What about spam traps? It seems to me that parsing of the spam traps for the BL still uses the old algorithm. Any plans to convert the spamtraps parsing to the mailhost-aware system as well?

Actually, the spam traps were some of the first "users" to convert to the new system. What gave you the impressiong they did not? All the spamtraps are using it already, with great success. I wanted to do that before I let users loose on the system. Spamtraps generally have a much simpler configuration.

-=Julian=-

Well, the reason I asked is that a llot of spam I receive now (after I switched to mailhost system) point to what appears to be adelphia.net (outgoing?) mail servers - 68.168.78.184, 68.168.78.190, 68.168.78.196, 68.168.78.199, 68.168.78.202, etc. Looking up the BL data for thise mailhosts, I have never seen any spamtrap activity, but if I look up the previous hop, then often it does have spamtrap data.

For example,

• in http://www.spamcop.net/sc?id=z367611762z54...5eb73387691deez 68.168.78.202 does not have spamtrap data, but 68.235.34.92 does.
• in http://www.spamcop.net/sc?id=z367611756zb5...dd75c2e2fb1643z the last relay is 68.168.78.190 w/o spamtrap data and the previous one is 68.71.146.7 which does have it...
• same in http://www.spamcop.net/sc?id=z367611755zd1...b7394d2bf56ce3z with the last hop is again 68.168.78.202 (no spamtraps) and 68.66.117.48 (some stamtraps)...

In fact I quickly browsed my recent report and for all my recent adelphia.net spam the picture is like that - the last hop does not have spamtrap data and the previous one does...

Hmm. Very astute.

I spot checked one of the tap hits against one of those internal hosts, and it didn't go through the main MX at all. So it looks like it's working the way it should.

But the as-yet unspoken corellary with all of this is that ISP's main mxes, if used for spamming, will be more likely to wind up on the BL. I think that's probably a good thing, both for the BL users and for 3rd parties who want to get less spam (more preassure on ISPs to lock down).

-=Julian=-