aboutsummaryrefslogtreecommitdiffstats
path: root/maintenance/benchmarks/bench_utf8_title_check.php
diff options
context:
space:
mode:
Diffstat (limited to 'maintenance/benchmarks/bench_utf8_title_check.php')
-rw-r--r--maintenance/benchmarks/bench_utf8_title_check.php107
1 files changed, 0 insertions, 107 deletions
diff --git a/maintenance/benchmarks/bench_utf8_title_check.php b/maintenance/benchmarks/bench_utf8_title_check.php
deleted file mode 100644
index 2e4b90292d81..000000000000
--- a/maintenance/benchmarks/bench_utf8_title_check.php
+++ /dev/null
@@ -1,107 +0,0 @@
-<?php
-/**
- * @file
- * @ingroup Benchmark
- */
-
-require_once( dirname( __FILE__ ) . '/Benchmarker.php' );
-
-/**
- * This little benchmark executes the regexp used in Language->checkTitleEncoding() and compares its execution time
- * against that of mb_check_encoding, if available.
- */
-class bench_utf8_title_check extends Benchmarker {
-
- private $canRun;
-
- private $data;
-
- public function __construct() {
- parent::__construct();
-
- $this->data = array (
- "",
- "United States of America", // 7bit ASCII
- "S%C3%A9rie%20t%C3%A9l%C3%A9vis%C3%A9e",
- "Acteur%7CAlbert%20Robbins%7CAnglais%7CAnn%20Donahue%7CAnthony%20E.%20Zuiker%7CCarol%20Mendelsohn",
- // This comes from bug 36839
- "Acteur%7CAlbert%20Robbins%7CAnglais%7CAnn%20Donahue%7CAnthony%20E.%20Zuiker%7CCarol%20Mendelsohn%7C"
- . "Catherine%20Willows%7CDavid%20Hodges%7CDavid%20Phillips%7CGil%20Grissom%7CGreg%20Sanders%7CHodges%7C"
- . "Internet%20Movie%20Database%7CJim%20Brass%7CLady%20Heather%7C"
- . "Les%20Experts%20(s%C3%A9rie%20t%C3%A9l%C3%A9vis%C3%A9e)%7CLes%20Experts%20:%20Manhattan%7C"
- . "Les%20Experts%20:%20Miami%7CListe%20des%20personnages%20des%20Experts%7C"
- . "Liste%20des%20%C3%A9pisodes%20des%20Experts%7CMod%C3%A8le%20discussion:Palette%20Les%20Experts%7C"
- . "Nick%20Stokes%7CPersonnage%20de%20fiction%7CPersonnage%20fictif%7CPersonnage%20de%20fiction%7C"
- . "Personnages%20r%C3%A9currents%20dans%20Les%20Experts%7CRaymond%20Langston%7CRiley%20Adams%7C"
- . "Saison%201%20des%20Experts%7CSaison%2010%20des%20Experts%7CSaison%2011%20des%20Experts%7C"
- . "Saison%2012%20des%20Experts%7CSaison%202%20des%20Experts%7CSaison%203%20des%20Experts%7C"
- . "Saison%204%20des%20Experts%7CSaison%205%20des%20Experts%7CSaison%206%20des%20Experts%7C"
- . "Saison%207%20des%20Experts%7CSaison%208%20des%20Experts%7CSaison%209%20des%20Experts%7C"
- . "Sara%20Sidle%7CSofia%20Curtis%7CS%C3%A9rie%20t%C3%A9l%C3%A9vis%C3%A9e%7CWallace%20Langham%7C"
- . "Warrick%20Brown%7CWendy%20Simms%7C%C3%89tats-Unis"
- );
-
- $this->canRun = function_exists ( 'mb_check_encoding' );
-
- if ( $this->canRun ) {
- $this->mDescription = "Benchmark for using a regexp vs. mb_check_encoding to check for UTF-8 encoding.";
- mb_internal_encoding( 'UTF-8' );
- } else {
- $this->mDescription = "CANNOT RUN benchmark using mb_check_encoding: function not available.";
- }
- }
-
- public function execute() {
- if ( !$this->canRun ) {
- return;
- }
- $benchmarks = array();
- foreach ($this->data as $val) {
- $benchmarks[] = array(
- 'function' => array( $this, 'use_regexp' ),
- 'args' => array( rawurldecode ( $val ) )
- );
- $benchmarks[] = array(
- 'function' => array( $this, 'use_regexp_non_capturing' ),
- 'args' => array( rawurldecode ( $val ) )
- );
- $benchmarks[] = array(
- 'function' => array( $this, 'use_regexp_once_only' ),
- 'args' => array( rawurldecode ( $val ) )
- );
- $benchmarks[] = array(
- 'function' => array( $this, 'use_mb_check_encoding' ),
- 'args' => array( rawurldecode ( $val ) )
- );
- }
- $this->bench( $benchmarks );
- print $this->getFormattedResults();
- }
-
- private $isutf8;
-
- function use_regexp( $s ) {
- $this->isutf8 = preg_match( '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
- '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
- }
-
- function use_regexp_non_capturing( $s ) {
- // Same as above with a non-capturing subgroup.
- $this->isutf8 = preg_match( '/^(?:[\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
- '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
- }
-
- function use_regexp_once_only( $s ) {
- // Same as above with a once-only subgroup.
- $this->isutf8 = preg_match( '/^(?>[\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
- '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})+$/', $s );
- }
-
- function use_mb_check_encoding( $s ) {
- $this->isutf8 = mb_check_encoding( $s, 'UTF-8' );
- }
-
-}
-
-$maintClass = 'bench_utf8_title_check';
-require_once( RUN_MAINTENANCE_IF_MAIN );