From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934639AbXGKWne (ORCPT ); Wed, 11 Jul 2007 18:43:34 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1762863AbXGKWnZ (ORCPT ); Wed, 11 Jul 2007 18:43:25 -0400 Received: from one.firstfloor.org ([213.235.205.2]:46633 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1762529AbXGKWnY (ORCPT ); Wed, 11 Jul 2007 18:43:24 -0400 Date: Thu, 12 Jul 2007 00:43:22 +0200 From: Andi Kleen To: Linus Torvalds Cc: Andi Kleen , Martin Orr , Stefano Rivoir , Linux Kernel Mailing List Subject: Re: Linux 2.6.22 released Message-ID: <20070711224322.GC25510@one.firstfloor.org> References: <4693327E.6000006@gts.it> <46947ADF.4080605@gts.it> <46951892.20406@martinorr.name> <469544D3.90004@martinorr.name> <20070711221626.GA25510@one.firstfloor.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Jul 11, 2007 at 03:37:41PM -0700, Linus Torvalds wrote: > > > On Thu, 12 Jul 2007, Andi Kleen wrote: > > > > I just checked with today's gcc 4.2 (070711) freshly compiled and from a quick inspection > > the code looks correct again. So perhaps it has been already fixed? > > Did you see the breakage with the original compiler? It might be some No, i don't have a debian compiler and i normally use gcc 4.1 on my workstation. BTW I just heard that at least one package in suse STABLE (using a slightly older gcc 4.2 snapshot) apprently with inline assembly got miscompiled too > config setup or something. > > But yeah, if Debian/sid is just using random compiler snapshots of the > day, I htink we can just bury this as "pointless". > > Who the heck takes a compiler snapshot and runs with it? At least when > your kernel breaks, it seldom breaks subtly (but I would expect that most > distros would not pick random nightly kernel builds). When your compiler > breaks, you have random problems in totally unexpected places, the last > thing you want to have is a random nightly snapshot in a distro - even a > development one. I guess it's because gcc 4.2 is technically supposed to be a -stable tree. Also recompiling whole distros tends to find a lot of compiler bugs so I guess it's useful QA for them. -Andi