GCC 3.0 branch Java iconv linking problems

[David Edelsohn]

>>>>> Rainer Orth writes:


Rainer> David Edelsohn <dje@watson.ibm.com> writes:
>> I think the following definition of LIBICONV in Makefile.in is
>> necessary for the recent iconv change to be correct.

Rainer> Right.  The same patch is required to return GCC to bootstrap land on Tru64
Rainer> UNIX

Rainer> http://gcc.gnu.org/ml/gcc-patches/2001-07/msg01381.html

Rainer> but still hasn't been reviewed ;-(

	Well this is totally ridiculous.  I am approving this as obviously
correct and going to commit it on both the trunk and the branch.  GCC
should not remains broken for this long.

	If we need a more complicated patch for -I and -L, we can address
that later as further refinements.  Normally iconv is supplied by the
system, so the header file and library are in the default paths.

David