sparc, the bug here is *really* weird.  I'm wondering if codegen isn't just
   plain broken... or if our sparc/netbsd.h is missing something it should have
   for gcc 2.8.  If anyone has a clue, I'd love to hear it.


what bug?  i can not reproduce any problem with the current egcs code
in our tree that i've seen reported (the ones christos reported).


as far as i can tell, it's working just as well as any version of GCC
has for netbsd/sparc, and better for c++ things.