Den 2012-03-09 09:11, Jari Fredriksson skrev:
No, SA will scan messages even if they originate from
trusted_networks,
and X-Spam headers will be added.
most sites bypass scanning of there own mails since its just ham
(permit_mynetworks) in postfix, my point is why not learn ham in bayes ?
There is something else in your system deciding that the messages are
not passed to SpamAssassin.
here i have removed all signs of permit_mynetworks in postfix main.cf,
only keeping permit_sasl_autienticated
there is other route in mailflows that could be the problem in not seen
x-spam headers if using amavisd-new recipient domain must be local and
spam score should be highter then spam_tag level, but that said i dont
know if OP do use amavisd-new