[racket] Submodules vs hygiene

From: Matthew Flatt (mflatt at cs.utah.edu)
Date: Tue Oct 21 14:28:40 EDT 2014

Yes, that's awkward. The problem is that the `quote` wrapper in
`(require (quote m))` is macro-introduced, and the whole `(quote m)`
form's context is used for the imported identifier.

In contrast, the variant

 #lang racket

 (define-syntax foo
   (syntax-rules ()
     ((_ mod quoted-mod name)
        (module mod racket/base
          (define name 1)
          (provide name))
        (require quoted-mod)

 (foo m 'm x)

works the way that you'd expect, because the whole


is supplied to the macro for using in `require`.

At Tue, 21 Oct 2014 13:19:35 -0500, Brian Mastenbrook wrote:
> I am more than a little confused by the behavior of macros that expand to 
> submodules. In the following example, I don't have any introduced names, but 
> the `x' binding created by the submodule is not usable from the enclosing 
> module. What confuses me even further is that if I (require (rename-in 'mod 
> [name name])) instead, it works; I would expect that to always be a no-op. Can 
> anyone explain this behavior?
> #lang racket
> (define-syntax foo
>   (syntax-rules ()
>     ((_ mod name)
>      (begin
>        (module mod racket/base
>          (define name 1)
>          (provide name))
>        (require 'mod)
>        name))))
> (foo m x)
> ==> x: unbound identifier in module in: x
> --
> Brian Mastenbrook
> brian at mastenbrook.net
> http:/brian.mastenbrook.net/
> ____________________
>   Racket Users list:
>   http://lists.racket-lang.org/users

Posted on the users mailing list.