Re: [Gcl-devel] Clarification re define-method-combination test

[Paul F. Dietz]

Camm Maguire wrote:
> Greetings!  Paul, could you please clarify the test below?  clisp and
> cmucl both fail as well as gcl.
>
> (PROGN
>         (EVAL '(DEFMETHOD CL-TEST::DMC-LONG-GF-01
>                    ((CL-TEST::X T) (CL-TEST::Y T)) :FOO))
>         (CL-TEST::HANDLER-CASE
>             (CL-TEST::DMC-LONG-GF-01 'CL-TEST::A 'CL-TEST::B)
>             (ERROR NIL :CAUGHT)))

The comment with the test says:

;;; The list of method groups specifiers for this method combination
;;; is empty, so no methods are valid.

Lispworks passes this test, acl 6.2, sbcl and clisp don't.
clisp is reporting an error at method definition time, but
the error is supposed to be signalled when there an applicable
method has no valid method combination, and the notion of
'applicable' method is defined only in the context of
a particular call to the generic function.

        Paul