From 92ed21f0ab30611e49168cd5a2619d26474b35fb Mon Sep 17 00:00:00 2001
From: "Mark A. Hershberger" <mah@users.mediawiki.org>
Date: Tue, 9 Mar 2010 04:19:55 +0000
Subject: [PATCH] =?utf8?q?follow-up=20r61856=20=E2=80=94=20wordsegmentatio?=
 =?utf8?q?n=20should=20be=20done=20for=20all=20search=20engines,=20not=20j?=
 =?utf8?q?ust=20mysql?=
MIME-Version: 1.0
Content-Type: text/plain; charset=utf8
Content-Transfer-Encoding: 8bit

---
 includes/search/SearchEngine.php      | 5 ++++-
 includes/search/SearchMySQL.php       | 3 +--
 includes/search/SearchUpdate.php      | 6 +++---
 languages/classes/LanguageZh_hans.php | 1 +
 4 files changed, 9 insertions(+), 6 deletions(-)

diff --git a/includes/search/SearchEngine.php b/includes/search/SearchEngine.php
index 7f80447125..67790dd974 100644
--- a/includes/search/SearchEngine.php
+++ b/includes/search/SearchEngine.php
@@ -56,7 +56,10 @@ class SearchEngine {
 	 * @return string
 	 */
 	public function normalizeText( $string ) {
-		return $string;
+		global $wgContLang;
+
+		// Some languages such as Chinese require word segmentation
+		return $wgContLang->wordSegmentation( $string );
 	}
 
 	/**
diff --git a/includes/search/SearchMySQL.php b/includes/search/SearchMySQL.php
index 0c238be892..ff6ae7abcb 100644
--- a/includes/search/SearchMySQL.php
+++ b/includes/search/SearchMySQL.php
@@ -325,8 +325,7 @@ class SearchMySQL extends SearchEngine {
 
 		wfProfileIn( __METHOD__ );
 		
-		// Some languages such as Chinese require word segmentation
-		$out = $wgContLang->wordSegmentation( $string );
+		$out = parent::normalizeText( $string );
 
 		// MySQL fulltext index doesn't grok utf-8, so we
 		// need to fold cases and convert to hex
diff --git a/includes/search/SearchUpdate.php b/includes/search/SearchUpdate.php
index e30c70e6d1..7840dd82a3 100644
--- a/includes/search/SearchUpdate.php
+++ b/includes/search/SearchUpdate.php
@@ -37,7 +37,7 @@ class SearchUpdate {
 
 		if( $this->mText === false ) {
 			$search->updateTitle($this->mId,
-				Title::indexTitle( $this->mNamespace, $this->mTitle ));
+				$search->normalizeText( Title::indexTitle( $this->mNamespace, $this->mTitle ) ) );
 			wfProfileOut( $fname );
 			return;
 		}
@@ -97,8 +97,8 @@ class SearchUpdate {
 		wfRunHooks( 'SearchUpdate', array( $this->mId, $this->mNamespace, $this->mTitle, &$text ) );
 
 		# Perform the actual update
-		$search->update($this->mId, Title::indexTitle( $this->mNamespace, $this->mTitle ),
-				$text);
+		$search->update($this->mId, $search->normalizeText( Title::indexTitle( $this->mNamespace, $this->mTitle ) ),
+				$search->normalizeText( $text ) );
 
 		wfProfileOut( $fname );
 	}
diff --git a/languages/classes/LanguageZh_hans.php b/languages/classes/LanguageZh_hans.php
index 5b03d73145..20d34155b6 100644
--- a/languages/classes/LanguageZh_hans.php
+++ b/languages/classes/LanguageZh_hans.php
@@ -25,6 +25,7 @@ class LanguageZh_hans extends Language {
 		// Double-width roman characters
 		$s = self::convertDoubleWidth( $string );
 		$s = trim( $s );
+		$s = self::wordSegmentation( $s );
 		$s = parent::normalizeForSearch( $s );
 
 		wfProfileOut( __METHOD__ );
-- 
2.20.1