On 08/09/2007, David Gerard dgerard@gmail.com wrote:
On 07/09/07, Mark Ryan ultrablue@gmail.com wrote:
There is a way, buried deep within the admin interface, to auto-discard messages matching a specific Regexp. I know nothing about regexps, but frequently wish I did, because it would help us cut down on a lot of the V14GR4-style spam. Which almost never reaches the mailing list but outnumbers genuine messages from real people in the moderation queue by a factor of about 20 or 25 to one.
I put a couple of regexps on it already, to silently dispose of stuff that scores more than 7 points on the Wikimedia mailserver's spam detector. What you're seeing in the mod queue is the stuff that's *left*, about a third of what it was. Most mail, by message count or volume (certainly by volume) is spam.
- d.
Do you want me to try to improve on the regexps? If so, I need a large sample of the spam that is getting through, as well as a copy of the regexps you are currently using.