At 09:56 2005-12-12 -0800, Gary Funck did say:
of out-of-the-office and unknown-user messages that come directly back
to me any time I post a reply to that list. Worse, these direct replies
don't have anything in the headers, or subject, to tip me off to the
fact that the reply relates to my original posting to the list.
Yes, those are irritating. Lack of identifying information in the reply,
plus often a non-synchronous reply (the timing of the reply isn't anywhere
near when the message was posted, because the recipient is on a crappy mail
system and/or has a non-static connection to the net) can make it very
difficult to correlate the reply to which list their reply was generated
through. Fortunatley, in the past few years, I've curbed my list
participation down markedly on various lists, so the number of posts that
might generate an autoreply is significantly lower, which makes identifying
a culprit list much easier. Of course, then you still have to send a
message to the listadmin of that list to ask that they deal with the
problem user.
both have a Return-Path header, but that this other list has no
Sender or Errors-To header.
Well, they should have a Sender. Errors-To is less commonly used (as it
should be - it's use was never formally adopted, and it is actually
discouraged).
If we assume that the various vacation programs and MTA's out there play
by a somewhat consistent set of rules,
They don't. Particularly when they're integral to some corporate mail
system like MS Exchange or Lotus Notes/Domino/cc:Mail (uhm, which all
happen to be windoz based: draw your own conclusions). Those packages just
sort of make things up as they go. It would make a LOT of sense to have
logic built in for identifying likely list mail so that vacation messages
aren't fired off in response to list messages...
would the lack of Sender and Errors-To explain the observed errant
behavior? What role do these headers play? Which headers take precedence
over others when it comes to replying?
Cuddle up with a nice RFC sometime: RFC2822 (which superceedes 822):
<http://www.faqs.org/rfcs/rfc2822.html>
See particularly section 3.6.2
Technically, Sender: is required when there are multiple From: addresses
(which is legal, but you simply never see it these days). This would be
like when a committee authors a message, but of course only ONE account is
actually going to be sending it (and it might be a secretary anyway -
someone not even a member of the committee). Reply-To is optional, but
basically redirects where the reply is supposed to be sent rather than just
to the From: address(es).
Entirely too many programs just grab the From: address and use it for
everything, ignoring such specifications.
There's also a "Resent-Sender:" header, but one rarely sees that: I've seen
it regularly on one list which ceased to exist about *8* years ago, and a
couple of times on the PAM list, apparently NOT inserted by the list, but
by individual users relaying information.
Return-Path is where errors are supposed to be directed, and is required to
be inserted into messages whenever the message leaves the SMTP layer.
FTR, Errors-To isn't defined in 2821 or 2822 - it was a PROPOSED header
long ago, but never formally adopted. It's use is discouraged in RFC2076:
<http://www.faqs.org/rfcs/rfc2076.html>
And given that my note, as transmitted by the list manager has Precedence:
bulk, and various List- header lines, shouldn't the various vacation programs
and mail bouncers no better than to reply directly back to me?
They should, but they don't. See above.
---
Sean B. Straw / Professional Software Engineering
Procmail disclaimer: <http://www.professional.org/procmail/disclaimer.html>
Please DO NOT carbon me on list replies. I'll get my copy from the list.
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail