Hitashi mancho via Mailman-users writes:
One of our users posted an email using a Bluemail Desktop client on one of the mailing lists. This email bounced with the reason "501 Message has defects".
Bouncing the message is arguably an overreaction to this particular defect, but in fact Bluemail is violating a "MUST NOT" in RFC 2045. Mailman takes the position that in principle we should not forward non-conforming mail. At least one expert has told me he considers this defect harmless, but (a) it's not clear to me that it is harmless (especially in the hands of a malicious actor targeting carelessly- written applications) and (b) it's hard to get excited about adding complicated logic to the code to make life easier for an application that doesn't conform to a trivial 30-year-old restriction.
However, it gets accepted if the same user posts the same email on the same mailing list using Webmail (RoundCube) or Thunderbird.
I'm not familiar with the maintainers of RoundCube, but Thunderbird has a long history of respect for Internet standards.
There is nothing in mailman.log and bounce.log, even in debug mode.
Since this is happening during receipt of the message I would guess that all messages related to the problem would be in Mailman's smtp.log. I don't think you'd see anything about this in mailman.log unless the defect managed to crash the lmtp runner or all of Mailman. bounce.log reports bounces *received* by Mailman, not bounces caused by Mailman.
Apr 21 18:00:47 2025 (1979121) ('127.0.0.1', 44502) << b'501 Message has defects'
I seem to recall that Mark submitted a patch to add the defect list to this message. I will follow up on that (although it doesn't help get the message delivered to subscribers, it will make diagnosis a lot easier and faster).
Content-Type: multipart/alternative; boundary="----sinikael-?=_1-17452386441150.6962994414278503" Content-Transfer-Encoding: quoted-printable
This the problem. When Content-Type is 'multipart', the only values allowed for Content-Transfer-Encoding are '7-bit', '8-bit' and 'binary'. This is a "what were they thinking?!" level bug.
-- GNU Mailman consultant (installation, migration, customization) Sirus Open Source https://www.siriusopensource.com/ Software systems consulting in Europe, North America, and Japan