Re: [exim] LMTP over TCP - returning errors

Philip Hazel wrote...

> I'm starting to look at this.

Thanks very much Philip. I can't even pretend to be a programmer - I can
sometimes effect the change I want from someone else's code, but after an
hour or two of fiddling with this issue I was still lost.

> The first thing that occurs to me is that
> 4XX is a temporary error which will not of itself cause the message to
> be returned. The return will not happen until the retry time is
> exceeded. Of course, you can set the retry time to zero, but logically
> that's how Exim handles it. So I would have expected the message to be
> "retry time exceeded" if anything. Can you confirm whether or not you
> see that in your log? I will peer at the code and see if the 4XX error
> message is still preserved, and then see if anything can be done.

In the log we would see the actual error response - for example:

2006-03-01 17:50:30 1FEVSn-0002pP-Ll == someone@??? R=ldap_mailrcpt T=dllstx09 defer (-44): SMTP error from remote mail server after RCPT TO:<someone@???>: host 1.2.3.4 [1.2.3.4]: 452 4.2.2 Over quota

But, when the address eventually timed out, the bounce message would say
only "retry time exceeded" as you suggest above. I did notice some
strange behaviour with respect to retries while testing - some messages
returned immediately when I expected them to be left on the queue. Can't
really quantify that without more testing but that's of secondary concern
to us at the moment.