David Gardner wrote: > Hey gang, > > The company I work for generates email messages in automated processes and > send them only to customers that request them. This week, one of these > customers mentioned they were not getting the messages we were sending out > and in fact have not been getting them since December. (Thanks for the > heads-up.) > > This customer is using Symantec Mail Security for SMTP which logs the message > with "SMTP ID:M200702141421121234 Info: Address or content not allowed. > Info2: The message is malformed." Looking on the Symantec web site tells us > this is a generic RFC conpliance issue <period>. > > What I'd like to do is use SpamAssassin to tell us what part of our message > (if any) is malformed. I don't want to check its content for SPAM, just the > message structure compared to the RFC standard. We want to do the right thing > here but not leave any room for interpretation of the RFC by spam filters. > > Can SpamAssassin do this? > Can it? Yes, but not without substantial work, enough work that it would be easier to analyze your email by hand unless you have a lot of different emails to check.
As it stands right now, SA has some tests which are based on RFC non-compliance. However, it only uses those tests because they happen to be good spam signs. After all, SA's job is to find spam, not RFC violations. In some cases, it winds up finding both, but that's somewhat coincidental. So, there is no comprehensive checking for RFC compliance in the default SA ruleset. You could write your own set of rules that perform comprehensive RFC checks. At that point you'd need to be pretty well versed in the RFCs in the first place, and would probably be able to spot the problem for yourself. Of course, you're free to run your mail through SA and see what the rules do pick up, but the amount of RFC-based tests in a standard SA setup is pretty limited. It might find your problem, but probably not.
