[PING][PATCH]: multilib fixes for fortran and java

[Jack Howarth <howarth at bromo dot msbb dot uc dot edu>]

   Can I get a review from someone with global commit rights
on...

http://gcc.gnu.org/ml/gcc-patches/2006-11/msg00493.html

The natural candidate is Geoff since this patch represents
a completion of his addition of the multi.m4 changes from...

http://gcc.gnu.org/ml/gcc-patches/2006-10/msg00872.html

...so that the fortran and java languages can be built
as multilib on a G4 Darwin PPC machine. Also, the fix
that this patch achieves can only be tested on Darwin PPC
since it is the only platform building a 32-bit native
compiler with a 64-bit multilib on 32-bit processors
(that I know of).
   I believe this should go in both gcc trunk and 4.2
branch since it fixes the following regression. The gcc
4.2 release now builds multilib by default and causes
a build failure on 32-bit PPC processor machine unless
an additional flag, --disable-multilib, is passed to
configure. This compares to gcc 4.1.1 which doesn't
require any additional flags to be passed to configure.
             Jack