Issue #16975: Fix error handling bug in the escape-decode decoder.

serhiy-storchaka · serhiy-storchaka · commit 01b3a08f5ee6 · 2013-01-25T23:30:50.000+02:00
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
@@ -730,6 +730,50 @@ class EscapeDecodeTest(unittest.TestCase):
     def test_empty(self):
         self.assertEqual(codecs.escape_decode(""), ("", 0))
 
+    def test_raw(self):
+        for b in ''.join(map(chr, range(256))):
+            if b != '\\':
+                self.assertEqual(codecs.escape_decode(b + '0'),
+                                 (b + '0', 2))
+
+    def test_escape(self):
+        self.assertEqual(codecs.escape_decode(b"[\\\n]"), (b"[]", 4))
+        self.assertEqual(codecs.escape_decode(br'[\"]'), (b'["]', 4))
+        self.assertEqual(codecs.escape_decode(br"[\']"), (b"[']", 4))
+        self.assertEqual(codecs.escape_decode(br"[\\]"), (br"[\]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\a]"), (b"[\x07]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\b]"), (b"[\x08]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\t]"), (b"[\x09]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\n]"), (b"[\x0a]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\v]"), (b"[\x0b]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\f]"), (b"[\x0c]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\r]"), (b"[\x0d]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\7]"), (b"[\x07]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\8]"), (br"[\8]", 4))
+        self.assertEqual(codecs.escape_decode(br"[\78]"), (b"[\x078]", 5))
+        self.assertEqual(codecs.escape_decode(br"[\41]"), (b"[!]", 5))
+        self.assertEqual(codecs.escape_decode(br"[\418]"), (b"[!8]", 6))
+        self.assertEqual(codecs.escape_decode(br"[\101]"), (b"[A]", 6))
+        self.assertEqual(codecs.escape_decode(br"[\1010]"), (b"[A0]", 7))
+        self.assertEqual(codecs.escape_decode(br"[\501]"), (b"[A]", 6))
+        self.assertEqual(codecs.escape_decode(br"[\x41]"), (b"[A]", 6))
+        self.assertEqual(codecs.escape_decode(br"[\X41]"), (br"[\X41]", 6))
+        self.assertEqual(codecs.escape_decode(br"[\x410]"), (b"[A0]", 7))
+        for b in ''.join(map(chr, range(256))):
+            if b not in '\n"\'\\abtnvfr01234567x':
+                self.assertEqual(codecs.escape_decode('\\' + b),
+                                 ('\\' + b, 2))
+
+    def test_errors(self):
+        self.assertRaises(ValueError, codecs.escape_decode, br"\x")
+        self.assertRaises(ValueError, codecs.escape_decode, br"[\x]")
+        self.assertEqual(codecs.escape_decode(br"[\x]\x", "ignore"), (b"[]", 6))
+        self.assertEqual(codecs.escape_decode(br"[\x]\x", "replace"), (b"[?]?", 6))
+        self.assertRaises(ValueError, codecs.escape_decode, br"\x0")
+        self.assertRaises(ValueError, codecs.escape_decode, br"[\x0]")
+        self.assertEqual(codecs.escape_decode(br"[\x0]\x0", "ignore"), (b"[]", 8))
+        self.assertEqual(codecs.escape_decode(br"[\x0]\x0", "replace"), (b"[?]?", 8))
+
 class RecodingTest(unittest.TestCase):
     def test_recoding(self):
         f = StringIO.StringIO()
diff --git a/Misc/NEWS b/Misc/NEWS
@@ -9,6 +9,8 @@ What's New in Python 2.7.4
 Core and Builtins
 -----------------
 
+- Issue #16975: Fix error handling bug in the escape-decode decoder.
+
 - Issue #14850: Now a charmap decoder treats U+FFFE as "undefined mapping"
   in any mapping, not only in a Unicode string.
 
diff --git a/Objects/stringobject.c b/Objects/stringobject.c
@@ -726,6 +726,10 @@ PyObject *PyString_DecodeEscape(const char *s,
                              errors);
                 goto failed;
             }
+            /* skip \x */
+            if (s < end && isxdigit(Py_CHARMASK(s[0])))
+                s++; /* and a hexdigit */
+            break;
 #ifndef Py_USING_UNICODE
         case 'u':
         case 'U':
