Theo Van Dinter wrote:

Are there more than the listed count of tokens if you did a "sa-learn --dump"?

Not only is my newly-created database out of sync with regards to ntokens, nspam, and nham, my old one is even more out of sync when upgraded to 2.60rc2.

If I run sa-learn --dump data on my old database, the huge one, I see even more interesting statistics:

[EMAIL PROTECTED]:~/.spamassassin$ ls -l old
total 10128
-rw-------    1 ben      ben       1331200 Aug 27 10:52 bayes_seen
-rw-------    1 ben      ben      11468800 Aug 27 10:52 bayes_toks

[EMAIL PROTECTED]:~/.spamassassin$ sa-learn --dbpath old/bayes --dump magic | grep ntokens
0.000 0 3217 0 non-token data: ntokens


[EMAIL PROTECTED]:~/.spamassassin$ sa-learn --dbpath old/bayes --dump data | wc -l
347678

This one is off by more than a power of 10!

Ben




------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to