This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Mainline fails to bootstrap was Re: Mainline problems with libgcc2.c
On Thu, May 23, 2002 at 03:48:31PM +0200, Andreas Jaeger wrote:
> Gordon Sadler <gbsadler1@lcisp.com> writes:
>
> > I've been seeing this for many days now. I've tried to bootstrap with a
> > plain script with a basic configure and still no luck. Others are
> > reporting test results so... perhaps it's something local?
> >
> > I'm running Debian Unstable updated daily. 3.1 branch bootstraps daily
> > and test results are on the mailing list, so tool chain on my system
> > should be good?
> >
> > Attached are the most recent compile script and tail end of build log.
>
> Why are you using make -e? I would remove the -e, perhaps there's
> something in your environment that causes the problem.
>
I was originally intending its use to pass variables from the script to
configure and make (eg CFLAGS, TCFLAGS, etc).
Removing it appears to have corrected my problem. The question now is,
why would it work for months in that fashion then suddenly cease to work
only on CVS HEAD? It appears to be related to this patch:
2002-05-16 Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
* Makefile.in: Allow for PWDCMD to override hardcoded pwd.
* configure.in: Likewise.
* fixinc/check.tpl: Likewise.
* fixinc/fixinc.dgux: Likewise.
* fixinc/fixinc.svr4: Likewise.
* fixinc/fixinc.winnt: Likewise.
* fixinc/fixincl.sh: Likewise.
* fixproto: Likewise.
* configure: Regenerate.
I cannot be positive though.
--
Gordon Sadler