On Thu, 2007-07-05 at 23:04 +0200, Dirk Bonengel wrote: > Maybe just a few words to close that discussion here:
Dirk, I don't think this really puts an end to this discussion, and I believe what Per actually was wondering about are some precise statements about each of the iXhash lists sources. At the very least, that is what I am wondering about. ;) Unfortunately, the example iXhash.cf of (current) version 1.0 is rather scarce when it comes to the definitions. I'd wish for these to become as informative again as they used to be. FWIW, these verbose descriptions and comments have been the reason for me to pick 2 out of 3 lists in the first place, long ago. (Which doesn't mean I am not re-evaluating this decision. In fact, the third one seems to be highly accurate, too -- granted, based on some brief, non-exhaustive tests...) Revision 23 of your plugins old home claims this: http://wiki.apache.org/spamassassin/iXhash?action=diff * ix.dnsbl.manitu.net: Uses iX Magazine's spam as datasource. * nospam.login-solutions.de: *Manually* verified stuff, fed by spamtraps run by LogIn & Solutions AG. * nospam.login-solutions.ag: Lots of stuff, but *automatically* categorized and contributed. So, the most important question is: Does this still hold true? Also, am I correct in understanding, that all you mentioned in your previous post about the various sources applies to the last list only? I do love your plugin and these lists, using them personally for a long time -- and I find them to not only match a huge part of my spam, but to also be highly accurate. However, Per is right in that trust is an important piece of the puzzle whether to use a list or not. I'd really love to see this info probably being mentioned on the (new) project home, and definitely back in place in the example cf file... guenther -- char *t="[EMAIL PROTECTED]"; main(){ char h,m=h=*t++,*x=t+2*h,c,i,l=*x,s=0; for (i=0;i<l;i++){ i%8? c<<=1: (c=*++x); c&128 && (s+=h); if (!(h>>=1)||!t[s+h]){ putchar(t[s]);h=m;s=0; }}}
