Re: [exim] Return reciept on forward

Pàgina inicial
Delete this message
Reply to this message
Autor: Alan J. Flavell
Data:  
A: Exim users list
Assumpte: Re: [exim] Return reciept on forward
On Tue, 27 Dec 2005, Alan J. Flavell wrote:

> On Tue, 27 Dec 2005, Giuliano Gavazzi wrote:
>
> > I wonder if there is a list of hosts sending bogus virus alerts.
>
> Sort of. Try www.spambag.org , lists BACKSCATTER and BACKSCATTER2.


I've just been re-reading http://www.spambag.org/backscatter.html

It appears that this chap takes the view that there can be no
exceptions; but he offers no viable way of dealing with forwarded
addresses nor with vacation responses. While I have a great deal of
sympathy with the rest of what he says, I'm at a loss to know how to
deal with this aspect.

If he contrives to produce a probe which, as a result of forwarding or
vacation response at the probed MTA, results in a bounce being
delivered to his spamtrap, then he'll blacklist the site in question.
There's definitely a (small) risk that we could do that, if we ever
came to his attention (which we, fortunately, haven't).

Considering that there's a strict limit on the time available for any
MTA activities after the end of DATA, I'd rule out the possibility of
attempting to despatch forwarded mail or vacation responses during
this phase of the inbound SMTP transaction. I don't see any other way
of implementing this, other than screening all such bounces by
postmaster inspection, which we really cannot afford to do. And I
think I speak for many another site in this regard.

So, we do go to considerable lengths to avoid responding to spam or
other kinds of abusive mail with a bounce, but, in the final analysis,
if an inbound mail appears to be bona fide and meets our other
criteria, and if forwarding is unsuccessful or a vacation response has
been configured by/for this user, we *will* compose a bounce-type
response for it, addressed to its envelope sender (whether genuine or
not, which of course we don't know).

I can't see any viable alternative, despite what the spambag chap
says. Nevertheless, as I said earlier, the spambag backscatter/2 RBL
entries can be useful, so long as they aren't used alone, but in
conjunction with other criteria to avoid rejecting bona fide bounces.

comments?