2 # Copyright (C) 2003 Brion Vibber <brion@pobox.com>
3 # http://www.mediawiki.org/
5 # This program is free software; you can redistribute it and/or modify
6 # it under the terms of the GNU General Public License as published by
7 # the Free Software Foundation; either version 2 of the License, or
8 # (at your option) any later version.
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
15 # You should have received a copy of the GNU General Public License along
16 # with this program; if not, write to the Free Software Foundation, Inc.,
17 # 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
18 # http://www.gnu.org/copyleft/gpl.html
22 * @subpackage SpecialPage
26 require_once( 'Revision.php' );
31 function wfSpecialExport( $page = '' ) {
32 global $wgOut, $wgLang, $wgRequest;
34 if( $wgRequest->getVal( 'action' ) == 'submit') {
35 $page = $wgRequest->getText( 'pages' );
36 $curonly = $wgRequest->getCheck( 'curonly' );
38 # Pre-check the 'current version only' box in the UI
44 header( "Content-type: application/xml; charset=utf-8" );
45 $pages = explode( "\n", $page );
47 $db =& wfGetDB( DB_SLAVE
);
48 $history = $curonly ? MW_EXPORT_CURRENT
: MW_EXPORT_FULL
;
49 $exporter = new WikiExporter( $db, $history );
50 $exporter->openStream();
51 $exporter->pagesByName( $pages );
52 $exporter->closeStream();
56 $wgOut->addWikiText( wfMsg( "exporttext" ) );
57 $titleObj = Title
::makeTitle( NS_SPECIAL
, "Export" );
58 $action = $titleObj->escapeLocalURL( 'action=submit' );
60 <form method='post' action=\"$action\">
61 <input type='hidden' name='action' value='submit' />
62 <textarea name='pages' cols='40' rows='10'></textarea><br />
63 <label><input type='checkbox' name='curonly' value='true' checked='checked' />
64 " . wfMsgHtml( 'exportcuronly' ) . "</label><br />
65 <input type='submit' />
70 define( 'MW_EXPORT_FULL', 0 );
71 define( 'MW_EXPORT_CURRENT', 1 );
73 define( 'MW_EXPORT_BUFFER', 0 );
74 define( 'MW_EXPORT_STREAM', 1 );
78 * @subpackage SpecialPage
81 var $pageCallback = null;
82 var $revCallback = null;
85 * If using MW_EXPORT_STREAM to stream a large amount of data,
86 * provide a database connection which is not managed by
87 * LoadBalancer to read from: some history blob types will
88 * make additional queries to pull source data while the
89 * main query is still running.
92 * @param int $history one of MW_EXPORT_FULL or MW_EXPORT_CURRENT
93 * @param int $buffer one of MW_EXPORT_BUFFER or MW_EXPORT_STREAM
95 function WikiExporter( &$db, $history = MW_EXPORT_CURRENT
,
96 $buffer = MW_EXPORT_BUFFER
) {
98 $this->history
= $history;
99 $this->buffer
= $buffer;
103 * Set a callback to be called after each page in the output
104 * stream is closed. The callback will be passed a database row
105 * object with the last revision output.
107 * A set callback can be removed by passing null here.
109 * @param mixed $callback
111 function setPageCallback( $callback ) {
112 $this->pageCallback
= $callback;
116 * Set a callback to be called after each revision in the output
117 * stream is closed. The callback will be passed a database row
118 * object with the revision data.
120 * A set callback can be removed by passing null here.
122 * @param mixed $callback
124 function setRevisionCallback( $callback ) {
125 $this->revCallback
= $callback;
129 * Returns the export schema version.
132 function schemaVersion() {
137 * Opens the XML output stream's root <mediawiki> element.
138 * This does not include an xml directive, so is safe to include
139 * as a subelement in a larger XML stream. Namespace and XML Schema
140 * references are included.
142 * To capture the stream to a string, use PHP's output buffering
143 * functions. Output will be encoded in UTF-8.
145 function openStream() {
146 global $wgContLanguageCode;
147 $ver = $this->schemaVersion();
148 print wfElement( 'mediawiki', array(
149 'xmlns' => "http://www.mediawiki.org/xml/export-$ver/",
150 'xmlns:xsi' => "http://www.w3.org/2001/XMLSchema-instance",
151 'xsi:schemaLocation' => "http://www.mediawiki.org/xml/export-$ver/ " .
152 "http://www.mediawiki.org/xml/export-$ver.xsd",
154 'xml:lang' => $wgContLanguageCode ),
159 function siteInfo() {
164 $this->caseSetting(),
165 $this->namespaces() );
166 print "<siteinfo>\n";
167 foreach( $info as $item ) {
170 print "</siteinfo>\n";
173 function sitename() {
175 return wfElement( 'sitename', array(), $wgSitename );
178 function generator() {
180 return wfElement( 'generator', array(), "MediaWiki $wgVersion" );
183 function homelink() {
184 $page = Title
::newFromText( wfMsgForContent( 'mainpage' ) );
185 return wfElement( 'base', array(), $page->getFullUrl() );
188 function caseSetting() {
189 global $wgCapitalLinks;
190 // "case-insensitive" option is reserved for future
191 $sensitivity = $wgCapitalLinks ?
'first-letter' : 'case-sensitive';
192 return wfElement( 'case', array(), $sensitivity );
195 function namespaces() {
197 $spaces = "<namespaces>\n";
198 foreach( $wgContLang->getFormattedNamespaces() as $ns => $title ) {
199 $spaces .= ' ' . wfElement( 'namespace', array( 'key' => $ns ), $title ) . "\n";
201 $spaces .= " </namespaces>";
206 * Closes the output stream with the closing root element.
207 * Call when finished dumping things.
209 function closeStream() {
210 print "</mediawiki>\n";
214 * Dumps a series of page and revision records for all pages
215 * in the database, either including complete history or only
216 * the most recent version.
218 function allPages() {
219 return $this->dumpFrom( '' );
223 * Dumps a series of page and revision records for those pages
224 * in the database falling within the page_id range given.
225 * @param int $start Inclusive lower limit (this id is included)
226 * @param int $end Exclusive upper limit (this id is not included)
227 * If 0, no upper limit.
229 function pagesByRange( $start, $end ) {
230 $condition = 'page_id >= ' . intval( $start );
232 $condition .= ' AND page_id < ' . intval( $end );
234 return $this->dumpFrom( $condition );
238 * @param Title $title
240 function pageByTitle( $title ) {
241 return $this->dumpFrom(
242 'page_namespace=' . $title->getNamespace() .
243 ' AND page_title=' . $this->db
->addQuotes( $title->getDbKey() ) );
246 function pageByName( $name ) {
247 $title = Title
::newFromText( $name );
248 if( is_null( $title ) ) {
249 return new WikiError( "Can't export invalid title" );
251 return $this->pageByTitle( $title );
255 function pagesByName( $names ) {
256 foreach( $names as $name ) {
257 $this->pageByName( $name );
262 // -------------------- private implementation below --------------------
264 function dumpFrom( $cond = '' ) {
265 $fname = 'WikiExporter::dumpFrom';
266 wfProfileIn( $fname );
268 $page = $this->db
->tableName( 'page' );
269 $revision = $this->db
->tableName( 'revision' );
270 $text = $this->db
->tableName( 'text' );
272 if( $this->history
== MW_EXPORT_FULL
) {
273 $join = 'page_id=rev_page';
274 } elseif( $this->history
== MW_EXPORT_CURRENT
) {
275 $join = 'page_id=rev_page AND page_latest=rev_id';
277 wfProfileOut( $fname );
278 return new WikiError( "$fname given invalid history dump type." );
280 $where = ( $cond == '' ) ?
'' : "$cond AND";
282 if( $this->buffer
== MW_EXPORT_STREAM
) {
283 $prev = $this->db
->bufferResults( false );
286 // Optimization hack for full-database dump
287 $pageindex = 'FORCE INDEX (PRIMARY)';
288 $revindex = 'FORCE INDEX(page_timestamp)';
293 $result = $this->db
->query(
298 WHERE $where $join AND rev_text_id=old_id
299 ORDER BY page_id", $fname );
300 $wrapper = $this->db
->resultObject( $result );
301 $this->outputStream( $wrapper );
303 if( $this->buffer
== MW_EXPORT_STREAM
) {
304 $this->db
->bufferResults( $prev );
307 wfProfileOut( $fname );
311 * Runs through a query result set dumping page and revision records.
312 * The result set should be sorted/grouped by page to avoid duplicate
313 * page records in the output.
315 * The result set will be freed once complete. Should be safe for
316 * streaming (non-buffered) queries, as long as it was made on a
317 * separate database connection not managed by LoadBalancer; some
318 * blob storage types will make queries to pull source data.
320 * @param ResultWrapper $resultset
323 function outputStream( $resultset ) {
325 while( $row = $resultset->fetchObject() ) {
326 if( is_null( $last ) ||
327 $last->page_namespace
!= $row->page_namespace ||
328 $last->page_title
!= $row->page_title
) {
329 if( isset( $last ) ) {
330 $this->closePage( $last );
332 $this->openPage( $row );
335 $this->dumpRev( $row );
337 if( isset( $last ) ) {
338 $this->closePage( $last );
344 * Opens a <page> section on the output stream, with data
345 * from the given database row.
350 function openPage( $row ) {
352 $title = Title
::makeTitle( $row->page_namespace
, $row->page_title
);
353 print ' ' . wfElementClean( 'title', array(), $title->getPrefixedText() ) . "\n";
354 print ' ' . wfElement( 'id', array(), strval( $row->page_id
) ) . "\n";
355 if( '' != $row->page_restrictions
) {
356 print ' ' . wfElement( 'restrictions', array(),
357 strval( $row->page_restrictions
) ) . "\n";
362 * Closes a <page> section on the output stream.
363 * If a per-page callback has been set, it will be called
364 * and passed the last database row used for this page.
369 function closePage( $row ) {
371 if( isset( $this->pageCallback
) ) {
372 call_user_func( $this->pageCallback
, $row );
377 * Dumps a <revision> section on the output stream, with
378 * data filled in from the given database row.
383 function dumpRev( $row ) {
384 $fname = 'WikiExporter::dumpRev';
385 wfProfileIn( $fname );
387 print " <revision>\n";
388 print " " . wfElement( 'id', null, strval( $row->rev_id
) ) . "\n";
390 $ts = wfTimestamp2ISO8601( strval( $row->rev_timestamp
) );
391 print " " . wfElement( 'timestamp', null, $ts ) . "\n";
393 print " <contributor>\n";
394 if( $row->rev_user
) {
395 print " " . wfElementClean( 'username', null, strval( $row->rev_user_text
) ) . "\n";
396 print " " . wfElement( 'id', null, strval( $row->rev_user
) ) . "\n";
398 print " " . wfElementClean( 'ip', null, strval( $row->rev_user_text
) ) . "\n";
400 print " </contributor>\n";
402 if( $row->rev_minor_edit
) {
405 if( $row->rev_comment
!= '' ) {
406 print " " . wfElementClean( 'comment', null, strval( $row->rev_comment
) ) . "\n";
409 $text = strval( Revision
::getRevisionText( $row ) );
410 print " " . wfElementClean( 'text', array( 'xml:space' => 'preserve' ), $text ) . "\n";
412 print " </revision>\n";
414 wfProfileOut( $fname );
416 if( isset( $this->revCallback
) ) {
417 call_user_func( $this->revCallback
, $row );
423 function wfTimestamp2ISO8601( $ts ) {
424 #2003-08-05T18:30:02Z
425 return preg_replace( '/^(....)(..)(..)(..)(..)(..)$/', '$1-$2-$3T$4:$5:$6Z', wfTimestamp( TS_MW
, $ts ) );
428 function xmlsafe( $string ) {
430 wfProfileIn( $fname );
433 * The page may contain old data which has not been properly normalized.
434 * Invalid UTF-8 sequences or forbidden control characters will make our
435 * XML output invalid, so be sure to strip them out.
437 $string = UtfNormal
::cleanUp( $string );
439 $string = htmlspecialchars( $string );
440 wfProfileOut( $fname );