On 8/17/05, John.Cowan <[EMAIL PROTECTED]> wrote:
> felix winkelmann scripsit:
> 
> > Q1: should `signum' be exactness-preserving? (i.e. should inexactness
> >   be contagious)
> 
> Yes, it should.  Returning an exact 0 for the signum of 0.0 would suggest
> that 0.0 can only represent an exact 0, whereas it can in fact represent
> any number less than e and greater than -e, where e is the smallest
> strictly positive flonum.

Hm. I think I'll do change it accordingly, then. 

> 
> > Q2: Does "#.<EXP>" (read-time-eval) make sense? (controlled via a
> >   parameter to avoid unexpected securty issues)
> 
> IMHO #, (SRFI-10) provides a reasonable balance between flexibility and
> security already, because it limits the possibilities to a predefined
> list of reader-constructor procedures rather than allowing arbitrary evals.
> In particular, allowing full evaluation in *data* seems particularly 
> dangerous.
> I think SRFI-10 is quite correct in saying that a mere on-off switch is
> too crude.  So I'd say don't add this.

Yes, SRFI-10 is in fact sufficient.

> 
> While I'm at it, I think it would be useful for you to advertise that SWIG
> supports Chicken, as I only found this out at the SWIG site.  People are
> probably writing their own wrappers when they could and should be using
> SWIG at least for the lower-level part of the job.
> 

Will do, thanks for the suggestion.


cheers,
felix


_______________________________________________
Chicken-users mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/chicken-users

Reply via email to