X-Git-Url: http://git.heureux-cyclage.org/?a=blobdiff_plain;f=maintenance%2FimportDump.php;h=b6e0fe09e5e40199c0aff5773391b66ce0434162;hb=392af46809d831514f49618cdef1e1529d7fddf4;hp=00b9ccedb0b3d19e7cc9bdc3d1c5735ed50d2c35;hpb=85fc6845f39966d3508da5c05df473b2e1c7cc98;p=lhc%2Fweb%2Fwiklou.git

diff --git a/maintenance/importDump.php b/maintenance/importDump.php
index 00b9ccedb0..b6e0fe09e5 100644
--- a/maintenance/importDump.php
+++ b/maintenance/importDump.php
@@ -1,6 +1,8 @@
 <?php
 /**
- * Copyright (C) 2005 Brion Vibber <brion@pobox.com>
+ * Import XML dump files into the current wiki.
+ *
+ * Copyright © 2005 Brion Vibber <brion@pobox.com>
  * http://www.mediawiki.org/
  *
  * This program is free software; you can redistribute it and/or modify
@@ -22,30 +24,129 @@
  * @ingroup Maintenance
  */
 
-$optionsWithArgs = array( 'report' );
-
-require_once( dirname( __FILE__ ) . '/commandLine.inc' );
+require_once( dirname( __FILE__ ) . '/Maintenance.php' );
 
 /**
+ * Maintenance script that imports XML dump files into the current wiki.
+ *
  * @ingroup Maintenance
  */
-class BackupReader {
+class BackupReader extends Maintenance {
 	var $reportingInterval = 100;
-	var $reporting = true;
 	var $pageCount = 0;
 	var $revCount  = 0;
 	var $dryRun    = false;
-	var $debug     = false;
 	var $uploads   = false;
+	var $imageBasePath = false;
+	var $nsFilter  = false;
 
 	function __construct() {
+		parent::__construct();
+		$gz = in_array('compress.zlib', stream_get_wrappers()) ? 'ok' : '(disabled; requires PHP zlib module)';
+		$bz2 = in_array('compress.bzip2', stream_get_wrappers()) ? 'ok' : '(disabled; requires PHP bzip2 module)';
+
+		$this->mDescription = <<<TEXT
+This script reads pages from an XML file as produced from Special:Export or
+dumpBackup.php, and saves them into the current wiki.
+
+Compressed XML files may be read directly:
+  .gz $gz
+  .bz2 $bz2
+  .7z (if 7za executable is in PATH)
+
+Note that for very large data sets, importDump.php may be slow; there are
+alternate methods which can be much faster for full site restoration:
+<http://www.mediawiki.org/wiki/Manual:Importing_XML_dumps>
+TEXT;
 		$this->stderr = fopen( "php://stderr", "wt" );
+		$this->addOption( 'report',
+			'Report position and speed after every n pages processed', false, true );
+		$this->addOption( 'namespaces',
+			'Import only the pages from namespaces belonging to the list of ' .
+			'pipe-separated namespace names or namespace indexes', false, true );
+		$this->addOption( 'dry-run', 'Parse dump without actually importing pages' );
+		$this->addOption( 'debug', 'Output extra verbose debug information' );
+		$this->addOption( 'uploads', 'Process file upload data if included (experimental)' );
+		$this->addOption( 'no-updates', 'Disable link table updates. Is faster but leaves the wiki in an inconsistent state' );
+		$this->addOption( 'image-base-path', 'Import files from a specified path', false, true );
+		$this->addArg( 'file', 'Dump file to import [else use stdin]', false );
+	}
+
+	public function execute() {
+		if( wfReadOnly() ) {
+			$this->error( "Wiki is in read-only mode; you'll need to disable it for import to work.", true );
+		}
+
+		$this->reportingInterval = intval( $this->getOption( 'report', 100 ) );
+		if ( !$this->reportingInterval ) {
+			$this->reportingInterval = 100; // avoid division by zero
+		}
+
+		$this->dryRun = $this->hasOption( 'dry-run' );
+		$this->uploads = $this->hasOption( 'uploads' ); // experimental!
+		if ( $this->hasOption( 'image-base-path' ) ) {
+			$this->imageBasePath = $this->getOption( 'image-base-path' );
+		}
+		if ( $this->hasOption( 'namespaces' ) ) {
+			$this->setNsfilter( explode( '|', $this->getOption( 'namespaces' ) ) );
+		}
+
+		if( $this->hasArg() ) {
+			$this->importFromFile( $this->getArg() );
+		} else {
+			$this->importFromStdin();
+		}
+
+		$this->output( "Done!\n" );
+		$this->output( "You might want to run rebuildrecentchanges.php to regenerate RecentChanges\n" );
+	}
+
+	function setNsfilter( array $namespaces ) {
+		if ( count( $namespaces ) == 0 ) {
+			$this->nsFilter = false;
+			return;
+		}
+		$this->nsFilter = array_unique( array_map( array( $this, 'getNsIndex' ), $namespaces ) );
+	}
+
+	private function getNsIndex( $namespace ) {
+		global $wgContLang;
+		if ( ( $result = $wgContLang->getNsIndex( $namespace ) ) !== false ) {
+			return $result;
+		}
+		$ns = intval( $namespace );
+		if ( strval( $ns ) === $namespace && $wgContLang->getNsText( $ns ) !== false ) {
+			return $ns;
+		}
+		$this->error( "Unknown namespace text / index specified: $namespace", true );
+	}
+
+	/**
+	 * @param $obj Title|Revision
+	 * @return bool
+	 */
+	private function skippedNamespace( $obj ) {
+		if ( $obj instanceof Title ) {
+			$ns = $obj->getNamespace();
+		} elseif ( $obj instanceof Revision ) {
+			$ns = $obj->getTitle()->getNamespace();
+		} elseif ( $obj instanceof WikiRevision ) {
+			$ns = $obj->title->getNamespace();
+		} else {
+			echo wfBacktrace();
+			$this->error( "Cannot get namespace of object in " . __METHOD__, true );
+		}
+		return is_array( $this->nsFilter ) && !in_array( $ns, $this->nsFilter );
 	}
 
 	function reportPage( $page ) {
 		$this->pageCount++;
 	}
 
+	/**
+	 * @param $rev Revision
+	 * @return mixed
+	 */
 	function handleRevision( $rev ) {
 		$title = $rev->getTitle();
 		if ( !$title ) {
@@ -53,6 +154,10 @@ class BackupReader {
 			return;
 		}
 
+		if ( $this->skippedNamespace( $title ) ) {
+			return;
+		}
+
 		$this->revCount++;
 		$this->report();
 
@@ -60,13 +165,20 @@ class BackupReader {
 			call_user_func( $this->importCallback, $rev );
 		}
 	}
-	
+
+	/**
+	 * @param $revision Revision
+	 * @return bool
+	 */
 	function handleUpload( $revision ) {
 		if ( $this->uploads ) {
+			if ( $this->skippedNamespace( $revision ) ) {
+				return;
+			}
 			$this->uploadCount++;
 			// $this->report();
 			$this->progress( "upload: " . $revision->getFilename() );
-			
+
 			if ( !$this->dryRun ) {
 				// bluuuh hack
 				// call_user_func( $this->uploadCallback, $revision );
@@ -77,6 +189,9 @@ class BackupReader {
 	}
 
 	function handleLogItem( $rev ) {
+		if ( $this->skippedNamespace( $rev ) ) {
+			return;
+		}
 		$this->revCount++;
 		$this->report();
 
@@ -92,7 +207,7 @@ class BackupReader {
 	}
 
 	function showReport() {
-		if ( $this->reporting ) {
+		if ( !$this->mQuiet ) {
 			$delta = wfTime() - $this->startTime;
 			if ( $delta ) {
 				$rate = sprintf( "%.2f", $this->pageCount / $delta );
@@ -102,12 +217,15 @@ class BackupReader {
 				$revrate = '-';
 			}
 			# Logs dumps don't have page tallies
-			if ( $this->pageCount )
+			if ( $this->pageCount ) {
 				$this->progress( "$this->pageCount ($rate pages/sec $revrate revs/sec)" );
-			else
+			} else {
 				$this->progress( "$this->revCount ($revrate revs/sec)" );
+			}
 		}
-		wfWaitForSlaves( 5 );
+		wfWaitForSlaves();
+		// XXX: Don't let deferred jobs array get absurdly large (bug 24375)
+		DeferredUpdates::doUpdates( 'commit' );
 	}
 
 	function progress( $string ) {
@@ -117,11 +235,9 @@ class BackupReader {
 	function importFromFile( $filename ) {
 		if ( preg_match( '/\.gz$/', $filename ) ) {
 			$filename = 'compress.zlib://' . $filename;
-		}
-		elseif ( preg_match( '/\.bz2$/', $filename ) ) {
+		} elseif ( preg_match( '/\.bz2$/', $filename ) ) {
 			$filename = 'compress.bzip2://' . $filename;
-		}
-		elseif ( preg_match( '/\.7z$/', $filename ) ) {
+		} elseif ( preg_match( '/\.7z$/', $filename ) ) {
 			$filename = 'mediawiki.compress.7z://' . $filename;
 		}
 
@@ -131,6 +247,9 @@ class BackupReader {
 
 	function importFromStdin() {
 		$file = fopen( 'php://stdin', 'rt' );
+		if( self::posix_isatty( $file ) ) {
+			$this->maybeHelp( true );
+		}
 		return $this->importFromHandle( $file );
 	}
 
@@ -140,7 +259,12 @@ class BackupReader {
 		$source = new ImportStreamSource( $handle );
 		$importer = new WikiImporter( $source );
 
-		$importer->setDebug( $this->debug );
+		if( $this->hasOption( 'debug' ) ) {
+			$importer->setDebug( true );
+		}
+		if ( $this->hasOption( 'no-updates' ) ) {
+			$importer->setNoUpdates( true );
+		}
 		$importer->setPageCallback( array( &$this, 'reportPage' ) );
 		$this->importCallback =  $importer->setRevisionCallback(
 			array( &$this, 'handleRevision' ) );
@@ -148,7 +272,13 @@ class BackupReader {
 			array( &$this, 'handleUpload' ) );
 		$this->logItemCallback = $importer->setLogItemCallback(
 			array( &$this, 'handleLogItem' ) );
-			
+		if ( $this->uploads ) {
+			$importer->setImportUploads( true );
+		}
+		if ( $this->imageBasePath ) {
+			$importer->setImageBasePath( $this->imageBasePath );
+		}
+
 		if ( $this->dryRun ) {
 			$importer->setPageOutCallback( null );
 		}
@@ -157,39 +287,5 @@ class BackupReader {
 	}
 }
 
-if ( wfReadOnly() ) {
-	wfDie( "Wiki is in read-only mode; you'll need to disable it for import to work.\n" );
-}
-
-$reader = new BackupReader();
-if ( isset( $options['quiet'] ) ) {
-	$reader->reporting = false;
-}
-if ( isset( $options['report'] ) ) {
-	$reader->reportingInterval = intval( $options['report'] );
-}
-if ( isset( $options['dry-run'] ) ) {
-	$reader->dryRun = true;
-}
-if ( isset( $options['debug'] ) ) {
-	$reader->debug = true;
-}
-if ( isset( $options['uploads'] ) ) {
-	$reader->uploads = true; // experimental!
-}
-
-if ( isset( $args[0] ) ) {
-	$result = $reader->importFromFile( $args[0] );
-} else {
-	$result = $reader->importFromStdin();
-}
-
-if ( WikiError::isError( $result ) ) {
-	echo $result->getMessage() . "\n";
-} else {
-	echo "Done!\n";
-	echo "You might want to run rebuildrecentchanges.php to regenerate\n";
-	echo "the recentchanges page.\n";
-}
-
-
+$maintClass = 'BackupReader';
+require_once( RUN_MAINTENANCE_IF_MAIN );