dccifd (unmaintained project) local scan and exim

[anup123]

Hi.

This may please be considered as feature request. Integrating local scan in exim for using dccifd for spam control.

This should be possible as per following url which needs including dccifd-client.c in compile... Exim compile i believe.

The file is available at:

http://www.physicman.net/projects/exim/

TIA

Anup

[chirpy]

Why, when you can simple install DCC and SpamAssassin will immediately start using it?

[anup123]

Why, when you can simple install DCC and SpamAssassin will immediately start using it?

I am using dccifd.
SA automatically starts using it fine.
But all mails are delivered despite the dccifd status being "reject" as defined by the reject threshold. SA is just adding the header and the score...

I did spamassassin -D --lint and it picks up dccifd fine.
So this request.
Something to reject it like exiscan ...

Thanks
Anup

[chirpy]

Ah, I see. Personally, I prefer to keep it as an additional scoring mechanism for SA in addition to Vipul's Razor, both of which help better identify spam without actually blocking anything (I need potential spam to still come through for a lot of customers).

[anup123]

Ah, I see. Personally, I prefer to keep it as an additional scoring mechanism for SA in addition to Vipul's Razor, both of which help better identify spam without actually blocking anything (I need potential spam to still come through for a lot of customers).

Hmmm.

Score bumping would hardly make a diff if 20 is pushed to 22 with Aid if DCC. If it's allowed in, then i think all these are really not neessary as a good ruleset would just suffice to mark a SPAM as SPAM... Vipul's Razor + Pyzor +DCC at best would bump the score by about 5-7 at best which becomes a cosmetic stuff as long as SA rules and additional rules take care of ensuring that a SPAM is indeed marked as SPAM

Anup

[chirpy]

Ah! Well, that's because I'm using a different approach (through MailScanner) using two scoring levels - 5 for probably spam, 20 for definitely spam. 20 usually gets deleted, 5 gets through but filtered out elsewhere. That's why DCC and Razor are extremely useful to me for bumping up the score.

[anup123]

Ah! Well, that's because I'm using a different approach (through MailScanner) using two scoring levels - 5 for probably spam, 20 for definitely spam. 20 usually gets deleted, 5 gets through but filtered out elsewhere. That's why DCC and Razor are extremely useful to me for bumping up the score.

16+ blocked ... at SMTP level and would need dccifd bumping up to enable further blocking without really lowering the theshold.

Actually a lot of mails blocked are with invalid HELO/EHLO which are obvious spam out from ratwares, bounce message addressed to multiple reepients etc which i implemented recently and then found that there was a corresponding decrease in the RBL's block count.

Anup