Need a little help with find_as_inner_class

Alexandre Petit-Bianco
Mon May 7 21:35:00 GMT 2001

Mo DeJong writes:

> In the case of "[ArrayBug.Inner" the first call to

Yes, because it hasn't be set properly. The patch below fixes the
problem. I believe it also fixes java/2538, and I'm going to propose
it as a solution. Regardless, I tested the patch and it looks fine,
I'm going to check it in shortly, both in the trunk and the branch.

Thank you for the time you spent investigating.


2001-05-07  Alexandre Petit-Bianco  <>

	* parse.y (build_unresolved_array_type): Set
	EXPR_WFL_QUALIFICATION on the newly created wfl.

Index: parse.y
RCS file: /cvs/gcc/gcc/gcc/java/parse.y,v
retrieving revision
diff -u -p -r1.251.2.21 parse.y
--- parse.y     2001/05/01 22:09:16
+++ parse.y     2001/05/08 02:52:48
@@ -3305,6 +3305,7 @@ build_unresolved_array_type (type_or_wfl
      tree type_or_wfl;
   const char *ptr;
+  tree wfl;
   /* TYPE_OR_WFL might be an array on a resolved type. In this case,
      just create a array type */
@@ -3316,10 +3317,12 @@ build_unresolved_array_type (type_or_wfl
                 IDENTIFIER_POINTER (EXPR_WFL_NODE (type_or_wfl)),
                 IDENTIFIER_LENGTH (EXPR_WFL_NODE (type_or_wfl)));
   ptr = obstack_finish (&temporary_obstack);
-  return build_expr_wfl (get_identifier (ptr),
-                        EXPR_WFL_FILENAME (type_or_wfl),
-                        EXPR_WFL_LINENO (type_or_wfl),
-                        EXPR_WFL_COLNO (type_or_wfl));
+  wfl = build_expr_wfl (get_identifier (ptr),
+                       EXPR_WFL_FILENAME (type_or_wfl),
+                       EXPR_WFL_LINENO (type_or_wfl),
+                       EXPR_WFL_COLNO (type_or_wfl));
+  return wfl;
 static void

More information about the Java mailing list