|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
From: Nick Simicich (njs
scifi.squawk.com)Date: Mon Jun 03 2002 - 10:50:34 CDT
It is apparent, Hobbit, that you have never run a mailing list. :-)
I have been running them for many years now, since before address
verification was common.
Bounces come in about a thousand different formats. You can't depend on
*any* of the content coming back. It just will not happen. I was using a
header to match for recipient address in bounce processing, and it just
don't work. You get bounces that tell you that you sent a message to a
domain you never heard of, because someone has forwarded their mail, you
get bounces that do not contain any addresses that were ever on any of your
lists. None of the original headers are included. And so forth. Many
bounces can be parsed. Some can't.
You can't use a unique Message-ID to tell bouncers apart, even presuming
that bounces had the headers in them reliably. Everyone on your mailing
list has to get the same message ID so that In-Reply-To: and References:
work, including in private mail. Lately, I have had a number of bounces
come back where the entire message is in base64. So your scripts are all
stymied unless they can decode mime.
The most likely thing that will survive a bounce is the MAIL FROM:
address. No, not all of them will survive, I have gotten bounces where it
set an address on my system out of the RFC822 headers as the source and
another address on my system out of the 822 headers as the target. This
made me speculate that perhaps a license to run procmail should not be
granted until the applicant passed a written test proving that they know
the difference between the envelope and the header...I ended up bouncing
those to the domain's support address with their postmaster forged as the
origin after finding that he had used a standard facility from his site for
bouncing. But MAIL FROM:<> has the highest likelihood of surviving the
bounce process, and the joy is that it is out of band in a specified format
so that those inventive people have a hard time munging it. If they bounce
it at all, they probably did it right. And you do not have to worry about
parsing 10,000 different formats of responses.
I was using something like VERP long before there was support for it in the
MTAs. I'm now using Majordomo2 which automatically uses its own variation
on VERP to track bounces so that one does not have to read them, and so
that the bounce software can reliably tell a bounce from a manual e-mail to
the owner address. The bounce processing works...because it uses VERP.
Now the answer to the sender dilemma is that if you want to match on sender
you use a regexp.
At 10:38 PM 2002-06-01 +0000, *Hobbit* wrote:
>VERP, if that's what it implies about sender addresses, appears to be
>a miserable crock because in part nobody running such a mailing list
>ever READS THEIR GODDAMN BOUNCES and takes appropriate action. Maybe
>it's worth blocking all of it outright and telling mailing list maintainers
>to use a more real and consistent sender address. What does this VERP
>thing buy anyone, anyways, that something like a message-ID can't?
By the way, I don't care if you block the mailing list mail. I don't get
paid to deliver it, people want it. Block it, and it won't get through.
-- War is an ugly thing, but it is not the ugliest of things. The decayed and degraded state of moral and patriotic feeling which thinks that nothing is worth war is much worse. A man who has nothing for which he is willing to fight, nothing he cares about more than his own personal safety, is a miserable creature who has no chance of being free, unless made so by the exertions of better men than himself. -- John Stuart Mill Nick Simicich - njsscifi.squawk.com
- To unsubscribe, send mail to majordomo
postfix.org with content (not subject): unsubscribe postfix-users
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]