[racket-dev] `letrec' and continuations

At Fri, 20 May 2011 14:25:54 -0700, John Clements wrote:
> Perhaps this goes without saying, but I'm hoping that if internal
> defines don't expand into letrec any more, that they expand into some
> similar form that has syntactically obvious scoping; I like the fact
> that the scope of letrec-declared variables is delimited by the
> syntactic letrec term.

Yes. The goal as I see it is to preserve the syntax and scoping of
internal definitions (since lots of code relies on that), but to adjust
the semantics of locations for internal-definition identifiers (which
no code likely relies on).