Re: [exim] 4.87 and acl_check_data

Góra strony
Delete this message
Reply to this message
Autor: Evgeniy Berdnikov
Data:  
Dla: exim-users
Temat: Re: [exim] 4.87 and acl_check_data
On Wed, Apr 13, 2016 at 12:14:01PM +0200, Heiko Schlittermann wrote:
> Evgeniy Berdnikov <bd4@???> (Di 12 Apr 2016 13:37:37 CEST):
> > acl_check_data:
> >
> >   # Deny if the message contains an overlong line.  Per the standards
> >   # we should never receive one such via SMTP.
> >   #
> >   deny    condition  = ${if > {$max_received_linelength}{998}}

> >
> > "References:", 2. in my configuration bounces are sent through the same
> > mail relay and are blocked by this rule, because they include headers
> > of the original mail with oversized lines. So bounces are lost and
> > my users have no hints that their mails were dropped.
>
> With the above configuration it's not your system sending the bounces!
> So it's not your problem, but the senders problem, isn't it?


Yes, but... I have a chain of two "types" of SMTP servers inside office,
with different roles:

SMTP clients --> SMTP(A) --> SMTP(B) --> LMTP (mailboxes)

Servers B were upgraded to Exim-4.87, severs A were not. Currently
if B rejects mail, bounce is generated on A, but A can not deliver it
directly to mailbox (it has no mappings "address->mbox"), it should
forward bounce through B. If bounce is rejected by B on the same reason,
it simply disappears.

Well, I can push the rejection edge towards SMTP client, rejecting
malformed mails on A. But problem was that after a "blind" upgrade
of servers B to Exim-4.87 I suddenly got RT tickets about lost mails.
If new exim acquire some new restrictions, the same situation could
happen again. I consider this situation as a problem of interoperability.
IMHO, the best solution today is to accept and pass all bounces.
--
Eugene Berdnikov