> > > i've seen an interesting feature in other white-list based spam filter > > systems: if a special token is found in the mail, the sender is assumed to > > be "trusted" and gets whitelisted. > > This sounds like a specific case of a feature I requested several months ago. > On the outgoing filter, there is a "bare=append" action which says that the > recipient of a message meeting certain criteria recieves the e-mail unaltered > and is added to the whitelist. This is most generally used to auto-whitelist > anyone you address e-mail to.
...which I don't use. I currently do not see any benefit in tagging outgoing mail. I have to read some more docs to understand what tagging this mail brings back to me. > My request was for an incoming filter rule, say "ok=append", which does the same > thing for an incoming message. Meeting certain criteria, the sender is > auto-whitelisted and the message passes. It becomes just another action in > your incoming filter, albeit the only one which affects the list files. > > It's not implemented yet, but it's on the TODO list for the 1.1 versions. Is > this something like what you're after? Yes, exactly. That would bet *it*. I got an advice that I could pass mails by adding something like body 'some match code' ok to my input filter, but this leads to the problem that all mail is passed, even if 'some match code' is not included in the mail. I think I made something wrong, but did not have the time to analyse it. I've been thinking to use an external program to check the incoming mail for the token and let it add it the sender address to my whitelist. An example how to call an external program is given in the docs with the "razor-check" call. This method would be a replacement for "ok=append", which is basically my need. Peter _____________________________________________ tmda-users mailing list ([EMAIL PROTECTED]) http://tmda.net/lists/listinfo/tmda-users
