[proposed but not yet applied fix for a typo in the Java FE <mikpe@it.uu.se>]

Date: Thu, 11 Oct 2012 23:13:21 -0700
From: Andrew Haley <aph at redhat dot com>
Subject: Re: possible typo in gcc/java/expr.c at line 1053
List-Archive: <http://gcc.gnu.org/ml/gcc/>

On 10/11/2012 06:59 PM, Kenneth Zadeck wrote:
> this code looks bogus, i think that the "== INTEGER_CST" needs to disappear.

> 
> tree
> build_newarray (int atype_value, tree length)
> {
>    tree type_arg;
> 
>    tree prim_type = decode_newarray_type (atype_value);
>    tree type
>      = build_java_array_type (prim_type,
>                   host_integerp (length, 0) == INTEGER_CST
>                   ? tree_low_cst (length, 0) : -1);
> 

Thanks, you're right.  This bug was introduced by Kenner in March 2000!

Andrew.

--- gcc-4.6.3/gcc/java/expr.c.~1~	2010-11-30 12:41:24.000000000 +0100
+++ gcc-4.6.3/gcc/java/expr.c	2012-10-13 10:03:21.000000000 +0200
@@ -1044,7 +1044,7 @@ build_newarray (int atype_value, tree le
   tree prim_type = decode_newarray_type (atype_value);
   tree type
     = build_java_array_type (prim_type,
-			     host_integerp (length, 0) == INTEGER_CST
+			     host_integerp (length, 0)
 			     ? tree_low_cst (length, 0) : -1);
 
   /* Pass a reference to the primitive type class and save the runtime