This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: 2.95.3 status
- To: "Bernd Schmidt" <bernds at cambridge dot redhat dot com>,"Joe Buck" <jbuck at synopsys dot COM>
- Subject: Re: 2.95.3 status
- From: "Geert Bosch" <bosch at gnat dot com>
- Date: Sat, 10 Feb 2001 12:13:51 -0500
- Cc: "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Reply-To: "Geert Bosch" <bosch at gnat dot com>
On Fri, 9 Feb 2001 15:48:53 -0800 (PST), Joe Buck wrote:
I don't think that we can drop HP with the native assembler, and I do
think that the comparison failures are false due to timestamps. It
doesn't feel right to say 2.95.3 won't work on a platform where 2.95.2
worked.
I found with other versions of GCC that the differences were due
to some random/uninitialized data (not a time-stamp IMO) that was
included in one of the assembler-generated sections (unwind info,
or something similar).
The text and data segments were identical. A possible solution would be
to use objcopy or a similar tool to remove the offending sections before
comparing. Actually, you'd only have to do this with the final linked
executable.
-Geert