public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
From: Albert Chin <gcc@lists.thewrittenword.com>
To: gcc@gcc.gnu.org
Subject: AIX 5.1 and g++ 3.3
Date: Fri, 30 May 2003 08:09:00 -0000	[thread overview]
Message-ID: <20030530035424.GA4958@spuckler.il.thewrittenword.com> (raw)

I've built GCC 3.3 on AIX 4.3.2 and AIX 5.1 and tried to build
gperf-3.0 against it. Building gperf succeeded on 4.3.2 but I ran into
an odd error on 5.1. We build GCC twice, once to /opt/TWWfsw/gcc33 and
again to /opt/TWWfsw/gcc33r. The second build installs only the
runtime libraries to /opt/TWWfsw/gcc33r while /opt/TWWfsw/gcc33 is a
full install. The idea behind this is that we -blibpath the runtime
library directory so the full GCC build isn't needed as a dependency
for programs that link against it.

On AIX 5.1, running the final gperf binary gave linker errors about
unresolved symbols. Running nm on libstdc++.a showed the problem:
  $ cd /opt/build/gperf-3.0/src
  $ ./gperf -h
  exec(): 0509-036 Cannot load program ./gperf because of the following
  errors:
          0509-130 Symbol resolution failed for gperf because:
          0509-136   Symbol _GLOBAL__F__GLOBAL__I__ZNSt21__ctype_abstract_baseIcED0Ev_opt_build_gcc_3.3_libstdc___v3_src_locale_inst.cchvgpjb (number 46) is not exported from

  $ cd /opt/TWWfsw/gcc33/lib
  $ nm libstdc++.a | grep _GLOBAL__F__GLOBAL__I__ | \
    grep ctype_abstract_base | \
    grep opt_build_gcc_3.3_libstdc___v3_src_locale_inst.cc
  _GLOBAL__F__GLOBAL__I__ZNSt21__ctype_abstract_baseIcED0Ev_opt_build_gcc_3.3_libstdc___v3_src_locale_inst.cchvgpjb D   536882124
  _GLOBAL__F__GLOBAL__I__ZNSt21__ctype_abstract_baseIcED0Ev_opt_build_gcc_3.3_libstdc___v3_src_locale_inst.cchvgpjb D      102860

  $ cd /opt/TWWfsw/gcc33r/lib
  $ nm libstdc++.a | grep _GLOBAL__F__GLOBAL__I__ | \
    grep ctype_abstract_base | \
    grep opt_build_gcc_3.3_libstdc___v3_src_locale_inst.cc
  _GLOBAL__F__GLOBAL__I__ZNSt21__ctype_abstract_baseIcED0Ev_opt_build_gcc_3.3_libstdc___v3_src_locale_inst.ccJEqnjb D   536882124
  _GLOBAL__F__GLOBAL__I__ZNSt21__ctype_abstract_baseIcED0Ev_opt_build_gcc_3.3_libstdc___v3_src_locale_inst.ccJEqnjb D      102860

The difference is the final 6 characters of the symbol. Is it odd that
a rebuild with a different installation prefix would have the final 6
characters different?

-- 
albert chin (china@thewrittenword.com)

             reply	other threads:[~2003-05-30  3:54 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-05-30  8:09 Albert Chin [this message]
2003-05-30 16:31 ` Mark Mitchell

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=20030530035424.GA4958@spuckler.il.thewrittenword.com \
    --to=gcc@lists.thewrittenword.com \
    --cc=gcc@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: link
Be 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).