From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 22140 invoked by alias); 31 Aug 2004 21:54:46 -0000 Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org Received: (qmail 22121 invoked from network); 31 Aug 2004 21:54:45 -0000 Received: from unknown (HELO colo.khms.westfalen.de) (213.239.196.208) by sourceware.org with SMTP; 31 Aug 2004 21:54:45 -0000 Received: from khms.vpn ([10.172.192.2]:38056 helo=khms.westfalen.de ident=Debian-exim) by colo.khms.westfalen.de with asmtp (TLS-1.0:RSA_ARCFOUR_SHA:16) (Exim 4.34) id 1C2GaS-0000Q2-OG for gcc@gcc.gnu.org; Tue, 31 Aug 2004 23:54:45 +0200 Received: from root (helo=khms.westfalen.de) by khms.westfalen.de with local-bsmtp (Exim 4.34) id 1C2GHY-0007Fa-AX for gcc@gcc.gnu.org; Tue, 31 Aug 2004 23:35:12 +0200 Received: by khms.westfalen.de (CrossPoint v3.12d.kh14 R/C435); 31 Aug 2004 23:23:36 +0200 Date: Tue, 31 Aug 2004 22:06:00 -0000 From: kaih@khms.westfalen.de (Kai Henningsen) To: gcc@gcc.gnu.org Message-ID: <9FuB74S1w-B@khms.westfalen.de> In-Reply-To: <87acwc3qq6.fsf@codesourcery.com> Subject: Re: Ada policy MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Organization: Organisation? Me?! Are you kidding? References: <10408310328.AA02238@vlsi1.ultra.nyu.edu> <10408310328.AA02238@vlsi1.ultra.nyu.edu> <87acwc3qq6.fsf@codesourcery.com> X-No-Junk-Mail: I do not want to get *any* junk mail. Comment: Unsolicited commercial mail will incur an US$100 handling fee per received mail. X-Fix-Your-Modem: +++ATS2=255&WO1 X-SW-Source: 2004-08/txt/msg01684.txt.bz2 zack@codesourcery.com (Zack Weinberg) wrote on 30.08.04 in <87acwc3qq6.fsf@codesourcery.com>: > That is a little different from 'a deliberate policy of breaking > source compatibility between releases', I admit. I think that a > reasonable person could reach that conclusion too, but only having > read messages which I remember reading but cannot presently find in > the archives. Hmm. I certainly seem to remember reading those messages, too. Not that it were a plan to always break compatibility, but that incompatible features were pretty much always introduced and they wanted to use them in the compiler, which then produced that incompatibility; and as long as the previous version was still compatible, that was not considered a problem. And that in fact only-the-previous-version had been true for a large number of versions. Hmm ... this seems relevant: http://gcc.gnu.org/ml/gcc/2002-03/msg01197.html http://gcc.gnu.org/ml/gcc/2002-03/msg01245.html http://gcc.gnu.org/ml/gcc/2002-03/msg01268.html ... and probably more in that thread. It seems completely obvious from that thread that ACT's ideas of how much old version bootstrap compatibility was worth, was vastly different from either the FSF's, or the one described in the current thread (that is, lots of old versions used to check exactly that internally). Furthermore, there's mention of a plan to document which versions work for 3.1 or the next version. We have 3.4 out, and now it seems we finally get such a list ... (I wonder if I should also mention the "either take our contributions the way they are, or we'll take our toys and go home" thing ...) MfG Kai