From 8570761928d996713f7fceb0a61a161b09914ec1 Mon Sep 17 00:00:00 2001
From: James Roseborough <james.roseborough@luaj.org>
Date: Mon, 14 Jan 2008 23:11:09 +0000
Subject: [PATCH] Make utf-8 conversion more robust to bad input.

---
 src/core/org/luaj/vm/LString.java          | 13 ++++++++-----
 src/test/java/org/luaj/vm/LStringTest.java |  1 +
 2 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/src/core/org/luaj/vm/LString.java b/src/core/org/luaj/vm/LString.java
index 5ae6065f..8c3f0280 100644
--- a/src/core/org/luaj/vm/LString.java
+++ b/src/core/org/luaj/vm/LString.java
@@ -102,21 +102,24 @@ public class LString extends LValue {
 	 */
 	public String toJavaString() {
 		char[] c = new char[m_length];
-		int n = 0;
+		int n=0, p=0;
 		int b;
 		for ( int i=0; i<m_length; i++ ) {
 			switch ( (b = m_bytes[m_offset+i]) & 0xe0 ) {
 			default:
 				if ( b == 0 )
 					return new String( c, 0, n );
-				c[n++] = (char) (0xff & b);
+				c[p=n++] = (char) (0xff & b);
+				break;
+			case 0x80:
+			case 0xa0:
+				c[p] = (char) ((c[p] << 6) | ( b & 0x3f ));
 				break;
 			case 0xc0:
-				c[n++] = (char) (((b & 0x1f) << 6) | ( m_bytes[m_offset+(++i)] & 0x3f));
+				c[p=n++] = (char) (b & 0x1f);
 				break;
 			case 0xe0:
-				c[n++] = (char) (((b & 0xf) << 12) | ((m_bytes[m_offset+i+1] & 0x3f) << 6) | (m_bytes[m_offset+i+2] & 0x3f));
-				i += 2;
+				c[p=n++] = (char) (b & 0xf);
 				break;
 			}			
 		}
diff --git a/src/test/java/org/luaj/vm/LStringTest.java b/src/test/java/org/luaj/vm/LStringTest.java
index 2e5b4a16..e66b9485 100644
--- a/src/test/java/org/luaj/vm/LStringTest.java
+++ b/src/test/java/org/luaj/vm/LStringTest.java
@@ -76,6 +76,7 @@ public class LStringTest extends TestCase {
 		assertEquals( userFriendly( before ), userFriendly( after ) );
 		
 	}
+
 	public void testNullTerminated() {		
 		char[] c = { 'a', 'b', 'c', '\0', 'd', 'e', 'f' };
 		String before = new String(c);