Re: defining module from macro

On Thu, Aug 17, 2006 at 06:23:42AM -0700, Dan wrote:
> It turns out that the error appears even without a
> macro; the following loads but doesn't sc-compile:
> 
> (begin
>   (module tmpmod (tmp)
>     (define (tmp) 1))
>   (module mod (f)
>     (import tmpmod)
>     (define (f x) x)))
> 
> Surely this is not a feature... It does compile in
> chicken, for instance.

The problem is that compile-file does not evaluate syntactic definitions 
as it compiles them, as "eval" in SISC does.  So, since that is one 
expression, it will attempt to perform the import when syntax-expanding 
it, even though tmpmod isn't defined until after the expression is 
*evaluated*.  If the two were not in the same expression there wouldn't 
be this problem.

> 
> On another note, is there a macroexpand function that
> doesn't insert random annotation stuff (so not
> sc-expand)? If there isn't, please add one. With all
> the problems that Scheme macros have (see my earlier
> question about meta and psyntax v. 6.9c), at least a
> two-pass compilation procedure (first expand, then
> compile) might be viable.

You can do this by turning off annotations, eg:

(emit-annotations #f)

To create an sc-expand that Just Works that way, use SRFI-39's 
parameterize (supported in SISC without any imports):

(define (syntax-expand expr)
  (parameterize ((emit-annotations #f))
    (sc-expand expr)))


-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642