libstdc++/4435

[Craig Rodrigues]

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&pr=4435&database=gcc

This is forwarded from PR 4473.
==================================================================================================================


               George sent me a followup, but he forgot to cc the GNATS
address.
               I've tried to reply to him since, but there's a mail loop
at his end;
               everything bounces. (The sysadmin there has been
informed.)


               > Date: Wed, 3 Oct 2001 11:32:01 -0400
               > From: george@elity.com
               > To: pme@gcc.gnu.org
               > Subject: libstdc++/4435
               > Reply-To: gsicherm@mail.elity.com
               >
               > Thanks for your response. I see now that it's the
struct _Rep
               > itself that is private, not its members.
               >
               > I set o-r on all the *.a files -- and the *.la files
too just to
               > be safe. The bug persisted.
               >
               > >From the assembler code, the reference to _S_max_size
seems to be
               > in _Rep::_S_create(). It's the same on the Sun, but
there I also
               > find what seem to be linkage declarations for
_S_max_size. It's
               > been a long time since I wrestled with C++; I can't
tell whether
               > _S_max_size really ought to be local or external.
               >
               > Maybe it's a thread problem. I didn't mention it in my
bug report,
               > but I'm getting two other undefined externals:
.pthread_mutex_lock
               > and .pthread_mutex_unlock.
               >
               > --
               > George L. Sicherman
               > work: gsicherm@elity.com
               > home: colonel@mail.monmouth.com
               >