[Mailman-Users] Reply-to options not working

[Stephen J. Turnbull]

Dimitri Maziuk writes:

 > Does it ave the same Message-ID though?

According to RFC, Message-ID is an originator field, and MUST be
present and MUST be unique.  The MUA or submission agent should add it
before handing off to the MTA.  As a last resort the MTA may add it.
If it gets past the MTA without it, it's non-conforming.  Mediator
software (such as mailing list managers) MAY add a Resent-Message-ID
field, which is not restricted in number.

In some cases it makes sense for a Mediator such as Mailman to change
the Message-ID, and it will always add one if not present.  Mark is
authoritative on when Mailman does it.  However, normally Mailman (and
other mailing list managers) will not change it, indicating that the
list considers the outgoing message to be the same from the author's
point of view as the incoming message.

Of course this is a judgment call.  Obviously *some* changes such as
adding Received fields to the header don't change "the message".  On
the other hand, I think it's reasonable for authors to claim that
mailing lists that go stripping attachments or HTML parts, or
translating text/html to text/plain, as Mailman can be configured to
do, have edited the message enough that it's a new message.  Stuff in
the middle (list tags and serial numbers in Subject, headers and
footers on the body) I would *never* consider to make a new message,
but some people claim they think so.

I don't recall ever seeing such a complaint from authors, even from
people who want the HTML preserved.  Authors don't question that it's
the same message, they just want the presentation preserved.  The
people who do question it generally do it in service of claims that
the mailing list "owns" the message so Reply-To munging is
RFC-conforming, etc.

The RFCs punt on "when is it a new message" in exactly the same way,
by the way.  "It's your call, just be sure to change the Message-ID if
you think it's a new message now."