On Monday, November 11 2013, I wrote: > Hi there,
Hi, again! I am sorry to ressurect this thread, but after some time, investigation and fixes, I would like to share what I did and ask for more opinions. First, I have fixed the previous warnings that I was seeing on the messages. URI_BLOCKED was easily fixed by setting up my own named (which, I confess, I should have done right after installing my server, but I was unfortunately postponing it...). UNPARSABLE_RELAY was happening because I modify the headers of every message sent through my server in order to anonymize the sender's IP address; however, SA has a strict rule for checking the "Received:" header, and I needed to adapt my modifications to that rule. Anyway, now everything's OK. Having said that, my SA is still missing lots of spams. For example, take a look at: <http://sergiodj.net/~sergio/sa/spam.txt> This is a spam message I have just received. SA did not recognize that as spam, and put the message on the INBOX folder. This has been happening for all spam messages I receive. I understand that a message like the one mentioned above doesn't have many terms for SA to work, so I assume it's OK for it to classify it as ham even when it isn't. But take a look at this other message, for example: <http://sergiodj.net/~sergio/sa/spam2.txt> It's a classical spam, I think. The score is even higher than the first spam. But it's still not catching it. I've already looked at the tests performed by SA, but couldn't find anything suspicious. So I'd like to ask for opinions here... Does anybody see anything wrong/suspicious in those messages? BTW, it's worth mentioning that my Bayes databse still has too few spam entries (17, as of now), so it's not being used for the tests, of course. Thanks a lot! -- Sergio
