No, it does not handle it correctly, so you should not learn messages that have already been detected as spam. This is one thing that should be improved (the VB6 source code is there).
SpamAddIn could also work like SpamBayes: Put the spam status and analysis into user definable fields when it is used as filter (SpamBayes puts the probability into Spam field, which you can make visible in Outlook). Currently it creates a Spam Status text attachment. One more unrelated thing comes to my mind now: Create a simple application for Outlook Express, that converters user-definable folders into mboxes and calls sa-learn. My batch file does that already, but it would be easier for non-technical users. I wonder how the dbx-mbox conversion could also handle only the original e-mail encapsulated as attachments by SAproxy? How easy it is to extract the original message when it is in mbox format? Have to check that. ----- Original Message ----- From: "Dean Gallea" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Saturday, August 09, 2003 8:02 AM Subject: [Spamassassin-saproxy] Outlook Add-In for SpamAssassin is here > Harri, > > Does the mbox conversion from Outlook allow SA-learn to properly parse spam > messages that are encapsulated as attachments by SAProxy? Or, will learning > those "pollute" the Bayes statistics because they all have the common > SAProxy header "This mail is probably spam..."? Should I only learn missed > spam (which is not encapsulated) and ham? > > -- Dean ------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01 _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
