This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Better relocation handling for nonstandard --libdirs
- From: Daniel Jacobowitz <drow at false dot org>
- To: gcc-patches at gcc dot gnu dot org, richard at codesourcery dot com
- Date: Thu, 10 May 2007 14:14:45 -0400
- Subject: Re: Better relocation handling for nonstandard --libdirs
- References: <87abwn1ok8.fsf@firetop.home> <87abwdcddz.fsf@firetop.home>
This looks sane, but I have two questions.
On Thu, May 10, 2007 at 09:05:12AM +0100, Richard Sandiford wrote:
> > +# $(exec_prefix), expressed as a path relative to $(prefix).
> > +prefix_to_exec_prefix = \
> > + `echo "$(exec_prefix)" | \
> > + sed -e 's|^$(prefix)||' -e 's|^/||' -e '/./s|$$|/|'`
If I understand correctly, this require that exec_prefix be the same
as, or a subdirectory of, prefix. If --prefix=/shared-usr
--exec-prefix=/local-usr then this will yield "local-usr/" which isn't
very useful. Is that restriction documented or enforced anywhere?
My other question is, why do we still need this mess at all? We have
make_relative_prefix nowadays and I'm confident we've gotten the kinks
out of it. If you want to get from the configured bindir to the
configured tooldir, then give it the bindir and the tooldir.
--
Daniel Jacobowitz
CodeSourcery