From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 1753 invoked by alias); 17 Nov 2004 23:17:15 -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 1229 invoked from network); 17 Nov 2004 23:17:08 -0000 Received: from unknown (HELO mail.codesourcery.com) (65.74.133.9) by sourceware.org with SMTP; 17 Nov 2004 23:17:08 -0000 Received: (qmail 22583 invoked from network); 17 Nov 2004 23:17:08 -0000 Received: from localhost (HELO digraph.polyomino.org.uk) (joseph@127.0.0.1) by mail.codesourcery.com with DES-CBC3-SHA encrypted SMTP; 17 Nov 2004 23:17:08 -0000 Received: from jsm28 (helo=localhost) by digraph.polyomino.org.uk with local-esmtp (Exim 4.42) id 1CUZ2w-0006eV-5X; Wed, 17 Nov 2004 23:17:06 +0000 Date: Wed, 17 Nov 2004 23:31:00 -0000 From: "Joseph S. Myers" X-X-Sender: jsm28@digraph.polyomino.org.uk To: Nicolas Roard cc: Gregory John Casamento , Lars Sonchocky-Helldorf , Mike Stump , gcc@gcc.gnu.org, discuss-gnustep@gnu.org Subject: Re: Is ObjC++ still in time for 4.0? In-Reply-To: <9F62E546-38EB-11D9-B4B5-0003934632AA@roard.com> Message-ID: References: <20041117224033.7788.qmail@web41601.mail.yahoo.com> <9F62E546-38EB-11D9-B4B5-0003934632AA@roard.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-SW-Source: 2004-11/txt/msg00588.txt.bz2 On Wed, 17 Nov 2004, Nicolas Roard wrote: > And now it seems that the problems are becoming more political than technical > ?? The problems are technical. It is for the ObjC++ maintainers to produce patches that do not adversely affect the memory consumption, compile time performance or maintainability of the C or C++ front ends. These are technical issues. If existing slots in datastructures are to be reused, then all uses of those datastructures in the existing front ends need analysing to make sure this is safe (probably with checking added to the accessors to ensure it remains safe). If new slots are added, statistics of performance on real code are needed to show no statistically significant adverse effect. Similarly, if conditionals are added anywhere someone thinks might be a hot spot, profiling results are needed to show there is no performance impact. -- Joseph S. Myers http://www.srcf.ucam.org/~jsm28/gcc/ jsm@polyomino.org.uk (personal mail) joseph@codesourcery.com (CodeSourcery mail) jsm28@gcc.gnu.org (Bugzilla assignments and CCs)