egcs 1.1.1 extremely slow

Martin v. Loewis martin@mira.isdn.cs.tu-berlin.de
Sun Jan 24 14:09:00 GMT 1999


> By this point, you've probably read Mark's and my later retractions.

Yes, it is sometimes difficult to read messages in the right order :-(

> However, I'm still not sure this is correct; as I recall, I changed
> ovl_member to use decls_match because different decls of the same
> function have to be unified.

After this discussion, I tried to change your code back, and
(apparently) nothing broke in the current test-suite (*). So I revised
my patch below.

Martin

(*) ns/using12.C and g++.jason/ctor1.C fail for different reasons.

1999-01-23  Martin von Löwis  <loewis@informatik.hu-berlin.de>

	* tree.c (ovl_member): Revert 1998-07-17 change.

Index: tree.c
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/cp/tree.c,v
retrieving revision 1.92
diff -u -p -r1.92 tree.c
--- tree.c	1999/01/21 21:16:22	1.92
+++ tree.c	1999/01/24 22:05:11
@@ -1405,9 +1405,9 @@ ovl_member (fn, ovl)
   if (ovl == NULL_TREE)
     return 0;
   if (TREE_CODE (ovl) != OVERLOAD)
-    return decls_match (ovl, fn);
+    return ovl == fn;
   for (; ovl; ovl = OVL_CHAIN (ovl))
-    if (decls_match (OVL_FUNCTION (ovl), fn))
+    if (OVL_FUNCTION (ovl) == fn)
       return 1;
   return 0;
 }



More information about the Gcc-patches mailing list