>>>>> "an" == a nuzzo <[EMAIL PROTECTED]> writes:
an> I am having a problem running LISA expert system on cmucl 18d. I have
an> traced functions to try to pinpoint the exact point of the problem where
an> I get the error:
an>
an> Type-error in KERNEL::OBJECT-NOT-TYPE-ERROR-HANDLER:
an> #<Standard-Class LISA::INFERENCE-ENGINE-OBJECT
an> {4812DC25}> is not of type CLASS
this is a bug in the CMUCL implementation of CLOS: when you're
doing metaobject programming, you should use symbols from the PCL
package, rather than than those exported from the COMMON-LISP
package. See
<URL:http://www2.cons.org/cmucl/doc/pcl-mop-hints.html>
[note: this is a mirror of the CMUCL web site that may be useful while
www.cons.org is down]
The following patch (against LISA CVS) resolves this problem, but
you'll run into another problem to do with CLASS-SLOTS on metaclasses.
I'm not sure whether this problem is in LISA or in CMUCL.
--- pkgdecl.lisp.~1.72.~ 2003-01-06 19:04:53.000000000 +0100
+++ pkgdecl.lisp 2003-01-06 21:05:50.000000000 +0100
@@ -53,6 +53,14 @@
"MAKE-INFERENCE-ENGINE" "RETE" "DUPLICATE-FACT"
"STANDARD-KB-CLASS" "USE-FANCY-ASSERT" "RULE-DEFAULT-NAME"
"SLOT-VALUE-OF-INSTANCE")
+ #+cmu ;; emarsden2003-01-06
+ (:shadowing-import-from "CLOS"
+ "FIND-CLASS"
+ "BUILT-IN-CLASS"
+ "CLASS-OF"
+ "CLASS-NAME"
+ "CLASS-SLOTS"
+ "SLOT-DEFINITION-NAME")
(:shadow "ASSERT"))
(defpackage "LISA-USER"
@@ -90,6 +98,14 @@
"ENSURE-CLASS"
"CLASS-DIRECT-SUPERCLASSES"
"CLASS-FINALIZED-P")
+ #+cmu ;; emarsden2003-01-06
+ (:shadowing-import-from "CLOS"
+ "FIND-CLASS"
+ "BUILT-IN-CLASS"
+ "CLASS-OF"
+ "CLASS-NAME"
+ "CLASS-SLOTS"
+ "SLOT-DEFINITION-NAME")
#+CMU
(:import-from "CLOS"
"FINALIZE-INHERITANCE"
--
Eric Marsden <URL:http://www.laas.fr/~emarsden/>
