Fix base64-url parsing errors in email headers

Author: matthieucan

Lib/email/_encoded_words.py

@@ -113,17 +113,22 @@ def decode_b(encoded):
         # The non-alphabet characters are ignored as far as padding
         # goes, but we don't know how many there are.  So try without adding
         # padding to see if it works.
+        #
+        # We use urlsafe_b64decode here because some mailers apparently use the
+        # urlsafe b64 alphabet, and urlsafe_b64decode will correctly decode
+        # both the urlsafe and regular alphabets.
+
         try:
             return (
-                base64.b64decode(encoded, validate=False),
+                base64.urlsafe_b64decode(encoded),
                 [errors.InvalidBase64CharactersDefect()],
             )
         except binascii.Error:
             # Add as much padding as could possibly be necessary (extra padding
             # is ignored).
             try:
                 return (
-                    base64.b64decode(encoded + b'==', validate=False),
+                    base64.urlsafe_b64decode(encoded + b'=='),
                     [errors.InvalidBase64CharactersDefect(),
                      errors.InvalidBase64PaddingDefect()],
                 )

Lib/email/base64mime.py

@@ -34,8 +34,8 @@
     ]
 
 
-from base64 import b64encode
-from binascii import b2a_base64, a2b_base64
+from base64 import b64encode, urlsafe_b64decode
+from binascii import b2a_base64
 
 CRLF = '\r\n'
 NL = '\n'
@@ -102,12 +102,15 @@ def decode(string):
     base64 (like =?iso-8859-1?b?bmloISBuaWgh?=) -- please use the high
     level email.header class for that functionality.
     """
+    # We use urlsafe_b64decode here because some mailers apparently use the
+    # urlsafe b64 alphabet, and urlsafe_b64decode will correctly decode both
+    # the urlsafe and regular alphabets.
     if not string:
         return bytes()
     elif isinstance(string, str):
-        return a2b_base64(string.encode('raw-unicode-escape'))
+        return urlsafe_b64decode(string.encode('raw-unicode-escape'))
     else:
-        return a2b_base64(string)
+        return urlsafe_b64decode(string)
 
 
 # For convenience and backwards compatibility w/ standard base64 module

Lib/test/test_email/test__encoded_words.py

@@ -38,6 +38,12 @@ def test_missing_padding(self):
         # 2 missing padding characters
         self._test(b'dg', b'v', [errors.InvalidBase64PaddingDefect])
 
+    def test_urlsafe_alphabet(self):
+        self._test(
+          b'QW5tZWxkdW5nIE5ldHphbnNjaGx1c3MgU_xkcmluZzNwLmpwZw==',
+          b'Anmeldung Netzanschluss S\xfcdring3p.jpg',
+          [errors.InvalidBase64CharactersDefect])
+
     def test_invalid_character(self):
         self._test(b'dm\x01k===', b'vi', [errors.InvalidBase64CharactersDefect])
 

Lib/test/test_email/test_email.py

@@ -4663,6 +4663,9 @@ def test_decode(self):
         eq = self.assertEqual
         eq(base64mime.decode(''), b'')
         eq(base64mime.decode('aGVsbG8='), b'hello')
+        eq(base64mime.decode(
+           'QW5tZWxkdW5nIE5ldHphbnNjaGx1c3MgU_xkcmluZzNwLmpwZw=='),
+           b'Anmeldung Netzanschluss S\xfcdring3p.jpg')
 
     def test_encode(self):
         eq = self.assertEqual