[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] Division by zero

On Thu, Aug 25, 2011 at 3:16 AM, Peter Bex <Peter.Bex@x> wrote:
> For the  floor/ ... euclidian-remainder, the document mentions that
> n2 *should not be* 0, but it doesn't really say what happens when it is.
> Since the standard now knows about exceptions, why not specify that an
> exception should be signaled?

This is how additional domain restrictions beyond
the naming conventions ("n" being an integer) are
handled (although "n2 should not be 0" should be the
first sentence of the paragraph).  The description
of the entry format in the "Notation and Terminology"
section describes these all as errors.  I think requiring
errors to be signalled in all of these cases is too big
a change, and prevents many useful extensions, but
we could consider it when and if we get a standard
condition system.


Scheme-reports mailing list