RE: [AMaViS-user] How to use sa-learn to manually train?

[SamW <sw...@sv...>]

So my understanding on this is that I have to remove all the headers
that amavisd-new/sa added before I feed them back to sa-learn? Is that a
correct assumption? (I have sendmail/sa 2.63 / amavisd-new-p9 setup.
Clients access their mail using pop3 clients only)

Example:

If spam detected on a mail, amavisd-new adds X-Spam-* headers and
changes the Subject. (Good that is what I want).

Now I looking at the mail I find that the mail is not SPAM although it
is marked as spam. Now I need to tell the SA/BAYES, this mail is not
SPAM. So I will forward the mail to no...@my...  and crontab
should take care of sa-learn --ham  command. But will SA learn this
correctly? 

I have also have read that SA remember the MessageID numbers of leart
messages, So will this forwarded mail have a new MessageID number and
training is going to be useless? If sa-learn, looks at only the
message-body this scheme may be reasonable? Any help/comments
appreciated?


I also have auto-learn turned on, so for manual training should I run a 

	sa-learn --forget <mbox>
before I run 
	sa-learn --ham <mbox>

- SamW

> > So simply saying I cannot use, for example, forward option in
> > Outlook Express to forward spam messages into other e-mail
> > box where I can run sa-learn ?
> >
> > Bart
> >
> >
> 
> Outlook has a resend function that works one message at a time.  It
does
> allow you to delete the "***SPAM***" in the subject line.  I don't
like to
> use it because you have to view the email in order to resend it.
> 
> Outlook express does have "Forward as an Attachment" that I think
would
> work.
> 
> In Outlook I have been experimenting with highlighting a group of
spam,
> clicking "Forward" and the sending to sp...@lo....  That
way
> all
> are attachments.  It seems to work ok.
>