[closer-devel] Re: [Ecls-list] Closer to MOP for ecl

[Greg Pfeil]

On 11 May 2006, at 11:15, Gary King wrote:

> I'm not sure what to do. Some Lisps (such as OpenMCL and SBCL) define 
> eql-specializer as a class. Others (such as LispWorks) do not 
> (AFAICT). Also, I don't see eql-specializer-p being exported or 
> defined in Closer to MOP. It's used in ECL and LispWorks but not 
> defined.

In the current darcs for Closer-MOP, I see this in ecl/closer-mop.lisp:

(defun eql-specializer-p (cons)
   (and (consp cons)
        (eq (car cons) 'eql)
        (null (cddr cons))))

(deftype eql-specializer ()
   '(or eql-specializer*
        (satisfies eql-specializer-p)))

(cl:defgeneric eql-specializer-object (eql-specializer)
   (:method ((cons cons))    (if (eql-specializer-p cons)
        (cadr cons)
      (error "~S is not an eql-specializer." cons))))

and a few other definitions. I didn't look at it for other impls, 
though.

I just changed the version in Moptilities to what was in your last 
mail, and now I'm back to the cl-containers error I was getting 
yesterday:

;;; Compiling (DEFCLASS CONTENTS-AS-HASHTABLE-MIXIN ...).
;;; Warning: Class KEY-VALUE-ITERATABLE-CONTAINER-MIXIN has been 
forward referenced.
There is no method on the generic function CONTENTS that agrees on 
qualifiers NIL and specializers (#<The STANDARD-CLASS 
CONTENTS-AS-HASHTABLE-MIXIN>)
Broken at LAMBDA.

which I'm going to consider a good thing ;) Continuing our conversation 
from IRC ... I don't see a problem with the defclass at all ... it 
seems like ECL is seeing the :accessor argument, and looking up the 
method as opposed to defining it. I'm not sure why this is, as other 
DEFCLASS* forms are behaving.

(defclass* contents-as-hashtable-mixin (uses-contents-mixin
                                           
key-value-iteratable-container-mixin
                                           findable-container-mixin)
   ((contents :unbound :accessor contents)  ;; unbound for 
make-ht-for-container
    ))