From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-49893-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 9748 invoked by alias); 13 Apr 2002 20:18:17 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 9741 invoked from network); 13 Apr 2002 20:18:17 -0000
Received: from unknown (HELO nile.gnat.com) (205.232.38.5)
  by sources.redhat.com with SMTP; 13 Apr 2002 20:18:17 -0000
Received: by nile.gnat.com (Postfix, from userid 338)
	id B29AFF28EB; Sat, 13 Apr 2002 16:18:16 -0400 (EDT)
To: mark@codesourcery.com, tromey@redhat.com
Subject: Re: GCC 3.1 Release
Cc: gcc@gcc.gnu.org
Message-Id: <20020413201816.B29AFF28EB@nile.gnat.com>
Date: Sat, 13 Apr 2002 13:51:00 -0000
From: dewar@gnat.com (Robert Dewar)
X-SW-Source: 2002-04/txt/msg00545.txt.bz2

<<In fact what I'd really like to see is a requirement for such a
regression tester for every platform that is considered release
critical.  (I'm sure the logistics of this are difficult.)
>>

I strongly support this. It is certainly what we do at ACT (we build all
versions every night, and immediately fix or backout any changes that break
any target).

Mark> The amount of breakage since GCC 3.0 continues to amaze me.

It does not surprise me so much. At this stage, it would somewhat surprise me
if we were NOT running into this kind of breakage in the absence of systematic
multi-platform regression testing.

I know that it would be near to impossible to keep GNAT highly stable without
this kind of regression testing.