On 2016-08-18 2:10 PM, Jerry Malcolm wrote:
Thanks for the quick response. I'll try to reply with what I know. But
I purchased a package "SpamAssassin In A Box" from JAM Software. I ran
the installer, and that's it. I'm sorry that I don't know more. But I
don't know much about the inner workings. I was just hoping it would work.
I am running Windows Server 2008 r2. I purchased the latest version of
SpamAssassin In A Box yesterday, Full x64. It had a build date of April
2016. I can't find the SpamAssassin version number. But I'm assuming
it had close to the latest build as of April.
I am a hosting server environment with several domains and several ids
per domain. I am running the Apache JAMES 3.0 mail server, using the
JAMES-provided mailet to invoke SA. I have been a JAMES user for over
10 years.
I understand the training of Bayes. But I'm having a problem with
that. I have been asking on the JAMES forum about how to run sa-learn
if I'm using a mySQL JAMES repository for mail. No responses as yet.
So I don't know how I can train Bayes at this time. The SpamAssassin In
a Box install asked me if I wanted them to install a pre-defined Bayes
ruleset. I did, but clue how old it is.
So first question that I can't seem to get a response to from JAMES....
is there any code that I can use that allows sa-learn to work with JAMES
db? Or do I have to write code that will take every spam I want to
train with and write it to an MBOX file and run sa-learn that way? I'm
really surprised if I'm the first JAMES user that needs to run sa-learn.
The second question... in the interim is there a 'better' set of Bayes
rules that I can just download and install until I can get the first
issue resolved and can start training myself?
Here is a pastebin.com link to an example uncaught spam message. SA
scored it a 4.7. http://pastebin.com/T1CfVgP4
Hi,
We are going to need the full headers, not just the email body.
Regards,
Rick