From: Chad Horohoe <chadh@wikimedia.org>
Date: Wed, 19 Jun 2013 17:55:33 +0000 (-0400)
Subject: Pass content to SearchUpdate rather than raw text
X-Git-Tag: 1.31.0-rc.0~19390^2
X-Git-Url: http://git.cyclocoop.org/%40spipnet%40?a=commitdiff_plain;h=539ec5ffa3c0c66c8fa382448b741dbc9e2902f0;p=lhc%2Fweb%2Fwiklou.git

Pass content to SearchUpdate rather than raw text

This is still less than ideal, but at least consolidates the
logic so we can later implement more content-specific handling.

Change-Id: Ic24a7ea2818f6eeb67453877dc5a31f110541ecf
---

diff --git a/includes/WikiPage.php b/includes/WikiPage.php
index 9e6a0c8a77..398a424cae 100644
--- a/includes/WikiPage.php
+++ b/includes/WikiPage.php
@@ -2109,8 +2109,7 @@ class WikiPage implements Page, IDBAccessObject {
 		}
 
 		DeferredUpdates::addUpdate( new SiteStatsUpdate( 0, 1, $good, $total ) );
-		DeferredUpdates::addUpdate( new SearchUpdate( $id, $title, $content->getTextForSearchIndex() ) );
-		// @todo let the search engine decide what to do with the content object
+		DeferredUpdates::addUpdate( new SearchUpdate( $id, $title, $content ) );
 
 		// If this is another user's talk page, update newtalk.
 		// Don't do this if $options['changed'] = false (null-edits) nor if
diff --git a/includes/search/SearchEngine.php b/includes/search/SearchEngine.php
index 00f7923257..bbd888698e 100644
--- a/includes/search/SearchEngine.php
+++ b/includes/search/SearchEngine.php
@@ -816,7 +816,8 @@ class SearchResult {
 		if ( !isset( $this->mText ) ) {
 			if ( $this->mRevision != null ) {
 				//TODO: if we could plug in some code that knows about special content models *and* about
-				//      special features of the search engine, the search could benefit.
+				//      special features of the search engine, the search could benefit. See similar
+				//      comment in SearchUpdate's constructor
 				$content = $this->mRevision->getContent();
 				$this->mText = $content ? $content->getTextForSearchIndex() : '';
 			} else { // TODO: can we fetch raw wikitext for commons images?
diff --git a/includes/search/SearchUpdate.php b/includes/search/SearchUpdate.php
index 759c7b9495..a0e1ccfd5f 100644
--- a/includes/search/SearchUpdate.php
+++ b/includes/search/SearchUpdate.php
@@ -33,7 +33,16 @@ class SearchUpdate implements DeferrableUpdate {
 	private $mId = 0, $mNamespace, $mTitle, $mText;
 	private $mTitleWords;
 
-	public function __construct( $id, $title, $text = false ) {
+	/**
+	 * Constructor
+	 *
+	 * @param int $id Page id to update
+	 * @param Title|string $title Title of page to update
+	 * @param Content|string|false $content Content of the page to update.
+	 *  If a Content object, text will be gotten from it. String is for back-compat.
+	 *  Passing false tells the backend to just update the title, not the content
+	 */
+	public function __construct( $id, $title, $content = false ) {
 		if ( is_string( $title ) ) {
 			$nt = Title::newFromText( $title );
 		} else {
@@ -42,7 +51,14 @@ class SearchUpdate implements DeferrableUpdate {
 
 		if ( $nt ) {
 			$this->mId = $id;
-			$this->mText = $text;
+			// @todo This isn't ideal, we'd really like to have content-specific
+			// handling here. See similar content in SearchEngine::initText().
+			if( is_string( $content ) ) {
+				// b/c for ApprovedRevs
+				$this->mText = $content;
+			} else {
+				$this->mText = $content ? $content->getTextForSearchIndex() : false;
+			}
 
 			$this->mNamespace = $nt->getNamespace();
 			$this->mTitle = $nt->getText(); # Discard namespace
diff --git a/maintenance/Maintenance.php b/maintenance/Maintenance.php
index a2bbec5312..e29ffd992b 100644
--- a/maintenance/Maintenance.php
+++ b/maintenance/Maintenance.php
@@ -1155,8 +1155,7 @@ abstract class Maintenance {
 			$title = $titleObj->getPrefixedDBkey();
 			$this->output( "$title..." );
 			# Update searchindex
-			# TODO: pass the Content object to SearchUpdate, let the search engine decide how to deal with it.
-			$u = new SearchUpdate( $pageId, $titleObj->getText(), $rev->getContent()->getTextForSearchIndex() );
+			$u = new SearchUpdate( $pageId, $titleObj->getText(), $rev->getContent() );
 			$u->doUpdate();
 			$this->output( "\n" );
 		}
diff --git a/maintenance/rebuildtextindex.php b/maintenance/rebuildtextindex.php
index a31aba2612..c651f720b5 100644
--- a/maintenance/rebuildtextindex.php
+++ b/maintenance/rebuildtextindex.php
@@ -115,9 +115,8 @@ class RebuildTextIndex extends Maintenance {
 
 					$rev = new Revision( $s );
 					$content = $rev->getContent();
-					$text = $content->getTextForSearchIndex();
 
-					$u = new SearchUpdate( $s->page_id, $title, $text );
+					$u = new SearchUpdate( $s->page_id, $title, $content );
 					$u->doUpdate();
 				} catch ( MWContentSerializationException $ex ) {
 					$this->output( "Failed to deserialize content of revision {$s->rev_id} of page "