Flow of addresses through the rewrite rules in sendmail 8.8
Last Update 1999-11-24
This is a draft!
Please
send me corrections/suggestions.
Introduction
The
Sendmail Installation and Operation Guide
explains in
Semantics of rewriting rule sets
the flow of addresses through the different
rewriting rules
.
If you are not familar with the way
sendmail
works,
read these documents first,
before you try to understand this!
However,
this guide
isn't up-to-date.
Therefore,
I'll try to outline the flow of addresses through
the rewrite rules
based on debugging
sendmail 8.8
with
-bs -d21.1
.
Just as a reminder, the parts of an
(E)SMTP
dialogue
which are important here are:
MAIL From:<sender@address>
to specify the sender's address,
and
RCPT To:<rcpt@address>
to specify the recipient's address,
of which multiple can be given.
The sender's address is rewritten by:
-
3, 0, 4
-
to do a basic check of the address and find the mailer
for the statistic program.
If the address resolves to the
$#error
mailer, the address is rejected.
-
3, 1, 4
-
This is stuff as explained in the
doc.
But why doesn't it use the outcome of ruleset 3 from the previous
rewriting (as done below)?
Is this something special with my test addresses?
(I used local and off-site addresses).
-
check_mail
-
This is the
new hook
of
sendmail 8.8.
The recipient's address is rewritten by:
-
3, 0
-
to do a basic check of the address and find the mailer
to use for delivery.
-
2, R= , 4
-
The input for these rules is the result of ruleset 3.
R= is taken from the mailer found before.
This is again the usual stuff as explained in the
doc.
-
check_rcpt
-
This is the
new hook
of
sendmail 8.8.
[(links)]
[Hints]
[Avoiding UBE]
[cf/README]
[New]
Copyright ©
Claus Aßmann
Please send comments to:
<ca at sendmail.org>
Disclaimer: the information provided may be inaccurate or outdated
or incomplete.
Please
contact me
if you find an error.