Re: [WL] [SAtalk] Yikes.. rules_du_jour

2004-01-19 Thread Jay Levitt
One problem: If a spamassassin --lint fails (because if you, oh, had outdated directives in your sa-mimedefang.cf file), then once you correct that, on the next run, rules_du_jour won't update anything, because it thinks everything is up to date. - Original Message - From: "Chris Thielen"

Re: [SAtalk] THIRD request: Someone please help with AWL oddness

2003-11-04 Thread Jay Levitt
> I've read all three of your posts.. and don't have any suggestions to help > you.. it looks like a bug to me. OK, I've opened bug #2734 on it... > What really makes me wonder is why the ip=none entry exists at all. > > Have you been doing something like spamassassin --add-addr-to-whitelist? >

[SAtalk] THIRD request: Someone please help with AWL oddness

2003-11-04 Thread Jay Levitt
- Original Message - From: "Jay Levitt" <[EMAIL PROTECTED]> To: "Jay Levitt" <[EMAIL PROTECTED]> Sent: Sunday, November 02, 2003 9:29 AM Subject: Second request: AWL false hits > Hi.. does anyone have any ideas on this? I have read the FAQ, and this d

Re: [SAtalk] Understanding AWL processing

2003-10-31 Thread Jay Levitt
> Each day I see a new message by this guy and add a new rule to find his > tracking codes. Your original un-munged e-mail address was in that base64 > text of that header. Since only a spam or two a day gets through from him, it's probably not worth it - one of the disadvantages of being a one-p

[SAtalk] Understanding AWL processing

2003-10-30 Thread Jay Levitt
My server handles exactly one mailbox: my own. I am running SA 2.60 through MimeDefang, with some MimeDefang changes to enable AWL processing. I've noticed an increasing amount of spam getting through that uses the age-old trick of forging my address in the From: line. This somehow triggers the

[SAtalk] 2.60: Bayes false positives - better training?

2003-08-16 Thread Jay Levitt
ese are spam.)   I'm using a site-wide bayes file, since I'm the only user.  Every time I upgrade SA, I wipe out the bayes database, and re-train SA on my inbox, some spam-free archives of my mailing lists, on the "probably spam" folder (5   How can I do a better job of training bayes?   Jay Levitt

Re: [SAtalk] [RD] AOL 7 but not from AOL.

2003-08-11 Thread Jay Levitt
are, not the client itself. We always used X-Mailer, and I don't believe that's changed. Likewise, since the headers are put there by the server, if it didn't go through the server, the header must be forged. Jay Levitt --- This S

[SAtalk] LOTS_OF_STUFF bug?

2003-07-16 Thread Jay Levitt
The LOTS_OF_STUFF rule will match URLs of the following type:   http://1.2.123.3/images/something.gif   I'm not sure what, if anything, would be a better solution....   Jay Levitt

Re: [SAtalk] meaning of rule descriptions

2003-07-01 Thread Jay Levitt
click on them, but take a look at the initials of each college and see if you detect a (somewhat juvenile) pattern... Jay Levitt --- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and F

[SAtalk] 2.60 RC1 make test errors (not Invalid Date)

2003-06-29 Thread Jay Levitt
.  Any ideas, anyone?   Jay Levitt

Re: [SAtalk] MIMEdefang and autowhitelist

2003-03-19 Thread Jay Levitt
Matt Kettler wrote: > Sorry, but the AWL is only conceptually valid when applied to a single > user's mail at a time. Applying it globally to multiple users at a time has > a tendency to cause problems, particularly if you ALL_SPAM_TO some users in > your network. In my case, sitewide IS a single

Re: [SAtalk] what is bayes_journal for ?

2003-02-25 Thread Jay Levitt
And conversely, why don't I have one? (Mandrake 9.0, built 2.50 myself with default configure) Not even an empty file. I gather that one should be created by the --rebuild option. However, on my Athlon 1.7GHz, the "fast" sa-learn --no-rebuild option takes half an hour to scan my 7200-message in