I have a general question about development versus using spamassassin.
Does the developer mailing list do anything other than bugs or is
that the wrong list for detailed discussions on how SA works under
the hood?
I'm interested right now in how to use the Mail::SpamAssassin suite
in some applications rather than the spamc/spamd or spamassassin
products. But I'm having some difficulty with the care and feeding
of some elements of the code. Yes, I want to write my own because
there is something that these other products just don't do. That's
what this Open Source stuff is all about, right? :)
As a specific example I'm having a lot of trouble identifying how you
might initialize a user database (Mail::SpamAssassin::Bayes:PgSQL)
for BayesStore use.
I can "force" it to happen by calling sa-learn --sync --username...
but I'm trying to do this from inside my own perl code.
It's things like this that are very difficult to cull out of the
perldocs, but with sufficient time, patience, and caffiene I'm making
real progress.
Similarly, I know that you can set up Mail::SpamAssassin to not try
and use/create the ~/.spamassassin per-user directories for
user_prefs and scores because that's how spampd works.
But there isn't anything that tells you how to do that as a programmer.
How can this be handled?
BTW -- what the heck is $spamtest->init(1)????
It's a completely undocumented method in Mail::SpamAssassin that's
related to the sa-learn --sync --username and sa-learn --spam --
username < sample-spam.txt calls that I've used previously to force
the initialization of a Bayesian database entry. But it's
undocumented!!! Are there others like this?
