Then the Authentication-Results: is written to SYSLOG,
it is formated differently depending on the outcome of
the domainkeys test.
The normal out formated like this:
Authentication-Results: $MTA $SENDER/FROM $DK-RESULT
($TESTING)
But if dk-filter gets a "permerror", output will
formated like this:
Authentication-Results: $MTA $DK-RESULT
Would it be possible to change the last one to:
Authentication-Results: $MTA $SENDER/FROM $DK-RESULT
($TESTING)
where $SENDER/FROM could be set to ERROR, format error,
or equal. This would help parsing the logfiles much
(since SenderID Milter uses the same format...)
I've included a sample of the log output...
/P
Some SYSLOG entries...
Logged In: YES
user_id=1048957
The logging is done by libmilter, not by dk-filter, so I
can't really fix it. However, it's only logging exactly the
header being added to such messages, so this also means the
header is also missing the sender/from information.
Logged In: YES
user_id=1048957
Looked at this a little further, and the absence of such
header information is proper since libdk reported "bad
format". It's not possible then to report on which header
or other message property was used to authenticate the
message, because authentication didn't happen.
Logged In: YES
user_id=370342
(I didn't think of this as a bug, it was supose to be a RFE.)
Yes, if from/sender is missing, then it's hard to add them.
But what i meant was that if it's impossible to print
from/sender, then print out the cause of the error something
equal like this:
Milter insert (1): header: Authentication-Results:
mx1.vattenfall.se; from="can't parse From: header";
domainkeys=permerror (bad format)
Logged In: YES
user_id=1048957
This would violate the current format defined in section 2
of the Authentication-Results: draft.
(Moving this back to a Feature Request.)