3 * Base classes for dumps and export
5 * Copyright © 2003, 2005, 2006 Brion Vibber <brion@pobox.com>
6 * http://www.mediawiki.org/
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License along
19 * with this program; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
21 * http://www.gnu.org/copyleft/gpl.html
31 * @ingroup SpecialPage Dump
34 var $list_authors = false ; # Return distinct author list (when not returning full history)
35 var $author_list = "" ;
37 var $dumpUploads = false;
38 var $dumpUploadFileContents = false;
42 const STABLE
= 4; // extension defined
53 * If using WikiExporter::STREAM to stream a large amount of data,
54 * provide a database connection which is not managed by
55 * LoadBalancer to read from: some history blob types will
56 * make additional queries to pull source data while the
57 * main query is still running.
59 * @param $db DatabaseBase
60 * @param $history Mixed: one of WikiExporter::FULL, WikiExporter::CURRENT,
61 * WikiExporter::RANGE or WikiExporter::STABLE,
62 * or an associative array:
63 * offset: non-inclusive offset at which to start the query
64 * limit: maximum number of rows to return
65 * dir: "asc" or "desc" timestamp order
66 * @param $buffer Int: one of WikiExporter::BUFFER or WikiExporter::STREAM
67 * @param $text Int: one of WikiExporter::TEXT or WikiExporter::STUB
69 function __construct( &$db, $history = WikiExporter
::CURRENT
,
70 $buffer = WikiExporter
::BUFFER
, $text = WikiExporter
::TEXT
) {
72 $this->history
= $history;
73 $this->buffer
= $buffer;
74 $this->writer
= new XmlDumpWriter();
75 $this->sink
= new DumpOutput();
80 * Set the DumpOutput or DumpFilter object which will receive
81 * various row objects and XML output for filtering. Filters
82 * can be chained or used as callbacks.
86 public function setOutputSink( &$sink ) {
90 public function openStream() {
91 $output = $this->writer
->openStream();
92 $this->sink
->writeOpenStream( $output );
95 public function closeStream() {
96 $output = $this->writer
->closeStream();
97 $this->sink
->writeCloseStream( $output );
101 * Dumps a series of page and revision records for all pages
102 * in the database, either including complete history or only
103 * the most recent version.
105 public function allPages() {
106 return $this->dumpFrom( '' );
110 * Dumps a series of page and revision records for those pages
111 * in the database falling within the page_id range given.
112 * @param $start Int: inclusive lower limit (this id is included)
113 * @param $end Int: Exclusive upper limit (this id is not included)
114 * If 0, no upper limit.
116 public function pagesByRange( $start, $end ) {
117 $condition = 'page_id >= ' . intval( $start );
119 $condition .= ' AND page_id < ' . intval( $end );
121 return $this->dumpFrom( $condition );
125 * Dumps a series of page and revision records for those pages
126 * in the database with revisions falling within the rev_id range given.
127 * @param $start Int: inclusive lower limit (this id is included)
128 * @param $end Int: Exclusive upper limit (this id is not included)
129 * If 0, no upper limit.
131 public function revsByRange( $start, $end ) {
132 $condition = 'rev_id >= ' . intval( $start );
134 $condition .= ' AND rev_id < ' . intval( $end );
136 return $this->dumpFrom( $condition );
140 * @param $title Title
142 public function pageByTitle( $title ) {
143 return $this->dumpFrom(
144 'page_namespace=' . $title->getNamespace() .
145 ' AND page_title=' . $this->db
->addQuotes( $title->getDBkey() ) );
148 public function pageByName( $name ) {
149 $title = Title
::newFromText( $name );
150 if ( is_null( $title ) ) {
151 throw new MWException( "Can't export invalid title" );
153 return $this->pageByTitle( $title );
157 public function pagesByName( $names ) {
158 foreach ( $names as $name ) {
159 $this->pageByName( $name );
163 public function allLogs() {
164 return $this->dumpFrom( '' );
167 public function logsByRange( $start, $end ) {
168 $condition = 'log_id >= ' . intval( $start );
170 $condition .= ' AND log_id < ' . intval( $end );
172 return $this->dumpFrom( $condition );
175 # Generates the distinct list of authors of an article
176 # Not called by default (depends on $this->list_authors)
177 # Can be set by Special:Export when not exporting whole history
178 protected function do_list_authors( $cond ) {
179 wfProfileIn( __METHOD__
);
180 $this->author_list
= "<contributors>";
183 $res = $this->db
->select(
184 array( 'page', 'revision' ),
185 array( 'DISTINCT rev_user_text', 'rev_user' ),
187 $this->db
->bitAnd( 'rev_deleted', Revision
::DELETED_USER
) . ' = 0',
194 foreach ( $res as $row ) {
195 $this->author_list
.= "<contributor>" .
197 htmlentities( $row->rev_user_text
) .
204 $this->author_list
.= "</contributors>";
205 wfProfileOut( __METHOD__
);
208 protected function dumpFrom( $cond = '' ) {
209 wfProfileIn( __METHOD__
);
210 # For logging dumps...
211 if ( $this->history
& self
::LOGS
) {
212 if ( $this->buffer
== WikiExporter
::STREAM
) {
213 $prev = $this->db
->bufferResults( false );
215 $where = array( 'user_id = log_user' );
217 $hideLogs = LogEventsList
::getExcludeClause( $this->db
);
218 if ( $hideLogs ) $where[] = $hideLogs;
219 # Add on any caller specified conditions
220 if ( $cond ) $where[] = $cond;
221 # Get logging table name for logging.* clause
222 $logging = $this->db
->tableName( 'logging' );
223 $result = $this->db
->select( array( 'logging', 'user' ),
224 array( "{$logging}.*", 'user_name' ), // grab the user name
227 array( 'ORDER BY' => 'log_id', 'USE INDEX' => array( 'logging' => 'PRIMARY' ) )
229 $wrapper = $this->db
->resultObject( $result );
230 $this->outputLogStream( $wrapper );
231 if ( $this->buffer
== WikiExporter
::STREAM
) {
232 $this->db
->bufferResults( $prev );
236 $tables = array( 'page', 'revision' );
237 $opts = array( 'ORDER BY' => 'page_id ASC' );
238 $opts['USE INDEX'] = array();
240 if ( is_array( $this->history
) ) {
241 # Time offset/limit for all pages/history...
242 $revJoin = 'page_id=rev_page';
244 if ( $this->history
['dir'] == 'asc' ) {
246 $opts['ORDER BY'] = 'rev_timestamp ASC';
249 $opts['ORDER BY'] = 'rev_timestamp DESC';
252 if ( !empty( $this->history
['offset'] ) ) {
253 $revJoin .= " AND rev_timestamp $op " .
254 $this->db
->addQuotes( $this->db
->timestamp( $this->history
['offset'] ) );
256 $join['revision'] = array( 'INNER JOIN', $revJoin );
258 if ( !empty( $this->history
['limit'] ) ) {
259 $opts['LIMIT'] = intval( $this->history
['limit'] );
261 } elseif ( $this->history
& WikiExporter
::FULL
) {
262 # Full history dumps...
263 $join['revision'] = array( 'INNER JOIN', 'page_id=rev_page' );
264 } elseif ( $this->history
& WikiExporter
::CURRENT
) {
265 # Latest revision dumps...
266 if ( $this->list_authors
&& $cond != '' ) { // List authors, if so desired
267 $this->do_list_authors( $cond );
269 $join['revision'] = array( 'INNER JOIN', 'page_id=rev_page AND page_latest=rev_id' );
270 } elseif ( $this->history
& WikiExporter
::STABLE
) {
271 # "Stable" revision dumps...
272 # Default JOIN, to be overridden...
273 $join['revision'] = array( 'INNER JOIN', 'page_id=rev_page AND page_latest=rev_id' );
274 # One, and only one hook should set this, and return false
275 if ( wfRunHooks( 'WikiExporter::dumpStableQuery', array( &$tables, &$opts, &$join ) ) ) {
276 wfProfileOut( __METHOD__
);
277 throw new MWException( __METHOD__
. " given invalid history dump type." );
279 } elseif ( $this->history
& WikiExporter
::RANGE
) {
280 # Dump of revisions within a specified range
281 $join['revision'] = array( 'INNER JOIN', 'page_id=rev_page' );
282 $opts['ORDER BY'] = 'rev_page ASC, rev_id ASC';
284 # Uknown history specification parameter?
285 wfProfileOut( __METHOD__
);
286 throw new MWException( __METHOD__
. " given invalid history dump type." );
288 # Query optimization hacks
290 $opts[] = 'STRAIGHT_JOIN';
291 $opts['USE INDEX']['page'] = 'PRIMARY';
293 # Build text join options
294 if ( $this->text
!= WikiExporter
::STUB
) { // 1-pass
296 $join['text'] = array( 'INNER JOIN', 'rev_text_id=old_id' );
299 if ( $this->buffer
== WikiExporter
::STREAM
) {
300 $prev = $this->db
->bufferResults( false );
303 wfRunHooks( 'ModifyExportQuery',
304 array( $this->db
, &$tables, &$cond, &$opts, &$join ) );
307 $result = $this->db
->select( $tables, '*', $cond, __METHOD__
, $opts, $join );
308 $wrapper = $this->db
->resultObject( $result );
309 # Output dump results
310 $this->outputPageStream( $wrapper );
312 if ( $this->buffer
== WikiExporter
::STREAM
) {
313 $this->db
->bufferResults( $prev );
316 wfProfileOut( __METHOD__
);
320 * Runs through a query result set dumping page and revision records.
321 * The result set should be sorted/grouped by page to avoid duplicate
322 * page records in the output.
324 * The result set will be freed once complete. Should be safe for
325 * streaming (non-buffered) queries, as long as it was made on a
326 * separate database connection not managed by LoadBalancer; some
327 * blob storage types will make queries to pull source data.
329 * @param $resultset ResultWrapper
331 protected function outputPageStream( $resultset ) {
333 foreach ( $resultset as $row ) {
334 if ( is_null( $last ) ||
335 $last->page_namespace
!= $row->page_namespace ||
336 $last->page_title
!= $row->page_title
) {
337 if ( isset( $last ) ) {
339 if ( $this->dumpUploads
) {
340 $output .= $this->writer
->writeUploads( $last, $this->dumpUploadFileContents
);
342 $output .= $this->writer
->closePage();
343 $this->sink
->writeClosePage( $output );
345 $output = $this->writer
->openPage( $row );
346 $this->sink
->writeOpenPage( $row, $output );
349 $output = $this->writer
->writeRevision( $row );
350 $this->sink
->writeRevision( $row, $output );
352 if ( isset( $last ) ) {
354 if ( $this->dumpUploads
) {
355 $output .= $this->writer
->writeUploads( $last, $this->dumpUploadFileContents
);
357 $output .= $this->author_list
;
358 $output .= $this->writer
->closePage();
359 $this->sink
->writeClosePage( $output );
363 protected function outputLogStream( $resultset ) {
364 foreach ( $resultset as $row ) {
365 $output = $this->writer
->writeLogItem( $row );
366 $this->sink
->writeLogItem( $row, $output );
374 class XmlDumpWriter
{
376 * Returns the export schema version.
379 function schemaVersion() {
384 * Opens the XML output stream's root <mediawiki> element.
385 * This does not include an xml directive, so is safe to include
386 * as a subelement in a larger XML stream. Namespace and XML Schema
387 * references are included.
389 * Output will be encoded in UTF-8.
393 function openStream() {
394 global $wgLanguageCode;
395 $ver = $this->schemaVersion();
396 return Xml
::element( 'mediawiki', array(
397 'xmlns' => "http://www.mediawiki.org/xml/export-$ver/",
398 'xmlns:xsi' => "http://www.w3.org/2001/XMLSchema-instance",
399 'xsi:schemaLocation' => "http://www.mediawiki.org/xml/export-$ver/ " .
400 "http://www.mediawiki.org/xml/export-$ver.xsd",
402 'xml:lang' => $wgLanguageCode ),
408 function siteInfo() {
413 $this->caseSetting(),
414 $this->namespaces() );
415 return " <siteinfo>\n " .
416 implode( "\n ", $info ) .
420 function sitename() {
422 return Xml
::element( 'sitename', array(), $wgSitename );
425 function generator() {
427 return Xml
::element( 'generator', array(), "MediaWiki $wgVersion" );
430 function homelink() {
431 return Xml
::element( 'base', array(), Title
::newMainPage()->getCanonicalUrl() );
434 function caseSetting() {
435 global $wgCapitalLinks;
436 // "case-insensitive" option is reserved for future
437 $sensitivity = $wgCapitalLinks ?
'first-letter' : 'case-sensitive';
438 return Xml
::element( 'case', array(), $sensitivity );
441 function namespaces() {
443 $spaces = "<namespaces>\n";
444 foreach ( $wgContLang->getFormattedNamespaces() as $ns => $title ) {
446 Xml
::element( 'namespace',
448 'case' => MWNamespace
::isCapitalized( $ns ) ?
'first-letter' : 'case-sensitive',
451 $spaces .= " </namespaces>";
456 * Closes the output stream with the closing root element.
457 * Call when finished dumping things.
461 function closeStream() {
462 return "</mediawiki>\n";
466 * Opens a <page> section on the output stream, with data
467 * from the given database row.
473 function openPage( $row ) {
475 $title = Title
::makeTitle( $row->page_namespace
, $row->page_title
);
476 $out .= ' ' . Xml
::elementClean( 'title', array(), self
::canonicalTitle( $title ) ) . "\n";
477 $out .= ' ' . Xml
::element( 'ns', array(), strval( $row->page_namespace
) ) . "\n";
478 $out .= ' ' . Xml
::element( 'id', array(), strval( $row->page_id
) ) . "\n";
479 if ( $row->page_is_redirect
) {
480 $page = WikiPage
::factory( $title );
481 $redirect = $page->getRedirectTarget();
482 if ( $redirect instanceOf Title
&& $redirect->isValidRedirectTarget() ) {
483 $out .= ' ' . Xml
::element( 'redirect', array( 'title' => self
::canonicalTitle( $redirect ) ) ) . "\n";
487 if ( $row->page_restrictions
!= '' ) {
488 $out .= ' ' . Xml
::element( 'restrictions', array(),
489 strval( $row->page_restrictions
) ) . "\n";
492 wfRunHooks( 'XmlDumpWriterOpenPage', array( $this, &$out, $row, $title ) );
498 * Closes a <page> section on the output stream.
503 function closePage() {
508 * Dumps a <revision> section on the output stream, with
509 * data filled in from the given database row.
515 function writeRevision( $row ) {
516 wfProfileIn( __METHOD__
);
518 $out = " <revision>\n";
519 $out .= " " . Xml
::element( 'id', null, strval( $row->rev_id
) ) . "\n";
521 $out .= $this->writeTimestamp( $row->rev_timestamp
);
523 if ( $row->rev_deleted
& Revision
::DELETED_USER
) {
524 $out .= " " . Xml
::element( 'contributor', array( 'deleted' => 'deleted' ) ) . "\n";
526 $out .= $this->writeContributor( $row->rev_user
, $row->rev_user_text
);
529 if ( $row->rev_minor_edit
) {
530 $out .= " <minor/>\n";
532 if ( $row->rev_deleted
& Revision
::DELETED_COMMENT
) {
533 $out .= " " . Xml
::element( 'comment', array( 'deleted' => 'deleted' ) ) . "\n";
534 } elseif ( $row->rev_comment
!= '' ) {
535 $out .= " " . Xml
::elementClean( 'comment', null, strval( $row->rev_comment
) ) . "\n";
539 if ( $row->rev_deleted
& Revision
::DELETED_TEXT
) {
540 $out .= " " . Xml
::element( 'text', array( 'deleted' => 'deleted' ) ) . "\n";
541 } elseif ( isset( $row->old_text
) ) {
542 // Raw text from the database may have invalid chars
543 $text = strval( Revision
::getRevisionText( $row ) );
544 $out .= " " . Xml
::elementClean( 'text',
545 array( 'xml:space' => 'preserve', 'bytes' => intval( $row->rev_len
) ),
546 strval( $text ) ) . "\n";
549 $out .= " " . Xml
::element( 'text',
550 array( 'id' => $row->rev_text_id
, 'bytes' => intval( $row->rev_len
) ),
554 if ( $row->rev_sha1
&& !( $row->rev_deleted
& Revision
::DELETED_TEXT
) ) {
555 $out .= " " . Xml
::element('sha1', null, strval( $row->rev_sha1
) ) . "\n";
557 $out .= " <sha1/>\n";
560 wfRunHooks( 'XmlDumpWriterWriteRevision', array( &$this, &$out, $row, $text ) );
562 $out .= " </revision>\n";
564 wfProfileOut( __METHOD__
);
569 * Dumps a <logitem> section on the output stream, with
570 * data filled in from the given database row.
576 function writeLogItem( $row ) {
577 wfProfileIn( __METHOD__
);
579 $out = " <logitem>\n";
580 $out .= " " . Xml
::element( 'id', null, strval( $row->log_id
) ) . "\n";
582 $out .= $this->writeTimestamp( $row->log_timestamp
);
584 if ( $row->log_deleted
& LogPage
::DELETED_USER
) {
585 $out .= " " . Xml
::element( 'contributor', array( 'deleted' => 'deleted' ) ) . "\n";
587 $out .= $this->writeContributor( $row->log_user
, $row->user_name
);
590 if ( $row->log_deleted
& LogPage
::DELETED_COMMENT
) {
591 $out .= " " . Xml
::element( 'comment', array( 'deleted' => 'deleted' ) ) . "\n";
592 } elseif ( $row->log_comment
!= '' ) {
593 $out .= " " . Xml
::elementClean( 'comment', null, strval( $row->log_comment
) ) . "\n";
596 $out .= " " . Xml
::element( 'type', null, strval( $row->log_type
) ) . "\n";
597 $out .= " " . Xml
::element( 'action', null, strval( $row->log_action
) ) . "\n";
599 if ( $row->log_deleted
& LogPage
::DELETED_ACTION
) {
600 $out .= " " . Xml
::element( 'text', array( 'deleted' => 'deleted' ) ) . "\n";
602 $title = Title
::makeTitle( $row->log_namespace
, $row->log_title
);
603 $out .= " " . Xml
::elementClean( 'logtitle', null, self
::canonicalTitle( $title ) ) . "\n";
604 $out .= " " . Xml
::elementClean( 'params',
605 array( 'xml:space' => 'preserve' ),
606 strval( $row->log_params
) ) . "\n";
609 $out .= " </logitem>\n";
611 wfProfileOut( __METHOD__
);
615 function writeTimestamp( $timestamp ) {
616 $ts = wfTimestamp( TS_ISO_8601
, $timestamp );
617 return " " . Xml
::element( 'timestamp', null, $ts ) . "\n";
620 function writeContributor( $id, $text ) {
621 $out = " <contributor>\n";
622 if ( $id ||
!IP
::isValid( $text ) ) {
623 $out .= " " . Xml
::elementClean( 'username', null, strval( $text ) ) . "\n";
624 $out .= " " . Xml
::element( 'id', null, strval( $id ) ) . "\n";
626 $out .= " " . Xml
::elementClean( 'ip', null, strval( $text ) ) . "\n";
628 $out .= " </contributor>\n";
633 * Warning! This data is potentially inconsistent. :(
636 function writeUploads( $row, $dumpContents = false ) {
637 if ( $row->page_namespace
== NS_IMAGE
) {
638 $img = wfLocalFile( $row->page_title
);
639 if ( $img && $img->exists() ) {
641 foreach ( array_reverse( $img->getHistory() ) as $ver ) {
642 $out .= $this->writeUpload( $ver, $dumpContents );
644 $out .= $this->writeUpload( $img, $dumpContents );
653 * @param $dumpContents bool
656 function writeUpload( $file, $dumpContents = false ) {
657 if ( $file->isOld() ) {
659 Xml
::element( 'archivename', null, $file->getArchiveName() ) . "\n";
663 if ( $dumpContents ) {
664 # Dump file as base64
665 # Uses only XML-safe characters, so does not need escaping
666 $contents = ' <contents encoding="base64">' .
667 chunk_split( base64_encode( file_get_contents( $file->getPath() ) ) ) .
672 return " <upload>\n" .
673 $this->writeTimestamp( $file->getTimestamp() ) .
674 $this->writeContributor( $file->getUser( 'id' ), $file->getUser( 'text' ) ) .
675 " " . Xml
::elementClean( 'comment', null, $file->getDescription() ) . "\n" .
676 " " . Xml
::element( 'filename', null, $file->getName() ) . "\n" .
678 " " . Xml
::element( 'src', null, $file->getCanonicalUrl() ) . "\n" .
679 " " . Xml
::element( 'size', null, $file->getSize() ) . "\n" .
680 " " . Xml
::element( 'sha1base36', null, $file->getSha1() ) . "\n" .
681 " " . Xml
::element( 'rel', null, $file->getRel() ) . "\n" .
687 * Return prefixed text form of title, but using the content language's
688 * canonical namespace. This skips any special-casing such as gendered
689 * user namespaces -- which while useful, are not yet listed in the
690 * XML <siteinfo> data so are unsafe in export.
692 * @param Title $title
696 public static function canonicalTitle( Title
$title ) {
697 if ( $title->getInterwiki() ) {
698 return $title->getPrefixedText();
702 $prefix = str_replace( '_', ' ', $wgContLang->getNsText( $title->getNamespace() ) );
704 if ( $prefix !== '' ) {
708 return $prefix . $title->getText();
714 * Base class for output stream; prints to stdout or buffer or whereever.
718 function writeOpenStream( $string ) {
719 $this->write( $string );
722 function writeCloseStream( $string ) {
723 $this->write( $string );
726 function writeOpenPage( $page, $string ) {
727 $this->write( $string );
730 function writeClosePage( $string ) {
731 $this->write( $string );
734 function writeRevision( $rev, $string ) {
735 $this->write( $string );
738 function writeLogItem( $rev, $string ) {
739 $this->write( $string );
743 * Override to write to a different stream type.
746 function write( $string ) {
751 * Close the old file, move it to a specified name,
752 * and reopen new file with the old name. Use this
753 * for writing out a file in multiple pieces
754 * at specified checkpoints (e.g. every n hours).
755 * @param $newname mixed File name. May be a string or an array with one element
757 function closeRenameAndReopen( $newname ) {
762 * Close the old file, and move it to a specified name.
763 * Use this for the last piece of a file written out
764 * at specified checkpoints (e.g. every n hours).
765 * @param $newname mixed File name. May be a string or an array with one element
766 * @param $open bool If true, a new file with the old filename will be opened again for writing (default: false)
768 function closeAndRename( $newname, $open = false ) {
773 * Returns the name of the file or files which are
774 * being written to, if there are any.
777 function getFilenames() {
783 * Stream outputter to send data to a file.
786 class DumpFileOutput
extends DumpOutput
{
787 protected $handle, $filename;
789 function __construct( $file ) {
790 $this->handle
= fopen( $file, "wt" );
791 $this->filename
= $file;
794 function writeCloseStream( $string ) {
795 parent
::writeCloseStream( $string );
796 fclose( $this->handle
);
799 function write( $string ) {
800 fputs( $this->handle
, $string );
803 function closeRenameAndReopen( $newname ) {
804 $this->closeAndRename( $newname, true );
807 function renameOrException( $newname ) {
808 if (! rename( $this->filename
, $newname ) ) {
809 throw new MWException( __METHOD__
. ": rename of file {$this->filename} to $newname failed\n" );
813 function checkRenameArgCount( $newname ) {
814 if ( is_array( $newname ) ) {
815 if ( count( $newname ) > 1 ) {
816 throw new MWException( __METHOD__
. ": passed multiple arguments for rename of single file\n" );
818 $newname = $newname[0];
824 function closeAndRename( $newname, $open = false ) {
825 $newname = $this->checkRenameArgCount( $newname );
827 fclose( $this->handle
);
828 $this->renameOrException( $newname );
830 $this->handle
= fopen( $this->filename
, "wt" );
835 function getFilenames() {
836 return $this->filename
;
841 * Stream outputter to send data to a file via some filter program.
842 * Even if compression is available in a library, using a separate
843 * program can allow us to make use of a multi-processor system.
846 class DumpPipeOutput
extends DumpFileOutput
{
847 protected $command, $filename;
849 function __construct( $command, $file = null ) {
850 if ( !is_null( $file ) ) {
851 $command .= " > " . wfEscapeShellArg( $file );
854 $this->startCommand( $command );
855 $this->command
= $command;
856 $this->filename
= $file;
859 function writeCloseStream( $string ) {
860 parent
::writeCloseStream( $string );
861 proc_close( $this->procOpenResource
);
864 function startCommand( $command ) {
866 0 => array( "pipe", "r" ),
869 $this->procOpenResource
= proc_open( $command, $spec, $pipes );
870 $this->handle
= $pipes[0];
873 function closeRenameAndReopen( $newname ) {
874 $this->closeAndRename( $newname, true );
877 function closeAndRename( $newname, $open = false ) {
878 $newname = $this->checkRenameArgCount( $newname );
880 fclose( $this->handle
);
881 proc_close( $this->procOpenResource
);
882 $this->renameOrException( $newname );
884 $command = $this->command
;
885 $command .= " > " . wfEscapeShellArg( $this->filename
);
886 $this->startCommand( $command );
894 * Sends dump output via the gzip compressor.
897 class DumpGZipOutput
extends DumpPipeOutput
{
898 function __construct( $file ) {
899 parent
::__construct( "gzip", $file );
904 * Sends dump output via the bgzip2 compressor.
907 class DumpBZip2Output
extends DumpPipeOutput
{
908 function __construct( $file ) {
909 parent
::__construct( "bzip2", $file );
914 * Sends dump output via the p7zip compressor.
917 class Dump7ZipOutput
extends DumpPipeOutput
{
918 function __construct( $file ) {
919 $command = $this->setup7zCommand( $file );
920 parent
::__construct( $command );
921 $this->filename
= $file;
924 function setup7zCommand( $file ) {
925 $command = "7za a -bd -si " . wfEscapeShellArg( $file );
926 // Suppress annoying useless crap from p7zip
927 // Unfortunately this could suppress real error messages too
928 $command .= ' >' . wfGetNull() . ' 2>&1';
932 function closeAndRename( $newname, $open = false ) {
933 $newname = $this->checkRenameArgCount( $newname );
935 fclose( $this->handle
);
936 proc_close( $this->procOpenResource
);
937 $this->renameOrException( $newname );
939 $command = $this->setup7zCommand( $this->filename
);
940 $this->startCommand( $command );
949 * Dump output filter class.
950 * This just does output filtering and streaming; XML formatting is done
951 * higher up, so be careful in what you do.
955 function __construct( &$sink ) {
956 $this->sink
=& $sink;
959 function writeOpenStream( $string ) {
960 $this->sink
->writeOpenStream( $string );
963 function writeCloseStream( $string ) {
964 $this->sink
->writeCloseStream( $string );
967 function writeOpenPage( $page, $string ) {
968 $this->sendingThisPage
= $this->pass( $page, $string );
969 if ( $this->sendingThisPage
) {
970 $this->sink
->writeOpenPage( $page, $string );
974 function writeClosePage( $string ) {
975 if ( $this->sendingThisPage
) {
976 $this->sink
->writeClosePage( $string );
977 $this->sendingThisPage
= false;
981 function writeRevision( $rev, $string ) {
982 if ( $this->sendingThisPage
) {
983 $this->sink
->writeRevision( $rev, $string );
987 function writeLogItem( $rev, $string ) {
988 $this->sink
->writeRevision( $rev, $string );
991 function closeRenameAndReopen( $newname ) {
992 $this->sink
->closeRenameAndReopen( $newname );
995 function closeAndRename( $newname, $open = false ) {
996 $this->sink
->closeAndRename( $newname, $open );
999 function getFilenames() {
1000 return $this->sink
->getFilenames();
1004 * Override for page-based filter types.
1007 function pass( $page ) {
1013 * Simple dump output filter to exclude all talk pages.
1016 class DumpNotalkFilter
extends DumpFilter
{
1017 function pass( $page ) {
1018 return !MWNamespace
::isTalk( $page->page_namespace
);
1023 * Dump output filter to include or exclude pages in a given set of namespaces.
1026 class DumpNamespaceFilter
extends DumpFilter
{
1027 var $invert = false;
1028 var $namespaces = array();
1030 function __construct( &$sink, $param ) {
1031 parent
::__construct( $sink );
1034 "NS_MAIN" => NS_MAIN
,
1035 "NS_TALK" => NS_TALK
,
1036 "NS_USER" => NS_USER
,
1037 "NS_USER_TALK" => NS_USER_TALK
,
1038 "NS_PROJECT" => NS_PROJECT
,
1039 "NS_PROJECT_TALK" => NS_PROJECT_TALK
,
1040 "NS_FILE" => NS_FILE
,
1041 "NS_FILE_TALK" => NS_FILE_TALK
,
1042 "NS_IMAGE" => NS_IMAGE
, // NS_IMAGE is an alias for NS_FILE
1043 "NS_IMAGE_TALK" => NS_IMAGE_TALK
,
1044 "NS_MEDIAWIKI" => NS_MEDIAWIKI
,
1045 "NS_MEDIAWIKI_TALK" => NS_MEDIAWIKI_TALK
,
1046 "NS_TEMPLATE" => NS_TEMPLATE
,
1047 "NS_TEMPLATE_TALK" => NS_TEMPLATE_TALK
,
1048 "NS_HELP" => NS_HELP
,
1049 "NS_HELP_TALK" => NS_HELP_TALK
,
1050 "NS_CATEGORY" => NS_CATEGORY
,
1051 "NS_CATEGORY_TALK" => NS_CATEGORY_TALK
);
1053 if ( $param { 0 } == '!' ) {
1054 $this->invert
= true;
1055 $param = substr( $param, 1 );
1058 foreach ( explode( ',', $param ) as $key ) {
1059 $key = trim( $key );
1060 if ( isset( $constants[$key] ) ) {
1061 $ns = $constants[$key];
1062 $this->namespaces
[$ns] = true;
1063 } elseif ( is_numeric( $key ) ) {
1064 $ns = intval( $key );
1065 $this->namespaces
[$ns] = true;
1067 throw new MWException( "Unrecognized namespace key '$key'\n" );
1072 function pass( $page ) {
1073 $match = isset( $this->namespaces
[$page->page_namespace
] );
1074 return $this->invert
xor $match;
1080 * Dump output filter to include only the last revision in each page sequence.
1083 class DumpLatestFilter
extends DumpFilter
{
1084 var $page, $pageString, $rev, $revString;
1086 function writeOpenPage( $page, $string ) {
1087 $this->page
= $page;
1088 $this->pageString
= $string;
1091 function writeClosePage( $string ) {
1093 $this->sink
->writeOpenPage( $this->page
, $this->pageString
);
1094 $this->sink
->writeRevision( $this->rev
, $this->revString
);
1095 $this->sink
->writeClosePage( $string );
1098 $this->revString
= null;
1100 $this->pageString
= null;
1103 function writeRevision( $rev, $string ) {
1104 if ( $rev->rev_id
== $this->page
->page_latest
) {
1106 $this->revString
= $string;
1112 * Base class for output stream; prints to stdout or buffer or whereever.
1115 class DumpMultiWriter
{
1116 function __construct( $sinks ) {
1117 $this->sinks
= $sinks;
1118 $this->count
= count( $sinks );
1121 function writeOpenStream( $string ) {
1122 for ( $i = 0; $i < $this->count
; $i++
) {
1123 $this->sinks
[$i]->writeOpenStream( $string );
1127 function writeCloseStream( $string ) {
1128 for ( $i = 0; $i < $this->count
; $i++
) {
1129 $this->sinks
[$i]->writeCloseStream( $string );
1133 function writeOpenPage( $page, $string ) {
1134 for ( $i = 0; $i < $this->count
; $i++
) {
1135 $this->sinks
[$i]->writeOpenPage( $page, $string );
1139 function writeClosePage( $string ) {
1140 for ( $i = 0; $i < $this->count
; $i++
) {
1141 $this->sinks
[$i]->writeClosePage( $string );
1145 function writeRevision( $rev, $string ) {
1146 for ( $i = 0; $i < $this->count
; $i++
) {
1147 $this->sinks
[$i]->writeRevision( $rev, $string );
1151 function closeRenameAndReopen( $newnames ) {
1152 $this->closeAndRename( $newnames, true );
1155 function closeAndRename( $newnames, $open = false ) {
1156 for ( $i = 0; $i < $this->count
; $i++
) {
1157 $this->sinks
[$i]->closeAndRename( $newnames[$i], $open );
1161 function getFilenames() {
1162 $filenames = array();
1163 for ( $i = 0; $i < $this->count
; $i++
) {
1164 $filenames[] = $this->sinks
[$i]->getFilenames();
1171 function xmlsafe( $string ) {
1172 wfProfileIn( __FUNCTION__
);
1175 * The page may contain old data which has not been properly normalized.
1176 * Invalid UTF-8 sequences or forbidden control characters will make our
1177 * XML output invalid, so be sure to strip them out.
1179 $string = UtfNormal
::cleanUp( $string );
1181 $string = htmlspecialchars( $string );
1182 wfProfileOut( __FUNCTION__
);