Great Circle Associates List-Managers
(July 1998)

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: individualized probe messages
From: Dave Sill <de5-list-managers @ sws5 . ctd . ornl . gov>
Organization: Oak Ridge National Lab, Oak Ridge, Tenn., USA
Date: Fri, 31 Jul 98 14:31:16 EST
To: List-Managers @ GreatCircle . COM
In-reply-to: <000101bdbc98$d33c7ed0$017b7b0a@gillette>
References: <><000101bdbc98$d33c7ed0$017b7b0a@gillette>

"Tom Neff" <> wrote:
>Instead of (or in addition to) serializing the Subject header in each
>probe message, you also serialize the From: and Errors-To: headers as
>	From:
>	Errors-To:
>	Subject: List Test #001347
>	This is a list test message - please ignore it.
>Now even the most uncooperative Prodigy-class mail agent can't help revealing
>which message it's bouncing, because each one will be delivered to a unique
>"userid."  Even if they omit a To: header in the bounce, you can examine the
>Received: headers for the delivery address your local MTA saw.

This is very similar to VERP, which encodes the recipient's address in 
the SMTP envelope return path. The advantage of VERP is that it can be 
used for each delivery--obviating the need for special probes, and
allowing the list manager to keep track of which messages have bounced 
to which subscribers.

The advantages of your method over VERP are that it will work with
idiotic daemons that send bounces to the From: address, and it'll work
with MTA's that can't do VERP.

>This is particularly easy to do with today's "virtual servers" that forward
> to the same address.

qmail's wildcardable user-defined extension addressing makes VERP'ed
bounces easy to deal with. And qmail supports sending VERP'ed
messages. Of course, ezmlm uses VERP to implement automatic bounce
handling. There's also a VERP bounce handler for Majordomo under
qmail that works nicely.


Indexed By Date Previous: Re: Unhelpful Bounce Of The Week
From: Norbert Bollow <>
Next: Re: individualized probe messages
From: murr rhame <>
Indexed By Thread Previous: Re: individualized probe messages
From: "Tom Neff" <>
Next: Re: individualized probe messages
From: murr rhame <>

Search Internet Search