> I've seen other people say this. My question is "why NOT
> send the listserv traffic through sa-learn as well?". After
> all, it is legit email you get, right?
I would NOT recommend feeding SAtalk automatically through sa-learn,
because people frequently include portions of spam in discussio
> Can anyone please tell me generally how frenquently they're
> DCC is hitting with SA?
A quick grep from last month's logfile reveals DCC hits
on 3248 emails out of 4365 quarantined => a good 74%
--
[EMAIL PROTECTED] GSM +358-40-767 8282
Oy Arrak Software Ab http://www.arrak.fi
---
> I'm still not sure if I should run through "ham" as well as "spam" or
> not. Is there a specific answer for this?
Yes you should run ham as well, because the Bayes algorithm needs to
compare againt both "spammy" and "hammy" words to determine how to
classify new mail. You cannot train it with
then
able to share the database better.
--
[EMAIL PROTECTED] GSM +358-40-767 8282
Oy Arrak Software Ab http://www.arrak.fi
> -Original Message-
> From: Theo Van Dinter [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, October 07, 2003 7:39 PM
> To: Kai Risku
> Cc: [
Enabling the Bayes journal should help ("bayes_learn_to_journal 1"),
as that allows concurrent processes to open the bayes database in
read-only mode most of time, with only an occasional journal syncing
needing a full write lock on the database.
--
[EMAIL PROTECTED] GSM +358-40-767 8282
Learning from 1000 messages takes a while. Perhaps you didn't wait long enough?
Try running
sa-learn --spam -D --file /tmp/spam.mbx
so you see what is happening. *Then* you know if it really hangs..
- Original Message -
From: "Robert Wornan" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED
I am currently running SpamAssassin 2.60 and I did migrate the bayes
database from 2.55 according to the instructions. However, I still have
problems with SA getting timeouts while using the Bayes database, and
someone else had earlier concluded that the Bayes checks get really slow
when there are
65.489 using SA 2.50 and default scores.
tests=AS_SEEN_ON, BANG_EXERCISE, BANG_GUARANTEE, BANG_OPRAH, BAYES_99,
BIZ_TLD, CLICK_BELOW_CAPS, DATE_IN_PAST_06_12, DATE_SPAMWARE_Y2K,
DCC_CHECK, DNS_FROM_RFCI_DSN, FORGED_IMS_HTML, FORGED_IMS_TAGS,
FORGED_MUA_IMS, FORGED_RCVD_NET_HELO, GUARANTEE, HAIR_LO