With all of the discussion on rbl's and utilization of such, also with
the recent changes
in spammers counter measures, in particular, the recent 'bayes
poisoning' attempts (specifically
random text with an image, if this is run through spamassassin, and
learned as spam, the text
portion will poison your bayes db). I would like to open the following
discussion in terms of
the toaster's methodology.
What, in terms of tcpserver's rules, could be done to this. Obviously,
blacklists have their
pro's and con's, i.e. what is one man's garbage is another man's gold.
In particular, if one is going to check blacklists, what could be done
to do this at the
tcpserver level, for instance, when the run script is run for smtp, the
options, '-R' and '-H'
is set, so if one is going to check a blacklist, and do a dns lookup
anyway, what would be the
pro's and con's of removing '-R' and '-H' and changing the program from
'true' to something else?
If certain tests weren't met at that level send back the appropriate
error message, and terminate
the connection.
Ideas?
---------------------------------------------------------------------
QmailToaster hosted by: VR Hosted <http://www.vr.org>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
