New SpamAssassin rules

[vividere]

I guess I still am not clear enough....

The spam Assassin software was done almost a year ago, last June to be exact. It worked fun for nine months and started messing up in March for no apparent reason. I had everyone involved that worked on it involved in a solution and they all denied having anything to do with my current problem and that made sense since how could something break now that worked fine for 9 months.

When I came here and found that changes were made to the code at the same time the problems started it seemed like I had suddenly found the reason.

When something breaks you look at what changed. Since nothing was changed on the server but the software was changed then that seemed to be the reason.

Frankly I was not told anyting about training the bayesian filter. I know people who do that when they run the filtering software on their local PC but I don't know how to do that with the server-based software.

For now the software is working like before. It seems to have gotten over the mode where it marked all new email as spam. It is back to catching some, misidentifying others, and missing a lot of the spam.

Lew

[groupw]

I've reported, and have had held, several thousand spams since you removed SpamCop from the SpamAssassin criteria. I've been running SpamAssassin at a threshold of 4. Out of all those spams, not a single one was falsely flagged positive by SpamAssassin. So for me, the fix worked great -- thanks!

I've had maybe a dozen or so legitimate emails held due to listings on block lists other than SpamCop (whose blocklist I don't use as a filter). Something like a couple of hundred spams have gotten through the filter, but Mozilla catches those with very near 100% reliability. So the system working well for me. Thanks again.

Roy Lewallen

[Enrico_C]

Maybe it would work fine if SpamAssassin added just 1,5 / 2 to its scoring when an IP is listed in the SpamCop BL.

[PeterJ]

JT --

You probably already know, but Chris Santerre has created a new custom rules website located here:

http://www.rulesemporium.com/

Also, I have read recently that the anti-drug rule has had several improvements added to it, following snipped from Matt Kettler regarding version 0.60:

Changes.:

    Accented characters are now in many rules for popular drugs, and a wider variety of accented characters are now used.

There are more changes but I think this one alone warrants updating to the 0.60 version of antidrug.

Can you comment on whether or not rules dujour is implemented for SpamCop's SA implementation to automatically update our custon rules? (Here: http://sandgnat.com/rdj/rules_du_jour)

[jefft]

JT --

You probably already know, but Chris Santerre has created a new custom rules website located here:

http://www.rulesemporium.com/

Also, I have read recently that the anti-drug rule has had several improvements added to it, following snipped from Matt Kettler regarding version 0.60:

Changes.:

    Accented characters are now in many rules for popular drugs, and a wider variety of accented characters are now used.

There are more changes but I think this one alone warrants updating to the 0.60 version of antidrug.

Can you comment on whether or not rules dujour is implemented for SpamCop's SA implementation to automatically update our custon rules? (Here: http://sandgnat.com/rdj/rules_du_jour)

No, we're not using the scripting to update the SA rules. I'm a little uncomfortable with stuff like that happening automatically without any intervention here.

I will, though, look at incorporating the updated drugs rule.

JT