Bug report

Bug description:

email.utils.encode_rfc2231() substitutes an empty string when language is omitted, but not when charset is omitted. When only language is given, None is interpolated into the result as the literal charset name:

>>> import email.utils
>>> email.utils.encode_rfc2231('abc', language='en')
"None'en'abc"

The same happens when a (None, language, value) tuple is passed as a parameter value to Message.add_header():

>>> from email.message import Message
>>> m = Message()
>>> m.add_header('Content-Disposition', 'attachment', filename=(None, 'en', 'report.txt'))
>>> m['Content-Disposition']
"attachment; filename*=None'en'report.txt"

RFC 2231 section 4 says "it is perfectly permissible to leave either the character set or language field blank", so the expected result is "'en'abc". The email package's own parser reads the value back correctly only because None is not a known codec, so it falls back to the undecoded text. Other consumers see a bogus charset name.

This is separate from gh-85014, which covers the case where both charset and language are None.

CPython versions tested on:

3.15, CPython main branch

Operating systems tested on:

Linux

Linked PRs

• gh-161222: Fix email.utils.encode_rfc2231() writing "None" as the charset #161225