RFA: Changes to interpreter to avoid non-debugging slowdown

[Keith Seitz]

Tom Tromey wrote:

> So, I looked at JVMTI a little more.  It seems that we only need to
> know the types of locals, is that correct?

Yeah, that's all that I can see/remember...

> Are there other reasons we may want a special debug-only interpreter?
> I thought Keith said there were, but I don't recall why exactly...

Yes, there is. While breakpointing won't interfere with the speed of the 
interpreter, I expect stepping support will be a much larger impact. I 
was planning on separating this into a debug interpreter to minimize the 
impact on the non-debug case.

Keith