Changeset 14891

Timestamp:

Jul 19, 2011, 9:07:53 AM (13 years ago)

Author:

Gary Byers

Message:

In CASE-KEY-TESTERS: don't try to reduce EQL to EQ here. That should
happen elsewhere and should deal with other types where EQL differs
from EQ. (We've had the notion of pointer-typed constants for a while;
if someone uses #. or otherwise introduces such a constant into a CASE
clause, we need to use EQL. Yes, that'd be a strange thing to do.)

File:

• trunk/source/lib/macros.lisp (modified) (1 diff)

trunk/source/lib/macros.lisp

@@ -931 +931 @@
         (warn "Duplicate keyform ~s in ~s statement." atom-or-list statement-type)
         (setq used-keys (nconc used-keys (list (cons atom-or-list t)))))
-      `((,(if (typep atom-or-list '(and number (not fixnum)))
-              'eql
-              'eq)
-         ,symbol ',atom-or-list)))
+      `((eql ,symbol ',atom-or-list)))
     (nconc (case-key-testers symbol used-keys (car atom-or-list) statement-type t)
            (when (cdr atom-or-list)