From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 30105 invoked by alias); 12 Oct 2005 08:46:16 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 30070 invoked by uid 48); 12 Oct 2005 08:46:13 -0000 Date: Wed, 12 Oct 2005 08:46:00 -0000 Message-ID: <20051012084613.30069.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug bootstrap/24297] [4.1 Regression] libtool: link: unable to infer tagged configuration In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "bonzini at gcc dot gnu dot org" X-SW-Source: 2005-10/txt/msg01489.txt.bz2 List-Id: ------- Comment #7 from bonzini at gcc dot gnu dot org 2005-10-12 08:46 ------- I'm starting to think that the way we detect target tools is incredibly overengineered -> and hence buggy. We should just *know* that an in-tree "as" has already been compiled wherever AS_FOR_TARGET is needed! We have dependencies to ensure that, everything else is unnecessary. (Similarly for CC_FOR_TARGET, which in this PR is being misdetected as not-yet-available). I have a prototype patch to do so, but it is quite big and I have to test it really a lot. On the other hand, the patch has 3 files changed, 198 insertions(+), 367 deletions(-) and this intrigues me, especially because it removes a lot of cruft from Makefile.tpl. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24297