M Bernasconi wrote: > The program calls "spamassassin". And in "spamassassin" has no user/domain > capability (correct me if i'm wrong).
You are correct. If you want to just scan one-user's mail, this is better handled by only calling SA for the mail you want to scan. This way you actually save the CPU overhead of loading perl, starting spamassassin, parsing the configuration files, parsing the message headers, etc. Also, unless your MTA inserts hints into the message headers, spamassassin cannot reliably determine who the message recipient is. Recipients are fundamentally handled by the message envelope, not the message headers. SA can guess based on To: and Cc: headers. It can also use the "for" clause of Received: headers, but unless those exist SA won't really know who a BCC'ed message is being delivered to. > And I thought it's just a piece of cake ...!
