]> scripts.mit.edu Git - autoinstallsdev/mediawiki.git/blobdiff - includes/search/SearchUpdate.php
MediaWiki 1.30.2
[autoinstallsdev/mediawiki.git] / includes / search / SearchUpdate.php
diff --git a/includes/search/SearchUpdate.php b/includes/search/SearchUpdate.php
deleted file mode 100644 (file)
index 5262faa..0000000
+++ /dev/null
@@ -1,124 +0,0 @@
-<?php
-/**
- * Search index updater
- *
- * See deferred.txt
- *
- * @file
- * @ingroup Search
- */
-
-/**
- * Database independant search index updater
- *
- * @ingroup Search
- */
-class SearchUpdate {
-
-       /* private */ var $mId = 0, $mNamespace, $mTitle, $mText;
-       /* private */ var $mTitleWords;
-
-       function __construct( $id, $title, $text = false ) {
-               $nt = Title::newFromText( $title );
-               if( $nt ) {
-                       $this->mId = $id;
-                       $this->mText = $text;
-
-                       $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() {
-               global $wgContLang, $wgDisableSearchUpdate;
-
-               if( $wgDisableSearchUpdate || !$this->mId ) {
-                       return false;
-               }
-
-               wfProfileIn( __METHOD__ );
-
-               $search = SearchEngine::create();
-               $lc = SearchEngine::legalSearchChars() . '&#;';
-
-               if( $this->mText === false ) {
-                       $search->updateTitle($this->mId,
-                               $search->normalizeText( Title::indexTitle( $this->mNamespace, $this->mTitle ) ) );
-                       wfProfileOut( __METHOD__ );
-                       return;
-               }
-
-               # Language-specific strip/conversion
-               $text = $wgContLang->normalizeForSearch( $this->mText );
-
-               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",
-                 "\\1\\2 \\2 \\2\\3", $text ); # Emphasize headings
-
-               # Strip external URLs
-               $uc = "A-Za-z0-9_\\/:.,~%\\-+&;#?!=()@\\xA0-\\xFF";
-               $protos = "http|https|ftp|mailto|news|gopher";
-               $pat = "/(^|[^\\[])({$protos}):[{$uc}]+([^{$uc}]|$)/";
-               $text = preg_replace( $pat, "\\1 \\3", $text );
-
-               $p1 = "/([^\\[])\\[({$protos}):[{$uc}]+]/";
-               $p2 = "/([^\\[])\\[({$protos}):[{$uc}]+\\s+([^\\]]+)]/";
-               $text = preg_replace( $p1, "\\1 ", $text );
-               $text = preg_replace( $p2, "\\1 \\3 ", $text );
-
-               # Internal image links
-               $pat2 = "/\\[\\[image:([{$uc}]+)\\.(gif|png|jpg|jpeg)([^{$uc}])/i";
-               $text = preg_replace( $pat2, " \\1 \\3", $text );
-
-               $text = preg_replace( "/([^{$lc}])([{$lc}]+)]]([a-z]+)/",
-                 "\\1\\2 \\2\\3", $text ); # Handle [[game]]s
-
-               # Strip all remaining non-search characters
-               $text = preg_replace( "/[^{$lc}]+/", " ", $text );
-
-               # Handle 's, s'
-               #
-               #   $text = preg_replace( "/([{$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 );
-               wfProfileOut( __METHOD__ . '-regexps' );
-
-               wfRunHooks( 'SearchUpdate', array( $this->mId, $this->mNamespace, $this->mTitle, &$text ) );
-
-               # Perform the actual update
-               $search->update($this->mId, $search->normalizeText( Title::indexTitle( $this->mNamespace, $this->mTitle ) ),
-                               $search->normalizeText( $text ) );
-
-               wfProfileOut( __METHOD__ );
-       }
-}
-
-/**
- * Placeholder class
- *
- * @ingroup Search
- */
-class SearchUpdateMyISAM extends SearchUpdate {
-       # Inherits everything
-}