On Mon, Mar 01, 2004 at 10:00:52PM -0800, Jeff Makey wrote: > Does sa-learn not update the atime on the Bayes tokens that it > modifies? My full procedure includes using sa-learn on all > non-whitelisted mail, so I would expect that the LRU aspect would be > preserved.
Yes, sa-learn will update atimes. If you learn from every piece of mail that gets scanned, yes, you keep the LRU behavior during expiry. The less mails you train on, the more you head towards FIFO. (as a reminder for folks: we're talking about if autolearning and token atime updates are disabled -- in case anyone digs this mail up in an archive and quotes me out of context. ;)) -- Randomly Generated Tagline: le died...
pgpg0XP9EYI3d.pgp
Description: PGP signature
