Re: [Exim] Callout timeouts: opinions sought

On Fri, 20 Aug 2004, David Woodhouse wrote:

> On Fri, 2004-08-20 at 14:58 +0100, Nigel Metheringham wrote:
> > On Fri, 2004-08-20 at 14:46, Alan J. Flavell wrote:
> > > I'm convinced that I've seen an MTA respond OK to MAIL FROM:<> , and
> > > then respond to the RCPT TO with a diagnostic which rejected the null
> > > sender address.
>
> <...>
>
> > That might well happen if someone forges a message from (say) the
> > exim-users list (using the list address as envelope sender - since the
> > list sends no mail itself it is configured to reject bounce messages),
> > or from people using one-time or keyed sender addresses - David
> > Woodhouse is, I believe, one person doing that.
>
> Indeed -- but those are situations where it's perfectly acceptable to
> reject the mail, and nobody needs to see a bounce because you _know_
> it's a forgery.

Indeed, I agree with you both, and there are addresses for which we do
that here also, for various reasons (as explained in the message which
we return, for those whose MTAs don't invent their own explanation
instead).

But it seems that ole.com (and terra.es, which is evidently the same
thing) responds in this way to every callout request. Possibly they
are looking for some specific address (abuse, postmaster etc?) before
deciding not to accept <> as envelope sender; I've no idea what their
algorithm is.

cheers