See the attachment to bugzilla # 2508. It came in with a very low score because the header (the entire header, not just a particular field) was mangled, so most of what would have been header spamsign showed up in the body, and the base64'd body content didn't get decoded.
My question is, should I feed this to sa-learn? It won't be able to parse it properly, so I'm wondering whether it'll do an effective tokenization. Of course the spam isn't very effective in this condition either, but I'd still rather it weren't in my inbox. ------------------------------------------------------- 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
