public inbox for gcc-prs@sourceware.org help / color / mirror / Atom feed
From: Hallvard B Furuseth <h.b.furuseth@usit.uio.no> To: nobody@gcc.gnu.org Cc: gcc-prs@gcc.gnu.org, Subject: Re: c++/10268: C++ executables fail: libstdc++.so.5 not found Date: Mon, 31 Mar 2003 12:16:00 -0000 [thread overview] Message-ID: <20030331101601.23326.qmail@sources.redhat.com> (raw) The following reply was made to PR c++/10268; it has been noted by GNATS. From: Hallvard B Furuseth <h.b.furuseth@usit.uio.no> To: ehrhardt@mathematik.uni-ulm.de, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org, nobody@gcc.gnu.org, gcc-gnats@gcc.gnu.org Cc: Subject: Re: c++/10268: C++ executables fail: libstdc++.so.5 not found Date: Mon, 31 Mar 2003 12:14:10 +0200 ehrhardt@mathematik.uni-ulm.de writes: > This behaviour is intentional. -R adds paths to the executable and > gcc should not add system specific paths to an executable. > You can either set LD_RUN_PATH at compile time or LD_LIBRARY_PATH > at run time. *What*? Why is it better to have an executable which doesn't work than one where GCC has added a path which is needed to make it work? In particular since the user must add the same path to make it work anyway. If you don't want to add to the executable's path, I think you should only build static libraries. If one configures gcc to build dynamic libraries, add the path and warn about whatever the problem is with adding such a path, or don't add it and give a very loud warning that g++ will build executables that don't work. If you don't want to do that, I it would be better to remove the -L<path> too from g++'s defaults, at least that way the user won't think he has built a working executable just because he compiled with standard options and the compilation succeeded. And document this, and how one is supposed to make it work. Requiring users to set LD_LIBRARY_PATH at run time is a silly default, and setting environment variables doesn't work well in Makefiles anyway, but I suppose one could add this to Makefile: LDFLAGS=-R`g++ -print-file-name=libstdc++.so.5 | sed 's%/[^/]*$%%'` (Does that work if there are other libstdc++'s in the path?) Or better, some new command which doesn't require the user to understand sed and pipes. I've noticed there is a libgcc_s.so as well which only seems to be used by g++, is that always in the same location as libstdc++.so or can one need another -R for that? BTW, can I configure gcc to put that directory in the path for g++ but not for gcc, or edit the specs file to do so? -- Hallvard
next reply other threads:[~2003-03-31 10:16 UTC|newest] Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top 2003-03-31 12:16 Hallvard B Furuseth [this message] -- strict thread matches above, loose matches on Subject: below -- 2003-03-31 15:29 Phil Edwards 2003-03-31 15:16 Hallvard B Furuseth 2003-03-31 15:06 Hallvard B Furuseth 2003-03-31 14:53 Christian Ehrhardt 2003-03-30 17:44 ehrhardt 2003-03-30 11:56 Hallvard B Furuseth
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20030331101601.23326.qmail@sources.redhat.com \ --to=h.b.furuseth@usit.uio.no \ --cc=gcc-prs@gcc.gnu.org \ --cc=nobody@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).