From 6751276083b97a0a4acce574c84c1c1f85fe943d Mon Sep 17 00:00:00 2001
From: Markus Scherer <markus.icu@gmail.com>
Date: Thu, 10 May 2007 20:40:35 +0000
Subject: [PATCH] ICU-5707 fix u_strFromUTF8Lenient() for 2-byte UTF-8 in last
 3 bytes of input

X-SVN-Rev: 21510
---
 icu4c/source/common/ustrtrns.c        | 4 ++--
 icu4c/source/test/cintltst/custrtrn.c | 9 ++++++++-
 2 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/icu4c/source/common/ustrtrns.c b/icu4c/source/common/ustrtrns.c
index 1801d8aeb64..5f013c10063 100644
--- a/icu4c/source/common/ustrtrns.c
+++ b/icu4c/source/common/ustrtrns.c
@@ -1,7 +1,7 @@
 /*
 ******************************************************************************
 *
-*   Copyright (C) 2001-2006, International Business Machines
+*   Copyright (C) 2001-2007, International Business Machines
 *   Corporation and others.  All Rights Reserved.
 *
 ******************************************************************************
@@ -839,7 +839,7 @@ u_strFromUTF8Lenient(UChar *dest,
             } else if(ch < 0xe0) { /* U+0080..U+07FF */
                 if(pSrc < pSrcLimit) {
                     /* 0x3080 = (0xc0 << 6) + 0x80 */
-                    *pDest++ = (UChar)(((ch & 0x1f) << 6) + *pSrc++ - 0x3080);
+                    *pDest++ = (UChar)((ch << 6) + *pSrc++ - 0x3080);
                     continue;
                 }
             } else if(ch < 0xf0) { /* U+0800..U+FFFF */
diff --git a/icu4c/source/test/cintltst/custrtrn.c b/icu4c/source/test/cintltst/custrtrn.c
index 0335a005575..90cc00ae5c4 100644
--- a/icu4c/source/test/cintltst/custrtrn.c
+++ b/icu4c/source/test/cintltst/custrtrn.c
@@ -1,6 +1,6 @@
 /********************************************************************
  * COPYRIGHT:
- * Copyright (c) 2001-2006, International Business Machines Corporation and
+ * Copyright (c) 2001-2007, International Business Machines Corporation and
  * others. All Rights Reserved.
  ********************************************************************/
 /********************************************************************************
@@ -643,6 +643,10 @@ Test_FromUTF8Lenient(void) {
         0xf0, 0x90, 0,
         0xf0, 0x90, 0x80, 0,
 
+        /* non-ASCII characters in the last few bytes */
+        0x61,  0xc3, 0x9f,  0xe0, 0xa0, 0x80, 0,
+        0x61,  0xe0, 0xa0, 0x80,  0xc3, 0x9f, 0,
+
         /* empty string */
         0,
 
@@ -664,6 +668,9 @@ Test_FromUTF8Lenient(void) {
         0xfffd, 0,
         0xfffd, 0,
 
+        0x61, 0xdf, 0x800,  0,
+        0x61, 0x800, 0xdf,  0,
+
         0,
 
         0