Re: [SAtalk] sa-learn & spamassassin question
At Wed Jan 28 09:39:06 2004, jean-christophe valiere wrote: > I agree with what you said, the probleme is that in the mail I attached the word > have some swaped letters like: > Hi, > > Genierc and Sepur Viarga (Caiils) available online! > Most trusted online source! > > > Cilais or (Spuer Vagira) > takes affect right away & lasts 24-36 huors! > FOR SUEPR VAIRGA TOCUH HERE > > I wanted to know if somebody has already received such mail and if it was > tagged as spam. I've received lots of them, but I've had to add specific rules to catch it. Once you've trained Bayes with a couple of these (assuming you're using Bayes, that is), it shouldn't have any problems with it. Check the list archives for my mail of 23rd Jan, in the thread "Re: [SAtalk] help pleasecant stop them at all." for the rules. Martin -- Martin Radford | "Only wimps use tape backup: _real_ [EMAIL PROTECTED] | men just upload their important stuff -o) Registered Linux user #9257 | on ftp and let the rest of the world /\\ - see http://counter.li.org | mirror it ;)" - Linus Torvalds _\_V --- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn ___ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
Re: [SAtalk] sa-learn & spamassassin question
On Wed, Jan 28 2004 (10:39:06 +0100), jean-christophe valiere wrote: > Cilais or (Spuer Vagira) > takes affect right away & lasts 24-36 huors! > FOR SUEPR VAIRGA TOCUH HERE > > I wanted to know if somebody has already received such mail and if it was > tagged as spam. I get quite a lot of these recently. They consist of a plaintext part containing some random words and of a HTML part containing the actual spam content. However, a well-trained bayes filter will spot them easily: X-Spam-Flag: YES X-Spam-Status: Yes, hits=5.5 required=5.0 tests=BAYES_99,HTML_MESSAGE autolearn=no version=2.63 Cheers, Toni --- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn ___ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
Re: [SAtalk] sa-learn & spamassassin question
On Tue, Jan 27, 2004 at 11:00:07AM -0500, Matt Kettler wrote: > At 12:22 PM 1/27/04 +0100, jean-christophe valiere wrote: > > >Hi, > > > >I've got a little problem with the mail that is attached. > >When I try spamassassin -t -D rulesrun=255 < mail.txt it is > >nor considered as spam. > >So I do spamassassin -r -D rulesrun=255 < mail.txt ans it > >sayes me that Razor already learnt this message but it is not > >in the report when I test it. > >I finally decide to do sa-learn --spam mail.txt and it says me > >that it doesn't learn anything from the mail. > > > >I just use razor2 with spamassassin and would like the mail to > >be tagged as spam from my bayes and from razor2. > > > >Thanks. > > > 1) Training one message alone is not usualy enough to "flip" the bayes > score of the message if there is strong evidence in dozens of other > messages that the tokens are nonspam.. Look at the bayes token output of > spamassassin -D. > > 2) if the spam in question has a habeas warant mark, add the habeas swe > headers to your bayes_ignore. > > 3) No one person (short of a cloudmark employee) can declare a message to > be spam in razor. Your report counts, but one person alone is generally not > enough. Also, make sure you've set up for razor reporting using razor-admin > --register. > I agree with what you said, the probleme is that in the mail I attached the word have some swaped letters like: Hi, Genierc and Sepur Viarga (Caiils) available online! Most trusted online source! Cilais or (Spuer Vagira) takes affect right away & lasts 24-36 huors! FOR SUEPR VAIRGA TOCUH HERE I wanted to know if somebody has already received such mail and if it was tagged as spam. Thx. -- Jean-Christophe VALIERE EPITA Paris Sud, 14-16 Rus Voltaire, 94270 Le Kremlin Bicetre Tel: +33 1 44786872 E-Mail : [EMAIL PROTECTED] --- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn ___ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
Re: [SAtalk] sa-learn & spamassassin question
At 12:22 PM 1/27/04 +0100, jean-christophe valiere wrote: Hi, I've got a little problem with the mail that is attached. When I try spamassassin -t -D rulesrun=255 < mail.txt it is nor considered as spam. So I do spamassassin -r -D rulesrun=255 < mail.txt ans it sayes me that Razor already learnt this message but it is not in the report when I test it. I finally decide to do sa-learn --spam mail.txt and it says me that it doesn't learn anything from the mail. I just use razor2 with spamassassin and would like the mail to be tagged as spam from my bayes and from razor2. Thanks. 1) Training one message alone is not usualy enough to "flip" the bayes score of the message if there is strong evidence in dozens of other messages that the tokens are nonspam.. Look at the bayes token output of spamassassin -D. 2) if the spam in question has a habeas warant mark, add the habeas swe headers to your bayes_ignore. 3) No one person (short of a cloudmark employee) can declare a message to be spam in razor. Your report counts, but one person alone is generally not enough. Also, make sure you've set up for razor reporting using razor-admin --register. --- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn ___ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
[SAtalk] sa-learn & spamassassin question
Hi, I've got a little problem with the mail that is attached. When I try spamassassin -t -D rulesrun=255 < mail.txt it is nor considered as spam. So I do spamassassin -r -D rulesrun=255 < mail.txt ans it sayes me that Razor already learnt this message but it is not in the report when I test it. I finally decide to do sa-learn --spam mail.txt and it says me that it doesn't learn anything from the mail. I just use razor2 with spamassassin and would like the mail to be tagged as spam from my bayes and from razor2. Thanks. -- Jean-Christophe VALIERE EPITA Paris Sud, 14-16 Rus Voltaire, 94270 Le Kremlin Bicetre Tel: +33 1 44786872 E-Mail : [EMAIL PROTECTED] Return-Path: <[EMAIL PROTECTED]> Delivered-To: [EMAIL PROTECTED] Received: (qmail 2334 invoked from network); 27 Jan 2004 06:44:18 - Received: from unknown (HELO mx-01.25tech.net) (212.37.195.56) by gladia.oxalide.net with SMTP; 27 Jan 2004 06:44:18 - Received: from localhost (localhost [127.0.0.1]) by mx-01.25tech.net (25TechMTA) with ESMTP id 81D453B109 for <[EMAIL PROTECTED]>; Tue, 27 Jan 2004 07:42:34 +0100 (CET) Received: from deliver.epitech.net (deliver.epitech.net [163.5.0.25]) by mx-01.25tech.net (25TechMTA) with SMTP id 310DA3B108 for <[EMAIL PROTECTED]>; Tue, 27 Jan 2004 07:42:27 +0100 (CET) Received: from epita.fr ([10.42.1.60]) by deliver.epitech.net (SAVSMTP 3.1.2.35) with SMTP id M2004012707422613148 ; Tue, 27 Jan 2004 07:42:26 +0100 Received: from colt.epita.fr (colt.epita.fr [62.23.180.27]) by epita.fr id i0R6i7E27523 Tue, 27 Jan 2004 07:44:07 +0100 (CET) Received: from ti100710a080-0025.bb.online.no (ti100710a080-0025.bb.online.no [80.213.224.25]) by colt.epita.fr id i0R6hPS08073 Tue, 27 Jan 2004 07:43:30 +0100 (CET) Received: from [80.226.27.218] by 80.213.224.25 with HTTP; Tue, 27 Jan 2004 01:35:15 -0400 From: "Alana Grady" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] Subject: apocrypha boa Mime-Version: 1.0 X-Mailer: byrd derisive Date: Tue, 27 Jan 2004 00:39:15 -0500 Reply-To: "Alana Grady" <[EMAIL PROTECTED]> Content-Type: multipart/alternative; boundary="36294582722586859646" Message-Id: <[EMAIL PROTECTED]> --36294582722586859646 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit gentility drool glacial asparagine fleabane bandwidth codify degeneracy audrey calder christmas ashmen counselor malt approve barrow diatomaceous --36294582722586859646 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 8bit Message Hi, Genierc Vigara and Sepur Viarga (Ciails) available onilne! Most trusted online source! Cilais and Vagira takes afecft right away & lasts 24-36 huors! http://www.ktzn.whpenn.com=www.sycgumk.pibian.seeitnowhere.com/cv/?AFF_ID=a3&jfqaoh=nwmm";>FOR SUEPR VAIRGA TOCUH HERE http://www.fcxxelz.czefz.com=www.wjbf.igrsqt.seeitnowhere.com/homepage/?bvevbarl=kzekx";>Not itnreseted plush occultate please carbone poisson cheesy alistair resolve bourbon elisha comprehensible agouti dragnet grosset handle buzzard briggs dram recitative malady aristotelian confectionery hough orb murre hellenic portentous finny dutiable frowzy fiddle gavel animate bucket burglary frugal aubrey hun housework highlight constipate dairy huck perch dress courtier mace attendant quadric begotten casteth erupt aching sahara instep antithetic resistor frame jason animosity harriman abstract bijection alaska orleans especial midway approximate cockle doorknob gloss include iraq broadcast pullover dramaturgy rampage extract ewing diebold hangar boylston fraser frilly mine hydrate acidulous insurgent cohosh accost eluate foote accretion huffman dental blest dalhousie shay fullback check inimitable equilibrium ames sick flute halo apparition chautauqua america dooley chancellor garrett detract horsemen diaphanous catchup martian austria compton raindrop divert bloodline chadwick facultative fruehauf lotus more frigid anything endomorphism co fiftieth articulate ethology inner repository bran optimistic behest cretaceous miranda cascara dc conception convulsion bladdernut catherine imprudent conferrable orderly hogan chimique fred necromantic dissuade maurice halfback rattail gallus say geography preparatory ligature junco ethiopia resident inflammatory shrink climactic schoolteacher menopause exculpatory medial culbertson ghana councilwomen aperture giles eugenic elude injunct farfetched chairmen extreme frankfurter arbiter advisor inbreed olsen scandalous largesse bereft byers flame headstone introject nil poseidon decadent deltoid pinnacle mad hoover indigene circus equate historiography prophetic canvas anonymity broccoli deceitful juniper conspire basin heroism barnar