Re: bug w/ jade 1.3

Re: bug w/ jade 1.3

[Martin v. Loewis]

> Hi!  I think this is the same problem as appeared for blitz, but figured it
> wouldn't hurt to give you two examples of it, on the off chance there's some
> subtle difference.
> 
> As with the other, it hits the 980401 abort in instantiate_type.

In your example, the original code would have been unnecessary, since
we already have a baselink. As there is an assertion for that before,
I just remove this code. With that change, your example compiles.

Martin

Wed May 13 16:38:47 1998  Martin v. Loewis  <loewis@informatik.hu-berlin.de>

	* class.c (instantiate_type): Always assume we already have baselinks.

Index: class.c
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/cp/class.c,v
retrieving revision 1.47
diff -u -r1.47 class.c
- --- class.c	1998/05/11 07:25:20	1.47
+++ class.c	1998/05/13 14:42:28
@@ -5245,12 +5245,13 @@
 	if (lhstype == error_mark_node)
 	  return lhstype;
 
+#if 0
+        /* We already have baselinks, here. MvL */
 	if (TREE_NONLOCAL_FLAG (rhs))
 	  {
- -	    my_friendly_abort (980401);
 	    /* Got to get it as a baselink.  */
 	    rhs = lookup_fnfields (TYPE_BINFO (current_class_type),
- -				   DECL_NAME (OVL_FUNCTION (rhs)), 0);
+				   TREE_PURPOSE (rhs), 0);
 	  }
 	else
 	  {
@@ -5259,6 +5260,7 @@
 				|| TREE_CODE (TREE_VALUE (rhs)) == OVERLOAD,
 				182);
 	  }
+#endif
 
 	for (baselink = rhs; baselink;
 	     baselink = next_baselink (baselink))