Juan Manuel <[EMAIL PROTECTED]> wrote: > Elias Oltmanns wrote: >> Juan Manuel <[EMAIL PROTECTED]> wrote: [...] >>> I want to achieve that, even if I retrain all the False Positive, the >>> mail will only arrive once to each mailbox. >>> >> >> Are you talking about retraining from within the webue? Since I don't >> use the webui, I have no idea how the mechanism works. How does it >> retrain messages? Does it call dspam directly or does it forward the >> message to the retrain-ham alias? And how are retrained false positives >> moved to the INBOX of the user? I suppose this latter task has to be >> done using some IMAP methods or some such. Just feeding the message >> back to exim without modifying the headers is most definitely not the >> right thing to do. >> >> Regards, >> >> Elias >> >> > Yes, I'm using Web UI.. As far as I see, the UI calls directly dspam > to retrain (I haven't seen forward when I do retrain).
Is the message retrained successfully according to dspam_stats? > As for the FPs, it mails (like mail() funcion in PHP) the user(s) > address (as if the mail would've come again from the outside, only > locally generated). Sorry, I still don't quite get it, I'm afraid. If there are multiple addresses in the To: header of a message to Bob, which has been wrongly classified as spam, and if Bob retrains that message afterwards, whom does the Web UI send this message to? Obviously, it should send it to a preconfigured address (i.e. Bob's) regardless what addresses are given in the To:, Cc: or any other header. And if so, you should configure exim to recognise messages sent by the Web UI in order to bypass the dspam routers. Regards, Elias
