Matt Kettler wrote: > Given that I'm looking for maximum message scanning speed and lowest RAM > overhead on a single server, is there any reason for me to prefer using SQL > over > SDBM?
Replying to myself I found one issue. sa-learn --restore doesn't work properly for SD BM: http://issues.apache.org/SpamAssassin/show_bug.cgi?id=4670 However, I found a work-around for that (I did this on a test-user's bayes DB, and it worked). I also found a comment by Michael Parker that SDBM isn't well tested in production: http://issues.apache.org/SpamAssassin/show_bug.cgi?id=4530 Has that changed at all? So aside from those two, any other drawbacks to SDBM?
