gerd.moellmann@... (Gerd Moellmann) writes:
> Antonio Martinez <tonyms@...> writes:
>> I also asked whether it was possible/desirable to store the generic
>> function object in the slot definition rather than the symbol, to try
>> and interact better with symbol-function-replacement mechanisms
>> (profile, memo, trace, etc.). /Is/ this possible/desirable?
> I think it should be possible, but interestingly, pasting the code
> from bug 241 into CMUCL HEAD results in no error, so it seems
> something fixed that in CMUCL.
> I'm not sure what exactly was the fix, but I think this having to do
> with symbols changing packages might point to the use of generalized
> function names for various stuff in CMUCL's PCL. Does SBCL's PCL
> use generalized function names?
Not completely, but to a limited extent, yes. However, ISTR that you
special cased profile/unprofile to cope with this, did you not? Last
time we discussed this bug, I think you even gave us this snippet:
(defun gdefinition (name)
(let ((fdefn (fdefinition name))
(info (gethash name profile::*profile-info*)))
(if (and info
(eq fdefn (profile::profile-info-new-definition info)))
(10th May, sbcl-devel; "plus some changes of FDEFINITION to
GDEFINITION", you said).
http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757
(set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b)))
(defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge)