Close #17839: support bytes-like objects in base64 module

This mostly affected the encodebytes and decodebytes function
(which are used by base64_codec)

Also added a test to ensure all bytes-bytes codecs can handle
memoryview input and tests for handling of multidimensional
and non-bytes format input in the modern base64 API.

3 files changed, 129 insertions, 39 deletions

diff --git a/Lib/base64.py b/Lib/base64.py
index 9c15752..0a93f2e 100755
--- a/Lib/base64.py
+++ b/Lib/base64.py
@@ -35,11 +35,13 @@ def _bytes_from_decode_data(s):
             return s.encode('ascii')
         except UnicodeEncodeError:
             raise ValueError('string argument should contain only ASCII characters')
-    elif isinstance(s, bytes_types):
+    if isinstance(s, bytes_types):
         return s
-    else:
-        raise TypeError("argument should be bytes or ASCII string, not %s" % s.__class__.__name__)
-
+    try:
+        return memoryview(s).tobytes()
+    except TypeError:
+        raise TypeError("argument should be a bytes-like object or ASCII "
+                        "string, not %r" % s.__class__.__name__) from None
 
 
 # Base64 encoding/decoding uses binascii
@@ -54,14 +56,9 @@ def b64encode(s, altchars=None):
 
     The encoded byte string is returned.
     """
-    if not isinstance(s, bytes_types):
-        raise TypeError("expected bytes, not %s" % s.__class__.__name__)
     # Strip off the trailing newline
     encoded = binascii.b2a_base64(s)[:-1]
     if altchars is not None:
-        if not isinstance(altchars, bytes_types):
-            raise TypeError("expected bytes, not %s"
-                            % altchars.__class__.__name__)
         assert len(altchars) == 2, repr(altchars)
         return encoded.translate(bytes.maketrans(b'+/', altchars))
     return encoded
@@ -149,7 +146,7 @@ def b32encode(s):
     s is the byte string to encode.  The encoded byte string is returned.
     """
     if not isinstance(s, bytes_types):
-        raise TypeError("expected bytes, not %s" % s.__class__.__name__)
+        s = memoryview(s).tobytes()
     leftover = len(s) % 5
     # Pad the last quantum with zero bits if necessary
     if leftover:
@@ -250,8 +247,6 @@ def b16encode(s):
 
     s is the byte string to encode.  The encoded byte string is returned.
     """
-    if not isinstance(s, bytes_types):
-        raise TypeError("expected bytes, not %s" % s.__class__.__name__)
     return binascii.hexlify(s).upper()
 
 
@@ -306,12 +301,26 @@ def decode(input, output):
         s = binascii.a2b_base64(line)
         output.write(s)
 
+def _input_type_check(s):
+    try:
+        m = memoryview(s)
+    except TypeError as err:
+        msg = "expected bytes-like object, not %s" % s.__class__.__name__
+        raise TypeError(msg) from err
+    if m.format not in ('c', 'b', 'B'):
+        msg = ("expected single byte elements, not %r from %s" %
+                                          (m.format, s.__class__.__name__))
+        raise TypeError(msg)
+    if m.ndim != 1:
+        msg = ("expected 1-D data, not %d-D data from %s" %
+                                          (m.ndim, s.__class__.__name__))
+        raise TypeError(msg)
+
 
 def encodebytes(s):
     """Encode a bytestring into a bytestring containing multiple lines
     of base-64 data."""
-    if not isinstance(s, bytes_types):
-        raise TypeError("expected bytes, not %s" % s.__class__.__name__)
+    _input_type_check(s)
     pieces = []
     for i in range(0, len(s), MAXBINSIZE):
         chunk = s[i : i + MAXBINSIZE]
@@ -328,8 +337,7 @@ def encodestring(s):
 
 def decodebytes(s):
     """Decode a bytestring of base-64 data into a bytestring."""
-    if not isinstance(s, bytes_types):
-        raise TypeError("expected bytes, not %s" % s.__class__.__name__)
+    _input_type_check(s)
     return binascii.a2b_base64(s)
 
 def decodestring(s):
diff --git a/Lib/test/test_base64.py b/Lib/test/test_base64.py
index 13695de..54f392d 100644
--- a/Lib/test/test_base64.py
+++ b/Lib/test/test_base64.py
@@ -5,10 +5,21 @@ import binascii
 import os
 import sys
 import subprocess
-
+import struct
+from array import array
 
 
 class LegacyBase64TestCase(unittest.TestCase):
+
+    # Legacy API is not as permissive as the modern API
+    def check_type_errors(self, f):
+        self.assertRaises(TypeError, f, "")
+        self.assertRaises(TypeError, f, [])
+        multidimensional = memoryview(b"1234").cast('B', (2, 2))
+        self.assertRaises(TypeError, f, multidimensional)
+        int_data = memoryview(b"1234").cast('I')
+        self.assertRaises(TypeError, f, int_data)
+
     def test_encodebytes(self):
         eq = self.assertEqual
         eq(base64.encodebytes(b"www.python.org"), b"d3d3LnB5dGhvbi5vcmc=\n")
@@ -24,7 +35,9 @@ class LegacyBase64TestCase(unittest.TestCase):
            b"Y3ODkhQCMwXiYqKCk7Ojw+LC4gW117fQ==\n")
         # Non-bytes
         eq(base64.encodebytes(bytearray(b'abc')), b'YWJj\n')
-        self.assertRaises(TypeError, base64.encodebytes, "")
+        eq(base64.encodebytes(memoryview(b'abc')), b'YWJj\n')
+        eq(base64.encodebytes(array('B', b'abc')), b'YWJj\n')
+        self.check_type_errors(base64.encodebytes)
 
     def test_decodebytes(self):
         eq = self.assertEqual
@@ -41,7 +54,9 @@ class LegacyBase64TestCase(unittest.TestCase):
         eq(base64.decodebytes(b''), b'')
         # Non-bytes
         eq(base64.decodebytes(bytearray(b'YWJj\n')), b'abc')
-        self.assertRaises(TypeError, base64.decodebytes, "")
+        eq(base64.decodebytes(memoryview(b'YWJj\n')), b'abc')
+        eq(base64.decodebytes(array('B', b'YWJj\n')), b'abc')
+        self.check_type_errors(base64.decodebytes)
 
     def test_encode(self):
         eq = self.assertEqual
@@ -73,6 +88,38 @@ class LegacyBase64TestCase(unittest.TestCase):
 
 
 class BaseXYTestCase(unittest.TestCase):
+
+    # Modern API completely ignores exported dimension and format data and
+    # treats any buffer as a stream of bytes
+    def check_encode_type_errors(self, f):
+        self.assertRaises(TypeError, f, "")
+        self.assertRaises(TypeError, f, [])
+
+    def check_decode_type_errors(self, f):
+        self.assertRaises(TypeError, f, [])
+
+    def check_other_types(self, f, bytes_data, expected):
+        eq = self.assertEqual
+        eq(f(bytearray(bytes_data)), expected)
+        eq(f(memoryview(bytes_data)), expected)
+        eq(f(array('B', bytes_data)), expected)
+        self.check_nonbyte_element_format(base64.b64encode, bytes_data)
+        self.check_multidimensional(base64.b64encode, bytes_data)
+
+    def check_multidimensional(self, f, data):
+        padding = b"\x00" if len(data) % 2 else b""
+        bytes_data = data + padding # Make sure cast works
+        shape = (len(bytes_data) // 2, 2)
+        multidimensional = memoryview(bytes_data).cast('B', shape)
+        self.assertEqual(f(multidimensional), f(bytes_data))
+
+    def check_nonbyte_element_format(self, f, data):
+        padding = b"\x00" * ((4 - len(data)) % 4)
+        bytes_data = data + padding # Make sure cast works
+        int_data = memoryview(bytes_data).cast('I')
+        self.assertEqual(f(int_data), f(bytes_data))
+
+
     def test_b64encode(self):
         eq = self.assertEqual
         # Test default alphabet
@@ -90,13 +137,16 @@ class BaseXYTestCase(unittest.TestCase):
            b"Y3ODkhQCMwXiYqKCk7Ojw+LC4gW117fQ==")
         # Test with arbitrary alternative characters
         eq(base64.b64encode(b'\xd3V\xbeo\xf7\x1d', altchars=b'*$'), b'01a*b$cd')
-        # Non-bytes
-        eq(base64.b64encode(bytearray(b'abcd')), b'YWJjZA==')
         eq(base64.b64encode(b'\xd3V\xbeo\xf7\x1d', altchars=bytearray(b'*$')),
            b'01a*b$cd')
-        # Check if passing a str object raises an error
-        self.assertRaises(TypeError, base64.b64encode, "")
-        self.assertRaises(TypeError, base64.b64encode, b"", altchars="")
+        eq(base64.b64encode(b'\xd3V\xbeo\xf7\x1d', altchars=memoryview(b'*$')),
+           b'01a*b$cd')
+        eq(base64.b64encode(b'\xd3V\xbeo\xf7\x1d', altchars=array('B', b'*$')),
+           b'01a*b$cd')
+        # Non-bytes
+        self.check_other_types(base64.b64encode, b'abcd', b'YWJjZA==')
+        self.check_encode_type_errors(base64.b64encode)
+        self.assertRaises(TypeError, base64.b64encode, b"", altchars="*$")
         # Test standard alphabet
         eq(base64.standard_b64encode(b"www.python.org"), b"d3d3LnB5dGhvbi5vcmc=")
         eq(base64.standard_b64encode(b"a"), b"YQ==")
@@ -110,15 +160,15 @@ class BaseXYTestCase(unittest.TestCase):
            b"RUZHSElKS0xNTk9QUVJTVFVWV1hZWjAxMjM0NT"
            b"Y3ODkhQCMwXiYqKCk7Ojw+LC4gW117fQ==")
         # Non-bytes
-        eq(base64.standard_b64encode(bytearray(b'abcd')), b'YWJjZA==')
-        # Check if passing a str object raises an error
-        self.assertRaises(TypeError, base64.standard_b64encode, "")
+        self.check_other_types(base64.standard_b64encode,
+                               b'abcd', b'YWJjZA==')
+        self.check_encode_type_errors(base64.standard_b64encode)
         # Test with 'URL safe' alternative characters
         eq(base64.urlsafe_b64encode(b'\xd3V\xbeo\xf7\x1d'), b'01a-b_cd')
         # Non-bytes
-        eq(base64.urlsafe_b64encode(bytearray(b'\xd3V\xbeo\xf7\x1d')), b'01a-b_cd')
-        # Check if passing a str object raises an error
-        self.assertRaises(TypeError, base64.urlsafe_b64encode, "")
+        self.check_other_types(base64.urlsafe_b64encode,
+                               b'\xd3V\xbeo\xf7\x1d', b'01a-b_cd')
+        self.check_encode_type_errors(base64.urlsafe_b64encode)
 
     def test_b64decode(self):
         eq = self.assertEqual
@@ -141,7 +191,8 @@ class BaseXYTestCase(unittest.TestCase):
             eq(base64.b64decode(data), res)
             eq(base64.b64decode(data.decode('ascii')), res)
         # Non-bytes
-        eq(base64.b64decode(bytearray(b"YWJj")), b"abc")
+        self.check_other_types(base64.b64decode, b"YWJj", b"abc")
+        self.check_decode_type_errors(base64.b64decode)
 
         # Test with arbitrary alternative characters
         tests_altchars = {(b'01a*b$cd', b'*$'): b'\xd3V\xbeo\xf7\x1d',
@@ -160,7 +211,8 @@ class BaseXYTestCase(unittest.TestCase):
             eq(base64.standard_b64decode(data), res)
             eq(base64.standard_b64decode(data.decode('ascii')), res)
         # Non-bytes
-        eq(base64.standard_b64decode(bytearray(b"YWJj")), b"abc")
+        self.check_other_types(base64.standard_b64decode, b"YWJj", b"abc")
+        self.check_decode_type_errors(base64.standard_b64decode)
 
         # Test with 'URL safe' alternative characters
         tests_urlsafe = {b'01a-b_cd': b'\xd3V\xbeo\xf7\x1d',
@@ -170,7 +222,9 @@ class BaseXYTestCase(unittest.TestCase):
             eq(base64.urlsafe_b64decode(data), res)
             eq(base64.urlsafe_b64decode(data.decode('ascii')), res)
         # Non-bytes
-        eq(base64.urlsafe_b64decode(bytearray(b'01a-b_cd')), b'\xd3V\xbeo\xf7\x1d')
+        self.check_other_types(base64.urlsafe_b64decode, b'01a-b_cd',
+                               b'\xd3V\xbeo\xf7\x1d')
+        self.check_decode_type_errors(base64.urlsafe_b64decode)
 
     def test_b64decode_padding_error(self):
         self.assertRaises(binascii.Error, base64.b64decode, b'abc')
@@ -205,8 +259,8 @@ class BaseXYTestCase(unittest.TestCase):
         eq(base64.b32encode(b'abcd'), b'MFRGGZA=')
         eq(base64.b32encode(b'abcde'), b'MFRGGZDF')
         # Non-bytes
-        eq(base64.b32encode(bytearray(b'abcd')), b'MFRGGZA=')
-        self.assertRaises(TypeError, base64.b32encode, "")
+        self.check_other_types(base64.b32encode, b'abcd', b'MFRGGZA=')
+        self.check_encode_type_errors(base64.b32encode)
 
     def test_b32decode(self):
         eq = self.assertEqual
@@ -222,7 +276,8 @@ class BaseXYTestCase(unittest.TestCase):
             eq(base64.b32decode(data), res)
             eq(base64.b32decode(data.decode('ascii')), res)
         # Non-bytes
-        eq(base64.b32decode(bytearray(b'MFRGG===')), b'abc')
+        self.check_other_types(base64.b32decode, b'MFRGG===', b"abc")
+        self.check_decode_type_errors(base64.b32decode)
 
     def test_b32decode_casefold(self):
         eq = self.assertEqual
@@ -277,8 +332,9 @@ class BaseXYTestCase(unittest.TestCase):
         eq(base64.b16encode(b'\x01\x02\xab\xcd\xef'), b'0102ABCDEF')
         eq(base64.b16encode(b'\x00'), b'00')
         # Non-bytes
-        eq(base64.b16encode(bytearray(b'\x01\x02\xab\xcd\xef')), b'0102ABCDEF')
-        self.assertRaises(TypeError, base64.b16encode, "")
+        self.check_other_types(base64.b16encode, b'\x01\x02\xab\xcd\xef',
+                               b'0102ABCDEF')
+        self.check_encode_type_errors(base64.b16encode)
 
     def test_b16decode(self):
         eq = self.assertEqual
@@ -293,7 +349,15 @@ class BaseXYTestCase(unittest.TestCase):
         eq(base64.b16decode(b'0102abcdef', True), b'\x01\x02\xab\xcd\xef')
         eq(base64.b16decode('0102abcdef', True), b'\x01\x02\xab\xcd\xef')
         # Non-bytes
-        eq(base64.b16decode(bytearray(b"0102ABCDEF")), b'\x01\x02\xab\xcd\xef')
+        self.check_other_types(base64.b16decode, b"0102ABCDEF",
+                               b'\x01\x02\xab\xcd\xef')
+        self.check_decode_type_errors(base64.b16decode)
+        eq(base64.b16decode(bytearray(b"0102abcdef"), True),
+           b'\x01\x02\xab\xcd\xef')
+        eq(base64.b16decode(memoryview(b"0102abcdef"), True),
+           b'\x01\x02\xab\xcd\xef')
+        eq(base64.b16decode(array('B', b"0102abcdef"), True),
+           b'\x01\x02\xab\xcd\xef')
 
     def test_decode_nonascii_str(self):
         decode_funcs = (base64.b64decode,
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index 2f3cf4d..99d928d 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -2285,6 +2285,24 @@ class TransformCodecTest(unittest.TestCase):
             sout = reader.readline()
             self.assertEqual(sout, b"\x80")
 
+    def test_buffer_api_usage(self):
+        # We check all the transform codecs accept memoryview input
+        # for encoding and decoding
+        # and also that they roundtrip correctly
+        original = b"12345\x80"
+        for encoding in bytes_transform_encodings:
+            data = original
+            view = memoryview(data)
+            data = codecs.encode(data, encoding)
+            view_encoded = codecs.encode(view, encoding)
+            self.assertEqual(view_encoded, data)
+            view = memoryview(data)
+            data = codecs.decode(data, encoding)
+            self.assertEqual(data, original)
+            view_decoded = codecs.decode(view, encoding)
+            self.assertEqual(view_decoded, data)
+
+
 
 @unittest.skipUnless(sys.platform == 'win32',
                      'code pages are specific to Windows')