On 21/10/16 16:22, John Hardin wrote:
I was going to say: you can't write a rule based on the *current* AWL
adjustment because that's calculated after all the rules have hit. But
SA *could* potentially have a rule that checks the current historical
average that AWL uses...
I suggest you file a New Feature bug to expose a mechanism to use the
current AWL average (not the per-message adjustment) in a rule.
A plugin I've developed could be handy here:
https://github.com/fmbla/spamassassin-tagmatch
tagmatch TAGMATCH_TXREP_IP_HIGHSCORE _TXREP_IP_MEAN_
/^[1-9][0-9]+(?:\.[0-9]+)?$/
describe TAGMATCH_TXREP_IP_HIGHSCORE TXRep mean score quite large
score TAGMATCH_TXREP_IP_HIGHSCORE 0.1
tagmatch TAGMATCH_TXREP_IP_LOWSCORE _TXREP_IP_MEAN_ /^\-[0-9]{2,}(?:\.[0-9]+)?$/
describe TAGMATCH_TXREP_IP_LOWSCORE TxRep mean score quite low
score TAGMATCH_TXREP_IP_HIGHSCORE -0.1
tagmatch TAGMATCH_TXREP_IP_UNKNOWN _TXREP_IP_UNKNOWN_ /^1$/
describe TAGMATCH_TXREP_IP_UNKNOWN IP unknown to TxRep
score TAGMATCH_TXREP_IP_UNKNOWN 0.1
You can use the other TXREP Tags of course (and any other tags provided
by plugins):
https://spamassassin.apache.org/full/3.4.x/doc/Mail_SpamAssassin_Plugin_TxRep.html#template_tags
Paul
--
Paul Stead
Systems Engineer
Zen Internet