Re: [squid-users] delay_access url_regex acl

On 28/05/11 04:21, Marcus Kool wrote:
>
> Amos Jeffries wrote:
>> On 28/05/11 00:46, Marc Nil wrote:
<snip, snip crazy>
>>>
>>> I also tested this option (with the sale result) I even removed the
>>> InternetAdmin line (still the same result)
>>>
>>
>> Definitely the regex bits then.
>
> If you post the whitelist.no_limit we are able to see if there is
> room for improvement.
> Usually there is.

FYI he did. It was two domain names :(

>
> As a side note for Amos: I made the URL filter ufdbGuard 70% faster
> with a regex optimiser that combines N REs into one large RE composed
> of the N original REs with OR logic. i.e.
> RE1
> RE2
> ...
> REn
> becomes
> RE1|RE2|...|REn
> The speed increase was measured with overall throughput so the actual
> improvement for RE processing must be much higher than 70%.
>

I know that is one of the speed improvements Squid needs. Another is
pruning away useless "." or ".*" prefixes and suffixes.

Feel like making a patch?
  Regex ACL in Squid are loaded by one very small function in
src/acl/RegexData.cc. All that has to be careful of is the -i/+i flags
which toggle case sensitivity between sets of |'able patterns.

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.12
   Beta testers wanted for 3.2.0.7 and 3.1.12.1