From: Wietse Venema (wietseporcupine.org)
Date: Mon Mar 05 2007 - 18:56:54 CST

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Ed Schouten:
> * Wietse Venema <wietseporcupine.org> wrote:
> > The problem was not as bad as I thought, because the _TODO case
> > happens only when a probe takes longer than 6 seconds by default
> > (see address_verify_poll_count and address_verify_poll_delay).
> >
> > However, be aware that the _DEFER status does not distinguish
> > between different forms of "try again" errors (for example, a 4xx
> > server reply, or a failed attempt to reach a server). This is
> > because normal mail does not need the distinction, and because
> > address verification probes use the same infrastructure as normal
> > mail.
>
> Yes, indeed. But this is exactly what I need, because I'd rather live
> with a small amount of spam, instead of emails that could possibly be
> deferred for an excessive amount of time. Is there a chance that options
> like these will be part of the 2.4 release?

Postfix 2.4 was feature frozen almost two weeks ago, and I won't
be ready to discuss 2.5 designs for another month or so.

Ignoring "defer" errors with only one feature and not all the other
ones makes Postfix less consistent. Obviously, adding a _defer_code
to every _reject_code is not a sustainable model.

Apart from that, ignoring "defer" errors makes the result of address
verification less predictable. The same address will or won't pass
depending on the phase of the moon and the state of the network.

        Wietse

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]