gh-98188: Fix EmailMessage.get_payload to decode data

Lib/email/message.py

@@ -14,6 +14,7 @@
 # Intrapackage imports
 from email import utils
 from email import errors
+from email import headerregistry
 from email._policybase import compat32
 from email import charset as _charset
 from email._encoded_words import decode_b
@@ -286,8 +287,11 @@ def get_payload(self, i=None, decode=False):
         if i is not None and not isinstance(self._payload, list):
             raise TypeError('Expected list, got %s' % type(self._payload))
         payload = self._payload
+        cte = self.get('content-transfer-encoding', '')
+        if isinstance(cte, headerregistry.ContentTransferEncodingHeader):
+            cte = cte.cte
         # cte might be a Header, so for now stringify it.
-        cte = str(self.get('content-transfer-encoding', '')).lower()
+        cte = str(cte).strip().lower()
         # payload may be bytes here.
         if not decode:
             if isinstance(payload, str) and utils._has_surrogates(payload):

Lib/test/test_email/test_headerregistry.py

@@ -837,6 +837,11 @@ def cte_as_value(self,
             '7bit',
             [errors.InvalidHeaderDefect]),
 
+        'extra_space_after_cte': (
+            'base64 ',
+            'base64',
+            []),
+
     }
 
 

Lib/test/test_email/test_message.py

@@ -1,3 +1,4 @@
+from email.errors import InvalidHeaderDefect
 import textwrap
 import unittest
 from email import message_from_bytes, message_from_string, policy
@@ -1054,6 +1055,37 @@ def test_string_payload_with_multipart_content_type(self):
         attachments = msg.iter_attachments()
         self.assertEqual(list(attachments), [])
 
+    def test_string_payload_with_base64_cte(self):
+        msg = message_from_string(textwrap.dedent("""\
+        Content-Transfer-Encoding: base64
+
+        SGVsbG8uIFRlc3Rpbmc=
+        """), policy=policy.default)
+        self.assertEqual(msg.get_payload(decode=True), b"Hello. Testing")
+        self.assertDefectsEqual(msg['content-transfer-encoding'].defects, [])
+
+    def test_string_payload_with_extra_space_after_cte(self):
+        # https://github.com/python/cpython/issues/98188
+        cte = "base64 "
+        msg = message_from_string(textwrap.dedent(f"""\
+        Content-Transfer-Encoding: {cte}
+
+        SGVsbG8uIFRlc3Rpbmc=
+        """), policy=policy.default)
+        self.assertEqual(msg.get_payload(decode=True), b"Hello. Testing")
+        self.assertDefectsEqual(msg['content-transfer-encoding'].defects, [])
+
+    def test_string_payload_with_extra_text_after_cte(self):
+        msg = message_from_string(textwrap.dedent("""\
+        Content-Transfer-Encoding: base64 some text
+
+        SGVsbG8uIFRlc3Rpbmc=
+        """), policy=policy.default)
+        self.assertEqual(msg.get_payload(decode=True), b"Hello. Testing")
+        cte = msg['content-transfer-encoding']
+        self.assertDefectsEqual(cte.defects, [InvalidHeaderDefect])
+
+
 
 if __name__ == '__main__':
     unittest.main()

Misc/ACKS

@@ -1127,6 +1127,7 @@ Gregor Lingl
 Everett Lipman
 Mirko Liss
 Alexander Liu
+Hui Liu
 Yuan Liu
 Nick Lockwood
 Stephanie Lockwood

Misc/NEWS.d/next/Library/2024-12-03-14-45-16.gh-issue-98188.GX9i2b.rst

@@ -0,0 +1,2 @@
+Fix an issue in :meth:`email.message.Message.get_payload` where data
+cannot be decoded if there is a trailing whitespace and/or texts following the mechanism.