I've recently started playing with Spamassassin, and I've hit a problem that I can't find in the Apache documentation.
The RCVD_IN_DNSWL_MED test obviously checks to see if a server is whitelisted, and reduces the spam score by 4 if it is. Leaving the merits or otherwise of this whitelist's contents to one side, I'm getting an undesirable effect that must impact other people: I've got lots of mail servers. It's a hobby, okay? I just like mail servers. Spam arrives at server A. Server A is whitelisted. Server A sends it (with a .forward) to server B, which is running Spamassassin. Server B looks up server A in the whitelist and knocks 4 of the spam score. What's the cure? There's an implication that adding the IP of server A to 'internal_networks' will cause it to be ignored, but it doesn't seem to be working in this case. I checked the source code and there's something in _check_whitelist_rcvd() that appears intended to get around this problem (WLBLEval.pm). This suggests that either it's broken, or I'm not configuring it correctly (i.e. setting the wrong option entirely?) Before I launch into rewriting this module in a language I only have a passing acquaintance with, can anyone point me in the right direction? I've found and read things like http://wiki.apache.org/spamassassin/TrustedRelays http://wiki.apache.org/spamassassin/TrustPath and http://spamassassin.apache.org/full/3.2.x/doc/Mail_SpamAssassin_Conf.html but I guess I'm missing something fundamental. Thanks, Frank. -- View this message in context: http://www.nabble.com/RCVD_IN_DNSWL_MED-test---I%27m-on-the-whitelist-so-spam-gets-through-tp23231790p23231790.html Sent from the SpamAssassin - Users mailing list archive at Nabble.com.
