Re: [Chicken-hackers] CR: Division by inexact zero should return infinity/nan

[John Cowan]

Peter Bex scripsit:

> I'm working on fixing several bugs in numbers, and now I stumbled upon
> Chicken's behaviour with zero-flonums:
>
> (/  1.  -1.  0.0 0.0) => error: division by zero
>
> However, an inexact zero is not exactly zero; it could be some very
> small value unrepresentable by the internal floating-point storage.
>
> This is really annoying, as you'd get "random" errors when you're
> working with numbers in the extreme ranges representable by flonums
> even if you absolutely know you're dividing by nonzero values.

I agree on all points, and have said so in the past.

> All other Schemes I tested with (Guile, Gauche, Scheme48, Racket
> and Gambit) return +inf.0 for (/ 1.0 0.0) and -inf.0 for (/ -1.0
> 0.0).  They return +nan.0 for (/ 0.0 0.0).  Extra data points would be
> welcome.

Currently, Chicken is the only Scheme in my usual list that provides
inexact infinity but signals an error when dividing by zero.  See
http://trac.sacrideo.us/wg/wiki/Zero . Note that this page does not
show which Schemes actually provide infinities and NaNs, nor is it
necessarily up to date with the latest versions.

-- 
The first thing you learn in a lawin' family    John Cowan
is that there ain't no definite answers         address@hidden
to anything.  --Calpurnia in To Kill A Mockingbird