On Sun, 11 Jan 2004 10:55:58 +0100, Joerg Rossdeutscher <[EMAIL PROTECTED]> wrote in message <[EMAIL PROTECTED]>:
> Am So, den 11.01.2004 schrieb Wayne Topa um 00:56: > > Joerg Rossdeutscher([EMAIL PROTECTED]) is reported to have said: > > > Am Sa, den 10.01.2004 schrieb Wayne Topa um 19:52: > > > > > If you remove the Message-Id: you might run into some problems. > > > > I have found that a lot of spam 'does not' include that header. > > > > > > That will not happen, since every mailserver that gets a mail in > > > his fingers is adviced to generate an ID if missing. > > > > Thats right, and my point is: > > I find that when mail, without a Message-Id:, is received by my ISP, > > 'they' add that header. So if I find a Message-Id with my ISP in > > that header, its deleted. I have not had a false positive using > > that rule. It's been working now for 3 months and catches an average > > of 10 spam mails per day. > > Certainly you can filter what you want to - but you're taking people > the chance away to write a valid MsgID. > > The background is: > > In RFC is written you're MsgID must be globally unique. So this can > only be done on a mailserver, since two, three, four... clients don't > communicate IDs. > > That's difficult enough, but even worse: evolution doesn't let you > configure the domain-part of the MsgID, so i.e. my messages always > have"@local". Half the world uses that. :-( > > So, the correct behaviour would be to send my messages without an ID. > My providers mailserver would generate a guaranteed unique one. ..exactly like it does with much of the spam I too receive: If I combine Wayne's rule with a check for my popserver sp's* ip-range, I too get zero false positives off it. ..[*]: I only use c2i.net's popservers to fetch mail. ;-) -- ..med vennlig hilsen = with Kind Regards from Arnt... ;-) ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]