Re: [Exim] sender callout failing

Top Page
Delete this message
Reply to this message
Author: David Saez
Date:  
To: Philip Hazel
CC: Nico Erfurth, exim-users@exim.org
Subject: Re: [Exim] sender callout failing
Hi !!

> I have two reasons for not implementing different senders for callouts:
>
> 1. Callout is supposed to be checking that a bounce message can be
> delivered to the sender. That is the whole idea of sender verification;
> it implements the rule "I won't accept a message unless I can deliver a
> bounce if I can't deliver it."


It has sense when verifying the envelope sender. Maybe I misunderstood
that concept when applying it to the header senders. My purpose on doing
a callout on the header sender is to be sure that the recipient of that
message will be able to reply to that message by enforcing a the use of
real email adresses in the headers. The fact that the server for the
header address accepts bounces or not is not relevant for me in that
case. The concept "I won't accept a message unless I can deliver a
bounce if I can't deliver it." losses it significance when applied to
the header sender as exim will never deliver a bounce to the header
sender (maybe only to Error-To:?)

> 2. Allowing different senders is far too risky because of the problem of
> loops. I suspect (actually, I *know* :-) that people would turn it on
> without considering the possibility of loops, and there would be endless
> trouble.


For sure that somebody will do it. But now it's also possible to
configure exim to do very bad things, and at the end the loop will
break due to timeout and/or process limits.

> The only possiblity would be to disable callout verification
> in the code for the special address (i.e. don't rely on the admin
> configuring it). I do not like this (a) because it is so hacky and
> special-case, but also (b) because it doesn't conform to my point 1
> about bounce messages, as explained above.


Ok, I have re-read the documentation and it's clear how callout works
but maybe where it says "Exim makes SMTP connections to the remote
hosts, to test whether the address is acceptable" it will be more
concise to say "whether the address is acceptable and the remote
server accepts bounces" (yes, i never tought a server could not
accept bounces)

--
Thanx & best regards ...

Self-made man: A horrible example of unskilled labor.

----------------------------------------------------------------
   David Saez Padros                http://www.ols.es
   On-Line Services 2000 S.L.       e-mail  david@???
   Pintor Vayreda 1                 telf    +34 902 50 29 75
   08184 Palau-Solita i Plegamans   movil   +34 670 35 27 53
----------------------------------------------------------------