Re: [Chicken-users] tinyclos and the repl

[Elf]

(also, if you (declare (uses eval)) , theres a (repl) procedure that will 
handle most all of this for you, along with hooks if you want to do your
own modifications... read eval.scm in the chicken source for more details. :)

im not familiar with srfi-49, but the behaviour that i think you were 
describing contradicts the (read) procedure as described in r5rs: read should
return the next scheme object available on the port, updating the port's
location to exactly one char after the character ending the object. (ie, 
the next char, not the beginning of the next object, unless the next char
is also the next object.)

what exactly are you trying to do, if i may ask?


-elf



On Thu, 24 Jul 2008, Jörg F. Wittenberger wrote:

> Am Donnerstag, den 24.07.2008, 07:19 -0700 schrieb Elf:
> > are you using the eval unit and using the repl function contained
> > therein, or
> > did you roll your own repl?  also, you should be able to require the
> > chicken-more-macros via (require 'chicken-more-macros) at the top.
> Thanks a lot, the additional (require 'chicken-more-macros) right after
> the (declare ...) did the trick.
>
> (I'm using "load" as it is, so probably from eval.)
>
> BTW: I noticed, that the SRFI-49 reference implementation (no matter how
> bad that may feel on it's own - though I have a lot of single, simple
> expressions embedded in XML attributes - there's it is a win for the
> lazy programmer) - yeah, that implementation does not play nice with
> chicken.  It assumes that it can call "read" on a port when peek-char
> returns #\( and assumes that, once "read" read the expression, can
> continue to read from the port right after the corresponding #\) .
> Unfortunately chicken's read has already read from the port until the
> next #\( - so the SRFI-49 code gets confused.  Does anybody have a nice
> Scheme reader in pure Scheme or Chicken-Scheme at hand?
>
> best regards
>
> /Jörg