Mark Sapiro writes:
However, I did a test injecting directly to Mailman a message with
Subject: [SAV-LINUX] Notice from Sophos_Anti-Virus on server.example.com [DEMO1]on a German language test list with subject prefix [Janinsliste] and I get the same result, I.e. the RFC 2047 encoded
Subject: =?utf-8?q?=5BJaninsliste=5D=5BSAV-LINUX=5D_Notice_from_Sophos=5FAnti-Virus_on_server=2Eexample=2Ecom?=so this is definitely a Mailman bug (or a bug in the underlying Python email package.
There seem to be other bugs here.
First, I can't see any content-based reason for encoding this 100% ASCII content. I guess the list is set to always encode the subject prefix, and we just encode the whole field body if any part is encoded?
Perhaps this is a quick workaround for the OP: set MIME-encoding for the subject prefix to "as needed".
Second (and this is important because of RFC conformance!), the encoded word is over 75 characters. I've filed https://gitlab.com/mailman/mailman/-/issues/1021 for this.