Re: [Exim] Re: RBL checking

On Tue, 25 Sep 2001, Dave C. wrote:

> On Mon, 24 Sep 2001, Alan J. Flavell wrote:

> > Doing it this way does, however, cause us a certain amount of tedium,
> > since most of our attempts to return a rejection report for spam are
> > then refused by their alleged reporting address (which is typically
> > counterfeit, or deliberately deaf to reports), and the rejection
> > reports then get frozen.
>
> [..] Ugh. thats why its better to reject the message at the time of the
> SMTP session, rather than accepting it and trying to bounce it.

Indeed - so if there seems any prospect of it being successful (and
not cause false positives), each of these jokers lands up in the
sender_reject or hostip_reject databases, or both, in pretty short
order.

But the scoring _does_ seem to have been a useful mechanism for
eliminating a worthwhile additional proportion of spam that would
otherwise get through, and with relatively little risk of false
positives. it's also fine-tunable on the basis of the scoring,
whereas sender_reject or hostip_reject are all-or-nothing switches.

Incidentally, since I posted that earlier report, there's been a
sudden outbreak of spams being sent out from somewhere with my address
counterfeited as the sender, so I'm getting the rejections. Hmmm,
coincidence or not?

cheers