> Jim Archer asked:
> I'm trying to figure out how, if it is possible, to use an
> application I wrote to validate a recipient address in an
> ACL. After going through the book and online, it seems that
> there is not a way for me to do this, although it does seem
> that I can use pipe in the system filter.
>
> I also saw that I can get an ACL to call a router, but I am
> confused about how that works.
>
> Does anyone know if it is possible to do this and if so what
> the trick is?
http://www.exim.org/exim-html-4.50/doc/html/spec_11.html#IX853
${run{<command> <args>}{<string1>}{<string2>}}
This requires starting a new process so if you need high performance
it isn't the best way -- but it can work for small volume or
testing purposes.
If it is a listening daemon process you can use the readsocket
(UNIX or INET sockets):
${readsocket{<name>}{<request>}{<timeout>}{<eol string>}{<fail string>}}
I have both GreyListing and SPF running this way
(i.e., readsocket) since I had problems with getting
any codes that implements them directly in Exim to
compile (on CygWin.)
Note: We do NOT block mail on SPF, but rather use it
to score in SpamAssassin, and intend to use it to drive
Greylist decisions. (In doubt, greylist -- pass SPF, no
point in Greylisting as they have likely setup real MX
SMTP servers anyway. That is, the mail is nor 'forged',
even if it may be spam.)
Greylist is using the Python Daemon and SPF is using the
Perl SPF spfd.
You could, for instance, get SPF working using the program
"spfquery" for each call, instead of running the daemon
and using the readsocket.
--
Herb Martin
