libjava build failure on branch for sparc-sun-solaris2.7

[Jeff Sturm]

On 23 May 2001, Tom Tromey wrote:
> Jeff> FAIL: Invoke_1 -O execution from bytecode->native test
> 
> These are probably problems with exception handling.

Yes, SIGSEGV handling in this case.

> Jeff> FAIL: Thread_Alive -O execution from bytecode->native test
> 
> I don't know what these are.

Probably EH too.  This failure is within a monitor cleanup during an
unwind.  It is as though local variables aren't restored.

Both need some investigation.

> Yes, please check this in.  Thanks.

Done.

> While this is ugly, I don't have a big problem with it even for the
> trunk.  As long as it is documented, I don't see why it would be a
> problem.  I think we'll see problems like this on various platforms
> where a Java method happens to have the same name as some C macro.

I haven't built the trunk recently.  If it turns out to need this patch,
shall I just commit it?

Jeff