Am 24.07.2017 um 05:41 schrieb Al Petrofsky:

As long as implementors are tidying up this kind of thing, I suggest they also adopt the rule that in an r7rs syntax-rules form that specifies the identifier that will be interpreted in the rules as an ellipsis, it is an error if the same (bound-identifier=?) identifier is simultaneously specified to be interpreted as a literal.

So,

(let-syntax
    ((m (syntax-rules ... (...)
          ((m x y ...) 'ellipsis)
          ((m x ...) 'literal))))
  (m x ...))
=> error

I'm wondering whether this is actually an error by the R7RS as written. It seems to be allowed 4.3.2. It explicitly says: "A subpattern followed by <ellipsis> can match zero or more elements of the input, unless <ellipsis> appears in the
<literal>s, in which case it is matched as a literal." And, <ellipsis> is defined by: "An <ellipsis> is the identifier specified in the second form of syntax-rules, or the default identifier ... (three consecutive periods) otherwise."

So the above expression is covered by the R7RS and should evaluate to 'literal. So in the R7RS, conceptually, literals generate a syntactic bounding contour inside the conceptual contour that binds the ellipsis.

Marc


(let-syntax
    ((m (syntax-rules ::: ()
          ((m dots) (let-syntax
                        ((n (syntax-rules ... (dots)
                              ((n dots ...) 1))))
                      (n dots))))))
  (m ...))
=> 1

-al