diff options
author | Pierre Schmitz <pierre@archlinux.de> | 2011-06-22 11:28:20 +0200 |
---|---|---|
committer | Pierre Schmitz <pierre@archlinux.de> | 2011-06-22 11:28:20 +0200 |
commit | 9db190c7e736ec8d063187d4241b59feaf7dc2d1 (patch) | |
tree | 46d1a0dee7febef5c2d57a9f7b972be16a163b3d /includes/search | |
parent | 78677c7bbdcc9739f6c10c75935898a20e1acd9e (diff) |
update to MediaWiki 1.17.0
Diffstat (limited to 'includes/search')
-rw-r--r-- | includes/search/SearchEngine.php | 761 | ||||
-rw-r--r-- | includes/search/SearchIBM_DB2.php | 53 | ||||
-rw-r--r-- | includes/search/SearchMssql.php | 254 | ||||
-rw-r--r-- | includes/search/SearchMySQL.php | 52 | ||||
-rw-r--r-- | includes/search/SearchMySQL4.php | 34 | ||||
-rw-r--r-- | includes/search/SearchOracle.php | 62 | ||||
-rw-r--r-- | includes/search/SearchPostgres.php | 64 | ||||
-rw-r--r-- | includes/search/SearchSqlite.php | 82 | ||||
-rw-r--r-- | includes/search/SearchUpdate.php | 31 |
9 files changed, 873 insertions, 520 deletions
diff --git a/includes/search/SearchEngine.php b/includes/search/SearchEngine.php index f4ca700d..17482da2 100644 --- a/includes/search/SearchEngine.php +++ b/includes/search/SearchEngine.php @@ -1,12 +1,16 @@ <?php /** - * @defgroup Search Search + * Basic search engine * * @file * @ingroup Search */ /** + * @defgroup Search Search + */ + +/** * Contain a class for special pages * @ingroup Search */ @@ -18,6 +22,14 @@ class SearchEngine { var $namespaces = array( NS_MAIN ); var $showRedirects = false; + function __construct($db = null) { + if ( $db ) { + $this->db = $db; + } else { + $this->db = wfGetDB( DB_SLAVE ); + } + } + /** * Perform a full text search query and return a result set. * If title searches are not supported or disabled, return null. @@ -41,12 +53,12 @@ class SearchEngine { function searchTitle( $term ) { return null; } - + /** If this search backend can list/unlist redirects */ function acceptListRedirects() { return true; } - + /** * When overridden in derived class, performs database-specific conversions * on text to be used for searching or updating search index. @@ -56,7 +68,10 @@ class SearchEngine { * @return string */ public function normalizeText( $string ) { - return $string; + global $wgContLang; + + // Some languages such as Chinese require word segmentation + return $wgContLang->segmentByWord( $string ); } /** @@ -66,7 +81,7 @@ class SearchEngine { function transformSearchTerm( $term ) { return $term; } - + /** * If an exact title match can be found, or a very slightly close match, * return the title. If no match, returns NULL. @@ -76,41 +91,53 @@ class SearchEngine { */ public static function getNearMatch( $searchterm ) { $title = self::getNearMatchInternal( $searchterm ); - + wfRunHooks( 'SearchGetNearMatchComplete', array( $searchterm, &$title ) ); return $title; } /** + * Do a near match (see SearchEngine::getNearMatch) and wrap it into a + * SearchResultSet. + * + * @param $searchterm string + * @return SearchResultSet + */ + public static function getNearMatchResultSet( $searchterm ) { + return new SearchNearMatchResultSet( self::getNearMatch( $searchterm ) ); + } + + /** * Really find the title match. */ private static function getNearMatchInternal( $searchterm ) { global $wgContLang; - $allSearchTerms = array($searchterm); + $allSearchTerms = array( $searchterm ); if ( $wgContLang->hasVariants() ) { - $allSearchTerms = array_merge($allSearchTerms,$wgContLang->convertLinkToAllVariants($searchterm)); + $allSearchTerms = array_merge( $allSearchTerms, $wgContLang->autoConvertToAllVariants( $searchterm ) ); } - if( !wfRunHooks( 'SearchGetNearMatchBefore', array( $allSearchTerms, &$titleResult ) ) ) { + $titleResult = null; + if ( !wfRunHooks( 'SearchGetNearMatchBefore', array( $allSearchTerms, &$titleResult ) ) ) { return $titleResult; } - foreach($allSearchTerms as $term) { + foreach ( $allSearchTerms as $term ) { # Exact match? No need to look further. $title = Title::newFromText( $term ); - if (is_null($title)) + if ( is_null( $title ) ) return null; if ( $title->getNamespace() == NS_SPECIAL || $title->isExternal() || $title->exists() ) { return $title; } - + # See if it still otherwise has content is some sane sense $article = MediaWiki::articleFromTitle( $title ); - if( $article->hasViewableContent() ) { + if ( $article->hasViewableContent() ) { return $title; } @@ -136,14 +163,14 @@ class SearchEngine { } # Now try Word-Caps-Breaking-At-Word-Breaks, for hyphenated names etc - $title = Title::newFromText( $wgContLang->ucwordbreaks($term) ); + $title = Title::newFromText( $wgContLang->ucwordbreaks( $term ) ); if ( $title && $title->exists() ) { return $title; } // Give hooks a chance at better match variants $title = null; - if( !wfRunHooks( 'SearchGetNearMatch', array( $term, &$title ) ) ) { + if ( !wfRunHooks( 'SearchGetNearMatch', array( $term, &$title ) ) ) { return $title; } } @@ -151,7 +178,7 @@ class SearchEngine { $title = Title::newFromText( $searchterm ); # Entering an IP address goes to the contributions page - if ( ( $title->getNamespace() == NS_USER && User::isIP($title->getText() ) ) + if ( ( $title->getNamespace() == NS_USER && User::isIP( $title->getText() ) ) || User::isIP( trim( $searchterm ) ) ) { return SpecialPage::getTitleFor( 'Contributions', $title->getDBkey() ); } @@ -165,22 +192,22 @@ class SearchEngine { # Go to images that exist even if there's no local page. # There may have been a funny upload, or it may be on a shared # file repository such as Wikimedia Commons. - if( $title->getNamespace() == NS_FILE ) { + if ( $title->getNamespace() == NS_FILE ) { $image = wfFindFile( $title ); - if( $image ) { + if ( $image ) { return $title; } } # MediaWiki namespace? Page may be "implied" if not customized. # Just return it, with caps forced as the message system likes it. - if( $title->getNamespace() == NS_MEDIAWIKI ) { + if ( $title->getNamespace() == NS_MEDIAWIKI ) { return Title::makeTitle( NS_MEDIAWIKI, $wgContLang->ucfirst( $title->getText() ) ); } # Quoted term? Try without the quotes... $matches = array(); - if( preg_match( '/^"([^"]+)"$/', $searchterm, $matches ) ) { + if ( preg_match( '/^"([^"]+)"$/', $searchterm, $matches ) ) { return SearchEngine::getNearMatch( $matches[1] ); } @@ -219,28 +246,28 @@ class SearchEngine { * * @param $query String */ - function replacePrefixes( $query ){ + function replacePrefixes( $query ) { global $wgContLang; $parsed = $query; - if( strpos($query,':') === false ) { // nothing to do + if ( strpos( $query, ':' ) === false ) { // nothing to do wfRunHooks( 'SearchEngineReplacePrefixesComplete', array( $this, $query, &$parsed ) ); return $parsed; } - - $allkeyword = wfMsgForContent('searchall').":"; - if( strncmp($query, $allkeyword, strlen($allkeyword)) == 0 ){ + + $allkeyword = wfMsgForContent( 'searchall' ) . ":"; + if ( strncmp( $query, $allkeyword, strlen( $allkeyword ) ) == 0 ) { $this->namespaces = null; - $parsed = substr($query,strlen($allkeyword)); - } else if( strpos($query,':') !== false ) { - $prefix = substr($query,0,strpos($query,':')); - $index = $wgContLang->getNsIndex($prefix); - if($index !== false){ - $this->namespaces = array($index); - $parsed = substr($query,strlen($prefix)+1); + $parsed = substr( $query, strlen( $allkeyword ) ); + } else if ( strpos( $query, ':' ) !== false ) { + $prefix = substr( $query, 0, strpos( $query, ':' ) ); + $index = $wgContLang->getNsIndex( $prefix ); + if ( $index !== false ) { + $this->namespaces = array( $index ); + $parsed = substr( $query, strlen( $prefix ) + 1 ); } } - if(trim($parsed) == '') + if ( trim( $parsed ) == '' ) $parsed = $query; // prefix was the whole query wfRunHooks( 'SearchEngineReplacePrefixesComplete', array( $this, $query, &$parsed ) ); @@ -255,16 +282,16 @@ class SearchEngine { public static function searchableNamespaces() { global $wgContLang; $arr = array(); - foreach( $wgContLang->getNamespaces() as $ns => $name ) { - if( $ns >= NS_MAIN ) { + foreach ( $wgContLang->getNamespaces() as $ns => $name ) { + if ( $ns >= NS_MAIN ) { $arr[$ns] = $name; } } - + wfRunHooks( 'SearchableNamespaces', array( &$arr ) ); return $arr; } - + /** * Extract default namespaces to search from the given user's * settings, returning a list of index numbers. @@ -274,78 +301,78 @@ class SearchEngine { */ public static function userNamespaces( $user ) { global $wgSearchEverythingOnlyLoggedIn; - + // get search everything preference, that can be set to be read for logged-in users $searcheverything = false; - if( ( $wgSearchEverythingOnlyLoggedIn && $user->isLoggedIn() ) + if ( ( $wgSearchEverythingOnlyLoggedIn && $user->isLoggedIn() ) || !$wgSearchEverythingOnlyLoggedIn ) - $searcheverything = $user->getOption('searcheverything'); - - // searcheverything overrides other options - if( $searcheverything ) - return array_keys(SearchEngine::searchableNamespaces()); - + $searcheverything = $user->getOption( 'searcheverything' ); + + // searcheverything overrides other options + if ( $searcheverything ) + return array_keys( SearchEngine::searchableNamespaces() ); + $arr = Preferences::loadOldSearchNs( $user ); $searchableNamespaces = SearchEngine::searchableNamespaces(); - - $arr = array_intersect( $arr, array_keys($searchableNamespaces) ); // Filter - + + $arr = array_intersect( $arr, array_keys( $searchableNamespaces ) ); // Filter + return $arr; } - + /** * Find snippet highlight settings for a given user * * @param $user User - * @return Array contextlines, contextchars + * @return Array contextlines, contextchars */ - public static function userHighlightPrefs( &$user ){ - //$contextlines = $user->getOption( 'contextlines', 5 ); - //$contextchars = $user->getOption( 'contextchars', 50 ); + public static function userHighlightPrefs( &$user ) { + // $contextlines = $user->getOption( 'contextlines', 5 ); + // $contextchars = $user->getOption( 'contextchars', 50 ); $contextlines = 2; // Hardcode this. Old defaults sucked. :) $contextchars = 75; // same as above.... :P - return array($contextlines, $contextchars); + return array( $contextlines, $contextchars ); } - + /** * An array of namespaces indexes to be searched by default - * - * @return Array + * + * @return Array */ - public static function defaultNamespaces(){ + public static function defaultNamespaces() { global $wgNamespacesToBeSearchedDefault; - - return array_keys($wgNamespacesToBeSearchedDefault, true); + + return array_keys( $wgNamespacesToBeSearchedDefault, true ); } - + /** * Get a list of namespace names useful for showing in tooltips * and preferences * * @param $namespaces Array */ - public static function namespacesAsText( $namespaces ){ + public static function namespacesAsText( $namespaces ) { global $wgContLang; - - $formatted = array_map( array($wgContLang,'getFormattedNsText'), $namespaces ); - foreach( $formatted as $key => $ns ){ - if ( empty($ns) ) + + $formatted = array_map( array( $wgContLang, 'getFormattedNsText' ), $namespaces ); + foreach ( $formatted as $key => $ns ) { + if ( empty( $ns ) ) $formatted[$key] = wfMsg( 'blanknamespace' ); } return $formatted; } - + /** * Return the help namespaces to be shown on Special:Search - * - * @return Array + * + * @return Array */ public static function helpNamespaces() { global $wgNamespacesToBeSearchedHelp; - + return array_keys( $wgNamespacesToBeSearchedHelp, true ); } - + /** * Return a 'cleaned up' search string * @@ -364,14 +391,15 @@ class SearchEngine { */ public static function create() { global $wgSearchType; - $dbr = wfGetDB( DB_SLAVE ); - if( $wgSearchType ) { + $dbr = null; + if ( $wgSearchType ) { $class = $wgSearchType; } else { + $dbr = wfGetDB( DB_SLAVE ); $class = $dbr->getSearchEngine(); } $search = new $class( $dbr ); - $search->setLimitOffset(0,0); + $search->setLimitOffset( 0, 0 ); return $search; } @@ -399,34 +427,34 @@ class SearchEngine { function updateTitle( $id, $title ) { // no-op } - + /** * Get OpenSearch suggestion template - * + * * @return String */ public static function getOpenSearchTemplate() { - global $wgOpenSearchTemplate, $wgServer, $wgScriptPath; - if( $wgOpenSearchTemplate ) { + global $wgOpenSearchTemplate, $wgServer; + if ( $wgOpenSearchTemplate ) { return $wgOpenSearchTemplate; - } else { + } else { $ns = implode( '|', SearchEngine::defaultNamespaces() ); - if( !$ns ) $ns = "0"; - return $wgServer . $wgScriptPath . '/api.php?action=opensearch&search={searchTerms}&namespace='.$ns; + if ( !$ns ) $ns = "0"; + return $wgServer . wfScript( 'api' ) . '?action=opensearch&search={searchTerms}&namespace=' . $ns; } } - + /** - * Get internal MediaWiki Suggest template - * + * Get internal MediaWiki Suggest template + * * @return String */ public static function getMWSuggestTemplate() { - global $wgMWSuggestTemplate, $wgServer, $wgScriptPath; - if($wgMWSuggestTemplate) + global $wgMWSuggestTemplate, $wgServer; + if ( $wgMWSuggestTemplate ) return $wgMWSuggestTemplate; - else - return $wgServer . $wgScriptPath . '/api.php?action=opensearch&search={searchTerms}&namespace={namespaces}&suggest'; + else + return $wgServer . wfScript( 'api' ) . '?action=opensearch&search={searchTerms}&namespace={namespaces}&suggest'; } } @@ -486,27 +514,27 @@ class SearchResultSet { /** * @return String: suggested query, null if none */ - function getSuggestionQuery(){ + function getSuggestionQuery() { return null; } /** * @return String: HTML highlighted suggested query, '' if none */ - function getSuggestionSnippet(){ + function getSuggestionSnippet() { return ''; } - + /** * Return information about how and from where the results were fetched, - * should be useful for diagnostics and debugging + * should be useful for diagnostics and debugging * * @return String */ function getInfo() { return null; } - + /** * Return a result set of hits on other (multiple) wikis associated with this one * @@ -515,7 +543,7 @@ class SearchResultSet { function getInterwikiResults() { return null; } - + /** * Check if there are results on other wikis * @@ -524,7 +552,6 @@ class SearchResultSet { function hasInterwikiResults() { return $this->getInterwikiResults() != null; } - /** * Fetches next search result, or false. @@ -558,20 +585,27 @@ class SqlSearchResultSet extends SearchResultSet { } function numRows() { + if ( $this->mResultSet === false ) + return false; + return $this->mResultSet->numRows(); } function next() { - if ($this->mResultSet === false ) + if ( $this->mResultSet === false ) return false; $row = $this->mResultSet->fetchObject(); - if ($row === false) + if ( $row === false ) return false; - return new SearchResult($row); + + return SearchResult::newFromRow( $row ); } function free() { + if ( $this->mResultSet === false ) + return false; + $this->mResultSet->free(); } } @@ -580,7 +614,7 @@ class SqlSearchResultSet extends SearchResultSet { * @ingroup Search */ class SearchResultTooMany { - ## Some search engines may bail out if too many matches are found + # # Some search engines may bail out if too many matches are found } @@ -594,32 +628,78 @@ class SearchResult { var $mRevision = null; var $mImage = null; - function __construct( $row ) { - $this->mTitle = Title::makeTitle( $row->page_namespace, $row->page_title ); - if( !is_null($this->mTitle) ){ + /** + * Return a new SearchResult and initializes it with a title. + * + * @param $title Title + * @return SearchResult + */ + public static function newFromTitle( $title ) { + $result = new self(); + $result->initFromTitle( $title ); + return $result; + } + /** + * Return a new SearchResult and initializes it with a row. + * + * @param $row object + * @return SearchResult + */ + public static function newFromRow( $row ) { + $result = new self(); + $result->initFromRow( $row ); + return $result; + } + + public function __construct( $row = null ) { + if ( !is_null( $row ) ) { + // Backwards compatibility with pre-1.17 callers + $this->initFromRow( $row ); + } + } + + /** + * Initialize from a database row. Makes a Title and passes that to + * initFromTitle. + * + * @param $row object + */ + protected function initFromRow( $row ) { + $this->initFromTitle( Title::makeTitle( $row->page_namespace, $row->page_title ) ); + } + + /** + * Initialize from a Title and if possible initializes a corresponding + * Revision and File. + * + * @param $title Title + */ + protected function initFromTitle( $title ) { + $this->mTitle = $title; + if ( !is_null( $this->mTitle ) ) { $this->mRevision = Revision::newFromTitle( $this->mTitle ); - if( $this->mTitle->getNamespace() === NS_FILE ) + if ( $this->mTitle->getNamespace() === NS_FILE ) $this->mImage = wfFindFile( $this->mTitle ); } } - + /** * Check if this is result points to an invalid title * * @return Boolean */ - function isBrokenTitle(){ - if( is_null($this->mTitle) ) + function isBrokenTitle() { + if ( is_null( $this->mTitle ) ) return true; return false; } - + /** * Check if target page is missing, happens when index is out of date - * + * * @return Boolean */ - function isMissingRevision(){ + function isMissingRevision() { return !$this->mRevision && !$this->mImage; } @@ -640,36 +720,36 @@ class SearchResult { /** * Lazy initialization of article text from DB */ - protected function initText(){ - if( !isset($this->mText) ){ - if($this->mRevision != null) + protected function initText() { + if ( !isset( $this->mText ) ) { + if ( $this->mRevision != null ) $this->mText = $this->mRevision->getText(); else // TODO: can we fetch raw wikitext for commons images? $this->mText = ''; - + } } - + /** * @param $terms Array: terms to highlight - * @return String: highlighted text snippet, null (and not '') if not supported + * @return String: highlighted text snippet, null (and not '') if not supported */ - function getTextSnippet($terms){ + function getTextSnippet( $terms ) { global $wgUser, $wgAdvancedSearchHighlighting; $this->initText(); - list($contextlines,$contextchars) = SearchEngine::userHighlightPrefs($wgUser); + list( $contextlines, $contextchars ) = SearchEngine::userHighlightPrefs( $wgUser ); $h = new SearchHighlighter(); - if( $wgAdvancedSearchHighlighting ) + if ( $wgAdvancedSearchHighlighting ) return $h->highlightText( $this->mText, $terms, $contextlines, $contextchars ); else return $h->highlightSimple( $this->mText, $terms, $contextlines, $contextchars ); } - + /** * @param $terms Array: terms to highlight * @return String: highlighted title, '' if not supported */ - function getTitleSnippet($terms){ + function getTitleSnippet( $terms ) { return ''; } @@ -677,46 +757,46 @@ class SearchResult { * @param $terms Array: terms to highlight * @return String: highlighted redirect name (redirect to this page), '' if none or not supported */ - function getRedirectSnippet($terms){ + function getRedirectSnippet( $terms ) { return ''; } /** * @return Title object for the redirect to this page, null if none or not supported */ - function getRedirectTitle(){ + function getRedirectTitle() { return null; } /** * @return string highlighted relevant section name, null if none or not supported */ - function getSectionSnippet(){ + function getSectionSnippet() { return ''; } /** * @return Title object (pagename+fragment) for the section, null if none or not supported */ - function getSectionTitle(){ + function getSectionTitle() { return null; } /** * @return String: timestamp */ - function getTimestamp(){ - if( $this->mRevision ) + function getTimestamp() { + if ( $this->mRevision ) return $this->mRevision->getTimestamp(); - else if( $this->mImage ) + else if ( $this->mImage ) return $this->mImage->getTimestamp(); - return ''; + return ''; } /** * @return Integer: number of words */ - function getWordCount(){ + function getWordCount() { $this->initText(); return str_word_count( $this->mText ); } @@ -724,38 +804,63 @@ class SearchResult { /** * @return Integer: size in bytes */ - function getByteSize(){ + function getByteSize() { $this->initText(); return strlen( $this->mText ); } - + /** * @return Boolean if hit has related articles */ - function hasRelated(){ + function hasRelated() { return false; } - + /** * @return String: interwiki prefix of the title (return iw even if title is broken) */ - function getInterwikiPrefix(){ + function getInterwikiPrefix() { return ''; } } +/** + * A SearchResultSet wrapper for SearchEngine::getNearMatch + */ +class SearchNearMatchResultSet extends SearchResultSet { + private $fetched = false; + /** + * @param $match mixed Title if matched, else null + */ + public function __construct( $match ) { + $this->result = $match; + } + public function hasResult() { + return (bool)$this->result; + } + public function numRows() { + return $this->hasResults() ? 1 : 0; + } + public function next() { + if ( $this->fetched || !$this->result ) { + return false; + } + $this->fetched = true; + return SearchResult::newFromTitle( $this->result ); + } +} /** * Highlight bits of wikitext - * + * * @ingroup Search */ -class SearchHighlighter { +class SearchHighlighter { var $mCleanWikitext = true; - - function SearchHighlighter($cleanupWikitext = true){ + + function __construct( $cleanupWikitext = true ) { $this->mCleanWikitext = $cleanupWikitext; } - + /** * Default implementation of wikitext highlighting * @@ -766,23 +871,23 @@ class SearchHighlighter { * @return String */ public function highlightText( $text, $terms, $contextlines, $contextchars ) { - global $wgLang, $wgContLang; + global $wgContLang; global $wgSearchHighlightBoundaries; $fname = __METHOD__; - - if($text == '') + + if ( $text == '' ) return ''; - + // spli text into text + templates/links/tables $spat = "/(\\{\\{)|(\\[\\[[^\\]:]+:)|(\n\\{\\|)"; // first capture group is for detecting nested templates/links/tables/references $endPatterns = array( 1 => '/(\{\{)|(\}\})/', // template 2 => '/(\[\[)|(\]\])/', // image - 3 => "/(\n\\{\\|)|(\n\\|\\})/"); // table - + 3 => "/(\n\\{\\|)|(\n\\|\\})/" ); // table + // FIXME: this should prolly be a hook or something - if(function_exists('wfCite')){ + if ( function_exists( 'wfCite' ) ) { $spat .= '|(<ref>)'; // references via cite extension $endPatterns[4] = '/(<ref>)|(<\/ref>)/'; } @@ -791,213 +896,215 @@ class SearchHighlighter { $otherExt = array(); // other extracts wfProfileIn( "$fname-split" ); $start = 0; - $textLen = strlen($text); + $textLen = strlen( $text ); $count = 0; // sequence number to maintain ordering - while( $start < $textLen ){ + while ( $start < $textLen ) { // find start of template/image/table - if( preg_match( $spat, $text, $matches, PREG_OFFSET_CAPTURE, $start ) ){ - $epat = ''; - foreach($matches as $key => $val){ - if($key > 0 && $val[1] != -1){ - if($key == 2){ + if ( preg_match( $spat, $text, $matches, PREG_OFFSET_CAPTURE, $start ) ) { + $epat = ''; + foreach ( $matches as $key => $val ) { + if ( $key > 0 && $val[1] != - 1 ) { + if ( $key == 2 ) { // see if this is an image link - $ns = substr($val[0],2,-1); - if( $wgContLang->getNsIndex($ns) != NS_FILE ) + $ns = substr( $val[0], 2, - 1 ); + if ( $wgContLang->getNsIndex( $ns ) != NS_FILE ) break; - + } $epat = $endPatterns[$key]; - $this->splitAndAdd( $textExt, $count, substr( $text, $start, $val[1] - $start ) ); + $this->splitAndAdd( $textExt, $count, substr( $text, $start, $val[1] - $start ) ); $start = $val[1]; break; } } - if( $epat ){ + if ( $epat ) { // find end (and detect any nested elements) - $level = 0; + $level = 0; $offset = $start + 1; $found = false; - while( preg_match( $epat, $text, $endMatches, PREG_OFFSET_CAPTURE, $offset ) ){ - if( array_key_exists(2,$endMatches) ){ + while ( preg_match( $epat, $text, $endMatches, PREG_OFFSET_CAPTURE, $offset ) ) { + if ( array_key_exists( 2, $endMatches ) ) { // found end - if($level == 0){ - $len = strlen($endMatches[2][0]); + if ( $level == 0 ) { + $len = strlen( $endMatches[2][0] ); $off = $endMatches[2][1]; - $this->splitAndAdd( $otherExt, $count, + $this->splitAndAdd( $otherExt, $count, substr( $text, $start, $off + $len - $start ) ); $start = $off + $len; $found = true; break; - } else{ + } else { // end of nested element $level -= 1; } - } else{ + } else { // nested $level += 1; } - $offset = $endMatches[0][1] + strlen($endMatches[0][0]); + $offset = $endMatches[0][1] + strlen( $endMatches[0][0] ); } - if( ! $found ){ + if ( ! $found ) { // couldn't find appropriate closing tag, skip - $this->splitAndAdd( $textExt, $count, substr( $text, $start, strlen($matches[0][0]) ) ); - $start += strlen($matches[0][0]); + $this->splitAndAdd( $textExt, $count, substr( $text, $start, strlen( $matches[0][0] ) ) ); + $start += strlen( $matches[0][0] ); } continue; } } // else: add as text extract - $this->splitAndAdd( $textExt, $count, substr($text,$start) ); + $this->splitAndAdd( $textExt, $count, substr( $text, $start ) ); break; } - + $all = $textExt + $otherExt; // these have disjunct key sets - + wfProfileOut( "$fname-split" ); - + // prepare regexps - foreach( $terms as $index => $term ) { + foreach ( $terms as $index => $term ) { // manually do upper/lowercase stuff for utf-8 since PHP won't do it - if(preg_match('/[\x80-\xff]/', $term) ){ - $terms[$index] = preg_replace_callback('/./us',array($this,'caseCallback'),$terms[$index]); + if ( preg_match( '/[\x80-\xff]/', $term ) ) { + $terms[$index] = preg_replace_callback( '/./us', array( $this, 'caseCallback' ), $terms[$index] ); } else { $terms[$index] = $term; } } $anyterm = implode( '|', $terms ); - $phrase = implode("$wgSearchHighlightBoundaries+", $terms ); + $phrase = implode( "$wgSearchHighlightBoundaries+", $terms ); // FIXME: a hack to scale contextchars, a correct solution // would be to have contextchars actually be char and not byte // length, and do proper utf-8 substrings and lengths everywhere, // but PHP is making that very hard and unclean to implement :( - $scale = strlen($anyterm) / mb_strlen($anyterm); + $scale = strlen( $anyterm ) / mb_strlen( $anyterm ); $contextchars = intval( $contextchars * $scale ); - + $patPre = "(^|$wgSearchHighlightBoundaries)"; - $patPost = "($wgSearchHighlightBoundaries|$)"; - - $pat1 = "/(".$phrase.")/ui"; - $pat2 = "/$patPre(".$anyterm.")$patPost/ui"; - + $patPost = "($wgSearchHighlightBoundaries|$)"; + + $pat1 = "/(" . $phrase . ")/ui"; + $pat2 = "/$patPre(" . $anyterm . ")$patPost/ui"; + wfProfileIn( "$fname-extract" ); - + $left = $contextlines; $snippets = array(); - $offsets = array(); - + $offsets = array(); + // show beginning only if it contains all words $first = 0; $firstText = ''; - foreach($textExt as $index => $line){ - if(strlen($line)>0 && $line[0] != ';' && $line[0] != ':'){ + foreach ( $textExt as $index => $line ) { + if ( strlen( $line ) > 0 && $line[0] != ';' && $line[0] != ':' ) { $firstText = $this->extract( $line, 0, $contextchars * $contextlines ); $first = $index; break; } } - if( $firstText ){ + if ( $firstText ) { $succ = true; // check if first text contains all terms - foreach($terms as $term){ - if( ! preg_match("/$patPre".$term."$patPost/ui", $firstText) ){ + foreach ( $terms as $term ) { + if ( ! preg_match( "/$patPre" . $term . "$patPost/ui", $firstText ) ) { $succ = false; break; } } - if( $succ ){ + if ( $succ ) { $snippets[$first] = $firstText; - $offsets[$first] = 0; + $offsets[$first] = 0; } } - if( ! $snippets ) { - // match whole query on text - $this->process($pat1, $textExt, $left, $contextchars, $snippets, $offsets); + if ( ! $snippets ) { + // match whole query on text + $this->process( $pat1, $textExt, $left, $contextchars, $snippets, $offsets ); // match whole query on templates/tables/images - $this->process($pat1, $otherExt, $left, $contextchars, $snippets, $offsets); + $this->process( $pat1, $otherExt, $left, $contextchars, $snippets, $offsets ); // match any words on text - $this->process($pat2, $textExt, $left, $contextchars, $snippets, $offsets); + $this->process( $pat2, $textExt, $left, $contextchars, $snippets, $offsets ); // match any words on templates/tables/images - $this->process($pat2, $otherExt, $left, $contextchars, $snippets, $offsets); - - ksort($snippets); + $this->process( $pat2, $otherExt, $left, $contextchars, $snippets, $offsets ); + + ksort( $snippets ); } - + // add extra chars to each snippet to make snippets constant size - $extended = array(); - if( count( $snippets ) == 0){ + $extended = array(); + if ( count( $snippets ) == 0 ) { // couldn't find the target words, just show beginning of article - $targetchars = $contextchars * $contextlines; - $snippets[$first] = ''; - $offsets[$first] = 0; - } else{ - // if begin of the article contains the whole phrase, show only that !! - if( array_key_exists($first,$snippets) && preg_match($pat1,$snippets[$first]) - && $offsets[$first] < $contextchars * 2 ){ - $snippets = array ($first => $snippets[$first]); + if ( array_key_exists( $first, $all ) ) { + $targetchars = $contextchars * $contextlines; + $snippets[$first] = ''; + $offsets[$first] = 0; } - + } else { + // if begin of the article contains the whole phrase, show only that !! + if ( array_key_exists( $first, $snippets ) && preg_match( $pat1, $snippets[$first] ) + && $offsets[$first] < $contextchars * 2 ) { + $snippets = array ( $first => $snippets[$first] ); + } + // calc by how much to extend existing snippets - $targetchars = intval( ($contextchars * $contextlines) / count ( $snippets ) ); - } + $targetchars = intval( ( $contextchars * $contextlines ) / count ( $snippets ) ); + } - foreach($snippets as $index => $line){ + foreach ( $snippets as $index => $line ) { $extended[$index] = $line; - $len = strlen($line); - if( $len < $targetchars - 20 ){ + $len = strlen( $line ); + if ( $len < $targetchars - 20 ) { // complete this line - if($len < strlen( $all[$index] )){ - $extended[$index] = $this->extract( $all[$index], $offsets[$index], $offsets[$index]+$targetchars, $offsets[$index]); + if ( $len < strlen( $all[$index] ) ) { + $extended[$index] = $this->extract( $all[$index], $offsets[$index], $offsets[$index] + $targetchars, $offsets[$index] ); $len = strlen( $extended[$index] ); } - + // add more lines $add = $index + 1; - while( $len < $targetchars - 20 - && array_key_exists($add,$all) - && !array_key_exists($add,$snippets) ){ + while ( $len < $targetchars - 20 + && array_key_exists( $add, $all ) + && !array_key_exists( $add, $snippets ) ) { $offsets[$add] = 0; - $tt = "\n".$this->extract( $all[$add], 0, $targetchars - $len, $offsets[$add] ); + $tt = "\n" . $this->extract( $all[$add], 0, $targetchars - $len, $offsets[$add] ); $extended[$add] = $tt; $len += strlen( $tt ); - $add++; + $add++; } - } + } } - - //$snippets = array_map('htmlspecialchars', $extended); + + // $snippets = array_map('htmlspecialchars', $extended); $snippets = $extended; - $last = -1; + $last = - 1; $extract = ''; - foreach($snippets as $index => $line){ - if($last == -1) + foreach ( $snippets as $index => $line ) { + if ( $last == - 1 ) $extract .= $line; // first line - elseif($last+1 == $index && $offsets[$last]+strlen($snippets[$last]) >= strlen($all[$last])) - $extract .= " ".$line; // continous lines + elseif ( $last + 1 == $index && $offsets[$last] + strlen( $snippets[$last] ) >= strlen( $all[$last] ) ) + $extract .= " " . $line; // continous lines else $extract .= '<b> ... </b>' . $line; $last = $index; } - if( $extract ) + if ( $extract ) $extract .= '<b> ... </b>'; - + $processed = array(); - foreach($terms as $term){ - if( ! isset($processed[$term]) ){ - $pat3 = "/$patPre(".$term.")$patPost/ui"; // highlight word + foreach ( $terms as $term ) { + if ( ! isset( $processed[$term] ) ) { + $pat3 = "/$patPre(" . $term . ")$patPost/ui"; // highlight word $extract = preg_replace( $pat3, "\\1<span class='searchmatch'>\\2</span>\\3", $extract ); $processed[$term] = true; } } - + wfProfileOut( "$fname-extract" ); - + return $extract; } - + /** * Split text into lines and add it to extracts array * @@ -1005,28 +1112,28 @@ class SearchHighlighter { * @param $count Integer * @param $text String */ - function splitAndAdd(&$extracts, &$count, $text){ - $split = explode( "\n", $this->mCleanWikitext? $this->removeWiki($text) : $text ); - foreach($split as $line){ - $tt = trim($line); - if( $tt ) + function splitAndAdd( &$extracts, &$count, $text ) { + $split = explode( "\n", $this->mCleanWikitext ? $this->removeWiki( $text ) : $text ); + foreach ( $split as $line ) { + $tt = trim( $line ); + if ( $tt ) $extracts[$count++] = $tt; } } - + /** * Do manual case conversion for non-ascii chars * * @param $matches Array */ - function caseCallback($matches){ + function caseCallback( $matches ) { global $wgContLang; - if( strlen($matches[0]) > 1 ){ - return '['.$wgContLang->lc($matches[0]).$wgContLang->uc($matches[0]).']'; + if ( strlen( $matches[0] ) > 1 ) { + return '[' . $wgContLang->lc( $matches[0] ) . $wgContLang->uc( $matches[0] ) . ']'; } else return $matches[0]; } - + /** * Extract part of the text from start to end, but by * not chopping up words @@ -1035,29 +1142,27 @@ class SearchHighlighter { * @param $end Integer * @param $posStart Integer: (out) actual start position * @param $posEnd Integer: (out) actual end position - * @return String + * @return String */ - function extract($text, $start, $end, &$posStart = null, &$posEnd = null ){ - global $wgContLang; - - if( $start != 0) + function extract( $text, $start, $end, &$posStart = null, &$posEnd = null ) { + if ( $start != 0 ) $start = $this->position( $text, $start, 1 ); - if( $end >= strlen($text) ) - $end = strlen($text); + if ( $end >= strlen( $text ) ) + $end = strlen( $text ); else $end = $this->position( $text, $end ); - - if(!is_null($posStart)) + + if ( !is_null( $posStart ) ) $posStart = $start; - if(!is_null($posEnd)) + if ( !is_null( $posEnd ) ) $posEnd = $end; - - if($end > $start) - return substr($text, $start, $end-$start); + + if ( $end > $start ) + return substr( $text, $start, $end - $start ); else return ''; - } - + } + /** * Find a nonletter near a point (index) in the text * @@ -1066,117 +1171,117 @@ class SearchHighlighter { * @param $offset Integer: offset to found index * @return Integer: nearest nonletter index, or beginning of utf8 char if none */ - function position($text, $point, $offset=0 ){ + function position( $text, $point, $offset = 0 ) { $tolerance = 10; $s = max( 0, $point - $tolerance ); - $l = min( strlen($text), $point + $tolerance ) - $s; + $l = min( strlen( $text ), $point + $tolerance ) - $s; $m = array(); - if( preg_match('/[ ,.!?~!@#$%^&*\(\)+=\-\\\|\[\]"\'<>]/', substr($text,$s,$l), $m, PREG_OFFSET_CAPTURE ) ){ + if ( preg_match( '/[ ,.!?~!@#$%^&*\(\)+=\-\\\|\[\]"\'<>]/', substr( $text, $s, $l ), $m, PREG_OFFSET_CAPTURE ) ) { return $m[0][1] + $s + $offset; - } else{ + } else { // check if point is on a valid first UTF8 char $char = ord( $text[$point] ); - while( $char >= 0x80 && $char < 0xc0 ) { + while ( $char >= 0x80 && $char < 0xc0 ) { // skip trailing bytes $point++; - if($point >= strlen($text)) - return strlen($text); + if ( $point >= strlen( $text ) ) + return strlen( $text ); $char = ord( $text[$point] ); } return $point; - + } } - + /** * Search extracts for a pattern, and return snippets * * @param $pattern String: regexp for matching lines - * @param $extracts Array: extracts to search + * @param $extracts Array: extracts to search * @param $linesleft Integer: number of extracts to make * @param $contextchars Integer: length of snippet * @param $out Array: map for highlighted snippets * @param $offsets Array: map of starting points of snippets * @protected */ - function process( $pattern, $extracts, &$linesleft, &$contextchars, &$out, &$offsets ){ - if($linesleft == 0) + function process( $pattern, $extracts, &$linesleft, &$contextchars, &$out, &$offsets ) { + if ( $linesleft == 0 ) return; // nothing to do - foreach($extracts as $index => $line){ - if( array_key_exists($index,$out) ) + foreach ( $extracts as $index => $line ) { + if ( array_key_exists( $index, $out ) ) continue; // this line already highlighted - + $m = array(); if ( !preg_match( $pattern, $line, $m, PREG_OFFSET_CAPTURE ) ) continue; - + $offset = $m[0][1]; - $len = strlen($m[0][0]); - if($offset + $len < $contextchars) - $begin = 0; - elseif( $len > $contextchars) + $len = strlen( $m[0][0] ); + if ( $offset + $len < $contextchars ) + $begin = 0; + elseif ( $len > $contextchars ) $begin = $offset; else - $begin = $offset + intval( ($len - $contextchars) / 2 ); - + $begin = $offset + intval( ( $len - $contextchars ) / 2 ); + $end = $begin + $contextchars; - + $posBegin = $begin; // basic snippet from this line - $out[$index] = $this->extract($line,$begin,$end,$posBegin); + $out[$index] = $this->extract( $line, $begin, $end, $posBegin ); $offsets[$index] = $posBegin; - $linesleft--; - if($linesleft == 0) + $linesleft--; + if ( $linesleft == 0 ) return; } } - - /** + + /** * Basic wikitext removal * @protected */ - function removeWiki($text) { + function removeWiki( $text ) { $fname = __METHOD__; wfProfileIn( $fname ); - - //$text = preg_replace("/'{2,5}/", "", $text); - //$text = preg_replace("/\[[a-z]+:\/\/[^ ]+ ([^]]+)\]/", "\\2", $text); - //$text = preg_replace("/\[\[([^]|]+)\]\]/", "\\1", $text); - //$text = preg_replace("/\[\[([^]]+\|)?([^|]]+)\]\]/", "\\2", $text); - //$text = preg_replace("/\\{\\|(.*?)\\|\\}/", "", $text); - //$text = preg_replace("/\\[\\[[A-Za-z_-]+:([^|]+?)\\]\\]/", "", $text); - $text = preg_replace("/\\{\\{([^|]+?)\\}\\}/", "", $text); - $text = preg_replace("/\\{\\{([^|]+\\|)(.*?)\\}\\}/", "\\2", $text); - $text = preg_replace("/\\[\\[([^|]+?)\\]\\]/", "\\1", $text); - $text = preg_replace_callback("/\\[\\[([^|]+\\|)(.*?)\\]\\]/", array($this,'linkReplace'), $text); - //$text = preg_replace("/\\[\\[([^|]+\\|)(.*?)\\]\\]/", "\\2", $text); - $text = preg_replace("/<\/?[^>]+>/", "", $text); - $text = preg_replace("/'''''/", "", $text); - $text = preg_replace("/('''|<\/?[iIuUbB]>)/", "", $text); - $text = preg_replace("/''/", "", $text); - + + // $text = preg_replace("/'{2,5}/", "", $text); + // $text = preg_replace("/\[[a-z]+:\/\/[^ ]+ ([^]]+)\]/", "\\2", $text); + // $text = preg_replace("/\[\[([^]|]+)\]\]/", "\\1", $text); + // $text = preg_replace("/\[\[([^]]+\|)?([^|]]+)\]\]/", "\\2", $text); + // $text = preg_replace("/\\{\\|(.*?)\\|\\}/", "", $text); + // $text = preg_replace("/\\[\\[[A-Za-z_-]+:([^|]+?)\\]\\]/", "", $text); + $text = preg_replace( "/\\{\\{([^|]+?)\\}\\}/", "", $text ); + $text = preg_replace( "/\\{\\{([^|]+\\|)(.*?)\\}\\}/", "\\2", $text ); + $text = preg_replace( "/\\[\\[([^|]+?)\\]\\]/", "\\1", $text ); + $text = preg_replace_callback( "/\\[\\[([^|]+\\|)(.*?)\\]\\]/", array( $this, 'linkReplace' ), $text ); + // $text = preg_replace("/\\[\\[([^|]+\\|)(.*?)\\]\\]/", "\\2", $text); + $text = preg_replace( "/<\/?[^>]+>/", "", $text ); + $text = preg_replace( "/'''''/", "", $text ); + $text = preg_replace( "/('''|<\/?[iIuUbB]>)/", "", $text ); + $text = preg_replace( "/''/", "", $text ); + wfProfileOut( $fname ); return $text; } - + /** * callback to replace [[target|caption]] kind of links, if * the target is category or image, leave it * * @param $matches Array */ - function linkReplace($matches){ - $colon = strpos( $matches[1], ':' ); - if( $colon === false ) + function linkReplace( $matches ) { + $colon = strpos( $matches[1], ':' ); + if ( $colon === false ) return $matches[2]; // replace with caption global $wgContLang; $ns = substr( $matches[1], 0, $colon ); - $index = $wgContLang->getNsIndex($ns); - if( $index !== false && ($index == NS_FILE || $index == NS_CATEGORY) ) - return $matches[0]; // return the whole thing + $index = $wgContLang->getNsIndex( $ns ); + if ( $index !== false && ( $index == NS_FILE || $index == NS_CATEGORY ) ) + return $matches[0]; // return the whole thing else return $matches[2]; - + } /** @@ -1190,11 +1295,11 @@ class SearchHighlighter { * @return String */ public function highlightSimple( $text, $terms, $contextlines, $contextchars ) { - global $wgLang, $wgContLang; + global $wgContLang; $fname = __METHOD__; $lines = explode( "\n", $text ); - + $terms = implode( '|', $terms ); $max = intval( $contextchars ) + 1; $pat1 = "/(.*)($terms)(.{0,$max})/i"; @@ -1213,7 +1318,7 @@ class SearchHighlighter { continue; } --$contextlines; - $pre = $wgContLang->truncate( $m[1], -$contextchars ); + $pre = $wgContLang->truncate( $m[1], - $contextchars ); if ( count( $m ) < 3 ) { $post = ''; @@ -1231,10 +1336,10 @@ class SearchHighlighter { $extract .= "${line}\n"; } wfProfileOut( "$fname-extract" ); - + return $extract; } - + } /** diff --git a/includes/search/SearchIBM_DB2.php b/includes/search/SearchIBM_DB2.php index d7587186..8cedd6f2 100644 --- a/includes/search/SearchIBM_DB2.php +++ b/includes/search/SearchIBM_DB2.php @@ -1,23 +1,25 @@ <?php -# Copyright (C) 2004 Brion Vibber <brion@pobox.com> -# http://www.mediawiki.org/ -# -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program; if not, write to the Free Software Foundation, Inc., -# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. -# http://www.gnu.org/copyleft/gpl.html - /** + * IBM DB2 search engine + * + * Copyright © 2004 Brion Vibber <brion@pobox.com> + * http://www.mediawiki.org/ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * http://www.gnu.org/copyleft/gpl.html + * * @file * @ingroup Search */ @@ -27,8 +29,13 @@ * @ingroup Search */ class SearchIBM_DB2 extends SearchEngine { + + /** + * Creates an instance of this class + * @param $db DatabaseIbm_db2: database object + */ function __construct($db) { - $this->db = $db; + parent::__construct( $db ); } /** @@ -102,8 +109,8 @@ class SearchIBM_DB2 extends SearchEngine { /** * Construct the full SQL query to do the search. * The guts shoulds be constructed in queryMain() - * @param string $filteredTerm String - * @param bool $fulltext Boolean + * @param $filteredTerm String + * @param $fulltext Boolean */ function getQuery( $filteredTerm, $fulltext ) { return $this->queryLimit($this->queryMain($filteredTerm, $fulltext) . ' ' . @@ -125,8 +132,8 @@ class SearchIBM_DB2 extends SearchEngine { /** * Get the base part of the search query. * - * @param string $filteredTerm String - * @param bool $fulltext Boolean + * @param $filteredTerm String + * @param $fulltext Boolean * @return String */ function queryMain( $filteredTerm, $fulltext ) { diff --git a/includes/search/SearchMssql.php b/includes/search/SearchMssql.php new file mode 100644 index 00000000..8b850fae --- /dev/null +++ b/includes/search/SearchMssql.php @@ -0,0 +1,254 @@ +<?php +/** + * Mssql search engine + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * http://www.gnu.org/copyleft/gpl.html + * + * @file + * @ingroup Search + */ + +/** + * Search engine hook base class for Mssql (ConText). + * @ingroup Search + */ +class SearchMssql extends SearchEngine { + + /** + * Creates an instance of this class + * @param $db DatabaseMssql: database object + */ + function __construct( $db ) { + parent::__construct( $db ); + } + + /** + * Perform a full text search query and return a result set. + * + * @param $term String: raw search term + * @return MssqlSearchResultSet + * @access public + */ + function searchText( $term ) { + $resultSet = $this->db->resultObject( $this->db->query( $this->getQuery( $this->filter( $term ), true ) ) ); + return new MssqlSearchResultSet( $resultSet, $this->searchTerms ); + } + + /** + * Perform a title-only search query and return a result set. + * + * @param $term String: raw search term + * @return MssqlSearchResultSet + * @access public + */ + function searchTitle( $term ) { + $resultSet = $this->db->resultObject( $this->db->query( $this->getQuery( $this->filter( $term ), false ) ) ); + return new MssqlSearchResultSet( $resultSet, $this->searchTerms ); + } + + + /** + * Return a partial WHERE clause to exclude redirects, if so set + * + * @return String + * @private + */ + function queryRedirect() { + if ( $this->showRedirects ) { + return ''; + } else { + return 'AND page_is_redirect=0'; + } + } + + /** + * Return a partial WHERE clause to limit the search to the given namespaces + * + * @return String + * @private + */ + function queryNamespaces() { + $namespaces = implode( ',', $this->namespaces ); + if ( $namespaces == '' ) { + $namespaces = '0'; + } + return 'AND page_namespace IN (' . $namespaces . ')'; + } + + /** + * Return a LIMIT clause to limit results on the query. + * + * @return String + * @private + */ + function queryLimit( $sql ) { + return $this->db->limitResult( $sql, $this->limit, $this->offset ); + } + + /** + * Does not do anything for generic search engine + * subclasses may define this though + * + * @return String + * @private + */ + function queryRanking( $filteredTerm, $fulltext ) { + return ' ORDER BY ftindex.[RANK] DESC'; // return ' ORDER BY score(1)'; + } + + /** + * Construct the full SQL query to do the search. + * The guts shoulds be constructed in queryMain() + * + * @param $filteredTerm String + * @param $fulltext Boolean + * @private + */ + function getQuery( $filteredTerm, $fulltext ) { + return $this->queryLimit( $this->queryMain( $filteredTerm, $fulltext ) . ' ' . + $this->queryRedirect() . ' ' . + $this->queryNamespaces() . ' ' . + $this->queryRanking( $filteredTerm, $fulltext ) . ' ' ); + } + + + /** + * Picks which field to index on, depending on what type of query. + * + * @param $fulltext Boolean + * @return string + */ + function getIndexField( $fulltext ) { + return $fulltext ? 'si_text' : 'si_title'; + } + + /** + * Get the base part of the search query. + * + * @param $filteredTerm String + * @param $fulltext Boolean + * @return String + * @private + */ + function queryMain( $filteredTerm, $fulltext ) { + $match = $this->parseQuery( $filteredTerm, $fulltext ); + $page = $this->db->tableName( 'page' ); + $searchindex = $this->db->tableName( 'searchindex' ); + + return 'SELECT page_id, page_namespace, page_title, ftindex.[RANK]' . + "FROM $page,FREETEXTTABLE($searchindex , $match, LANGUAGE 'English') as ftindex " . + 'WHERE page_id=ftindex.[KEY] '; + } + + /** @todo document */ + function parseQuery( $filteredText, $fulltext ) { + global $wgContLang; + $lc = SearchEngine::legalSearchChars(); + $this->searchTerms = array(); + + # FIXME: This doesn't handle parenthetical expressions. + $m = array(); + $q = array(); + + if ( preg_match_all( '/([-+<>~]?)(([' . $lc . ']+)(\*?)|"[^"]*")/', + $filteredText, $m, PREG_SET_ORDER ) ) { + foreach ( $m as $terms ) { + $q[] = $terms[1] . $wgContLang->normalizeForSearch( $terms[2] ); + + if ( !empty( $terms[3] ) ) { + $regexp = preg_quote( $terms[3], '/' ); + if ( $terms[4] ) + $regexp .= "[0-9A-Za-z_]+"; + } else { + $regexp = preg_quote( str_replace( '"', '', $terms[2] ), '/' ); + } + $this->searchTerms[] = $regexp; + } + } + + $searchon = $this->db->strencode( join( ',', $q ) ); + $field = $this->getIndexField( $fulltext ); + return "$field, '$searchon'"; + } + + /** + * Create or update the search index record for the given page. + * Title and text should be pre-processed. + * + * @param $id Integer + * @param $title String + * @param $text String + */ + function update( $id, $title, $text ) { + // We store the column data as UTF-8 byte order marked binary stream + // because we are invoking the plain text IFilter on it so that, and we want it + // to properly decode the stream as UTF-8. SQL doesn't support UTF8 as a data type + // but the indexer will correctly handle it by this method. Since all we are doing + // is passing this data to the indexer and never retrieving it via PHP, this will save space + $table = $this->db->tableName( 'searchindex' ); + $utf8bom = '0xEFBBBF'; + $si_title = $utf8bom . bin2hex( $title ); + $si_text = $utf8bom . bin2hex( $text ); + $sql = "DELETE FROM $table WHERE si_page = $id;"; + $sql .= "INSERT INTO $table (si_page, si_title, si_text) VALUES ($id, $si_title, $si_text)"; + return $this->db->query( $sql, 'SearchMssql::update' ); + } + + /** + * Update a search index record's title only. + * Title should be pre-processed. + * + * @param $id Integer + * @param $title String + */ + function updateTitle( $id, $title ) { + $table = $this->db->tableName( 'searchindex' ); + + // see update for why we are using the utf8bom + $utf8bom = '0xEFBBBF'; + $si_title = $utf8bom . bin2hex( $title ); + $sql = "DELETE FROM $table WHERE si_page = $id;"; + $sql .= "INSERT INTO $table (si_page, si_title, si_text) VALUES ($id, $si_title, 0x00)"; + return $this->db->query( $sql, 'SearchMssql::updateTitle' ); + } +} + +/** + * @ingroup Search + */ +class MssqlSearchResultSet extends SearchResultSet { + function __construct( $resultSet, $terms ) { + $this->mResultSet = $resultSet; + $this->mTerms = $terms; + } + + function termMatches() { + return $this->mTerms; + } + + function numRows() { + return $this->mResultSet->numRows(); + } + + function next() { + $row = $this->mResultSet->fetchObject(); + if ( $row === false ) + return false; + return new SearchResult( $row ); + } +} + + diff --git a/includes/search/SearchMySQL.php b/includes/search/SearchMySQL.php index 0c238be8..b92682ad 100644 --- a/includes/search/SearchMySQL.php +++ b/includes/search/SearchMySQL.php @@ -1,23 +1,25 @@ <?php -# Copyright (C) 2004 Brion Vibber <brion@pobox.com> -# http://www.mediawiki.org/ -# -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program; if not, write to the Free Software Foundation, Inc., -# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. -# http://www.gnu.org/copyleft/gpl.html - /** + * MySQL search engine + * + * Copyright (C) 2004 Brion Vibber <brion@pobox.com> + * http://www.mediawiki.org/ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * http://www.gnu.org/copyleft/gpl.html + * * @file * @ingroup Search */ @@ -30,9 +32,12 @@ class SearchMySQL extends SearchEngine { var $strictMatching = true; static $mMinSearchLength; - /** @todo document */ + /** + * Creates an instance of this class + * @param $db DatabaseMysql: database object + */ function __construct( $db ) { - $this->db = $db; + parent::__construct( $db ); } /** @@ -325,8 +330,7 @@ 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 @@ -401,7 +405,7 @@ class SearchMySQL extends SearchEngine { * @ingroup Search */ class MySQLSearchResultSet extends SqlSearchResultSet { - function MySQLSearchResultSet( $resultSet, $terms, $totalHits=null ) { + function __construct( $resultSet, $terms, $totalHits=null ) { parent::__construct( $resultSet, $terms ); $this->mTotalHits = $totalHits; } @@ -409,4 +413,4 @@ class MySQLSearchResultSet extends SqlSearchResultSet { function getTotalHits() { return $this->mTotalHits; } -}
\ No newline at end of file +} diff --git a/includes/search/SearchMySQL4.php b/includes/search/SearchMySQL4.php deleted file mode 100644 index 3e2bb2d1..00000000 --- a/includes/search/SearchMySQL4.php +++ /dev/null @@ -1,34 +0,0 @@ -<?php -# Copyright (C) 2004 Brion Vibber <brion@pobox.com> -# http://www.mediawiki.org/ -# -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program; if not, write to the Free Software Foundation, Inc., -# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. -# http://www.gnu.org/copyleft/gpl.html - -/** - * @file - * @ingroup Search - */ - -/** - * Search engine hook for MySQL 4+ - * This class retained for backwards compatibility... - * The meat's been moved to SearchMySQL, since the 3.x variety is gone. - * @ingroup Search - * @deprecated - */ -class SearchMySQL4 extends SearchMySQL { - /* whee */ -} diff --git a/includes/search/SearchOracle.php b/includes/search/SearchOracle.php index e4c5deee..15c386ce 100644 --- a/includes/search/SearchOracle.php +++ b/includes/search/SearchOracle.php @@ -1,23 +1,25 @@ <?php -# Copyright (C) 2004 Brion Vibber <brion@pobox.com> -# http://www.mediawiki.org/ -# -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program; if not, write to the Free Software Foundation, Inc., -# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. -# http://www.gnu.org/copyleft/gpl.html - /** + * Oracle search engine + * + * Copyright © 2004 Brion Vibber <brion@pobox.com> + * http://www.mediawiki.org/ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * http://www.gnu.org/copyleft/gpl.html + * * @file * @ingroup Search */ @@ -54,9 +56,13 @@ class SearchOracle extends SearchEngine { 'TRSYN' => 1, 'TT' => 1, 'WITHIN' => 1); - + + /** + * Creates an instance of this class + * @param $db DatabasePostgres: database object + */ function __construct($db) { - $this->db = $db; + parent::__construct( $db ); } /** @@ -240,16 +246,24 @@ class SearchOracle extends SearchEngine { 'si_title' => $title, 'si_text' => $text ), 'SearchOracle::update' ); - $dbw->query("CALL ctx_ddl.sync_index('si_text_idx')"); - $dbw->query("CALL ctx_ddl.sync_index('si_title_idx')"); + + // Sync the index + // We need to specify the DB name (i.e. user/schema) here so that + // it can work from the installer, where + // ALTER SESSION SET CURRENT_SCHEMA = ... + // was used. + $dbw->query( "CALL ctx_ddl.sync_index(" . + $dbw->addQuotes( $dbw->getDBname() . '.' . trim( $dbw->tableName( 'si_text_idx' ), '"' ) ) . ")" ); + $dbw->query( "CALL ctx_ddl.sync_index(" . + $dbw->addQuotes( $dbw->getDBname() . '.' . trim( $dbw->tableName( 'si_title_idx' ), '"' ) ) . ")" ); } /** * Update a search index record's title only. * Title should be pre-processed. * - * @param int $id - * @param string $title + * @param $id Integer + * @param $title String */ function updateTitle($id, $title) { $dbw = wfGetDB(DB_MASTER); diff --git a/includes/search/SearchPostgres.php b/includes/search/SearchPostgres.php index 0006fa82..9d6d1539 100644 --- a/includes/search/SearchPostgres.php +++ b/includes/search/SearchPostgres.php @@ -1,23 +1,25 @@ <?php -# Copyright (C) 2006-2007 Greg Sabino Mullane <greg@turnstep.com> -# http://www.mediawiki.org/ -# -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program; if not, write to the Free Software Foundation, Inc., -# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. -# http://www.gnu.org/copyleft/gpl.html - /** + * PostgreSQL search engine + * + * Copyright © 2006-2007 Greg Sabino Mullane <greg@turnstep.com> + * http://www.mediawiki.org/ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * http://www.gnu.org/copyleft/gpl.html + * * @file * @ingroup Search */ @@ -27,9 +29,12 @@ * @ingroup Search */ class SearchPostgres extends SearchEngine { - + /** + * Creates an instance of this class + * @param $db DatabaseSqlite: database object + */ function __construct( $db ) { - $this->db = $db; + parent::__construct( $db ); } /** @@ -129,24 +134,21 @@ class SearchPostgres extends SearchEngine { /** * Construct the full SQL query to do the search. - * @param $filteredTerm String + * @param $term String * @param $fulltext String + * @param $colname */ function searchQuery( $term, $fulltext, $colname ) { - global $wgDBversion; + $postgresVersion = $this->db->getServerVersion(); - if ( !isset( $wgDBversion ) ) { - $this->db->getServerVersion(); - $wgDBversion = $this->db->numeric_version; - } - $prefix = $wgDBversion < 8.3 ? "'default'," : ''; + $prefix = $postgresVersion < 8.3 ? "'default'," : ''; # Get the SQL fragment for the given term $searchstring = $this->parseQuery( $term ); ## We need a separate query here so gin does not complain about empty searches $SQL = "SELECT to_tsquery($prefix $searchstring)"; - $res = $this->db->doQuery($SQL); + $res = $this->db->query($SQL); if (!$res) { ## TODO: Better output (example to catch: one 'two) die ("Sorry, that was not a valid search string. Please go back and try again"); @@ -166,8 +168,8 @@ class SearchPostgres extends SearchEngine { } } - $rankscore = $wgDBversion > 8.2 ? 5 : 1; - $rank = $wgDBversion < 8.3 ? 'rank' : 'ts_rank'; + $rankscore = $postgresVersion > 8.2 ? 5 : 1; + $rank = $postgresVersion < 8.3 ? 'rank' : 'ts_rank'; $query = "SELECT page_id, page_namespace, page_title, ". "$rank($fulltext, to_tsquery($prefix $searchstring), $rankscore) AS score ". "FROM page p, revision r, pagecontent c WHERE p.page_latest = r.rev_id " . @@ -204,7 +206,7 @@ class SearchPostgres extends SearchEngine { $SQL = "UPDATE pagecontent SET textvector = NULL WHERE old_id IN ". "(SELECT rev_text_id FROM revision WHERE rev_page = " . intval( $pageid ) . " ORDER BY rev_text_id DESC OFFSET 1)"; - $this->db->doQuery($SQL); + $this->db->query($SQL); return true; } diff --git a/includes/search/SearchSqlite.php b/includes/search/SearchSqlite.php index fb55efec..6accc31b 100644 --- a/includes/search/SearchSqlite.php +++ b/includes/search/SearchSqlite.php @@ -1,22 +1,22 @@ <?php -# SQLite search backend, based upon SearchMysql -# -# This program is free software; you can redistribute it and/or modify -# it under the terms of the GNU General Public License as published by -# the Free Software Foundation; either version 2 of the License, or -# (at your option) any later version. -# -# This program is distributed in the hope that it will be useful, -# but WITHOUT ANY WARRANTY; without even the implied warranty of -# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the -# GNU General Public License for more details. -# -# You should have received a copy of the GNU General Public License along -# with this program; if not, write to the Free Software Foundation, Inc., -# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. -# http://www.gnu.org/copyleft/gpl.html - /** + * SQLite search backend, based upon SearchMysql + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. + * http://www.gnu.org/copyleft/gpl.html + * * @file * @ingroup Search */ @@ -26,15 +26,12 @@ * @ingroup Search */ class SearchSqlite extends SearchEngine { - // Cached because SearchUpdate keeps recreating our class - private static $fulltextSupported = null; - /** * Creates an instance of this class * @param $db DatabaseSqlite: database object */ function __construct( $db ) { - $this->db = $db; + parent::__construct( $db ); } /** @@ -42,18 +39,11 @@ class SearchSqlite extends SearchEngine { * @return Boolean */ function fulltextSearchSupported() { - if ( self::$fulltextSupported === null ) { - self::$fulltextSupported = $this->db->selectField( - 'updatelog', - 'ul_key', - array( 'ul_key' => 'fts3' ), - __METHOD__ ) !== false; - } - return self::$fulltextSupported; + return $this->db->checkForEnabledSearch(); } - /** - * Parse the user's query and transform it into an SQL fragment which will + /** + * Parse the user's query and transform it into an SQL fragment which will * become part of a WHERE clause */ function parseQuery( $filteredText, $fulltext ) { @@ -67,7 +57,7 @@ class SearchSqlite extends SearchEngine { $filteredText, $m, PREG_SET_ORDER ) ) { foreach( $m as $bits ) { @list( /* all */, $modifier, $term, $nonQuoted, $wildcard ) = $bits; - + if( $nonQuoted != '' ) { $term = $nonQuoted; $quote = ''; @@ -86,7 +76,7 @@ class SearchSqlite extends SearchEngine { } else { $variants = array( $term ); } - + // The low-level search index does some processing on input to work // around problems with minimum lengths and encoding in MySQL's // fulltext engine. @@ -94,12 +84,12 @@ class SearchSqlite extends SearchEngine { $strippedVariants = array_map( array( $wgContLang, 'normalizeForSearch' ), $variants ); - + // Some languages such as Chinese force all variants to a canonical // form when stripping to the low-level search index, so to be sure // let's check our variants list for unique items after stripping. $strippedVariants = array_unique( $strippedVariants ); - + $searchon .= $modifier; if( count( $strippedVariants) > 1 ) $searchon .= '('; @@ -114,7 +104,7 @@ class SearchSqlite extends SearchEngine { } if( count( $strippedVariants) > 1 ) $searchon .= ')'; - + // Match individual terms or quoted phrase in result highlighting... // Note that variants will be introduced in a later stage for highlighting! $regexp = $this->regexTerm( $term, $wildcard ); @@ -129,10 +119,10 @@ class SearchSqlite extends SearchEngine { $field = $this->getIndexField( $fulltext ); return " $field MATCH '$searchon' "; } - + function regexTerm( $string, $wildcard ) { global $wgContLang; - + $regex = preg_quote( $string, '/' ); if( $wgContLang->hasWordBreaks() ) { if( $wildcard ) { @@ -172,7 +162,7 @@ class SearchSqlite extends SearchEngine { function searchTitle( $term ) { return $this->searchInternal( $term, false ); } - + protected function searchInternal( $term, $fulltext ) { global $wgCountTotalSearchHits, $wgContLang; @@ -182,7 +172,7 @@ class SearchSqlite extends SearchEngine { $filteredTerm = $this->filter( $wgContLang->lc( $term ) ); $resultSet = $this->db->query( $this->getQuery( $filteredTerm, $fulltext ) ); - + $total = null; if( $wgCountTotalSearchHits ) { $totalResult = $this->db->query( $this->getCountQuery( $filteredTerm, $fulltext ) ); @@ -192,7 +182,7 @@ class SearchSqlite extends SearchEngine { } $totalResult->free(); } - + return new SqliteSearchResultSet( $resultSet, $this->searchTerms, $total ); } @@ -226,7 +216,7 @@ class SearchSqlite extends SearchEngine { /** * Returns a query with limit for number of results set. - * @param $sql String: + * @param $sql String: * @return String */ function limitResult( $sql ) { @@ -246,7 +236,7 @@ class SearchSqlite extends SearchEngine { $this->queryNamespaces() ); } - + /** * Picks which field to index on, depending on what type of query. * @param $fulltext Boolean @@ -300,7 +290,7 @@ class SearchSqlite extends SearchEngine { $dbw = wfGetDB( DB_MASTER ); $dbw->delete( 'searchindex', array( 'rowid' => $id ), __METHOD__ ); - + $dbw->insert( 'searchindex', array( 'rowid' => $id, @@ -333,7 +323,7 @@ class SearchSqlite extends SearchEngine { * @ingroup Search */ class SqliteSearchResultSet extends SqlSearchResultSet { - function SqliteSearchResultSet( $resultSet, $terms, $totalHits=null ) { + function __construct( $resultSet, $terms, $totalHits=null ) { parent::__construct( $resultSet, $terms ); $this->mTotalHits = $totalHits; } @@ -341,4 +331,4 @@ class SqliteSearchResultSet extends SqlSearchResultSet { function getTotalHits() { return $this->mTotalHits; } -}
\ No newline at end of file +} diff --git a/includes/search/SearchUpdate.php b/includes/search/SearchUpdate.php index e30c70e6..5262faa4 100644 --- a/includes/search/SearchUpdate.php +++ b/includes/search/SearchUpdate.php @@ -1,6 +1,16 @@ <?php /** + * Search index updater + * * See deferred.txt + * + * @file + * @ingroup Search + */ + +/** + * Database independant search index updater + * * @ingroup Search */ class SearchUpdate { @@ -8,7 +18,7 @@ class SearchUpdate { /* private */ var $mId = 0, $mNamespace, $mTitle, $mText; /* private */ var $mTitleWords; - function SearchUpdate( $id, $title, $text = false ) { + function __construct( $id, $title, $text = false ) { $nt = Title::newFromText( $title ); if( $nt ) { $this->mId = $id; @@ -29,23 +39,23 @@ class SearchUpdate { if( $wgDisableSearchUpdate || !$this->mId ) { return false; } - $fname = 'SearchUpdate::doUpdate'; - wfProfileIn( $fname ); + + wfProfileIn( __METHOD__ ); $search = SearchEngine::create(); $lc = SearchEngine::legalSearchChars() . '&#;'; if( $this->mText === false ) { $search->updateTitle($this->mId, - Title::indexTitle( $this->mNamespace, $this->mTitle )); - wfProfileOut( $fname ); + $search->normalizeText( Title::indexTitle( $this->mNamespace, $this->mTitle ) ) ); + wfProfileOut( __METHOD__ ); return; } # Language-specific strip/conversion $text = $wgContLang->normalizeForSearch( $this->mText ); - wfProfileIn( $fname.'-regexps' ); + wfProfileIn( __METHOD__ . '-regexps' ); $text = preg_replace( "/<\\/?\\s*[A-Za-z][^>]*?>/", ' ', $wgContLang->lc( " " . $text . " " ) ); # Strip HTML markup $text = preg_replace( "/(^|\\n)==\\s*([^\\n]+)\\s*==(\\s)/sD", @@ -92,20 +102,21 @@ class SearchUpdate { # Strip wiki '' and ''' $text = preg_replace( "/''[']*/", " ", $text ); - wfProfileOut( "$fname-regexps" ); + wfProfileOut( __METHOD__ . '-regexps' ); 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 ); + wfProfileOut( __METHOD__ ); } } /** * Placeholder class + * * @ingroup Search */ class SearchUpdateMyISAM extends SearchUpdate { |