[patch] java run-time stack trace

[Per Bothner]

Tom Tromey <tromey@redhat.com> writes:

> I think it is fine: it is low risk, correct, useful, etc.
> c++filt is part of gcc so I don't think there is any danger with
> version skew.

Actually, when I think of it:  Invoking an external c++filt is wrong,
since C++filt is a host program, not a target program.  The correct
solution is probably to link in cplus-dem.c from xiberty.

> Could you add a PR for this?

I just did - also mentioning the host/target problem, and C++ vs Java.
It is libgcj/1907: symbol demangling by Throwable.printStackTrace
-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/~per/