[Bug c++/14607] [3.3 Regression] Duplicate symbol "vtable for node" in files div.o and env.o

["dave at hiauly1 dot hia dot nrc dot ca" <gcc-bugzilla at gcc dot gnu dot org>]

------- Additional Comments From dave at hiauly1 dot hia dot nrc dot ca  2004-06-11 22:40 -------
Subject: Re:  [3.3 Regression] Duplicate symbol "vtable for nod

>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>    Target Milestone|3.3.4                       |3.3.5

I'm not sure this milestone is reasonable.  I did quite a bit
of investigation on this problem and came to the conclusion that
on targets that don't support one-only semantics we are more or
less scuppered by the C++ "ABI" used by GCC.

I have been testing a fix using the support for COMDAT that's
present in the hpux 10.20 linker to implement one-only support.
If I go ahead with this fix on [3.3/3.4], the new libstdc++
library will not be compatible with the previously installed
libraries.  I also note that using COMDAT instead on secondary
definition symbols (as in hpux 11) will make programs compiled
under hpux 10.20 incompatible with hpux 11.  As a result, any
existing old binaries will need recompilation and linking.  On
the other hand, C++ is more or less broken for any non trivial
programs at the moment.  So, implementing one-only support
would be a significant improvement.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14607