Versioning mess proved!!!

[Ulrich Drepper]

Geoff Keating <geoffk@cygnus.com> writes:

> [geoffk@thief geoffk]$ gcc tA2.c -o tA2
> [geoffk@thief geoffk]$ ./tA2
> (nil)
> [geoffk@thief geoffk]$ LD_PRELOAD=./tB.so ./tA2
> (nil)
> 
> on x86-linux.

X86 wasn't the problem.  At least the compilers I used never generated
the PLT.  The question is whether the PPC compiler which did generate
a PLT (since it somehow thought it's better instead of using the
pointer variable) does now stop doing this.

-- 
---------------.                          ,-.   1325 Chesapeake Terrace
Ulrich Drepper  \    ,-------------------'   \  Sunnyvale, CA 94089 USA
Red Hat          `--' drepper at redhat.com   `------------------------