From owner-freebsd-hackers Sat May 18 20:09:43 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id UAA02809 for hackers-outgoing; Sat, 18 May 1996 20:09:43 -0700 (PDT) Received: from xs1.simplex.nl (xs1.simplex.NL [193.78.46.10]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id UAA02792 for ; Sat, 18 May 1996 20:09:39 -0700 (PDT) X-Organisation-1: Simplex Networking Amsterdam X (Inter)Network X-Organisation-2: Kruislaan 419-38a 1098 VA Amsterdam X Solutions & X-Organisation-3: tel:+31(20)-6932433 fax:+31(20)-6685486 X Access Provider Received: (from uucp@localhost) by xs1.simplex.nl (8.7.5/8.7.3-RS) with UUCP id FAA06527 for freebsd-hackers@freebsd.org; Sun, 19 May 1996 05:09:37 +0200 (MET DST) Received: (from plm@localhost) by plm.simplex.nl (8.7.5/8.7.3) id JAA16166; Sat, 18 May 1996 09:55:23 +0200 (MET DST) To: freebsd-hackers@freebsd.org Subject: Re: internal compiler error References: <87loirxkdu.fsf@totally-fudged-out-message-id> From: Peter Mutsaers Date: 18 May 1996 09:55:21 +0200 In-Reply-To: "Gary Palmer"'s message of Fri, 17 May 1996 07:44:57 +0100 Message-ID: <87afz6o1p2.fsf@plm.simplex.nl> Lines: 36 X-Mailer: September Gnus v0.85/Emacs 19.30 Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk >> On Fri, 17 May 1996 07:44:57 +0100, "Gary Palmer" >> said: P> Peter Mutsaers wrote in message ID P> <87loisb709.fsf@plm.simplex.nl>: >> But the newer versions are quite a bit faster. I used to use Linux >> until recently and it uses 2.7.2 for everything, so it must be >> possible for FreeBSD too. It is a pity to ignore such easy performance >> gain. P> It's POSSIBLE. Just a LOT of work to ensure that any new 2.7.x P> ``features'' (read: bugs) don't cause problems for the LARGE ammount P> of code in FreeBSD (Remember, the /usr/src is on the order of 120Mb's P> ...) P> It's about the only advantage to the way Linux is packaged (i.e. the P> kernel is one bit, and you put the rest of what you want around P> it). It's a lot less work for it to be migrated to a new P> compiler, as the rest of the bits don't NECESSARILY have to be P> moved at the same time to using the same compiler version. But usuaaly with a new compiler version at least libc and the kernel are recompiled using it. So it makes not much difference. P> Since it is likely that there will be ANOTHER release of GCC before P> 2.2 is out the door (and well before the code freeze starts) I'd P> probably be more inclined to wait a while and take the new compiler P> (2.7.3 I believe) if that is the road that is chosen. I agree that if 2.7.3 appears soon that has all bugfixes from the intermediate (2.7.2.x) versions it is better to wait. -- ______________________________________________________________________ Peter Mutsaers | Abcoude (Utrecht), | "Quod licet bovis, plm@simplex.nl | the Netherlands | non licet Jovi."