Re: [Exim] callbacks and <>

On Mon, 8 Dec 2003, Russell King wrote:

> Why on earth would you want to accept mail from a site refusing null
> senders, thereby preventing the return of non-delivery messages?

Indeed

> It effectively means that such a site could not care whether the
> message gets through or not, so, IMHO it's far better to refuse the
> message until the postmaster at the site gets some clue.

Background: We only use callback if the domain is listed in a local
database. We add domains to that base primarily on the grounds that
they've been presented as counterfeited sender addresses, and a manual
check has indicated promising results.

Domains where null senders are refused at MAIL or RCPT time can also
be added to that list - that way, at least with versions of exim4 that
we've used recently, the callback results in them being the immediate
cause of our rejection (;-); and there's no maintenance needed, since
as soon as they decide to behave properly, they'll have no further
problems mailing to us. Seems fair to me? - modulo those folks who
oppose callout on principle, of course.