> Ok, subject doesn't say much, but here is the deal. IMGate always checks the > *last* sender IP address and looks up in the spam databases if it listed or > not. > Now, Check this header out (below), As you can see, 207.202.32.34 seems to > be the spammer right?,not at all. IMGate is checking the spam databases if > 207.202.32.34 is listed in any of them, negative, it's clean. However,the > real spammer relies behind 63.143.236.242. And it is listed in every spam > database I know of. > Then the question is, why does IMGate check only the last sender IP address > in the header? not the first one, where the email is originating from,? > where the real spammer is hiding behind. > That would definitely increase the accuracy. > Any idea if we could do that ?
Not with just Postfix. Well, you could with Postfix, bt header checks get to be a bit unwieldly. I am playing with a SpamAssassin install on a test machine to see what it is like getting up. So far I have had some interesting oddities in getting it working, but nothing horrible. For example, AMaViS, which I am using the daemonized version of to run SA, has some issues with older versions of Perl, like the 5.0.0.5 that came with FreeBSD. But SA has issues with 5.8.0, the newest version. So I have had to split the difference and install Perl 5.6.1. --Eric
