Hello again,

SRFI 152’s ‘string-segment’ is underspecified:

* Per the Notation, ‘k’ can be zero, but there is no indication
  of what happens if it is.  Clearly this case is “an error”.
  (The sample implementation signals one.)

* The specification seems to suggest that (string-segment "" 1)
  evaluates to (""), but the sample implementation returns ().
  Is this a special case?  I suspect so, since otherwise an empty
  final segment would have to be tacked on to every segment list.
  The sample implementation is correct, I think.

I’ve attached an erratum patch fixing both of these issues.  I hope
it’s acceptable.

Wolfgang

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