From mboxrd@z Thu Jan 1 00:00:00 1970 From: "John David Anglin" To: jason@cygnus.com (Jason Merrill) Cc: law@cygnus.com, gcc-patches@gcc.gnu.org, egcs-bugs@egcs.cygnus.com Subject: Re: g++-mike-eh8-C test failure Date: Wed, 03 May 2000 17:27:00 -0000 Message-id: <200005040027.UAA16050@hiauly1.hia.nrc.ca> References: X-SW-Source: 2000-05/msg00196.html > > Applications that depend on shared libraries should not use archive > > libraries to satisfy symbol imports in a shared library. > > Yes, absolutely. That's why we're talking about libgcc.sl. Are you > worried about symbols from libm? I looked at that again. It appears this isn't a problem since the pa1.1 archive version of libm is only used when the link is !shared. > > This suggests that only a shared version of libc should be used in > > applications using shared libraries. If an archive version of a > > dependent library must be used, all of its definitions should be > > explicitly exported with the -E or +e options to the linker to avoid > > multiple definitions. > > Which suggests that libstdc++.sl should be built with -E. It's worth a try. Possibly, this might cause libstdc++.sl to better advertise the symbols that it depends on from libgcc. Dave -- J. David Anglin dave.anglin@nrc.ca National Research Council of Canada (613) 990-0752 (FAX: 952-6605)