From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23760 invoked by alias); 14 Feb 2003 23:56:00 -0000 Mailing-List: contact gcc-prs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-prs-owner@gcc.gnu.org Received: (qmail 23746 invoked by uid 71); 14 Feb 2003 23:56:00 -0000 Date: Fri, 14 Feb 2003 23:56:00 -0000 Message-ID: <20030214235600.23745.qmail@sources.redhat.com> To: nobody@gcc.gnu.org Cc: gcc-prs@gcc.gnu.org, From: william@nscs.fast.net Subject: Re: target/9603: SCO 5 build of gcc-3.2.2 fails with i386.c: line 927: error: undefined symbol: DEFAULT_PCC_STRUCT_RETURN Reply-To: william@nscs.fast.net X-SW-Source: 2003-02/txt/msg00649.txt.bz2 List-Id: The following reply was made to PR target/9603; it has been noted by GNATS. From: william@nscs.fast.net To: florent.pillet@wanadoo.fr Cc: gcc-gnats@gcc.gnu.org, gcc-bugs@gcc.gnu.org, nobody@gcc.gnu.org, gcc-prs@gcc.gnu.org Subject: Re: target/9603: SCO 5 build of gcc-3.2.2 fails with i386.c: line 927: error: undefined symbol: DEFAULT_PCC_STRUCT_RETURN Date: Fri, 14 Feb 2003 18:52:15 EST I think that the problem was due using DEFAULT_PCC_STRUCT_RETURN in an os-independent file (i386.c) and not supplying definitions in all of the os-dependent places that use it. SCO OpenServer is Pentium-based. I submitted a patch for the SCO-specific header sco5.h. http://gcc.gnu.org/ml/gcc-bugs/2003-02/msg00300.html Someone else who had this problem with Solaris/Intel submitted a patch for i386.h. I think that most i386 builds eventually include i386.h eventually. http://gcc.gnu.org/ml/gcc-bugs/2003-02/msg00282.html I think that the each operating system needs a separate patch in its os-specific header because the value of the macro depends on the operating system specific rules on whether functions can return structures and unions in registers instead of memory. It is possible that iBCS2 defines this and applies to all i386 environments, so one value might work for everything. It is my theory that changes like this are someone's subtle method of finding out which build targets can be deprecated... Good luck with your Darwin-hosted i386 cross-compile! It sounds painful -- I'm afraid to ask why you need to do it. William http://williambader.com