From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 1319 invoked by alias); 10 May 2003 16:09:04 -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 1307 invoked from network); 10 May 2003 16:09:04 -0000 Received: from unknown (HELO igw2.watson.ibm.com) (129.34.20.6) by sources.redhat.com with SMTP; 10 May 2003 16:09:04 -0000 Received: from sp1n293en1.watson.ibm.com (sp1n293en1.watson.ibm.com [9.2.112.57]) by igw2.watson.ibm.com (8.11.7/8.11.4) with ESMTP id h4AG7YT205368; Sat, 10 May 2003 12:07:34 -0400 Received: from makai.watson.ibm.com (makai.watson.ibm.com [9.2.216.144] (may be forged)) by sp1n293en1.watson.ibm.com (AIX5.1/8.11.0/8.11.4) with ESMTP id h4AG7Yi55174; Sat, 10 May 2003 12:07:34 -0400 Received: from watson.ibm.com (localhost [127.0.0.1]) by makai.watson.ibm.com (AIX4.3/8.9.3/8.9.3/09-18-2002) with ESMTP id MAA26982; Sat, 10 May 2003 12:07:33 -0400 Message-Id: <200305101607.MAA26982@makai.watson.ibm.com> To: Zdenek Dvorak cc: Richard Kenner , gcc@gcc.gnu.org Subject: Re: An issue for the SC: horrible documentation quality of GCC In-Reply-To: Message from Zdenek Dvorak of "Sat, 10 May 2003 17:52:53 +0200." <20030510155253.GA2483@atrey.karlin.mff.cuni.cz> Date: Sat, 10 May 2003 16:09:00 -0000 From: David Edelsohn X-SW-Source: 2003-05/txt/msg01009.txt.bz2 >>>>> Zdenek Dvorak writes: Zdenek> The sane way Zdenek> how to write a compiler is to run these generic passes first, then Zdenek> run machine-dependent passes that may eventually fix the mistakes Zdenek> the previous optimizations have done (due to their generic ideas being Zdenek> not right in some corner cases). This way we would get approximately Zdenek> the same results, but in much more transparent way. This assumes that the mistakes can be corrected with a later machine-dependent pass, which is not always correct. Plus it takes more compile time to correct the mistakes instead of choosing the correct heuristics in the first place. The generic pass already needs to apply *some* heuristic. David