This is the mail archive of the
gcc-prs@gcc.gnu.org
mailing list for the GCC project.
Re: ada/6160: configure sets up bootstrap to build ada with too-oldgnat
- From: Christian Cornelssen <ccorn at cs dot tu-berlin dot de>
- To: nobody at gcc dot gnu dot org
- Cc: gcc-prs at gcc dot gnu dot org,
- Date: 24 Feb 2003 23:56:01 -0000
- Subject: Re: ada/6160: configure sets up bootstrap to build ada with too-oldgnat
- Reply-to: Christian Cornelssen <ccorn at cs dot tu-berlin dot de>
The following reply was made to PR ada/6160; it has been noted by GNATS.
From: Christian Cornelssen <ccorn at cs dot tu-berlin dot de>
To: <gcc-gnats at gcc dot gnu dot org>, <lucier at math dot purdue dot edu>,
<gcc-prs at gcc dot gnu dot org>
Cc:
Subject: Re: ada/6160: configure sets up bootstrap to build ada with too-old
gnat
Date: Tue, 25 Feb 2003 00:38:32 +0100 (CET)
Hi,
Brad, if the phenomenon is that all the *.ali compiled by gnatgcc
contain just a single linefeed or similarly truncated output, then
gnat1 may have been miscompiled (either now or formerly). I have
experienced that with the stage1 xgcc from the gcc-3_2-branch when I
used STAGE1_CFLAGS="-march=i686 -O2". Apparently, either Debian's
gcc-2.95.4 or the gnatgcc v2.8.1 (plus GNATBIND v3.14p (20010503))
used for bootstrapping did not do a good job with that option. When I
removed the "-march=i686", everything went fine.
NB: The problem occurs on gcc-3_2-branch only (including the 3.2.2
release); it has not shown up with checkouts of 3.3 or 3.4 (mainline).
See also: http://gcc.gnu.org/ml/gcc-bugs/2003-02/msg00015.html
Regards,
Christian Cornelssen