From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <java-patches-return-15794-listarch-java-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 11563 invoked by alias); 19 Jan 2006 18:13:33 -0000
Received: (qmail 11555 invoked by uid 22791); 19 Jan 2006 18:13:32 -0000
X-Spam-Check-By: sourceware.org
Received: from mx1.redhat.com (HELO mx1.redhat.com) (66.187.233.31)     by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 19 Jan 2006 18:13:28 +0000
Received: from int-mx1.corp.redhat.com (int-mx1.corp.redhat.com [172.16.52.254]) 	by mx1.redhat.com (8.12.11/8.12.11) with ESMTP id k0JIDRBb017070 	for <java-patches@gcc.gnu.org>; Thu, 19 Jan 2006 13:13:27 -0500
Received: from opsy.redhat.com (vpn50-124.rdu.redhat.com [172.16.50.124]) 	by int-mx1.corp.redhat.com (8.11.6/8.11.6) with ESMTP id k0JIDQ113441; 	Thu, 19 Jan 2006 13:13:26 -0500
Received: by opsy.redhat.com (Postfix, from userid 500) 	id E35062DC1A0; Thu, 19 Jan 2006 11:09:23 -0700 (MST)
To: Java Patch List <java-patches@gcc.gnu.org>
Subject: [gcjx] Patch: FYI: use canonical NaN values
From: Tom Tromey <tromey@redhat.com>
Reply-To: tromey@redhat.com
X-Attribution:  Tom
Date: Thu, 19 Jan 2006 18:13:00 -0000
Message-ID: <m3mzhsccoc.fsf@localhost.localdomain>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-IsSubscribed: yes
Mailing-List: contact java-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:java-patches-subscribe@gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/java-patches/>
List-Post: <mailto:java-patches@gcc.gnu.org>
List-Help: <mailto:java-patches-help@gcc.gnu.org>, <http://gcc.gnu.org/ml/#faqs>
Sender: java-patches-owner@gcc.gnu.org
X-SW-Source: 2006-q1/txt/msg00080.txt.bz2

I'm checking this in on the gcjx branch.

The japi run on builder.classpath.org pointed out a buglet in gcjx's
class generation code.  It wasn't emitting the canonical NaN values
into the constant pool.  Fixed as appended.

Tom

Index: ChangeLog
from  Tom Tromey  <tromey@redhat.com>

	* bytecode/outpool.cc (add(jdouble)): Use canonical NaN
	representation.
	(add(jfloat)): Likewise.

Index: bytecode/outpool.cc
===================================================================
--- bytecode/outpool.cc	(revision 108950)
+++ bytecode/outpool.cc	(working copy)
@@ -1,6 +1,6 @@
 // A constant pool being generated.
 
-// Copyright (C) 2004, 2005 Free Software Foundation, Inc.
+// Copyright (C) 2004, 2005, 2006 Free Software Foundation, Inc.
 //
 // This file is part of GCC.
 //
@@ -67,7 +67,7 @@
 template<typename T>
 int
 outgoing_constant_pool::check_qual (std::map<T, type_map_type> &the_map,
-				  model_type *the_type, T the_value)
+				    model_type *the_type, T the_value)
 {
   typename std::map<T, type_map_type>::iterator iter
     = the_map.find (the_value);
@@ -268,7 +268,12 @@
 int
 outgoing_constant_pool::add (jfloat val)
 {
-  jint w = float_to_word (val);
+  // Use Java's canonical NaN value.
+  jint w;
+  if (isnan (val))
+    w = 0x7fc00000;
+  else
+    w = float_to_word (val);
   int index = check (float_map, w);
   if (index < 0)
     {
@@ -286,7 +291,16 @@
 outgoing_constant_pool::add (jdouble val)
 {
   jint words[2];
-  double_to_words (words, val);
+
+  // Use Java's canonical NaN value.
+  if (isnan (val))
+    {
+      words[0] = 0x7ff80000;
+      words[1] = 0;
+    }
+  else
+    double_to_words (words, val);
+
   // Note that order of words here doesn't matter, since this
   // representation is for internal use only.
   jlong lval = words_to_long (words[0], words[1]);