[racket-dev] varrefs in teaching languages are actually top-level?

From: Stephen Chang (stchang at ccs.neu.edu)
Date: Wed Jun 29 14:53:00 EDT 2011

> I just observed something that gave me a bit of a start; it appears to me that bindings in the teaching languages are still parsed as top-level bindings, and not as module bindings.  This is weird, because I'm pretty confident that the definitions get wrapped in a module before expansion.  However, I'm seeing #f as the value for the 'identifier-binding' on variables in my beginner-language stepper test cases, and in fact in the real stepper as well.

When I was messing with the stepper, the bindings in the teaching
languages were not top-level (this was back in April). I dont know if
anything has changed since.

I did add something to the lazy language because the stepper doesnt
work with module languages, so its bindings are top-level, but I
wanted them to get a break-wrap, but this should not have affected the
beginning languages. And all the tests were passing at the time.

Posted on the dev mailing list.