Lots of "make check" regressions

Thu Feb 15 19:13:00 GMT 2001

Alexandre Petit-Bianco writes:
> So yes, there's something wrong. I'll look into it shortly. The
> branch is affected too.

That's because I checked in a patch that lays out types that are
resolved when processing operators that require type arguments (like
instanceof and casts.)
So here's a patch, I'm testing it.
./A
2001年02月15日 Alexandre Petit-Bianco <apbianco@redhat.com>
	* expr.c (build_instanceof): Check for arrays when trying fold to
	false.
Index: expr.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/java/expr.c,v
retrieving revision 1.102
diff -u -p -r1.102 expr.c

--- expr.c	2001年02月05日 23:27:46	1.102
+++ expr.c	2001年02月16日 03:01:46
@@ -1157,7 +1157,9 @@ build_instanceof (value, type)
 		 value,
 		 boolean_true_node, boolean_false_node);
 }
- else if (DECL_P (klass) && DECL_P (valclass)
+ else if (! TYPE_ARRAY_P (type)
+	 && ! TYPE_ARRAY_P (valtype)
+	 && DECL_P (klass) && DECL_P (valclass)
 	 && ! CLASS_INTERFACE (valclass)
 	 && ! CLASS_INTERFACE (klass)
 	 && ! inherits_from_p (type, valtype)
