[R6RS] draft module SRFI

[Matthew Flatt]

Thanks for the feedback!

At Thu, 04 Aug 2005 19:32:36 +0200, Michael Sperber wrote:
> - "Need better support": not a complete sentence.  It's also too
>   strong, unless the missing noun is "I" or somebody not me.

I'll revise it.

>   o The imports and exports are conceptually but not syntactically
>     separate from the body.  This permits implementations to allow
>     macro-expanding into import and export forms. 

Where the draft says "a macro cannot expand to an <impexp-form>", I'll
change it to "an error is signalled if a macro expands to an
<impexp-form>". The intent is that an implementation is non-conformant
if it allows a macro expandsion to `import' or `export'.

>     Alternatively, the module language could be syntactically
>     distinguished from the body language (by a pair of parentheses or
>     a keyword), and the module language could live in a separate
>     import phase.  This would allow macro expansion at the module
>     level, and not require full macro expansion of the body to find
>     out the exports and imports.

I'll add this to the list of issues, although I personally dislike
having a separate macro-expansion phase for the imports and exports.

I have no problem with delimiting the imports and exports from the
body, though. Anyone else in favor?

> - last issue bullet: "[...] determining the phase of an import would
>   require macro-expanding the module body." I don't understand this:
>   the phasing can influence the macro expansion, right?  Wouldn't this
>   introduce a circularity into the semantics?

Not as far as I can tell. Do you have an example in mind?

I'm not really inclined to debate it, though, since I'd rather not go
this route.

> - nit: I got slightly tripped up by the spelling of "signaled," which
>   is correct but different from R5RS's "signalled"

Hmm, my Mac ispell installation likes only "signaled", and my PC ispell
installation likes only "signalled". I'll change the draft to
"signalled" anyway.

Matthew