On Tue, 9 Jan 2007 20:17:39 +0100, Heiko Schlittermann
<hs@???> wrote:
>Not tested (and modulo syntax errors) *could* do the trick:
>
> hosts_require_auth = ${if exists{CONFIG_DIR/passwd.client} \
> { ${lookup{$host}lsearch{CONFIG_DIR/passwd.client}{$host_address}} }\
> {} \
> }
I changed this to
| hosts_try_auth = ${if exists{CONFDIR/passwd.client} \
| { ${lookup{$host}nwildlsearch{CONFDIR/passwd.client}{$host_address}} }\
| {} \
| }
and this seems to work fine at least fot the test cases I did. Thanks
for helping! I hope your wife and your kid are fine.
>(Or use hosts_try_auth because the hosts_require_auth isn't mentioned in
>30.2 of the spec.)
Debian exim4 uses hosts_try_auth. I do not have an idea whether this
makes sense or not.
>*If* this option is evaluated for each transport/delivery attempt in
>turn, it could work: if the $host (name) you try to connect to is in the
>passwd.client file, than use the $host_address as value for the option.
>The $host_address should come from the router (thus avoiding the extra
>DNS lookup).
Seems to be this way.
Can other people please comment?
This is getting more and more ugly. Another issue that could more
easily be solved by way of having more control about authentication in
the transport, as in bugzilla #66.
Greetings
Marc
--
-------------------------------------- !! No courtesy copies, please !! -----
Marc Haber | " Questions are the | Mailadresse im Header
Mannheim, Germany | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834
