Hi Guys, Thanks for your help, and I promise to send my mail in Text format next time.
So, for sure, I agree with you that it is not the right way to fight against SPAM, but my configuration is based on SpamAssassin+Amavis+ClamAV with DNSBL, DKIM implementation and SPF record to fight the Spoofing Attack again the domain I'm responsible for. And I'm very strict with this F**K activity with the spammer, so every email is banned immediately if the spam score is > 8 and a mail domain is banned after 10 occurrences of SPAM. With that, I can manage a BLACKLIST/GREYLIST/WHITELIST approach. With the domain I own for much more than 10 years, the spam number reach less than 10% of the daily mail traffic, compared to 90%, two years ago. I use the same approach for my customers without any issue, and the users are satisfied to be focused only on the relevant email. By the way, some malformed email address still passed through the mail gateway, so it's why I decided to implement this approach. Even if it is not the right way to fight the spam, the objective is to REJECT all the non compliant email at the entry point. SPAM is a mess today, and it is a continuous activity to fight those spammers. Franck ------------------------------------------- E: m...@civis.net -----Message d'origine----- De : owner-postfix-us...@postfix.org [mailto:owner-postfix-us...@postfix.org] De la part de Jeroen Geilman Envoyé : lundi 14 février 2011 02:22 À : postfix-users@postfix.org Objet : Re: Issue with header_checks On 02/14/2011 02:15 AM, Victor Duchovni wrote: > On Mon, Feb 14, 2011 at 01:45:28AM +0100, Jeroen Geilman wrote: > > >>> Any idea for the REJECT 4 to reject the header "From: us e...@domain.tld >>> <mailto:e...@domain.tld>", ie with a space in the first part. >>> >> The example I showed you, while being much more readable, already included >> the space. >> > Dare I suggest that the whole exercise is likely rather pointless. There > are surely better ways to deflect spam than attempting to use regexps > for RFC-enforcement on the From: header. > > This likely yields very little net benefit, and is rather fragile. The > OP is IMHO better off directing his energies elsewhere... > > I don't disagree. I was ready to drop it after my suggestion to use more readable and less error-prone character classes fell on deaf ears. But he hasn't rightly stated why he wants to do this, so I didn't want to flat-out say he shouldn't. -- J.