Raymond> While not opposing this change, I think the nice thing about
Raymond> using sloop is that everyone uses the same sloop
Raymond> implementation, unlike the maze of twisty little
Raymond> implementations, all different, that is ANSI loop.
Turning that around, I'd wonder if maxima could help in exposing CL
implementations deviating from the standard, and perhaps clarifying
what the standard is?
It would be interesting to know what implementations return for the
following:
(defun nleft (n x &optional tail)
(loop for v on (nthcdr n x)
for w on x
when (eq v tail) do (return w)
finally (return w)))
(nleft 1 '(0 1))
SLOOP says '(1), while MIT LOOP says '(0 1). I had assumed MIT LOOP is
correct but I haven't checked.
Ole