John Owens writes: >I've got a small problem with SpamAssassin that I hope you can help me >with. I receive a significant amount of email forwarded through >another account. I don't have any access to the way SpamAssassin is >run on this account; it's run on every piece of mail. And it wraps the >message in a "Start SpamAssassin results" wrapper. Again, I don't have >any ability to control how this forwarding site runs SA. > >Now, the forwarding site _does_ properly mark spam. The problem is >that when this message is then forwarded to my real account (which >also runs SA, or rather I run SA via procmail) it does NOT trigger as >spam. You will note in the message below that the original SA run had >11.50 hits and was flagged as spam, but my SA invocation finds 1.9 >hits and doesn't mark it as spam. This is probably because the >forwarding site wrapped it.
John -- there's a number of options: 1. if you can set up user preference data ( a ~/.spamassassin/user_prefs file) on the upstream site, you can use "report_safe 0" to turn this off. 2. you can just match on the "X-Spam-Flag: YES" header added by the upstream and save yourself the bother of running sa on those messages; they've already been marked as spam. 3. you could pipe the incoming mail through "spamassassin -d" as it arrives. At the same time, developers, this is a problem IMO. I think the upstream site should *not* use the default "report_safe 1" in this case, since the mails are being forwarded on. Perhaps we should add some doco to that effect to the sitewide.html webpage... --j. ------------------------------------------------------- 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