1 files changed, 41 insertions, 65 deletions

diff --git a/maintenance/backupPrefetch.inc b/maintenance/backupPrefetch.inc
index 512af1c7..9d743137 100644
--- a/maintenance/backupPrefetch.inc
+++ b/maintenance/backupPrefetch.inc
@@ -1,42 +1,10 @@
 <?php
-
-// Some smart guy removed XMLReader's global constants from PHP 5.1
-// and replaced them with class constants. Breaking source compatibility
-// is SUPER awesome, and I love languages which do this constantly!
-$xmlReaderConstants = array(
-	"NONE",
-	"ELEMENT",
-	"ATTRIBUTE", 
-	"TEXT",
-	"CDATA",
-	"ENTITY_REF",
-	"ENTITY",
-	"PI",
-	"COMMENT",
-	"DOC",
-	"DOC_TYPE",
-	"DOC_FRAGMENT",
-	"NOTATION",
-	"WHITESPACE",
-	"SIGNIFICANT_WHITESPACE",
-	"END_ELEMENT",
-	"END_ENTITY",
-	"XML_DECLARATION",
-	"LOADDTD",
-	"DEFAULTATTRS",
-	"VALIDATE",
-	"SUBST_ENTITIES" );
-foreach( $xmlReaderConstants as $name ) {
-	$fullName = "XMLREADER_$name";
-	$newName = "XMLReader::$name";
-	if( !defined( $fullName ) ) {
-		if( defined( $newName ) ) {
-			define( $fullName, constant( $newName ) );
-		} else {
-			// broken or missing the extension...
-		}
-	}
-}
+/**
+ * Helper class for the --prefetch option of dumpTextPass.php
+ *
+ * @file
+ * @ingroup Maintenance
+ */
 
 /**
  * Readahead helper for making large MediaWiki data dumps;
@@ -51,7 +19,6 @@ foreach( $xmlReaderConstants as $name ) {
  * - text contents are immutable and should not change once
  *   recorded, so the previous dump is a reliable source
  *
- * Requires PHP 5 and the XMLReader PECL extension.
  * @ingroup Maintenance
  */
 class BaseDump {
@@ -60,9 +27,12 @@ class BaseDump {
 	var $atPageEnd = false;
 	var $lastPage = 0;
 	var $lastRev = 0;
+	var $infiles = null;
 
 	function BaseDump( $infile ) {
+		$this->infiles = explode(';',$infile);
 		$this->reader = new XMLReader();
+		$infile = array_shift($this->infiles);
 		$this->reader->open( $infile );
 	}
 
@@ -71,26 +41,26 @@ class BaseDump {
 	 * from the dump stream. May return null if the page is
 	 * unavailable.
 	 *
-	 * @param int $page ID number of page to read
-	 * @param int $rev ID number of revision to read
+	 * @param $page Integer: ID number of page to read
+	 * @param $rev Integer: ID number of revision to read
 	 * @return string or null
 	 */
 	function prefetch( $page, $rev ) {
 		$page = intval( $page );
 		$rev = intval( $rev );
-		while( $this->lastPage < $page && !$this->atEnd ) {
+		while ( $this->lastPage < $page && !$this->atEnd ) {
 			$this->debug( "BaseDump::prefetch at page $this->lastPage, looking for $page" );
 			$this->nextPage();
 		}
-		if( $this->lastPage > $page || $this->atEnd ) {
+		if ( $this->lastPage > $page || $this->atEnd ) {
 			$this->debug( "BaseDump::prefetch already past page $page looking for rev $rev  [$this->lastPage, $this->lastRev]" );
 			return null;
 		}
-		while( $this->lastRev < $rev && !$this->atEnd && !$this->atPageEnd ) {
+		while ( $this->lastRev < $rev && !$this->atEnd && !$this->atPageEnd ) {
 			$this->debug( "BaseDump::prefetch at page $this->lastPage, rev $this->lastRev, looking for $page, $rev" );
 			$this->nextRev();
 		}
-		if( $this->lastRev == $rev && !$this->atEnd ) {
+		if ( $this->lastRev == $rev && !$this->atEnd ) {
 			$this->debug( "BaseDump::prefetch hit on $page, $rev [$this->lastPage, $this->lastRev]" );
 			return $this->nextText();
 		} else {
@@ -101,22 +71,27 @@ class BaseDump {
 
 	function debug( $str ) {
 		wfDebug( $str . "\n" );
-		//global $dumper;
-		//$dumper->progress( $str );
+		// global $dumper;
+		// $dumper->progress( $str );
 	}
 
 	/**
 	 * @access private
 	 */
 	function nextPage() {
-		if( $this->skipTo( 'page', 'mediawiki' ) ) {
-			if( $this->skipTo( 'id' ) ) {
+		if ( $this->skipTo( 'page', 'mediawiki' ) ) {
+			if ( $this->skipTo( 'id' ) ) {
 				$this->lastPage = intval( $this->nodeContents() );
 				$this->lastRev = 0;
 				$this->atPageEnd = false;
 			}
 		} else {
-			$this->atEnd = true;
+			$this->close();
+			if (count($this->infiles)) {
+				$infile = array_shift($this->infiles);
+				$this->reader->open( $infile );
+				$this->atEnd = false;
+			}
 		}
 	}
 
@@ -124,8 +99,8 @@ class BaseDump {
 	 * @access private
 	 */
 	function nextRev() {
-		if( $this->skipTo( 'revision' ) ) {
-			if( $this->skipTo( 'id' ) ) {
+		if ( $this->skipTo( 'revision' ) ) {
+			if ( $this->skipTo( 'id' ) ) {
 				$this->lastRev = intval( $this->nodeContents() );
 			}
 		} else {
@@ -144,16 +119,16 @@ class BaseDump {
 	/**
 	 * @access private
 	 */
-	function skipTo( $name, $parent='page' ) {
-		if( $this->atEnd ) {
+	function skipTo( $name, $parent = 'page' ) {
+		if ( $this->atEnd ) {
 			return false;
 		}
-		while( $this->reader->read() ) {
-			if( $this->reader->nodeType == XMLREADER_ELEMENT &&
+		while ( $this->reader->read() ) {
+			if ( $this->reader->nodeType == XMLReader::ELEMENT &&
 				$this->reader->name == $name ) {
 				return true;
 			}
-			if( $this->reader->nodeType == XMLREADER_END_ELEMENT &&
+			if ( $this->reader->nodeType == XMLReader::END_ELEMENT &&
 				$this->reader->name == $parent ) {
 				$this->debug( "BaseDump::skipTo found </$parent> searching for <$name>" );
 				return false;
@@ -166,25 +141,26 @@ class BaseDump {
 	 * Shouldn't something like this be built-in to XMLReader?
 	 * Fetches text contents of the current element, assuming
 	 * no sub-elements or such scary things.
-	 * @return string
+	 *
+	 * @return String
 	 * @access private
 	 */
 	function nodeContents() {
-		if( $this->atEnd ) {
+		if ( $this->atEnd ) {
 			return null;
 		}
-		if( $this->reader->isEmptyElement ) {
+		if ( $this->reader->isEmptyElement ) {
 			return "";
 		}
 		$buffer = "";
-		while( $this->reader->read() ) {
+		while ( $this->reader->read() ) {
 			switch( $this->reader->nodeType ) {
-			case XMLREADER_TEXT:
-//			case XMLREADER_WHITESPACE:
-			case XMLREADER_SIGNIFICANT_WHITESPACE:
+			case XMLReader::TEXT:
+//			case XMLReader::WHITESPACE:
+			case XMLReader::SIGNIFICANT_WHITESPACE:
 				$buffer .= $this->reader->value;
 				break;
-			case XMLREADER_END_ELEMENT:
+			case XMLReader::END_ELEMENT:
 				return $buffer;
 			}
 		}