[plt-scheme] Generating top-level definitions from inner syntax

On Oct  9, Lauri Alanko wrote:
> 
> I think I just expect self-evaluating values to be "trivial" to
> handle, and regexps don't quite fit the definition. Compiling a
> regexp is much more complicated than just creating a similar syntax
> object, and this is where it differs from numbers and strings. It's
> just work done at the wrong phase. Of course the matter would be
> different if an actual compiled state machine were directly
> externalized, instead of just the posix string.

Well, the way Matthew did it, it looks pretty simple, if you compile a
"foo" regexp, you get something that has #rx"foo" as its external
representation...

> But this is really just a matter of taste, nothing very fundamental.

Yeah...  This is more disturbing:

| Welcome to UMB Scheme, version 3.2  [...]
|
| ==> (lambda (x) (+ x 1))
| (lambda (x)
|   (+ x 1) )

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                  http://www.barzilay.org/                 Maze is Life!