http://bugzilla.spamassassin.org/show_bug.cgi?id=3225
------- Additional Comments From [EMAIL PROTECTED] 2004-04-23 18:54 ------- Subject: Re: RFE: Bayes optimizations -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 OK, I took a look to remind me ;) http://svn.apache.org/viewcvs.cgi/incubator/spamassassin/trunk/lib/Mail/SpamAssassin/BayesStore.pm?root=Apache-SVN Note that rev 5112, May 9 2003, was when atime was changed to refer to the message received time: 'bug 1821: we decided on a new bayes format again. this time it's token atime is based on message received time. this brought up a whole bunch of issues, but mostly required new pack/unpack, upgrade_db, expire, and a bunch of other changes. while I was milling about I fixed a few other random bugs. I don't quite remember what they are right now, so just assume everything's peachy.' Rev 5038 -- 'bugs 1523, 1666, 1775, and 1776: set a new bayes db format for 2.6x, version 1 (2.5x had version 0). atimes are now unsigned ints (32-bit) instead of a short (16-bit) so we won't get overflow/expire issues. code is now included to automatically upgrade to v1 when the db is opened for writing. v0 is still supported for readonly mode. bayes tokens are now packed in little-endian (VAX) format. this means that the same DB will work on different endian platforms.' rev 4722 -- 'fix for bug 1625. we only touch the token when syncing the journal, and the journal only has updates from scan() whereas tokens are automatically updated from learn(), so we will likely end up lowering the atime of the token when we sync. doh. :(' So I think this means we *can* use a 2-byte atime format safely, since the problems we had with that before were *definitely* caused by using message-scan and message-learn time as the atime, instead of message-received time. - --j. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Exmh CVS iD8DBQFAicjIQTcbUG5Y7woRAh/OAKDYl0gQQ0obM5jSipDoXVqGLLJLlwCdFtpU b3I9HZfe8PQKFDTYhhZaY94= =MT1f -----END PGP SIGNATURE----- ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.
