On November 26, 2003 12:58 pm, Kris Deugau wrote:
> Which is next to useless for even 30-40 rules. You would have to have
> n![1] states for n rules- there is NO way to determine which individual
> rules will trigger in advance.
Yeah, I wondered about that... I was just proposing how one may be
Pedro Sam wrote:
> hehe, we can give a state to each possible combination of HITS for the
> rules. So if rules 1, 3, 5, 7 hit, we give that a state, and if 2, 4,
> 6, 8 hit, we give it another state, and so on... I think they call it
> subset construction or something...
Which is next to useless
On November 25, 2003 10:31 pm, Alexander Litvinov wrote:
> Heh, it seems it would be nice to make SA scan messages fatser. If I
> undersand
your idea correctly, you want not to run regexp one by one, but
> write the state machine for all regepes and walk on this states by the
> mail, but... I und
Heh, it seems it would be nice to make SA scan messages fatser. If I undersand
your idea correctly, you want not to run regexp one by one, but write the
state machine for all regepes and walk on this states by the mail, but... I
undersand how this may be faster (liner time of the message size) i
> At 03:27 11/25/2003 -0600, Scott A Crosby wrote:
> ... except one caveat: What was the memory utilization like? That's *my*
> big problem with SA - I had to bump the RAM in my mailserver twice (256M
> ->
> 512M -> 1G) for SA alone, and it's still shakey to the point I cannot
> deploy it sitewide
At 03:27 11/25/2003 -0600, Scott A Crosby wrote:
I mentioned this about a year ago, but now that people are starting to
write rulesets with hundreds to thousands of new rules, I thought I'd
bring it up again.
How happy are people with the performance of SA, especially with all
of thee new rules? Th
On Tue, 25 Nov 2003, Scott A Crosby wrote:
> How happy are people with the performance of SA, especially with all
> of thee new rules? The reason I ask is that I'm on-again, off-again
I think a faster engine is a great idea, no matter what. Spam is growing --
even if our servers are big enough N
