----- Original Message -----
From: "Marc Perkel" <[EMAIL PROTECTED]>
Mark,
Since I don't use Exim, do you know how I can implement this to call
from SA?
Something like this would work:
header __RCVD_IN_JMFILTER eval:check_rbl('JMFILTER',
'hostkarma.junkemailfilter.com.')
describe __RCVD_IN_JMFILTER Sender listed in JMFILTER
tflags __RCVD_IN_JMFILTER net
header RCVD_IN_JMFILTER_W eval:check_rbl_sub('JMFILTER', '127.0.0.1')
describe RCVD_IN_JMFILTER_W Sender listed in JMFILTER-WHITE
tflags RCVD_IN_JMFILTER_W net nice
score RCVD_IN_JMFILTER_W -1.5
header RCVD_IN_JMFILTER_B eval:check_rbl_sub('JMFILTER', '127.0.0.2')
describe RCVD_IN_JMFILTER_B Sender listed in JMFILTER-BLACK
tflags RCVD_IN_JMFILTER_B net
score RCVD_IN_JMFILTER_B 1.0
header RCVD_IN_JMFILTER_Y eval:check_rbl_sub('JMFILTER', '127.0.0.3')
describe RCVD_IN_JMFILTER_Y Sender listed in JMFILTER-YELLOW
tflags RCVD_IN_JMFILTER_Y net nice
score RCVD_IN_JMFILTER_Y -0.5
Thanks Bill,
Yellow listing shouldn't get a score. It should be used in a way that says
not to run and IP based black lists because the black list will be wrong.
It's for avoiding false positives. I think that the white score could
be -5. I'm confident in that list.
I don't believe there is any facility for derailing tests once a message has
been passed to SA for examination. So, the best one can do once SA had
received the message is apply a negative weight if the particular test
result should help classify the message as ham.
Actual scores assigned to each test are really up to the SA administrator
since they are the only ones that really know how any particular test score
will affect their spam/ham detection rates. The scores I showed were simply
examples.
Bill
