Hi Mouss, Thanks for you answer.
Bellow some observations/questions. On Fri, Mar 19, 2010 at 6:08 PM, mouss <mo...@ml.netoyen.net> wrote: [...] >> This setup doesn't work well for me because it doesn't update the >> maildirsize (I'm using the VDA patches in Postfix) file in the user's >> Maildir and it doesn't automatically create the Maildir when needed. >> Without using procmail as a transport it does it very well. >> >> I am wondering if I can use procmail (or some other filtering >> software) in another way so it does it's filtering and then send the >> message back to Postfix, like my DSPAM setup, that uses the LMTP >> socket, filters the message and send it to the LMTP in the >> localhost:10026 port. > > you mean SMTP, not LMTP... Well, SMTP or LMTP, it's a instance to only receive local mails. But, I don't think it makes any difference here, right? ;) > > with procmail or maildrop, your best bet is to resubmit mail via the > sendmail command (postfix sendmail, not Sendmail sendmail). for this, > you must make sure that you don't have a content_filter for pickup: > > pickup fifo n - n 60 1 pickup > -o content_filter= > ... > > (the reason is to avoid an infinite loop: you filter mail, you pass it > to sendmail, it passes it to the filter...). Sure. That's why I had to use another SMTP to pick the DSPAM result. But I don't know yet how to do it with procmail (or similar), and that's what I'm asking. > > All that said, the VDA patch isn't supported here. so you're on your own. I know that, but I think it's unrelated to the real question. Let's just suppose I want to do it without those patches. > > An latenrative based on a policy service has been proposed on the list > (I'm really sorry, but I forgot who posted this. If the developer sees > this message, he'll reply. otherwise, google...). In any case, this is a > better approach than a patch. Alternative to that patch? I did some searches and as far as I could see, none of the alternatives (that I've found) could reject the message in SMTP. I think it's nice to not generate bouncing messages/backscatter. But if someone have an alternative besides that, it's always welcome. Once more, thanks for your suggestions. ;) Mauro