On Fri, 4 Jul 2003, Bryan Henderson wrote:
> But those standards don't seem to bear directly on this issue. They
> talk about communications between MTAs and interpretation of headers
> in emails.
Interpretation of envelopes is the main issue for routing.
> This question is just about me controlling exim. Options
> on the exim command line that simply override the exim configuration
> files would do the job.
-C allows you to use an alternate configuration file (if you are
suitably privileged). -D allows you to set macro values (ditto).
If what you are trying to do is to route certain addresses "unusually",
you may be able to do it by putting a special router in your normal
configuration file that detects these special addresses, or is triggered
by an if-then-else predicated on the setting of a macro.
> I have a related question about standards compliance: I notice in the
> logs lots of messages being rejected because of syntactically invalid
> HELO. Is this because Exim really doesn't know how to receive the
> message or just that Exim is enforcing the standard?
Exim is enforcing the standard. It will also reject syntactically
invalid MAIL, RCPT, etc. commands. In the case of HELO, the most common
syntax error is an underscore in the name. Exim does have a workaround
for this, because it is such a widespread error. See helo_accept_chars.
Use it if you feel you must. We don't.
> I ask because I
> have to believe that people wouldn't be sending messages like this
> unless most systems out there are accepting them.
Hollow laugh. It is my impression that some MTA software, usually
commercial, is released as soon as it can talk to itself. The vendors
don't bother (a) reading the RFCs very well or (b) acknowledging the
fact that there are other MTAs (both free and commercial) out there, so
they never bother to test with them.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book