[Python-checkins] cpython (2.7): Issue #19279: UTF-7 decoder no more produces illegal unicode strings.

[serhiy.storchaka]

http://hg.python.org/cpython/rev/214c0aac7540
changeset:   86476:214c0aac7540
branch:      2.7
parent:      86462:c207ac413457
user:        Serhiy Storchaka <storchaka at gmail.com>
date:        Sat Oct 19 20:37:49 2013 +0300
summary:
  Issue #19279: UTF-7 decoder no more produces illegal unicode strings.

files:
  Lib/test/test_codecs.py |  29 +++++++++++++++++++++++++++++
  Misc/NEWS               |   2 ++
  Objects/unicodeobject.c |   2 ++
  3 files changed, 33 insertions(+), 0 deletions(-)


diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -611,6 +611,35 @@
             ]
         )
 
+    def test_errors(self):
+        tests = [
+            ('a\xffb', u'a\ufffdb'),
+            ('a+IK', u'a\ufffd'),
+            ('a+IK-b', u'a\ufffdb'),
+            ('a+IK,b', u'a\ufffdb'),
+            ('a+IKx', u'a\u20ac\ufffd'),
+            ('a+IKx-b', u'a\u20ac\ufffdb'),
+            ('a+IKwgr', u'a\u20ac\ufffd'),
+            ('a+IKwgr-b', u'a\u20ac\ufffdb'),
+            ('a+IKwgr,', u'a\u20ac\ufffd'),
+            ('a+IKwgr,-b', u'a\u20ac\ufffd-b'),
+            ('a+IKwgrB', u'a\u20ac\u20ac\ufffd'),
+            ('a+IKwgrB-b', u'a\u20ac\u20ac\ufffdb'),
+            ('a+/,+IKw-b', u'a\ufffd\u20acb'),
+            ('a+//,+IKw-b', u'a\ufffd\u20acb'),
+            ('a+///,+IKw-b', u'a\uffff\ufffd\u20acb'),
+            ('a+////,+IKw-b', u'a\uffff\ufffd\u20acb'),
+        ]
+        for raw, expected in tests:
+            self.assertRaises(UnicodeDecodeError, codecs.utf_7_decode,
+                              raw, 'strict', True)
+            self.assertEqual(raw.decode('utf-7', 'replace'), expected)
+
+    def test_nonbmp(self):
+        self.assertEqual(u'\U000104A0'.encode(self.encoding), '+2AHcoA-')
+        self.assertEqual(u'\ud801\udca0'.encode(self.encoding), '+2AHcoA-')
+        self.assertEqual('+2AHcoA-'.decode(self.encoding), u'\U000104A0')
+
 class UTF16ExTest(unittest.TestCase):
 
     def test_errors(self):
diff --git a/Misc/NEWS b/Misc/NEWS
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -9,6 +9,8 @@
 Core and Builtins
 -----------------
 
+- Issue #19279: UTF-7 decoder no more produces illegal unicode strings.
+
 - Issue #18739: Fix an inconsistency between math.log(n) and math.log(long(n));
   the results could be off from one another by a ulp or two.
 
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -1671,6 +1671,7 @@
                                        (base64buffer >> (base64bits-16));
                     base64bits -= 16;
                     base64buffer &= (1 << base64bits) - 1; /* clear high bits */
+                    assert(outCh <= 0xffff);
                     if (surrogate) {
                         /* expecting a second surrogate */
                         if (outCh >= 0xDC00 && outCh <= 0xDFFF) {
@@ -1737,6 +1738,7 @@
                 inShift = 1;
                 shiftOutStart = p;
                 base64bits = 0;
+                base64buffer = 0;
             }
         }
         else if (DECODE_DIRECT(ch)) { /* character decodes as itself */

-- 
Repository URL: http://hg.python.org/cpython