On Thu, Jul 15, 2004 at 09:55:41AM +1000, Chris Drake wrote:
Look. The specification, in effect, says
If you get a Received-SPF: neutral header, you MUST
pretend that SPF evaluation had not occurred at all.
A Bayesian filter chooses to ignore that "MUST".
Only if you configure things that way. You could generate a
Received-SPF: neutral when there is no SPF record, and then
a bayes filter will treat them the same.
I do something a little different. When there is no SPF, I supply a default of
"v=spf1 a/24 mx/24 ptr" - and the result is always pass or neutral. I add the
keyword "guessing" to the Received-SPF header in this case, but that doesn't
seem to have the same problem, since it is associated with no SPF
record (both pass and neutral) rather than a neutral result.
--
Stuart D. Gathman <stuart(_at_)bmsi(_dot_)com>
Business Management Systems Inc. Phone: 703 591-0911 Fax: 703 591-6154
"Confutatis maledictis, flamis acribus addictis" - background song for
a Microsoft sponsored "Where do you want to go from here?" commercial.