Intentional top note: Now I'm the one who needs sleep. A couple of
episodes of Farscape has cleared my head a bit.
On Tue, 11 Jul 2006 23:34:25 -0500 (Central Standard Time) "Gerald V.
Livingston II" <debuser(_at_)sysmatrix(_dot_)net> wrote:
On Tue, 11 Jul 2006 09:52:56 -0700 Lev Lvovsky <lists2(_at_)sonous(_dot_)com>
wrote:
On Jul 10, 2006, at 4:25 PM, Dallman Ross wrote:
* ^X-Spam-Status.*YES
* !^X-Spam-Status.*NO
Finally, your conditions seem silly to me. Why have a condition that
says to look for a "YES", then have a second condition that says to
look for the absense of a "NO"? I would think you could do the whole
thing with only one condition, better formulated.
yeah, me too - I slapped that together really quick - I found that
the first rule failed some messages, but will revisit that after I
fix this "problem".
Actually, If there are any conditions where a message may bypass
SpamAssassin it would end up meeting that combination of conditions. It
would not have an X-Spam-Status: Yes header so would fail that condition if
it were alone and would not be classified into the "probably spam" folder.
It would also NOT have an X-Spam-Status: NO header. It would pass that
condition thus counteracting the lack of a YES flag.
Why it bypassed SA or why it has neither header is up to the fine admin to
discover.
Gerald
Ignore most of that. I knew what I was going to write and looked at it all
wrong after being sidetracked.
Conditions are ANDed. In the recipe above if EITHER of the conditions fails
the whole recipe fails and falls through. So my contention that it would
only fail if missing both headers is false. It will also fail if missing
the YES flag whether or not it has a NO flag and it will fail if it HAS a
NO flag whether or not it has a YES flag.
So, the recipe above will still catch the "oops" condition I mentioned
where a message has bypassed SpamAssassin and there are X-Spam-Status flags
at all. It will ALSO catch a condition where Spamassassin has been
improperly set up and is not stripping old SpamAssassin tags prior to
inserting new ones. If a message has BOTH X-Spam-Status: YES -- AND --
X-Spam-Status: NO headers the above recipe will catch that failure. If one
leaves out the test for the nonexistence of the NO flag then a message with
both would pass.
I'm still tired and probably digging a deeper hole.
Gerald
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail