On 2020-09-02 17:11 -0400, John Cowan wrote:
> > (define (string-range s)
> >   (vector-range (string->vector s)))
> >
>
> This requires that string-range be O(n).  Added.

This is already in the current draft of the SRFI document.  You may
want to do a merge, then change whatever you don't like.

> On testing: I think we should make sure that everything works with ranges
> of length 0.

Yes.  I may also add some more tests to try and ensure that an
implementation can't turn an empty range into a range with elements,
since this violates the principle that all empty ranges are equal.
The sample implementation consistently avoids this, I believe.

--
Wolfgang Corcoran-Mathe  <xxxxxx@sigwinch.xyz>

"In the presence of the unknown all have an equal right to
think." --Robert G. Ingersoll, "Individuality"