On Thu, 17 Oct 2002, Kenneth Chen wrote: > Hey Justin: > > Thanks for your answer! I'm curious about something else, though: does > your procmail recipe say (in words) "Take whatever has 5 stars OR more and > pipe it to /dev/null?" I'm wondering about that last part with the *.*.
That's what the 2nd recipe does. The first one delivers it to $SPAM_DIR which is /var/mail/spool/quarantine/spam in this case. When I actually go public with my SA setup, I'll just score messages with SA (alter the subject only if the score is >=10), and keep a copy of all messages hitting probably 10 for my reporting/logging. I won't be using procmail to reject mail unless I have to. I'm going to pass that duty to the users. > And what is the difference between your ".*\(\*\*\*\*\*.*)" and > "\*\*\*\*\*.*" examples? The default SpamAssassin X-Spam-Score header looks like X-Spam-Score: 7.2 (*******) I can't use that however. I have a lot of OutLook users and I haven't found a way to create a filter in LookOut that will match that (and higher scores). I came up with a solution though. Since I'm calling SA from MIMEDefang, MD makes all the changes to the message. SA can only score the message and return the results. Rearranging the header in MD is simple. Mine now looks like: X-Spam-Score: ******* (7.2) LookOut can match that. I can tell LookOut to search all headers for the entire string "X-Spam-Score: *******" and it will match all messages with 7 OR MORE stars. Works slick. Now if you want to have two actions for your spam, say delete above 10 and move to a "Possible Spam" folder for 5-9, you need to match the rule with the most stars first. Otherwise you match on the shorter string of stars will match from 5 to infinity and you other rule will never be met. > Thanks -- I really have no experience with syntax of this nature so > anybody's help would be much appreciated. The syntax is a royal pain, even for me. I'm not great at regexs. You can pick it up from looking at the example recipes, other peoples' recipes, or trial and error. The asterisk is a special character that has a certain meaning in a regex. To use it as a basic character you have to "escape" it. ie, put a slash in front of it. \* will match a single star. Now you also have to allow for other characters on the line. That's where the ".* comes in to play. X-Spam-Score: \*\*\*\*\*.* matches the string "X-Spam-Score: *****" and anything after it. The "^" means beginning of line. The "$" by itself means end of line. For example: grep bash$ /etc/passwd to see who all is smart and uses bash for their shell. :) I know enough regex stuff to be dangerous. It comes in time though. HTH Justin ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
