OT: performance smf-sav vs milter-ahead

[Scott Silva]

Ken A spake the following on 7/26/2007 7:10 AM:
> Bryan Guest wrote:
>> Hello
>>
>> With my apologies for the off topic nature of this question, has
>> anyone compared performance between SMF-SAV and Milter-Ahead?
>>
>> If so, is there an appreciable difference when using one or the other
>> in conjuction with MailScanner?
>>
>> Many thanks to Julian Field and everyone on this list for MailScanner
>> and the support it receives.  I sincerely appreciate any feedback
>> provided.
>>
>> Bryan Guest
>> Bruce Telecom
> 
> smf-sav seems okay, but watch out for:
> 
> // if (verify && strcmp(verify, "OK") == 0) return SMFIS_ACCEPT;
> 
> ..which says "If the client ssl client certificate verifies with an
> authority, then skip the milters". That's a bad assumption. I just
> commented out the line in the code and recompiled(Yes, I reported it). I
> have not looked at how milter-ahead handles this, or how well it works,
> though it's codebase is certainly more mature and well tested.
> 
> Ken
> 
> 
I agree. Even a spammer can buy a certificate. I have been meaning to give it
a shot, but spent so much time getting it going with mimedefang, I didn't want
to touch a running system. I will give it a shot when I build the new
mailservers probably next month.
I also wanted to try the patch for smf-sav to do the recipient verificatins
first, and not bother with sender verify if the recipient fails.

-- 

MailScanner is like deodorant...
You hope everybody uses it, and
you notice quickly if they don't!!!!