Jack wrote: > Hello, > > In SpamAssassin's FAQ, there is a page > "AvoidingFpsForSenders" > (http://wiki.apache.org/spamassassin/AvoidingFpsForSenders) > under the heading "I'm not a spammer!". That page > mentioned that "Emails with ... invalid or missing > message-ids ... are frequently signs of spam". As the > Message-ID header of my mail server looks invalid, I > have some questions concerning the definition of > invalid message-ids: > > Question 1. Suppose the part on the right of "@" in > the Message-ID header is not a valid domain name. For > instance, <xxxx... @abc>. Will SpamAssassin consider > this as an invalid message id and so consider the > email as possible spam? > Generally speaking, the "invalid message id" rules are looking mostly at what comes BEFORE the @, not after it.
The rules are also looking for messages with specific mail systems declared in the headers (ie: outlook) that have message id's in a format that such a system would not generate. That said, it is strongly suggested that the right-half of a message-id be the fully-qualified domain name of the system that generated the message-id. This helps ensure the message-id is globally unique. It would be exceptionally bad form to intentionally use the same right-half as another mail system unless you've co-ordinated some scheme between the two to ensure they'll never use the same left-half. > Question 2. Suppose the domain name in the Message-ID > header does not match the domain name in the From > header. For instance, the From header is "[EMAIL PROTECTED]" > and the Message-ID header is "<xxxx... @xyz.com>" . > Will SpamAssassin consider this as an invalid message > id and so consider the email as possible spam? > Personally, I would consider this a weak but possibly usable spam sign. That said, SA does not appear to do so.
