libjava test failures on some targets

[Tom Tromey]

>>>>> "Kaz" == Kaz Kojima <kkojima@rr.iij4u.or.jp> writes:

Kaz> and if it's reverted with adding #if 0 to natClassLoader.cc:
Kaz> _Jv_RegisterClassHookDefault:

Kaz> then the above testcases pass.  I don't understand what is going on
Kaz> here, though.

Is this a build from scratch?  Or a rebuild of an existing tree?

This is consistent with a rebuild of an existing tree.  After that
patch went in, the libgcj core has to be completely recompiled in
order to properly work.

Kaz> BTW, I couldn't find any PR for it in bugzilla.  Should I file it?

If that isn't it, then I don't have a theory.  File it and I'll try to
dig up an ia-64 or ppc box.

Tom