Bryce> OK. I was planning on changing the implementation of
Bryce> _Jv_IsInterpretedClass to use aux_id, but thinking about it more I'm
Bryce> starting to think the whole idea is not so great. If the interpreter's
Bryce> resolver, etc, were more cleanely organised then it should probably
Bryce> not even need to map from java.lang.Class to
Bryce> _Jv_InterpretedClass.
Where would we put the extra data that the interpreter needs?
E.g., a pointer to the bytecode for the method?
Just hang it off the closure and make sure that vtables are marked?
That may suffice, and it is basically the approach taken by JNI
stubs.