On Mon, 2004-02-23 at 13:50, Philip Hazel wrote:
> On Mon, 23 Feb 2004, David Woodhouse wrote:
> > For _recipient_ verification callouts I think Exim's current behaviour
> > is suboptimal.
> There is nothing in the ChangeLog that suggests this has ever been
> changed since recipient callouts were invented at the start of Exim 4.
> Now that I think about it some more, however, I suspect that "<>" was
> used so as to prevent the next server from doing yet another callout. If
> the current MTA has already called out to verify the sender address, it
> is wasteful - and discourteous - to callout-verify it again.
However recipient callouts would normally only be used in limited
situations - for example a gatekeeper MTA handling external mail which
will be forwarded onto a internal mail system (in which case it may not
have valid recipient information itself), or the case of a secondary MX
host.
In these cases the final destination MTA should inhibit sender callouts
for SMTP sessions originating from its gatekeeper/secondary. [In the
same way that SPF lookups would also need inhibiting]
One wrinkle is that recipient callouts that use the originating sender
can only be cached as a sender/recipient pair, so there may be rather
more callouts than you expect.
Nigel.
--
[ Nigel Metheringham Nigel.Metheringham@??? ]
[ - Comments in this message are my own and not ITO opinion/policy - ]
