[Python-checkins] cpython (merge 3.3 -> default): Issue #19279: UTF-7 decoder no more produces illegal strings.

[serhiy.storchaka]

http://hg.python.org/cpython/rev/7dde9c553f16
changeset:   86478:7dde9c553f16
parent:      86475:b8c5964cadc1
parent:      86477:f471f2f05621
user:        Serhiy Storchaka <storchaka at gmail.com>
date:        Sat Oct 19 20:39:28 2013 +0300
summary:
  Issue #19279: UTF-7 decoder no more produces illegal strings.

files:
  Lib/test/test_codecs.py |  30 +++++++++++++++++++++++++++++
  Misc/NEWS               |   2 +
  Objects/unicodeobject.c |   2 +
  3 files changed, 34 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
@@ -820,6 +820,36 @@
             ]
         )
 
+    def test_errors(self):
+        tests = [
+            (b'a\xffb', 'a\ufffdb'),
+            (b'a+IK', 'a\ufffd'),
+            (b'a+IK-b', 'a\ufffdb'),
+            (b'a+IK,b', 'a\ufffdb'),
+            (b'a+IKx', 'a\u20ac\ufffd'),
+            (b'a+IKx-b', 'a\u20ac\ufffdb'),
+            (b'a+IKwgr', 'a\u20ac\ufffd'),
+            (b'a+IKwgr-b', 'a\u20ac\ufffdb'),
+            (b'a+IKwgr,', 'a\u20ac\ufffd'),
+            (b'a+IKwgr,-b', 'a\u20ac\ufffd-b'),
+            (b'a+IKwgrB', 'a\u20ac\u20ac\ufffd'),
+            (b'a+IKwgrB-b', 'a\u20ac\u20ac\ufffdb'),
+            (b'a+/,+IKw-b', 'a\ufffd\u20acb'),
+            (b'a+//,+IKw-b', 'a\ufffd\u20acb'),
+            (b'a+///,+IKw-b', 'a\uffff\ufffd\u20acb'),
+            (b'a+////,+IKw-b', 'a\uffff\ufffd\u20acb'),
+        ]
+        for raw, expected in tests:
+            with self.subTest(raw=raw):
+                self.assertRaises(UnicodeDecodeError, codecs.utf_7_decode,
+                                raw, 'strict', True)
+                self.assertEqual(raw.decode('utf-7', 'replace'), expected)
+
+    def test_nonbmp(self):
+        self.assertEqual('\U000104A0'.encode(self.encoding), b'+2AHcoA-')
+        self.assertEqual('\ud801\udca0'.encode(self.encoding), b'+2AHcoA-')
+        self.assertEqual(b'+2AHcoA-'.decode(self.encoding), '\U000104A0')
+
 class UTF16ExTest(unittest.TestCase):
 
     def test_errors(self):
diff --git a/Misc/NEWS b/Misc/NEWS
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -10,6 +10,8 @@
 Core and Builtins
 -----------------
 
+- Issue #19279: UTF-7 decoder no more produces illegal strings.
+
 - Issue #16612: Add "Argument Clinic", a compile-time preprocessor for
   C files to generate argument parsing code.  (See PEP 436.)
 
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -4341,6 +4341,7 @@
                     Py_UCS4 outCh = (Py_UCS4)(base64buffer >> (base64bits-16));
                     base64bits -= 16;
                     base64buffer &= (1 << base64bits) - 1; /* clear high bits */
+                    assert(outCh <= 0xffff);
                     if (surrogate) {
                         /* expecting a second surrogate */
                         if (Py_UNICODE_IS_LOW_SURROGATE(outCh)) {
@@ -4408,6 +4409,7 @@
                 inShift = 1;
                 shiftOutStart = writer.pos;
                 base64bits = 0;
+                base64buffer = 0;
             }
         }
         else if (DECODE_DIRECT(ch)) { /* character decodes as itself */

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