diff -r 88a7b9c3b6c0 Lib/test/test_codecs.py
--- a/Lib/test/test_codecs.py	Sun Oct 07 18:04:38 2012 +0200
+++ b/Lib/test/test_codecs.py	Mon Oct 08 18:32:00 2012 +0300
@@ -313,7 +313,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u0100\uffff",
+            "\x00\xff\u0100\uffff\U00010000",
             [
                 "", # first byte of BOM read
                 "", # second byte of BOM read
@@ -335,6 +335,10 @@
                 "\x00\xff\u0100",
                 "\x00\xff\u0100",
                 "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff\U00010000",
             ]
         )
 
@@ -369,7 +373,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u0100\uffff",
+            "\x00\xff\u0100\uffff\U00010000",
             [
                 "",
                 "",
@@ -387,6 +391,10 @@
                 "\x00\xff\u0100",
                 "\x00\xff\u0100",
                 "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff\U00010000",
             ]
         )
 
@@ -409,7 +417,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u0100\uffff",
+            "\x00\xff\u0100\uffff\U00010000",
             [
                 "",
                 "",
@@ -427,6 +435,10 @@
                 "\x00\xff\u0100",
                 "\x00\xff\u0100",
                 "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff\U00010000",
             ]
         )
 
@@ -477,7 +489,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u0100\uffff",
+            "\x00\xff\u0100\uffff\U00010000",
             [
                 "", # first byte of BOM read
                 "", # second byte of BOM read => byteorder known
@@ -489,6 +501,10 @@
                 "\x00\xff\u0100",
                 "\x00\xff\u0100",
                 "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff\U00010000",
             ]
         )
 
@@ -526,7 +542,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u0100\uffff",
+            "\x00\xff\u0100\uffff\U00010000",
             [
                 "",
                 "\x00",
@@ -536,6 +552,10 @@
                 "\x00\xff\u0100",
                 "\x00\xff\u0100",
                 "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff\U00010000",
             ]
         )
 
@@ -565,7 +585,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u0100\uffff",
+            "\x00\xff\u0100\uffff\U00010000",
             [
                 "",
                 "\x00",
@@ -575,6 +595,10 @@
                 "\x00\xff\u0100",
                 "\x00\xff\u0100",
                 "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff",
+                "\x00\xff\u0100\uffff\U00010000",
             ]
         )
 
@@ -604,7 +628,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\x00\xff\u07ff\u0800\uffff",
+            "\x00\xff\u07ff\u0800\uffff\U00010000",
             [
                 "\x00",
                 "\x00",
@@ -617,6 +641,10 @@
                 "\x00\xff\u07ff\u0800",
                 "\x00\xff\u07ff\u0800",
                 "\x00\xff\u07ff\u0800\uffff",
+                "\x00\xff\u07ff\u0800\uffff",
+                "\x00\xff\u07ff\u0800\uffff",
+                "\x00\xff\u07ff\u0800\uffff",
+                "\x00\xff\u07ff\u0800\uffff\U00010000",
             ]
         )
 
@@ -690,7 +718,7 @@
 
     def test_partial(self):
         self.check_partial(
-            "\ufeff\x00\xff\u07ff\u0800\uffff",
+            "\ufeff\x00\xff\u07ff\u0800\uffff\U00010000",
             [
                 "",
                 "",
@@ -709,6 +737,10 @@
                 "\ufeff\x00\xff\u07ff\u0800",
                 "\ufeff\x00\xff\u07ff\u0800",
                 "\ufeff\x00\xff\u07ff\u0800\uffff",
+                "\ufeff\x00\xff\u07ff\u0800\uffff",
+                "\ufeff\x00\xff\u07ff\u0800\uffff",
+                "\ufeff\x00\xff\u07ff\u0800\uffff",
+                "\ufeff\x00\xff\u07ff\u0800\uffff\U00010000",
             ]
         )
 
diff -r 88a7b9c3b6c0 Objects/unicodeobject.c
--- a/Objects/unicodeobject.c	Sun Oct 07 18:04:38 2012 +0200
+++ b/Objects/unicodeobject.c	Mon Oct 08 18:32:00 2012 +0300
@@ -3572,8 +3572,11 @@
 
         /* UTF-16 code pair: */
         if (e - q < 2) {
+            q -= 2;
+            if (consumed)
+                break;
             errmsg = "unexpected end of data";
-            startinpos = (((const char *)q) - 2) - starts;
+            startinpos = ((const char *)q) - starts;
             endinpos = ((const char *)e) - starts;
             goto utf16Error;
         }