Re: Fallback Host & smartlist integration problems

Top Page
Delete this message
Reply to this message
Author: Philip Hazel
Date:  
To: Christoph Lameter
CC: exim-users
New-Topics: Re: Fallback Host, SmartList integration OK.
Subject: Re: Fallback Host & smartlist integration problems
On Mon, 23 Dec 1996, Christoph Lameter wrote:

> 1. When resending e-mail using smartlist exim generated a
>
> Resent-To:
>
> line and lists one by one all receipients on the list!


We had a discussion about this on the list a few weeks ago. (I know you
are new to the list since then.) The RFC mandates that if any resent-
headers are present, these should be treated as the working set of
headers, and then there has to be at least one of To, Cc or Bcc. Exim's
rules for a missing To, Cc, or Bcc header (at least one is required)
cause it to add a To header if (and only if) the envelope is coming in
from a list of addresses on the command line. In other circumstances it
adds an empty Bcc header. In the next version of Exim there will be an
option requesting that it always add an empty Bcc header in all cases.

I hadn't realized that mailing list software passed over large numbers
of addresses in the command line like that.

> I hacked exim by changing accept.c. It seems that exim only looks for
> Resent-To if any field beginning with "resent-*" is found into the header.
> This means it later concludes that there is no TO: field and supplies it.
> I dont know if the following solution is correct but I sure dont want to
> have all people on the list listed in each outgoing message:


Michelle Dick <artemis@???> was able to configure smartlist to
make even the current Exim behave the way she wanted, without hacking
exim. She might even have posted her solution. There's an archive of the
Exim list at

http://www.roads.lut.ac.uk/lists/exim-users/exim-users.archive

and in HTML via Hypermail at

http://www.roads.lut.ac.uk/lists/exim-users

You should be able to find the discussion there.

> 2. Setting the Fallback Timeout options does not work.


> I have tried the following configuration file. But when trying to deliver
> a test message to a system that does not support mail (there is nothing on
> port 25!) I still get an exim process that is hanging around for 15
> minutes before giving up and giving the message to the mail router.


If a system is up and running, but not listening on port 25, it should
given an immediate "connection refused". A 15-minute delay suggests a
different problem. If the system is switched off, the attempt to connect
to it will fail. In the current release of Exim, you get your system's
default time limit on the connect() call. I don't know about other
systems, but on Solaris 2.4 this time is around 3.3 minutes. In the next
version of Exim there is a configuration option to specify a shorter
timeout for connect().

> Also: How can I make sure that failing DNS lookups do not take too long?


I'm afraid you can't at the moment. You get whatever timeouts are built
into your system. I have made a note to investigate whether it is
possible to put a time limit on a DNS lookup. Nobody has ever suggested
this before; thanks for the idea.

--
Philip Hazel                   University Computing Service,
ph10@???             New Museums Site, Cambridge CB2 3QG,
P.Hazel@???          England.  Phone: +44 1223 334714