Re: [exim] when are exim vars first available? and ...

Dave Lugo wrote:
> On Fri, 30 Sep 2005, Wakko Warner wrote:
> > How about a fake reject in data acl if some users do not want the message and
> > some do that states the users who the message did not get delivered to? I
> > haven't thought it out enough to know if it's a good idea or not. And of
> > course some servers/clients will butcher the message to mean something else.
>
> Where I may go, is post-bounce the rejection(s) instead, but only in
> limited cases (blowback is something I'd like to try to avoid).

That's why I mentioned fake reject. It's basically an accept with a 5xx
return code. The server still delivers, but the sender still sees a hard
failure. the sending server is responcible for the bounce, not you.

> It does get much easier if you can limit the rcpt_to count to '1'. On
> my home vanity domains, I do so, and it's not a problem. I doubt that
> such a limit would be appropriate for larger domains.

Quite true. I could do it on my domain as well.

Where I work is bad for bounces. I took over the mail server at work
replacing an Xxtramail NT server (pop3 drop box pickup server) which if the
message couldn't be delivered, it was sent to the admin. When I setup exim
3.12 back then, it was setup to drop unknowns to a special mail box and I'd
forward as required and it bounced the message at the same time. Ok, it's a
bad idea and I'd love to change it but the PHB wants the "unknowns" to be
delivered incase of a mis-spelling or something. They don't want to change
and I can't convince them. I think a fake reject would be a good idea but
I'd have to upgrade exim to get it (currently something like 4.20) IIRC,
fake rejecct was part of exiscan.

--
Lab tests show that use of micro$oft causes cancer in lab animals
Got Gas???