From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 20663 invoked by alias); 3 Sep 2002 20:38:33 -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 20656 invoked from network); 3 Sep 2002 20:38:32 -0000 Received: from unknown (HELO disaster.jaj.com) (66.93.21.106) by sources.redhat.com with SMTP; 3 Sep 2002 20:38:32 -0000 Received: (from phil@localhost) by disaster.jaj.com (8.11.4/8.11.4) id g83KcId27664; Tue, 3 Sep 2002 15:38:18 -0500 Date: Tue, 03 Sep 2002 13:38:00 -0000 From: Phil Edwards To: Joe Buck Cc: gcc@gcc.gnu.org Subject: Re: Passing -Wabi to all target libraries Message-ID: <20020903163818.A27494@disaster.jaj.com> References: <20020829131151.A31575@disaster.jaj.com> <200208292028.NAA25210@atrus.synopsys.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200208292028.NAA25210@atrus.synopsys.com>; from Joe.Buck@synopsys.com on Thu, Aug 29, 2002 at 01:28:17PM -0700 X-SW-Source: 2002-09/txt/msg00100.txt.bz2 On Thu, Aug 29, 2002 at 01:28:17PM -0700, Joe Buck wrote: > > > I worked up a patch to always build libstdc++ with -Wabi, and then thought > > that perhaps all C++-using target libraries should have this on. What are > > people's thoughts on this? > > Yes, it's a good idea. Someone else will have to do it. I have been defeated by toplevel configury. - Makefile.in sets and passes both CXXFLAGS_FOR_TARGET and LIBCXXFLAGS_FOR_TARGET, but in target libraries, sets CXXFLAGS to the first, not the second. I don't know what the second variable is meant to be used for, given that it /isn't/ used even though all the targets are libraries. - Adding -Wabi to CXXFLAGS_FOR_TARGET doesn't change anything, because the assignment on line 141 of Makefile.in is later overridden (not appended to) by a fragment inserted into the generated Makefile. I don't know which fragment or where it comes from. The variable is not documented as "this is substituted by configure," so this took a while to debug. Phil -- I would therefore like to posit that computing's central challenge, viz. "How not to make a mess of it," has /not/ been met. - Edsger Dijkstra, 1930-2002