regexps in PHP,again

People keep on insisting,that preg_match is better for non-unicode lookups than mb_ereg. So,here are actual benchmarks to make it clear.

Here are results:

preg_match:19.8039090633mb_ereg:  15.9386620522mb_ereg_search:1.24934506416

Here is the source:

<?php$regexp = '[\w]+@[\w]+\.com';$pcre_regexp = '/'.$regexp.'/';$regexp2 = '[\s]+@[\s]+\.com';$pcre_regexp2 = '/'.$regexp2.'/';$text = 'blabla bla blbaaasdajkln dsfkl klewnjklfnjkne qwe123@gg.net adkljaskdlnkljnasdljk qwe@test.comasdjlajnsdklnasdklnjl';$t1 = microtime(true);for ($i = 0;$i < 100000;$i++){ $res1 = preg_match($pcre_regexp,$text); $res2 = preg_match($pcre_regexp2,$text)}$t2 = microtime(true);$t3 = microtime(true);for ($i = 0;$i < 100000;$i++){ $res3 = mb_ereg($regexp,$text); $res4 = mb_ereg($regexp2,$text)}$t4 = microtime(true);$t5 = microtime(true);mb_ereg_search_init($text);for ($i = 0;$i < 100000;$i++){ $res5 = mb_ereg_search($regexp); $res6 = mb_ereg_search($regexp2)}$t6 = microtime(true);echo 'preg_match:'.($t2 - $t1)."\n";echo 'mb_ereg:  '.($t4 - $t3)."\n";echo 'mb_ereg_search:'.($t6 - $t5)."\n";

Feel free to check it out yourself.