Overriding cl package symbols
Andy Peterson
andy.arvid at gmail.com
Mon Jun 10 15:37:00 UTC 2019
parenscript does not redefine "defun", it imports cl:defun and exports it
again.
CL-USER> (eq 'ps:defun 'cl:defun )
T
Exporting has nothing to do with the package lock. it refers to when you
try to redefine cl:defun.
You can define your own defun in your own package. The library *Series *does
this but it does not export it.
from s-package.lisp
(defpackage #:series
(:use #:cl)
(:export
...
)
(:shadow
#:let #:let* #:multiple-value-bind #:funcall #:defun
#+(or cmu scl) #:collect #+(or cmu scl) #:iterate)
....
)
from s-code.lisp:
;; EXTENSION
(defmacro defun (name lambda-list &environment *env* &body body)
(if (dolist (form body)
(cond ((and (stringp form) (eq form (car body))))
((and (consp form) (eq-car form 'declare))
(if (assoc 'optimizable-series-function (cdr form)) (return
t)))
(t (return nil))))
(define-optimizable-series-fn name lambda-list body)
(progn (undefine-optimizable-series-fn name)
`(cl:defun ,name ,lambda-list
. ,body))))
you could export it, put it would generate conflicts when you use it.
SERIES> (export 'defun)
T
SERIES> (in-package :cl-user)
#<PACKAGE "COMMON-LISP-USER">
CL-USER> (use-package :series)
USE-PACKAGE #<PACKAGE "SERIES"> causes name-conflicts in
#<PACKAGE "COMMON-LISP-USER"> between the following symbols:
SERIES:DEFUN, COMMON-LISP:DEFUN
[Condition of type NAME-CONFLICT]
See also:
Common Lisp Hyperspec, 11.1.1.2.5 [:section]
; Evaluation aborted on #<NAME-CONFLICT {1002576393}>.
CL-USER>
I hope this clarifies it for you.
andy peterson
On Sun, 9 Jun 2019 at 15:39, James Gunn <jgunn987 at gmail.com> wrote:
> Hello,
>
> I have been reading the source code for parenscript and I am keen to find
> out how you override locked symbols? In the parenscript package there are
> exports such as "#:defun". I have tried myself to create my own package
> that exports defun but i get the following error message(using SBCL)
>
> "Lock on package COMMON-LISP violated when proclaiming DEFUN as a function"
>
> I'd be very interested in finding out how you got around this.
>
> Thank you,
>
> James Gunn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/parenscript-devel/attachments/20190610/0bdf0aec/attachment.html>
More information about the parenscript-devel
mailing list