[plt-scheme] (require... ) expanded from a macro

Dan Muresan skrev:
>> Due to hygiene
>>
>>    If a macro transformer inserts a binding for an identifier (variable
>>    or keyword), the identifier will in effect be renamed throughout its
>>    scope to avoid conflicts with other identifiers.
>>
>> the names bound by the require are renamed.
> 
> Thanks. How do I debug that in PLT? 

Debug what exactly?

Start with the section "Macro-Generated Top-Level and Module Definitions":

<http://download.plt-scheme.org/doc/352/html/mzscheme/mzscheme-Z-H-12.html#node_sec_12.3.5>

> I tried switching to the Expander
> language, but it doesn't seem to show me anything useful when
> expanding sample expressions in the bottom window (I get
> #<syntax...>). 

I seem to recall there was even a macro stepper for
> PLT...

Yes - get a pre-release:

     <http://pre.plt-scheme.org/installers/>

>> SRFI-55 is the new kid on the block.
>>
>>    <http://srfi.schemers.org/srfi-55/mail-archive/msg00058.html>
>>    <http://srfi.schemers.org/srfi-55/mail-archive/msg00006.html>
> 
> Those messages are from 2004. I tend to agree more with some of the
> replies in those threads, such as those from Chicken's author...
> SRFI-55 is a portable way to specify that a program depends on certain
> SRFI's. It already works in Chicken, SISC and Guile. I don't see how
> its availability could hurt users, and it seems it wouldn't clash with
> PLT's design...

The last time someone asked srfi 55 was in 2004. Maybe that's
why no one bothered to put it in. With the advent of the R6RS
module system, srfi-55 and srfi-0 will not be needed anymore.

> I'm sure you will be horrified, but from my point of view, it's like
> this: why make me learn this new weird (require) syntax, when my I
> just want to run my code in PLT?
> 
>> SRFI-0 is only a partial solution.
>>     <http://srfi.schemers.org/srfi-0/mail-archive/msg00032.html>
> 
> But it already works across several implementations (Gambit, Chicken,
> Bigloo, SISC, Guile) *today*. Again, I don't see how building it in
> would hurt users (esthetical speculations aside).

   The require-extension form is used to make an extension available in
   the toplevel scope in which it appears. The definition of a "toplevel
   scope" and the exact meaning of what it means to make an extension
   available in one is implementation-defined, ...

Part of the problem is that top-level scope is ill-defined. Furthermore
none of the mentioned systems have a module system where the
runtime- and syntax-expansion environments are seperate. You may
also want to ask around, why Scheme48 is missing from the list :-)

> Again, from a user's perspective, what can I add to my existing code
> that will only be seen by PLT? With most other implementations, at
> least I could conditionally load chicken-prelude.scm, sisc-prelude.scm
> etc from a central prelude.scm. I hope you will at least see the
> inconvenience of a missing cond-expand.

Make a run-plt.scm that contains the PLT-specific things, and then
include the normal main file.

   ; run-plt.scm
   (require (lib "include.ss"))
   (include "main.scm")

Voila!

-- 
Jens Axel Søgaard