Re: [clisp bug] assoc code fails *only* when compiled...

[Sam S. <sd...@gn...>]

> * In message <ap31y9klv35.fsf@y.glue.umd.edu>
> * On the subject of "[clisp bug] assoc code fails *only* when compiled..."
> * Sent on 30 Jul 2002 18:17:02 -0400
> * Honorable "D. Goel" <de...@gl...> writes:
>
> [1] Here's a small piece of code, which works great in lisp, compiles
> fine, with no warnings or errors, but gives bizarre untraceable errors
> when running the compiled code.  IMHO, either it should run fine when
> compiled, or else---> it should break when uncompiled as well, and
> should give warnings or errors when being compiled.  FWIW, franz's CL
> runs this code just when even when compiled.  The code is:
> 
> (assoc (list 1 2) (list (list 1 2) 3) :test #'equal :test #'equal)
> 
> 
> BTW, the error message was really even more bizarre for the way i had
> actually implemeted the code ---- i had defined an assoc-equal --whcih
> was like assoc, except that the default was equal.  Then, i had called
> this assoc-equal with arguments :test #equal. 

patch appended -- thanks for the bug report!

-- 
Sam Steingold (http://www.podval.org/~sds) running RedHat7.3 GNU/Linux
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html>
Linux - find out what you've been missing while you've been rebooting Windows


--- compiler.lisp.~1.91.~	Fri Aug  2 17:53:37 2002
+++ compiler.lisp	Fri Aug  2 18:32:18 2002
@@ -3863,11 +3863,17 @@
                            ;; and the Slot is still empty
                            ;; for-value = NIL: ALLOW-allowed Keyword
                            ;; or Slot already filled
+                           (unless for-value
+                             (c-warn
+                              (TEXT "~S: ignored duplicate keyword ~S ~S")
+                              fun key arg))
                            (let* ((*stackz* (cons 0 *stackz*)) ; 0 will be replaced later
                                   (anode (c-form arg (if for-value
                                                          'ONE 'NIL))))
                              (seclass-or-f seclass anode)
-                             (push (list t (second tripel) anode *stackz*) L))
+                             (push (list t (second tripel) anode *stackz*
+                                         for-value)
+                                   L))
                            (setf (third tripel) nil)))
                         (nreverse L))))
                 (let ((depth1 0)
@@ -3942,7 +3948,7 @@
                           ;; correct stack-depth
                           (setf (first (fourth anodeetc)) depth-now)
                           (push (third anodeetc) codelist)
-                          (when (second anodeetc)
+                          (when (and (second anodeetc) (fifth anodeetc))
                             (push `(STORE ,(- (second anodeetc) depth2))
                                   codelist)))))
                     ;; now codelist-from-end: