Auto learn as Ham when ALL_TRUSTED?

[Nate Schindler <na...@riconcorp.com>]

Looks like others are posting this question as well, but I couldn't see it in the archives...

I recently upgraded to v3.  Everything looks great so far, except one change that I can't seem to figure out.

We have an internal Exchange server, and SpamAssassin running on a different machine, but on the same private network.

I use clear_trusted_networks, and clear_internal_networks in my config, because I want messages from the inside learned the same as messages from the outside.  I force outgoing mail to pass with a 'whitelist_from <my domain>' line so that the messages themselves are scored the same, but the threshold is just higher.

So... after upgrading to v3, I see a new test in the log - ALL_TRUSTED.  It seems that any message that matches this test (e.g. all outgoing mail) is auto-learned as ham, no matter what the score is.

I'd like to make it stop doing this, if possible, and make it learn outgoing mail based on the same criteria as incoming mail like it did in 2.64.


Otherwise, congratulations SA people on a job well done.  SA is the best thing since chocolate pop-tarts!

N8