bpo-22602: Raise an exception in the UTF-7 decoder for ill-formed sequences starting with "+". (GH-8741)

ZackerySpytz · serhiy-storchaka · commit e349bf23584e · 2018-08-19T07:43:38.000+03:00
The UTF-7 decoder now raises UnicodeDecodeError for ill-formed
sequences starting with "+" (as specified in RFC 2152).
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
@@ -1020,6 +1020,7 @@ def test_errors(self):
             (b'a+////,+IKw-b', 'a\uffff\ufffd\u20acb'),
             (b'a+IKw-b\xff', 'a\u20acb\ufffd'),
             (b'a+IKw\xffb', 'a\u20ac\ufffdb'),
+            (b'a+@b', 'a\ufffdb'),
         ]
         for raw, expected in tests:
             with self.subTest(raw=raw):
diff --git a/Lib/test/test_unicode.py b/Lib/test/test_unicode.py
@@ -1630,6 +1630,10 @@ def test_codecs_utf7(self):
         for c in set_o:
             self.assertEqual(c.encode('ascii').decode('utf7'), c)
 
+        with self.assertRaisesRegex(UnicodeDecodeError,
+                                    'ill-formed sequence'):
+            b'+@'.decode('utf-7')
+
     def test_codecs_utf8(self):
         self.assertEqual(''.encode('utf-8'), b'')
         self.assertEqual('\u20ac'.encode('utf-8'), b'\xe2\x82\xac')
diff --git a/Misc/NEWS.d/next/Library/2018-08-12-00-14-54.bpo-22602.ybG9K8.rst b/Misc/NEWS.d/next/Library/2018-08-12-00-14-54.bpo-22602.ybG9K8.rst
@@ -0,0 +1,3 @@
+The UTF-7 decoder now raises :exc:`UnicodeDecodeError` for ill-formed
+sequences starting with "+" (as specified in RFC 2152).  Patch by Zackery
+Spytz.
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
@@ -4479,6 +4479,11 @@ PyUnicode_DecodeUTF7Stateful(const char *s,
                 if (_PyUnicodeWriter_WriteCharInline(&writer, '+') < 0)
                     goto onError;
             }
+            else if (s < e && !IS_BASE64(*s)) {
+                s++;
+                errmsg = "ill-formed sequence";
+                goto utf7Error;
+            }
             else { /* begin base64-encoded section */
                 inShift = 1;
                 surrogate = 0;

Original file line number	Diff line number	Diff line change
`@@ -1020,6 +1020,7 @@ def test_errors(self):`
`1020`	`1020`	`(b'a+////,+IKw-b', 'a\uffff\ufffd\u20acb'),`
`1021`	`1021`	`(b'a+IKw-b\xff', 'a\u20acb\ufffd'),`
`1022`	`1022`	`(b'a+IKw\xffb', 'a\u20ac\ufffdb'),`
	`1023`	`+ (b'a+@b', 'a\ufffdb'),`
`1023`	`1024`	`]`
`1024`	`1025`	`for raw, expected in tests:`
`1025`	`1026`	`with self.subTest(raw=raw):`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	+The UTF-7 decoder now raises :exc:`UnicodeDecodeError` for ill-formed
	`2`	`+sequences starting with "+" (as specified in RFC 2152). Patch by Zackery`
	`3`	`+Spytz.`