testInterfacePost: ignore the difference of Message-ID header format on zope5py3.

bt5/erp5_interface_post/TestTemplateItem/portal_components/test.erp5.testInterfacePost.py

@@ -247,7 +247,7 @@ class TestInterfacePost(ERP5TypeTestCase):
     last_message, = self.portal.MailHost._message_list
     self.assertNotEqual((), last_message)
     _, _, message_text = last_message
-    self.assertIn(message_text, sequence['internet_message_post'].getData())
+    self.assertIn(message_text.replace(b'Message-ID: \n ', b'Message-ID: '), sequence['internet_message_post'].getData())
 
   def _getMailHostMessageForRecipient(self, recipient_email_address):
     message_list = self.portal.MailHost._message_list
@@ -268,7 +268,7 @@ class TestInterfacePost(ERP5TypeTestCase):
       self.assertEqual(len(message_list), 1)
       message = message_list[0]
       _, _, message_text = message
-      self.assertIn(message_text, post.getData())
+      self.assertIn(message_text.replace(b'Message-ID: \n ', b'Message-ID: '), post.getData())
 
   def stepCheckMailMessagePreviewDisplaysLatestInternetMessagePostData(self, sequence=None, sequence_list=None):
     mail_message = sequence['mail_message']

[Jérome Perrin]

I thought that this was a bug, but according to https://datatracker.ietf.org/doc/html/rfc5322#section-2.2.3 it might be OK. Is that also your understanding ? if yes, maybe we can add a reference to that ?

[Jérome Perrin]

mmh but still, it looks like a bug somewhere, if it's equivalent to have \n in the serialized message, then isn't the parser supposed to remove the extra \n ? I investigated a bit some days ago, these were my notes:

import email

mb = (
  b'Content-Type: multipart/mixed; boundary="===============boundary=="\n'
  b'Message-ID: <170831831588.80413.17161314261293097219.17161314261293097219@example.com>\n'
  b'--===============boundary==--'
)

m = email.message_from_bytes(mb)

print(repr(m.get('Message-ID')))
# OK '<170831831588.80413.17161314261293097219.17161314261293097219@example.com>'

m_from_bytes = email.message_from_bytes(m.as_bytes())
print(repr(m_from_bytes.get('Message-ID')))
# NOT OK '\n <170831831588.80413.17161314261293097219.17161314261293097219@example.com>'


# note that when using str it is OK:
m_from_str = email.message_from_string(m.as_string())
print(repr(m_from_str.get('Message-ID')))
# OK '<170831831588.80413.17161314261293097219.17161314261293097219@example.com>'


# already different here:
print(repr(m.as_bytes()))
print(repr(m.as_string()))

# the output is controlled by max_header_length of the `policy` argument.
# not sure if the problem is when emitting the email or parsing it