Autor: Alan J. Flavell Data: A: Exim users list Assumpte: Re: [Exim] 451-Invalid HELO
On Thu, 29 Apr 2004, Chris Edwards wrote:
> | RFC2821 S4.1.1.1 for syntax\n451-of the HELO command. This is a
> temporary error. You may try again | after\n451 this server
> (**my-server-ip**) has been fixed.
>
> ooh - didn't spot on first reading - this is a temp failure
Hmmm, I hadn't been sure whether to comment on it in my first reply...
> they're giving 4xx after RCPT for the broken HELO
>
> interesting...
It sort-of gives the site an opportunity to fix the problem, doesn't
it? I have to admit that if I'm going to do anything, then I'm
inclined to return a 5xx, on the supposition that the sending site is
unlikely to respond to a 4xx by fixing their problem on their own
initiative within our mailer's timeout of a few days.
OTOH in practice the majority of cases will be viruses (and exploited
proxies?) which will react to 4xx in much the same way as to 5xx i.e
abandon the current attempt.
TBH what I'm doing about unqualified HELO strings at the moment,
rather than an outright rejection, is to try a Long Wait at RCPT time
(somewhere above 60 secs, but well short of the 5min timeout that's
suggested in the RFC), which shakes off a quite a lot of viruses and
exploited proxies; and then toss a goodly score into the spamassassin
rating, in the hope that other spam-suspect features will take it
above our rejection threshold if it's not bona fide.
I still can't make up my mind about underscores, though. My personal
inclination would be to reject them outright. But there has been
pressure on us to tolerate them.