This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Unreviewed patches^2 / [PATCH] Fix specs processing
- From: Roger Sayle <roger at www dot eyesopen dot com>
- To: Joern Rennecke <joern dot rennecke at superh dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Zack Weinberg <zack at codesourcery dot com>, kaz Kojima <kkojima at rr dot iij4u dot or dot jp>
- Date: Tue, 4 Mar 2003 09:59:38 -0700 (MST)
- Subject: Re: Unreviewed patches^2 / [PATCH] Fix specs processing
On Tue, 4 Mar 2003, Joern Rennecke wrote:
> 2003-02-08 Roger Sayle <roger at eyesopen dot com>
>
> * gcc.c (do_spec_1) ['{']: Handle pending argument upon return
> from handle_braces in "%{...}".
>
> This breaks most of the sh targets (sh-elf, sh64-elf, sh3-elf, ...)
Sorry.
The problem was that unless a "%{...}" is followed by a space, it
would never get processed. RTH's review of my workaround to get
Tru64 bootstraping again suggested that this was indeed a bug.
http://gcc.gnu.org/ml/gcc-patches/2003-02/msg00368.html
Assuming that the SH back-end's use of %{...} is valid, we'll need
to find some other point in the code for the arg to be processed,
or some way to avoid spliting "sh elf" with the current mechanism.
As a temporary solution we could try reverting my 2003-02-08 patch
and using http://gcc.gnu.org/ml/gcc-patches/2003-02/msg00296.html
This should keep both SH and Tru64 bootstraping, whilst the correct
fix to the problem is investigated.
My apologies again, I really know very little about specs processing
and how they're supposed to work, just that it currently doesn't work
very well. I'm happy to test trial patches on alphaev67-dec-osf5.1.
Roger
--