Auteur: David Woodhouse Date: À: Matthias Waffenschmidt CC: exim-dev, Kai Risku Sujet: Re: [exim-dev] [PATCH] Rudimentary XFORWARD-support in
smtptransport
On Thu, 2006-05-18 at 13:55 +0200, Matthias Waffenschmidt wrote: > > That problem is fixed in Exim -- I don't really see much need for
> > anything like XFORWARD.
I think I'm missing part of the big picture here. I see those specific
possibilities, but I don't actually see how they'd fit in to a sane
system architecture.
> You could for example deny at recipient time (recipients selectively)
> instead of having only one deny/accept option after data.
So you have one dumb MTA which accepts mail and then tries to forward it
to a second, using XFORWARD, and the _second_ then has the option on
rejecting individual recipients that the _first_ should have rejected in
the first place? And the first then has to bounce them?
> But the most important point IMO is if callouts would send this
> information via XFORWARD, because there currently is no way of
> transfer such kind of data to MTA2 (as there is no data transferred).
So you have one dumb MTA which does callouts to the second _before_
accepting the mail?
Either way, I'm not really sure why you'd do any of this.