[milters] Archive

Lists Index Date Thread Search

Article: 1520
From: Anthony Howe
Date: 2007-04-01 02:10:37 -0400
Subject: Re: Does such a milter exist...

Removal...........: milters-request@milter.info?subject=remove
More information..: http://www.milter.info/#Support

Grant Taylor wrote:
> Anthony Howe wrote:
>> You could test this with milter-cli and a script. You could also try 
>> forcing sendmail into one message one recipient mode by specifying the 
>> max. number of recipients of 1. It should force the sending mailer to 
>> split out envelopes or take a long time to deliver a message destined to 
>> many people at the same site.
> I may consider Milter-CLI.  However I really do not like the idea of 
> only limiting all messages to one recipient per message.

Using a script with milter-cli, you wouldn't need to. I was just saying 
you could do it without a milter in sendmail if you wanted to go that route.

At least with milter-cli you can test the concept first then if you 
really like it, pester me for a milter.

> Grey listing is a function of temporarily failing senders, possibly in 
> conjunction with recipient(s).

Normally it is { IP, MAIL, RCPT }, though milter-gris allows other combos.

> What I am wanting is to temporarily fail recipients based on the 
> criteria of what OTHER recipients have been specified for a given 

I understood that.

> envelope.  Let's say that there is a group of protected email addresses 
> that includes postmaster@, abuse@, root@, etc.  Members of this group 
> will only (successfully) receive emails if the envelope recipients for 
> given messages are members of this group.  If an envelope recipient is 

Hmm. milter-group...

Target                                          Member of group
milter-group-to:fluffy-bunnies.com		pets, staff, bofh
milter-group-to:postmaster@			bofh, support
milter-group-to:abuse@				support

milter-group-from:prez@vicious-bunny-rabbit.com	phb

milter-group-connect:192.168			staff, marketing

Not sure the worth of -connect or -from variants and the precedence if 
used, but its an idea.

> not a member of this group, the recipient temp failed thus causing the 
> sending server to retry said recipient at another time.  Or if an 

OK. If you do this during RCPT phase, then the first RCPT issued defines 
the group and subsequent RCPT have to be members or rejected. My 
thoughts on it were to collect the RCPT and then at DATA reject (or 
temp.fail) the DATA command if all RCPT were not members of the same group.

In the case of postmaster/abuse I would have collected the whole list 
and at DATA if either of those appears with any other 
non-postmaster/abuse address, then temp.fail/reject.

> envelope recipient that is not a member of this group is specified first 
> and then an envelope recipient that is a member of this group is 
> specified the envelope recipient is temporarily failed.

How can you know the group before you get the first recipient?

> The main difference between what I am proposing and grey listing is that 
> grey listing is based on memory of previous envelopes and what I am 
> proposing is based on envelope recipients of the current envelope. 
> Granted grey listing and what I am proposing both temp fail some 
> recipients, but the way they decide to do it or not is completely 
> different based on a completely different set of criteria.

Yes. But the grey-list criteria might have been sufficient.

Anthony C Howe          Skype: SirWumpus                    SnertSoft
+33 6 11 89 73 78         AIM: SirWumpus    Sendmail Milter Solutions
http://www.snert.com/     ICQ: 7116561

Lists Index Date Thread Search