Sender: |
|
Subject: |
|
From: |
|
Date: |
Thu, 28 Aug 1997 00:26:24 +0200 |
In-Reply-To: |
Message of Wed, 27 Aug 1997 16:05:56 -0600 from LISTSERV list
owners' forum < [log in to unmask]> |
Reply-To: |
|
Parts/Attachments: |
|
|
On Wed, 27 Aug 1997 16:05:56 -0600 "Michael Loftis(ADMIN)"
<[log in to unmask]> said:
>Mostly sites seem to be using this to tell where mail comes from
>exclusively. If it doesn't resolve (as in the case of a blank) then it
>doesn't get there.
This is a violation of RFC821 reinforced in RFC1123. Certain messages
(bounces) are actually REQUIRED to use MAIL FROM:<>. Systems which reject
MAIL FROM:<> need to be fixed.
>1.8c seems to be "undecided" on when to use and when not to use a proper
>mail from.
MAIL FROM:<> is used when required by the standards and then in cases
where there is no point in being sent a bounce (which would have to be
passed on to the maintainer - that's you, that's a LOT of mail, and that
leads to a filter in your mail program which will also kill messages that
you SHOULD be reading). Support for MAIL FROM:<> has been mandatory for
15 years, there is nothing "broken" about it and it is a key element of
the SMTP standard (for loop prevention). Bouncing a message that has MAIL
FROM:<> is an inexcusable and extremely dangerous violation of the
standards as it has the potential to lead to a mailing loop, of which we
had one yesterday, to the tune of 2000 ever-increasing messages. Luckily
this was a Windows 3.1 site so when the bounce reached 64k their mail
program died and the loop ended :-) The very purpose of MAIL FROM:<> is
to suppress bounces and avoid endless loops, as such bouncing is one of
the most irresponsible acts an ISP can possibly commit (worse than
allowing spam, spams at least do not loop and increase with every
iteration).
Eric
|
|
|