Issue #4770: Restrict binascii module to accept only bytes (as specified).

And fix the email package to encode to ASCII instead of ``raw-unicode-escape`` before ASCII-to-binary decoding.

Doc/library/binascii.rst

@@ -18,6 +18,11 @@ use these functions directly but use wrapper modules like :mod:`uu`,
 low-level functions written in C for greater speed that are used by the
 higher-level modules.
 
+.. note::
+
+   Encoding and decoding functions do not accept Unicode strings.  Only bytestring
+   and bytearray objects can be processed.
+
 The :mod:`binascii` module defines the following functions:
 
 
@@ -54,6 +59,9 @@ The :mod:`binascii` module defines the following functions:
    data. More than one line may be passed at a time. If the optional argument
    *header* is present and true, underscores will be decoded as spaces.
 
+   .. versionchanged:: 3.2
+      accept only bytestring or bytearray object as input.
+
 
 .. function:: b2a_qp(data, quotetabs=False, istext=True, header=False)
 
@@ -83,6 +91,9 @@ The :mod:`binascii` module defines the following functions:
    decompressed data, unless data input data ends in an orphaned repeat indicator,
    in which case the :exc:`Incomplete` exception is raised.
 
+   .. versionchanged:: 3.2
+      accept only bytestring or bytearray object as input.
+
 
 .. function:: rlecode_hqx(data)
 
@@ -139,6 +150,9 @@ The :mod:`binascii` module defines the following functions:
    of hexadecimal digits (which can be upper or lower case), otherwise a
    :exc:`TypeError` is raised.
 
+   .. versionchanged:: 3.2
+      accept only bytestring or bytearray object as input.
+
 
 .. exception:: Error
 
@@ -164,4 +178,3 @@ The :mod:`binascii` module defines the following functions:
 
    Module :mod:`quopri`
       Support for quoted-printable encoding used in MIME email messages.
-

Lib/email/base64mime.py

@@ -74,12 +74,12 @@ def header_encode(header_bytes, charset='iso-8859-1'):
 
 
 def body_encode(s, maxlinelen=76, eol=NL):
-    """Encode a string with base64.
+    r"""Encode a string with base64.
 
     Each line will be wrapped at, at most, maxlinelen characters (defaults to
     76 characters).
 
-    Each line of encoded text will end with eol, which defaults to "\\n".  Set
+    Each line of encoded text will end with eol, which defaults to "\n".  Set
     this to "\r\n" if you will be using the result of this function directly
     in an email.
     """

Lib/email/message.py

@@ -198,17 +198,19 @@ class Message:
             return None
         cte = self.get('content-transfer-encoding', '').lower()
         if cte == 'quoted-printable':
+            if isinstance(payload, str):
+                payload = payload.encode('ascii')
             return utils._qdecode(payload)
         elif cte == 'base64':
             try:
                 if isinstance(payload, str):
-                    payload = payload.encode('raw-unicode-escape')
+                    payload = payload.encode('ascii')
                 return base64.b64decode(payload)
             except binascii.Error:
                 # Incorrect padding
                 pass
         elif cte in ('x-uuencode', 'uuencode', 'uue', 'x-uue'):
-            in_file = BytesIO(payload.encode('raw-unicode-escape'))
+            in_file = BytesIO(payload.encode('ascii'))
             out_file = BytesIO()
             try:
                 uu.decode(in_file, out_file, quiet=True)

Lib/test/test_binascii.py

@@ -55,7 +55,7 @@ class BinASCIITest(unittest.TestCase):
                              "{!r} != {!r}".format(fb, fa, res, raw))
             self.assertIsInstance(res, bytes)
             self.assertIsInstance(a, bytes)
-            self.assertLess(max(c for c in a), 128)
+            self.assertLess(max(a), 128)
         self.assertIsInstance(binascii.crc_hqx(raw, 0), int)
         self.assertIsInstance(binascii.crc32(raw), int)
 
@@ -167,7 +167,7 @@ class BinASCIITest(unittest.TestCase):
     def test_qp(self):
         # A test for SF bug 534347 (segfaults without the proper fix)
         try:
-            binascii.a2b_qp("", **{1:1})
+            binascii.a2b_qp(b"", **{1:1})
         except TypeError:
             pass
         else:
@@ -179,12 +179,10 @@ class BinASCIITest(unittest.TestCase):
         self.assertEqual(binascii.a2b_qp(b"=00\r\n=00"), b"\x00\r\n\x00")
         self.assertEqual(
             binascii.b2a_qp(b"\xff\r\n\xff\n\xff"),
-            b"=FF\r\n=FF\r\n=FF"
-        )
+            b"=FF\r\n=FF\r\n=FF")
         self.assertEqual(
             binascii.b2a_qp(b"0"*75+b"\xff\r\n\xff\r\n\xff"),
-            b"0"*75+b"=\r\n=FF\r\n=FF\r\n=FF"
-        )
+            b"0"*75+b"=\r\n=FF\r\n=FF\r\n=FF")
 
         self.assertEqual(binascii.b2a_qp(b'\0\n'), b'=00\n')
         self.assertEqual(binascii.b2a_qp(b'\0\n', quotetabs=True), b'=00\n')
@@ -210,13 +208,15 @@ class BinASCIITest(unittest.TestCase):
             except Exception as err:
                 self.fail("{}({!r}) raises {!r}".format(func, empty, err))
 
-    def test_no_binary_strings(self):
-        # b2a_ must not accept strings
-        for f in (binascii.b2a_uu, binascii.b2a_base64,
-                  binascii.b2a_hqx, binascii.b2a_qp,
-                  binascii.hexlify, binascii.rlecode_hqx,
-                  binascii.crc_hqx, binascii.crc32):
-            self.assertRaises(TypeError, f, "test")
+    def test_unicode_strings(self):
+        # Unicode strings are not accepted.
+        for func in all_functions:
+            try:
+                self.assertRaises(TypeError, getattr(binascii, func), "test")
+            except Exception as err:
+                self.fail('{}("test") raises {!r}'.format(func, err))
+        # crc_hqx needs 2 arguments
+        self.assertRaises(TypeError, binascii.crc_hqx, "test", 0)
 
 
 class ArrayBinASCIITest(BinASCIITest):

Lib/test/test_struct.py

@@ -213,6 +213,7 @@ class StructTest(unittest.TestCase):
                     expected = '%x' % expected
                     if len(expected) & 1:
                         expected = "0" + expected
+                    expected = expected.encode('ascii')
                     expected = unhexlify(expected)
                     expected = (b"\x00" * (self.bytesize - len(expected)) +
                                 expected)

Misc/NEWS

@@ -473,6 +473,10 @@ C-API
 Library
 -------
 
+- Issue #4770: Restrict binascii module to accept only bytes (as specified).
+  And fix the email package to encode to ASCII instead of
+  ``raw-unicode-escape`` before ASCII-to-binary decoding.
+
 - Issue #9384: python -m tkinter will now display a simple demo applet.
 
 - The default size of the re module's compiled regular expression cache has

Modules/binascii.c

@@ -546,7 +546,7 @@ binascii_a2b_hqx(PyObject *self, PyObject *args)
     Py_ssize_t len;
     int done = 0;
 
-    if ( !PyArg_ParseTuple(args, "s*:a2b_hqx", &pascii) )
+    if ( !PyArg_ParseTuple(args, "y*:a2b_hqx", &pascii) )
         return NULL;
     ascii_data = pascii.buf;
     len = pascii.len;
@@ -750,7 +750,7 @@ binascii_rledecode_hqx(PyObject *self, PyObject *args)
     PyObject *rv;
     Py_ssize_t in_len, out_len, out_len_left;
 
-    if ( !PyArg_ParseTuple(args, "s*:rledecode_hqx", &pin) )
+    if ( !PyArg_ParseTuple(args, "y*:rledecode_hqx", &pin) )
         return NULL;
     in_data = pin.buf;
     in_len = pin.len;
@@ -1121,7 +1121,7 @@ binascii_unhexlify(PyObject *self, PyObject *args)
     char* retbuf;
     Py_ssize_t i, j;
 
-    if (!PyArg_ParseTuple(args, "s*:a2b_hex", &parg))
+    if (!PyArg_ParseTuple(args, "y*:a2b_hex", &parg))
         return NULL;
     argbuf = parg.buf;
     arglen = parg.len;
@@ -1199,7 +1199,7 @@ binascii_a2b_qp(PyObject *self, PyObject *args, PyObject *kwargs)
     static char *kwlist[] = {"data", "header", NULL};
     int header = 0;
 
-    if (!PyArg_ParseTupleAndKeywords(args, kwargs, "s*|i", kwlist, &pdata,
+    if (!PyArg_ParseTupleAndKeywords(args, kwargs, "y*|i", kwlist, &pdata,
           &header))
         return NULL;
     data = pdata.buf;