http://issues.apache.org/SpamAssassin/show_bug.cgi?id=4674





------- Additional Comments From [EMAIL PROTECTED]  2008-01-22 02:03 -------
I agree.  It would be nice if the hashcash module could pick up some default
settings.  I can see a couple of options for how this would work.

1) look for headers left by the MTA.  The MTA knows what addresses it is allowed
to accept mail for, and should be trusted.  If the MTA adds a "Envelope-To:" or
"Received: ... for <[EMAIL PROTECTED]> header, look for stamps matching that.  
You would
have to be careful to make sure they were inserted by the local MTA, and not
spoofed of course.  This may not work in more complicated setups where mail is
always forwarded, but might work for simple setups.

2) Check the hostname, and accept stamps for [EMAIL PROTECTED]

3) Do statistical analysis.  If I have received lots of mail with a given stamp
that is classified as not-spam, add that address to the list.  This would allow
automatic classification of mailing lists.  

4) tell people who package spamassassin, such as linux distributions to figure
it out themselves.  If the installation process asked for a list of domains we
accept mail from, that would go a long way.  That is also information other
plugins may want, it doesn't need to be restricted to hashcash.

Of these, only (3) can catch people who have .forward files set up to relay
other mail to the spamassassin server.  For this, it would be nice to have some
warning sent to the user suggesting that "This email is stamped with hashcash,
but none match your address.  Please add any additional email addresses to
user_prefs".  Doing this in a non-annoying way would be tricky.  At the very
least though, there could be a test added: HASHCASH_WARN_INVALID_EMAIL=.001 to
let people know that there was something wrong, and something they could fix.




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

Reply via email to