Re: Dest-dependent source rewrite?

Página superior
Eliminar este mensaje
Responder a este mensaje
Autor: Dr. Rich Artym
Fecha:  
A: exim-users
Asunto: Re: Dest-dependent source rewrite?
In message <Pine.SOL.3.96.970410095212.25953N-100000@???>,
Philip Hazel writes:

> The only logical thing would presumably to send two different messages,
> i.e. to have different headers for different deliveries. I'm afraid Exim
> isn't up to that.


Exim already performs routing and transport per destination domain, so
it already has suitable multiple trigger and transformation points. The
missing ingredient seems to be lack of a mechanism by which the delivery
set could be split into subsets. If that existed then presumably it
would be possible to run an additional rewrite during the subsetting.

> Notwithstanding the comment I made about about separate programs, there
> is a related item on the Exim wish list which might help here. Somebody
> wanted to do automatic encryption on mail to certain domains. I have an
> intention (though no details or timescale) to implement a facility on
> Exim's transport drivers such that the outgoing message is filtered by a
> configured program on its way out. Such a program could encrypt the
> message, and could also do any other munging it liked. However, it would
> see only the body of the message, not the envelope.


Yes, filtering on the way out would be very useful, but as Piete says,
the whole item needs to be passed over, not just the body. Almost every
header line contains information that could be valuable to one kind of
filter program or another -- that's certainly the case for encryption,
which one would expect to be strongly determined by destination and
source addresses at the very least.

I think it would be inefficient to use that approach for the header
rewriting that I was describing though, simply because most of the
required machinery seems already to exist within Exim. The problem
is rearranging it into the right order. :-) Having said that, if such
external filtering were being implemented then the message would be
rewritten physically, and that would be an excellent time to invoke
the kind of internal header rewriting that I described, controlled by
conditions asserted by the routers. Almost two for the price of one!

Rich.
--
###########  Dr. Rich Artym  ================  PGP public key available
# galacta #  Email   : rich@???         158.152.156.137
# ->demon #  Web     : http://www.galacta.demon.co.uk  - temp page only
# ->ampr  #  AMPR    : rich@g7exm[.uk].ampr.org 44.131.164.1 BBS:GB7MSW
# ->NTS   #  Fun     : Unix, X, TCP/IP, kernel, O-O, C++, SoftEng, Nano
###########  More fun: Regional IP Coordinator Hertfordshire + N.London