bayes_dump_to_trusted_networks - generate configuration from Bayes database
sa-learn --dump | bayes_dump_to_trusted_networks [opts] > trust.cf bayes_dump_to_trusted_networks bayes.dump [opts] > trust.cf
options:
--minham n --rdns
This tool uses a dump of your Bayes database to determine which IP addresses are 'trustworthy', and therefore should be listed in 'trusted_networks' lines in your configuration.
This will reduce unneccesary DNSBL lookups, will whitelist mails from trustworthy sources, and allows several SpamAssassin rules to operate more effectively.
A 'trustworthy' IP is one that is trusted not to forge emails; in other words, it's not a subverted machine running a proxy, or one under spammer control.
As such, any IP that has relayed more than 3 ham mails is considered trustworthy. It doesn't matter if it has ever relayed spam mails to you, since large ISP smarthost relays will have done so -- relaying both ham and spam from their customer pool. (The important thing is that it relayed the mail without forging sender address information.)
Require n or more ham messages before considering an IP a candidate for trust. Default: 3.
Annotate with reverse-DNS for that IP address. Slows things down, but easier to read.
Net::CIDR::Lite
