On Wednesday, December 8, 2004, 9:49:55 AM, Daryl O'Shea wrote: > Additionally, assuming there isn't an extreme query frequency drop off > after the top 100 or 200 excluded domains, it would be nice to have > access to the rest of the exclusion list which wouldn't be realistic to > be storing (and currently copying around) in memory.
> There's got to be a reason why SpamAssassin currently only includes the > top 100 or whatever excluded domains... either the rest of the data > wasn't useful or it wasn't worth the performance hit having them in memory. I believe the 125 cutoff was entirely arbitrary, but it happens to correspond almost exactly with the 50th percentile of DNS queries against whitelisted domains, which is a happy coincidence and a perfectly reasonable cut off point. > New additions to the exclusion list would immediately be available too, > not that that is really a huge concern. Remember that the only reason to build this hard-coded exclusion list into SA was to prevent unnecessary DNS queries from happening in the first place: http://bugzilla.spamassassin.org/show_bug.cgi?id=3805 The much larger global whitelist is applied internally in SURBLs to prevent those domains from ever getting listed. It is an exclusion list there. Jeff C. -- Jeff Chan mailto:[EMAIL PROTECTED] http://www.surbl.org/
