Custom Query (1030 matches)

Just tried this ​https://bugs.launchpad.net/sbcl/+bug/492851 in CCL and got the same result after a while.

Here is the code:

(defgeneric blah (x))

(defun test ()
  (let ((object (make-list 10000)))
    (defmethod blah ((x (eql object)))
      (format t "~A~%" x))
    (remove-method #'blah (first (generic-function-methods #'blah))))
  (values))

Then do:

(loop (test))

..and memory use keeps rising.

There are inconsistencies in the documentation of the character sets ​http://ccl.clozure.com/ccl-documentation.html#Supported-Character-Encodings

:latin,3 instead of :latin-3 :cp68,7 instead of :cp68-8

on one hand, and on the other, inconsistencies in the naming of some character set, namely, why :latin-2 when the others are :latin1 :latin3 and :latin4 ?

cl-user> (dolist (cs '(:latin1 :latin2 :latin3 :latin4
                       :latin-1 :latin-2 :latin-3 :latin-4
                       :latin 3 
                       :cp63-7  :cp63 7))
           (handler-case 
               (progn
                 (close (open "/tmp/a.txt" :direction :probe :external-format cs))
                 (format t "~S~%" cs))
             (error (err)
               (format t "~S --> ~A~%" cs err))))
:latin1
:latin2 --> Invalid external-format: :latin2
:latin3
:latin4
:latin-1 --> Invalid external-format: :latin-1
:latin-2
:latin-3 --> Invalid external-format: :latin-3
:latin-4 --> Invalid external-format: :latin-4
:latin --> Invalid external-format: :latin
3 --> Invalid external-format: 3
:cp63-7 --> Invalid external-format: :cp63-7
:cp63 --> Invalid external-format: :cp63
7 --> Invalid external-format: 7
nil
cl-user> 

Another point, which is probably related, is that there's no introspection function to get the list of character sets and their aliases. Yes, there's a describe-character-encodings function but it gives a human readable string, that's in no way guaranteed to be parseable. I'd suggest a function returning a lisp a-list of name and aliases, so that programs can learn what encodings are supported by the current version of ccl. The documentation could then automatically be generated from this function, and no inconsistency would be to be expected.

? (decode-universal-time
    (encode-universal-time 10 10 10 1 1 2000 1787/360) 1787/360)
0
61/6
10
1
1
2000
5
NIL
1787/360

The first two values are wrong.