[dokuwiki] Re: Blacklist Performance
- From: Doogie <wiki@xxxxxxxxx>
- To: <dokuwiki@xxxxxxxxxxxxx>
- Date: Wed, 15 Sep 2010 08:45:27 +0200
On Wed, 15 Sep 2010 08:37:49 +0200, Doogie <wiki@xxxxxxxxx> wrote:
> So I think this can easily be optimized, for example by transforming all
> the lines into one big "OR'ed" regular expression:
> s/blockeded-domain1.com|anoter-bad-domain.net|and-so-on.de/
> Robert
Ups, I just realized, that this is what the code is already doing :-) But
still I think this can be optimized, by "anchoring" the regular expression
at the "http" prefix:
old code:
if(preg_match('#('.join('|',$re).')#si',$text, $match=array())) {
return true;
}
my suggestion: strip all http prefixes in wordblock.conf and put them as a
prefix:
if(preg_match('#https?:\/\/([^\/]*\.)?('.join('|',$re).')#si',$text,
$match=array())) {
return true;
}
Shouldn't this be much faster, because matching the prefix in the regular
expression's internal state machine fails much faster, when checking the
contant of a page?
Yours,
Robert
--
DokuWiki mailing list - more info at
http://www.dokuwiki.org/mailinglist
Other related posts:
