From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 31963 invoked by alias); 4 Jul 2002 17:19:44 -0000 Mailing-List: contact gdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sources.redhat.com Received: (qmail 31910 invoked from network); 4 Jul 2002 17:19:43 -0000 Received: from unknown (HELO localhost.localdomain) (66.60.148.227) by sources.redhat.com with SMTP; 4 Jul 2002 17:19:43 -0000 Received: from warlock.codesourcery.com (localhost.localdomain [127.0.0.1]) by localhost.localdomain (8.11.6/8.11.6) with ESMTP id g64HIBa02218; Thu, 4 Jul 2002 10:18:12 -0700 Date: Thu, 04 Jul 2002 10:19:00 -0000 From: Mark Mitchell To: DJ Delorie cc: "neroden@doctormoo.dyndns.org" , "gcc@gcc.gnu.org" , "binutils@sources.redhat.com" , "gdb@sources.redhat.com" Subject: Re: configure/make/make install with moving srcdir, builddir... Message-ID: <61310000.1025803091@warlock.codesourcery.com> In-Reply-To: <200207041636.g64GacS06439@greed.delorie.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline X-SW-Source: 2002-07/txt/msg00048.txt.bz2 --On Thursday, July 04, 2002 12:36:38 PM -0400 DJ Delorie wrote: > >> I think that's fine. And if we can really simplify our makefiles that's >> worth more than being able to change the $srcdir around. We can always >> add that later if someone really, really needs it. > > What about the case where you do a build on one machine, and do "make > install" on many others with different mount points? Doesn't that > need to know where srcdir is, yet srcdir is a different location for > them? Yes -- but this is exactly the kind of thing that I think we can live without. I know people do this; I know it's convenient. This is a special case of a general problem we have with GCC. There are some processes we have that we know are hard to maintain and error-prone (build is one, test is another). On the other hand, over the years, we've beaten on these issues to the point where there's support for lots and lots of somewhat obscure usages. In order to make progress, we know we can't make an incremental change; we need to make a drastic change. (In this case, autoconfiscation.) But, it's hard to do that and preserve every last feature of the old system, all at once. (For one thing, the compiler keeps changing as you go.) I think we willing to say "Yes, autoconfiscation is good; yes, that hits the common cases; yes that will reduce our overall development burden. Do it!" And then the (relatively few) people who do "make install" onto a bunch of machines can come back and agitate later to add that feature back. The feature is a good one; I just don't think we shouldn't set the bar too high. -- Mark Mitchell mark@codesourcery.com CodeSourcery, LLC http://www.codesourcery.com