Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:38174
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 66.249.92.172 as permitted sender)
DomainKey-Status: bad
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:to:subject:in-reply-to:mime-version
         :content-type:references;
        b=tarEyHGUwrWbmPmMLMQXCx9qvspYzxYFHrn1JV24C1gAE/pJ8SJQcyKds9utKw7hSi
         bCrRitXyd6LucOaKs63rWBaFD9lsh9Vw5LBqllWHNGqY1zMZ5Oj5NyD8Q+V974sgDvJx
         wLF8WhtLe0m+9lL7xw9tLAaO2jEXEhN32JtNw=
Message-ID: <31fe29920806111634l637f846es7227056264d16601@mail.gmail.com>
Date: Thu, 12 Jun 2008 00:34:35 +0100
To: "Stanislav Malyshev" <stas@zend.com>, internals@lists.php.net
In-Reply-To: <31fe29920806111444u250d933au36d824f48c78ff5@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/alternative; 
	boundary="----=_Part_199_13472389.1213227275058"
References: <7E62CA6E-83F4-4F9C-86FB-75EBE7D489C9@gmail.com>
	 <D406487CA14F443B81EAE37236CAD5E6@pc07654>
	 <484D36EB.9080202@macvicar.net>
	 <31fe29920806110114i263a47b4sd66d4849ceb22980@mail.gmail.com>
	 <484F910A.3080208@zend.com>
	 <31fe29920806111444u250d933au36d824f48c78ff5@mail.gmail.com>
Subject: Re: [PHP-DEV] Algorithm Optimizations - string search
From: michal.dziemianko@gmail.com ("Michal Dziemianko")

------=_Part_199_13472389.1213227275058
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

I;ve just noticed that the diff I posted is form my test-bed, not the actual
final version. the code should be:

   if (needle_len == 1){
        return (char *)memchr(p, *needle, (end-p));
    }

instead of

+       if (needle_len == 1){
+               return (char *)memchr(p, *needle, (end-p+1)); <- this fails
for strpos("", NULL);
+       }
+


Christopher Jones wrote:
* Hi Michal,
*
*This is always a fun topic.
*
*I think your stripos tests should show the no-match performance case.
*
*Chris

I haven't said the  implementation with lowering chars at the compare time
is better: I've just shown some comparison results. Of course it is slower
in no-match case - it is even slower when the match is closer to end of the
string (it is why I posted 2 different results for the same string length -
one for match in the middle, and one close to the end - which is actually
similar to no-match). I just said it MIGHT be worth considering. I will
analyse my logs to find out what is the average match position, to check if
it is really worth implementing.

------=_Part_199_13472389.1213227275058--