CLISP has a problem with the routine to lookup the integrals for the Bessel
functions. The following code is the reason:
;; If form is a function then evaluate it with actual args
(or (not (functionp form))
(setq form (apply form real-args))))
In CLISP the expression (functionp (lambda (x y) ... )) gives always NIL and not
T as is expected. The evaluation is not done and later in the code we get an
Lisp Error with the unevaluated lambda expression.
For GCL the test with functionp works.
Dieter Kaiser