2000 at 09:46:10AM +0100, Andreas Jaeger wrote:
>
> The current GCC version from CVS has done changes to CPP which break
> some assumptions in glibc. One place (I didn't find others so far but
> suspect that there might be more) has been reported by Stanislav via
> GNATS:
>
> > Preprocessing of
> > i.86-.*-linux.* ld=ld-linux.so.2
> > will cause
> > i.86 -.*-linux.* ld=ld-linux.so.2
> >
> > I didn't investigated more about reason, why space is added (see fix),
> > but problematic phase is cpp'ing.
> >
> > >How-To-Repeat:
> > echo "i.86-.*-linux.* ld=ld-linux.so.2" | gcc -E -undef -x c -
> > The result is the line above.
>
> The following patch fixes this and works also with gcc 2.95.2. Neil,
> can you send your explanation why this patch is needed?
>
> Please note that the current CVS version still will not compile glibc,
> the problems are:
> - zero size arrays don't work at all
> - -g -include does not work (there's a patch available but not applied)
> - some code generation problems
>
> Is it ok to commit this?
Isn't it better to use
echo "i.86-.*-linux.* ld=ld-linux.so.2" | gcc -E -undef -x assembler-with-cpp -
Works with all of:
gcc 2.95.2
RH gcc 2.96
CVS gcc 2.97
The language of the i.86-.*-linux.* string is definitely not C...
Jakub