I suspect that "comparison function" inadvertently referred to the srfi-114 term? 

I interpret srfi-128 that equality predicate is not optional in comparators, so we don't need to worry about how key equality is determined.


On Wed, Mar 7, 2018 at 7:13 PM, Alex Shinn <xxxxxx@gmail.com> wrote:
The current draft states:

  The library (srfi 146) described in this SRFI requires comparators
  to provide a comparison function.

I'm guessing this means the ordering predicate, and that the equality
predicate is not required/used, so that key equality is effectively
determined by

  (and (not (less? key1 key2)) (not (less? key2 key1)))

?  It's worth making this explicit.

I assume for mapping-adjoin, the previous key object is preserved.
For mapping-set/replace/intern/update we should state what the
behavior is, even if that behavior is unspecified.  Note in SRFI 113,
set/bag-replace specify that the new key object is used.

--
Alex

To unsubscribe from this list please go to http://www.simplelists.com/confirm.php?u=rBAxDxCgJFjJUssha9ixdMZU57qO20An