fix: redundant \n and WSP when using policy.compat32

Author: RanKKI

Lib/email/_policybase.py

@@ -307,7 +307,29 @@ def header_source_parse(self, sourcelines):
 
         """
         name, value = sourcelines[0].split(':', 1)
-        value = value.lstrip(' \t') + ''.join(sourcelines[1:])
+
+        # Fixed: https://github.com/python/cpython/issues/124452
+        #
+        # Root cause: The function '_refold_parse_tree' in '_header_value_parse.py'.
+        # If there is no WSP, it can't figure out how to wrap the text.
+        # Therefore, it places the entire value directly after '\n', and because
+        # there is a WSP after '<HeaderName>:', the WSP will be moved to the front
+        # of the value according to RFC5322, section 2.2.3.
+        #
+        # However, the WSP is not part of the value; therefore, we must 
+        # remove it.
+
+        no_first_value = value.strip() == '' and len(sourcelines) > 1
+
+        # When using the compat32 policy, the value is '\n'. Therefore,
+        # use an empty string if there is no value (without WSP and CRLF)
+        # on the first line
+        value = '' if no_first_value else value.lstrip(' \t')
+
+        if no_first_value and sourcelines[1][0] in ' \t':
+            sourcelines[1] = sourcelines[1][1:]
+
+        value += ''.join(sourcelines[1:])
         return (name, value.rstrip('\r\n'))
 
     def header_store_parse(self, name, value):

Lib/email/policy.py

@@ -125,24 +125,11 @@ def header_source_parse(self, sourcelines):
 
         """
         name, value = sourcelines[0].split(':', 1)
+        no_first_value = value.strip() == '' and len(sourcelines) > 1
 
-        # Fixed: https://github.com/python/cpython/issues/124452
-        #
-        # Root cause: The function '_refold_parse_tree' in '_header_value_parse.py'.
-        # If there is no WSP, it can't figure out how to wrap the text.
-        # Therefore, it places the entire value directly after '\n', and because
-        # there is a WSP after '<HeaderName>:', the WSP will be moved to the front
-        # of the value according to RFC5322, section 2.2.3.
-        #
-        # However, the WSP is not part of the value; therefore, we must 
-        # remove it.
-
-        # Remove leading WSP in the first line only if there no value in the
-        # first line, and has values after that
-        remove_wsp = not value.strip() and len(sourcelines) > 1
-
-        value = value.lstrip(' \t')
-        if remove_wsp and sourcelines[1][0] in ' \t':
+        value = '' if no_first_value else value.lstrip(' \t')
+
+        if no_first_value and sourcelines[1][0] in ' \t':
             sourcelines[1] = sourcelines[1][1:]
 
         value += ''.join(sourcelines[1:])

Lib/test/test_email/test_message.py

@@ -970,7 +970,7 @@ def test_folding_with_short_nospace_1(self):
         parsed_msg = message_from_bytes(m.as_bytes(), policy=policy.default)
         self.assertEqual(parsed_msg['Message-ID'], m['Message-ID'])
 
-    def test_folding_with_long_nospace_1(self):
+    def test_folding_with_long_nospace_default_policy_1(self):
         # Fixed: https://github.com/python/cpython/issues/124452
         # 
         # When the value is too long, it should be converted back 
@@ -986,6 +986,18 @@ def test_folding_with_long_nospace_1(self):
         parsed_msg = message_from_bytes(m.as_bytes(), policy=policy.default)
         self.assertEqual(parsed_msg['Message-ID'], m['Message-ID'])
 
+    def test_folding_with_long_nospace_compat32_policy_1(self):
+        # Fixed: https://github.com/python/cpython/issues/124452
+        # 
+        # When the value is too long, it should be converted back 
+        # to its original form without any modifications. 
+
+        m = EmailMessage(policy.compat32)
+        m['Message-ID'] = '12345678912345678123456789123456789123456789'\
+                        '12345678912345678123456789123456789123456789'
+        parsed_msg = message_from_bytes(m.as_bytes(), policy=policy.default)
+        self.assertEqual(parsed_msg['Message-ID'], m['Message-ID'])
+
 
     def test_get_body_malformed(self):
         """test for bpo-42892"""