From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23917 invoked by alias); 17 May 2003 08:06:01 -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 23840 invoked by uid 71); 17 May 2003 08:06:00 -0000 Date: Sat, 17 May 2003 08:06:00 -0000 Message-ID: <20030517080600.23837.qmail@sources.redhat.com> To: rsandifo@sources.redhat.com Cc: gcc-prs@gcc.gnu.org, From: Richard Sandiford Subject: Re: libstdc++/10801: ld bus error on linking .libs/libstdc++ under mips-irix-6.3 Reply-To: Richard Sandiford X-SW-Source: 2003-05/txt/msg01967.txt.bz2 List-Id: The following reply was made to PR libstdc++/10801; it has been noted by GNATS. From: Richard Sandiford To: "RenE J.V. Bertin" Cc: gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org, gcc-gnats@gcc.gnu.org Subject: Re: libstdc++/10801: ld bus error on linking .libs/libstdc++ under mips-irix-6.3 Date: 17 May 2003 08:54:14 +0100 "RenE J.V. Bertin" writes: > The linker crashes when I issue the -s option > in the build command. Hmm, this sounds similar to a bug mentioned here: http://ftp.mayn.de/pub/irix/updates/MIPSpro_7.2.1.3m-README Specifically: Ld32 seg fault when linking g++ objects. When the latest g++ 2.8.1 is used with the latest 7.2.1 SGI back-end (assembler/linker/strip) any invocation of g++ with the '-s' (strip) option seems to cause a SEGV in ld32. This has been fixed (630196). I suggest you try patching or upgrading your linker. Alternatively, binutils 2.14 is likely to be released in the next couple of weeks and it should handle n32 & n64 much better than previous versions. You could try using that. (HJ's 2.14.90.0.2 version is out now and may work OK.) Using binutils is obviously more bleeding edge, though. I can't guarantee it'll work. > Things work fine when I do not do that. To me, this does > suggest that gcc does something wrong somewhere when it > invokes the linker. There doesn't seem to be anything wrong with the failing command line. And if, as you say, removing -s causes the link to succeed, I really do think that this is a linker bug. Richard