|
From: Michel B. <mi...@bo...> - 2005-06-27 20:07:14
|
Le Lundi 27 Juin 2005 21:43, /dev/rob0 a =E9crit : > > On Monday 27 June 2005 14:33, Michel Bouissou wrote: > > Le Lundi 27 Juin 2005 21:27, /dev/rob0 a =E9crit : > > > Airmail.net is one of those stupid ISP's which uses a callback like > > > Postfix's verify(8) on all sender addresses ... > > > > I'm another one of these "stupid", event though I'm no ISP ;-) > > On a large scale I think it's quite reckless. Possibly, but I don't have a scale large enough here for it to pose=20 problems ;-) > Even on a small scale it ought to be limited in some way: perhaps do SP= F > checks and only verify SPF failures. There are too many spammers out there, and too few SPF-enabled domains fo= r SPF=20 checks to be enough... > Note the "WARNING" text in the ADDRESS_VERIFICATION_README. > > Does it work well for you? Perfectly, altough I had to build a small table of exceptions. Sometimes = it=20 collides with greylisting one way or the other, but it usually just delay= s=20 the first message a little longer (most often one hour or two). > Does it come after other restrictions?=20 Yes, it comes absolutely last, after all possible other 2821-level checks= have=20 been performed, just before considering accepting the message. And of cou= rse=20 I keep the results cached in a DB. --=20 Michel Bouissou <mi...@bo...> OpenPGP ID 0xDDE8AC6E |