From: Michel B. <mi...@bo...> - 2005-01-11 08:07:00
|
Le Lundi 10 Janvier 2005 21:22, Lionel Bouton a =E9crit : > > sender address verification isn't something people should use blindly, > if you do so you block some mails the recipient are willing to receive. Well, I've actually been using sender address verification on several ser= vers=20 in different environments for months, and it works rather good. The sende= r=20 addresses that cannot be verified but yet are "legitimate" are very rare,= and=20 in any case it's a sender system misconfiguration issue, generally for so= me=20 automated mail systems such as newsletter or website generated mail. My own opinion is that any mail should come with an envelope-sender which= is=20 replyable (bounceable), and, if not, the mail shouldn't be accepted. For = this=20 reason sender address verification makes much sense for me. > I'm not willing to poke holes in the greylisting process just to cope > with what I still think is a (small) defect in the verify daemon. In > fact if you really want to, you can configure Postfix itself to not > greylist mails coming from the verification addresses. Yes, I have already configured a workaround in Postfix. But I don't belie= ve=20 that adding configuration flexibility in a tool is "poking holes" into it= . I=20 believe that a tool should be as configurable and flexible as possible,=20 according to users demands, and provided it is feasible and the developpe= rs=20 have time to develop the requested features ;-) I'd rather have an sqlgrey that manages sender whitelisting by itself rat= her=20 than having to have to configure a workaround in Postfix, because the=20 workaround in Postfix "pokes bigger holes" in the filtering than what it=20 would if done only at sqlgrey level. I believe that each filter / policy server should be as configurable as=20 possible, leaving the the sysadmin the choice of what should be processe= d or=20 not by any given filter. Cheers. --=20 Michel Bouissou <mi...@bo...> OpenPGP ID 0xDDE8AC6E |