This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [toplevel] Import two libtool patches to simplify AC_DEPLIBS_CHECK_METHODunder Linux.
- From: Roger Sayle <roger at eyesopen dot com>
- To: Andreas Schwab <schwab at suse dot de>, Kelley Cook <kcook at gcc dot gnu dot org>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Alexandre Oliva <aoliva at redhat dot com>, "Frank Ch. Eigler" <fche at redhat dot com>
- Date: Thu, 25 Nov 2004 18:02:28 -0700 (MST)
- Subject: Re: [toplevel] Import two libtool patches to simplify AC_DEPLIBS_CHECK_METHODunder Linux.
On Fri, 26 Nov 2004, Andreas Schwab wrote:
> Kelley Cook <kcook@gcc.gnu.org> writes:
> > 2004-11-24 Kelley Cook <kcook@gcc.gnu.org>
> >
> > * libtool.m4 (AC_DEPLIBS_CHECK_METHOD): Import linux-gnu case
> > from CVS libtool to always pass_all.
>
> Why was this change reverted again?
I may owe Kelley an apology. There was a bad interaction between
libtool and Devang Patel's patch to change process_command line to
supply -v to the linker (which was committed at about the same time).
http://gcc.gnu.org/ml/gcc-patches/2004-11/msg01353.html
I'd figured out that the i686-pc-linux-gnu bootstrap failure was
caused by a change in the behaviour of libtool, and using a generated
libtool from a few hours earlier fixed the regression and allowed
x86/Linux bootstrap to complete. To test the hypothesis I did a
"cvs update", reverted the change to libtool.m4, and bootstrapped
and indeed everything worked again.
My apologies for the confusion, and my appreciation to Kelley Cook
for rapidly backing out his patch until he had time to analyse the
true cause of the failure. It looks like his change was completely
innocent after all, but just committed at a bad time for GCC's
"libtool".
Roger
--