X-Git-Url: http://git.heureux-cyclage.org/?a=blobdiff_plain;f=includes%2FSearchUpdate.php;h=3f268b86aba3244fa7cc10cbed4713ceff935f25;hb=45b3ac166d86c7c93d1b2298c733c83ce8b85844;hp=03978bea0461541b90ec1c1acd88372bac9297e4;hpb=46787701ce245af3f5ac15aabea7534d79c78749;p=lhc%2Fweb%2Fwiklou.git

diff --git a/includes/SearchUpdate.php b/includes/SearchUpdate.php
index 03978bea04..3f268b86ab 100644
--- a/includes/SearchUpdate.php
+++ b/includes/SearchUpdate.php
@@ -1,45 +1,65 @@
-<?
-# See deferred.doc
-
+<?php
+# $Id$
+/**
+ * See deferred.doc
+ * @package MediaWiki
+ */
+
+/**
+ *
+ * @package MediaWiki
+ */
 class SearchUpdate {
 
-	/* private */ var $mId, $mNamespace, $mTitle, $mText;
+	/* private */ var $mId = 0, $mNamespace, $mTitle, $mText;
 	/* private */ var $mTitleWords;
 
-	function SearchUpdate( $id, $title, $text = false )
-	{
-		$this->mId = $id;
-		$this->mText = $text;
-
+	function SearchUpdate( $id, $title, $text = false ) {
 		$nt = Title::newFromText( $title );
-		$this->mNamespace = $nt->getNamespace();
-		$this->mTitle = $nt->getText(); # Discard namespace
+		if( $nt ) {
+			$this->mId = $id;
+			$this->mText = $text;
 
-		$this->mTitleWords = $this->mTextWords = array();
+			$this->mNamespace = $nt->getNamespace();
+			$this->mTitle = $nt->getText(); # Discard namespace
+
+			$this->mTitleWords = $this->mTextWords = array();
+		} else {
+			wfDebug( "SearchUpdate object created with invalid title '$title'\n" );
+		}
 	}
 
-	function doUpdate()
-	{
+	function doUpdate() {
 		global $wgDBminWordLen, $wgLang, $wgDisableSearchUpdate;
 
-		if( $wgDisableSearchUpdate ) {
+		if( $wgDisableSearchUpdate || !$this->mId ) {
 			return false;
 		}
-		$lc = SearchEngine::legalSearchChars() . "&#;";
+		$fname = 'SearchUpdate::doUpdate';
+		wfProfileIn( $fname );
+
+		require_once( 'SearchEngine.php' );
+		$lc = SearchEngine::legalSearchChars() . '&#;';
+		$db =& wfGetDB( DB_MASTER );
+		$searchindex = $db->tableName( 'searchindex' );
+		
 		if( $this->mText == false ) {
 			# Just update the title
-			$sql = "UPDATE LOW_PRIORITY searchindex SET si_title='" .
-			  wfStrencode( Title::indexTitle( $this->mNamespace, $this->mTitle ) ) .
+			$lowpri = $db->lowPriorityOption();
+			$sql = "UPDATE $lowpri $searchindex SET si_title='" .
+			  $db->strencode( Title::indexTitle( $this->mNamespace, $this->mTitle ) ) .
 			  "' WHERE si_page={$this->mId}";
-			wfQuery( $sql, DB_WRITE, "SearchUpdate::doUpdate" );
+			$db->query( $sql, "SearchUpdate::doUpdate" );
+			wfProfileOut( $fname );
 			return;
 		}
 
 		# Language-specific strip/conversion
 		$text = $wgLang->stripForSearch( $this->mText );
 
+		wfProfileIn( $fname.'-regexps' );
 		$text = preg_replace( "/<\\/?\\s*[A-Za-z][A-Za-z0-9]*\\s*([^>]*?)>/",
-		  " ", strtolower( " " . $text /*$this->mText*/ . " " ) ); # Strip HTML markup
+		  ' ', strtolower( " " . $text /*$this->mText*/ . " " ) ); # Strip HTML markup
 		$text = preg_replace( "/(^|\\n)\\s*==\\s+([^\\n]+)\\s+==\\s/sD",
 		  "\\2 \\2 \\2 ", $text ); # Emphasize headings
 
@@ -65,16 +85,33 @@ class SearchUpdate {
 		$text = preg_replace( "/[^{$lc}]+/", " ", $text );
 
 		# Handle 's, s'
-		$text = preg_replace( "/([{$lc}]+)'s /", "\\1 \\1's ", $text );
-		$text = preg_replace( "/([{$lc}]+)s' /", "\\1s ", $text );
+		#
+		#   $text = preg_replace( "/([{$lc}]+)'s /", "\\1 \\1's ", $text );
+		#   $text = preg_replace( "/([{$lc}]+)s' /", "\\1s ", $text );
+		#
+		# These tail-anchored regexps are insanely slow. The worst case comes
+		# when Japanese or Chinese text (ie, no word spacing) is written on
+		# a wiki configured for Western UTF-8 mode. The Unicode characters are
+		# expanded to hex codes and the "words" are very long paragraph-length
+		# monstrosities. On a large page the above regexps may take over 20
+		# seconds *each* on a 1GHz-level processor.
+		#
+		# Following are reversed versions which are consistently fast
+		# (about 3 milliseconds on 1GHz-level processor).
+		#
+		$text = strrev( preg_replace( "/ s'([{$lc}]+)/", " s'\\1 \\1", strrev( $text ) ) );
+		$text = strrev( preg_replace( "/ 's([{$lc}]+)/", " s\\1", strrev( $text ) ) );
 
 		# Strip wiki '' and '''
 		$text = preg_replace( "/''[']*/", " ", $text );
-		
-		$sql = "REPLACE DELAYED INTO searchindex (si_page,si_title,si_text) VALUES ({$this->mId},'" .
-		  wfStrencode( Title::indexTitle( $this->mNamespace, $this->mTitle ) ) . "','" .
-		  wfStrencode( $text ) . "')";
-		wfQuery( $sql, DB_WRITE, "SearchUpdate::doUpdate" );
+		wfProfileOut( "$fname-regexps" );
+                $db->replace( $searchindex, array(array('si_page')),
+                  array(
+                    'si_page' => $this->mId,
+                    'si_title' => $db->strencode( Title::indexTitle( $this->mNamespace, $this->mTitle ) ),
+                    'si_text' => $db->strencode( $text )
+                  ), 'SearchUpdate::doUpdate' );
+		wfProfileOut( $fname );
 	}
 }