gcj on intel mac patches

[Tom Tromey]

>>>>> "Mohan" == Mohan Embar <gnustuff@thisiscool.com> writes:

Mohan> Thanks for these. The Win32 libgcj bits look very small and
Mohan> straightforward. Tom: if you want, I can commit these to the
Mohan> trunk.

I think we should wait for Sandro's paperwork to clear.
It shouldn't be long now.

Mohan> I don't know what the status of the trunk is.

For Java we're still putting in changes on the trunk.
I believe 4.1 is frozen at the moment.

This patch is small enough that you could put it in 4.0.x if you
like, and 4.1 when it reopens.  It is up to you.

Mohan> At one point in
Mohan> time, I'd like to commit this too if it hasn't been done already:

Mohan> - http://gcc.gnu.org/ml/java/2005-12/msg00137.html

Yeah, this is ok.  Please add an explanatory comment as mentioned in
one of the followups.

As to reading Apple's jni.h... I don't know.  You could ask FSF legal.
You could also write a little test program that prints the offset of
each of the function pointers, and then compare that to our jni.h.
That would let you know how much padding to insert.

Tom