John Hardin wrote: >> >> No, the message is queued first, then passed to spamd. > > In that case I can't understand why there would not be a local > Received: header. >
Yeah ... > MISSING_DATE,MISSING_HEADERS,MISSING_MID,MISSING_SUBJECT,NO_HEADERS_MESSAGE,NO_RECEIVED,NO_RELAYS > > Wow. > > I would suggest you look closely at your proxy. Are you _positive_ > it's in the mail stream at the point you think it is? Are you > _positive_ that it will always see a complete message (i.e. headers > and body)? Well, I'm beginning to have some doubts now, but it has been working fine for the last two-three years. I'll have to double check my code - maybe there's a weird corner case that's becoming more and more exposed (for whatever reason). >> That's why I'd like to have something like "score NO_RELAYS die" to >> make spamd quit processing it. > > That's probably not going to happen. The scan/no-scan decision has to > be made upstream. It was just a thought - although it doesn't really sound like a scan/noscan decision. I mean, the message has already been scanned, but it has a condition that - according to my rules - makes it an invalid scan/message. I'm looking through the code to see if I can the right place. /Per Jessen, Zürich
