Richard Pieri wrote:
> WB> But this raises another question. If someone goes on vacation right
> WB> after sending a mail to someone else on vacation, and both use
> WB> autoresponders, what's the protocol that prevents a mail loop there?
>
> What is supposed to happen is that for each incoming message, the
> vacation program compares the envelope sender with a log of previously
> received senders. If it finds a match it does nothing; if it does not
> find a match it adds the data to the log and sends out the response
> file. At worst, a mailing list loop should cease at the second
> iteration, as the user's own envelope finally gets into the log.
>
> Of course, just like the plethora of broken MUAs, there are a plethora
> of broken vacation and autoresponder programs out there.
So in the case at hand, the autoresponder would send a vacation message to
the list once for each poster, but the second post would not cause a message
to be sent. So there'd be a lot of traffic if the list was active with many
contibutors, but no actual loops. Or, if the list was set up to use bulk
precedence AND the autoresponder honored that, no mail would be sent at all.
Sounds like the list setup and the autoresponder are more to blame than
Reply-To.
> WB> Seems what's needed is a way to distinguish mail sent directly by a
> WB> person, from mail sent directly by a program.
>
> No, what is *really* needed is software that is not so thoroughly brain-
> damaged that workarounds need to be instituted to deal with them.
I don't think I suggested a workaround. I asked how to get the information
needed to properly handle the situation by either the autoresponder not
sending mail to a program address (the list), or majordomo resend not
responding to mail from a program (the autoresponder). Does that
information exist in any RFC-compliant header or not?
In my mind, it is a very useful thing for a list owner to indicate that all
replies should go publicly back to the list instead of privately to the
poster, and expecting everyone in the world to give up a useful setup is the
workaround to avoid. If I can't use Reply-To, what *IS* the RFC-compliant
way to do that?
-- Brad Whitlock
Follow-Ups:
|
|
