includes/SpecialImport.php

   1 <?php
   2 /**
   3  * MediaWiki page data importer
   4  * Copyright (C) 2003,2005 Brion Vibber <brion@pobox.com>
   5  * http://www.mediawiki.org/
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @addtogroup SpecialPage
  23  */
  24
  25 /**
  26  * Constructor
  27  */
  28 function wfSpecialImport( $page = '' ) {
  29         global $wgUser, $wgOut, $wgRequest, $wgTitle, $wgImportSources;
  30         global $wgImportTargetNamespace;
  31
  32         $interwiki = false;
  33         $articleName = '';
  34         $namespace = $wgImportTargetNamespace;
  35         $frompage = '';
  36         $history = true;
  37
  38         if( $wgRequest->wasPosted() && $wgRequest->getVal( 'action' ) == 'submit') {
  39                 $isUpload = false;
  40                 $articleName = $wgRequest->getText( 'articlename' );
  41                 $namespace = $wgRequest->getIntOrNull( 'namespace' );
  42
  43                 switch( $wgRequest->getVal( "source" ) ) {
  44                 case "upload":
  45                         $isUpload = true;
  46                         if( $wgUser->isAllowed( 'importupload' ) ) {
  47                                 $source = ImportStreamSource::newFromUpload( "xmlimport" );
  48                         } else {
  49                                 return $wgOut->permissionRequired( 'importupload' );
  50                         }
  51                         break;
  52                 case "interwiki":
  53                         $interwiki = $wgRequest->getVal( 'interwiki' );
  54                         $history = $wgRequest->getCheck( 'interwikiHistory' );
  55                         $frompage = $wgRequest->getText( "frompage" );
  56                         $source = ImportStreamSource::newFromInterwiki(
  57                                 $interwiki,
  58                                 $frompage,
  59                                 $history );
  60                         break;
  61                 default:
  62                         $source = new WikiErrorMsg( "importunknownsource" );
  63                 }
  64
  65                 if( WikiError::isError( $source ) ) {
  66                         $wgOut->addWikiText( '<p class="error">' . wfMsg( "importfailed", wfEscapeWikiText( $source->getMessage() ) ) . '</p>' );
  67                 } else {
  68                         $wgOut->addWikiText( wfMsg( "importstart" ) );
  69
  70                         $importer = new WikiImporter( $source );
  71                         if( !is_null( $namespace ) ) {
  72                                 $importer->setTargetNamespace( $namespace );
  73                         }
  74                         if( $articleName !== '' ) {
  75                                 $importer->setTargetArticleName( $articleName );
  76                         }
  77                         $reporter = new ImportReporter( $importer, $isUpload, $interwiki, $frompage );
  78
  79                         $reporter->open();
  80                         $result = $importer->doImport();
  81                         $resultCount = $reporter->close();
  82
  83                         if( WikiError::isError( $result ) ) {
  84                                 # No source or XML parse error
  85                                 $wgOut->addWikiText( '<p class="error">' . wfMsg( "importfailed", wfEscapeWikiText( $result->getMessage() ) ) . '</p>' );
  86                         } elseif( WikiError::isError( $resultCount ) ) {
  87                                 # Zero revisions
  88                                 $wgOut->addWikiText( '<p class="error">' . wfMsg( "importfailed", wfEscapeWikiText( $resultCount->getMessage() ) ) . '</p>' );
  89                         } else {
  90                                 # Success!
  91                                 $wgOut->addWikiText( wfMsg( "importsuccess" ) );
  92                         }
  93                         $wgOut->addWikiText( '<hr />' );
  94                 }
  95         }
  96
  97         $action = $wgTitle->escapeLocalUrl( 'action=submit' );
  98
  99         if( $wgUser->isAllowed( 'importupload' ) ) {
 100                 $wgOut->addWikiText( wfMsg( "importtext" ) );
 101                 $wgOut->addHTML(
 102                         Xml::openElement( 'fieldset' ).
 103                         Xml::element( 'legend', null, wfMsg( 'upload' ) ) .
 104                         Xml::openElement( 'form', array( 'enctype' => 'multipart/form-data', 'method' => 'post', 'action' => $action ) ) .
 105                         Xml::hidden( 'action', 'submit' ) .
 106                         Xml::hidden( 'source', 'upload' ) .
 107                         Xml::hidden( 'MAX_FILE_SIZE', '2000000' ) .
 108                         "<input type='file' name='xmlimport' value='' size='30' />" . // No Xml function for type=file? Todo?
 109                         Xml::submitButton( wfMsg( 'uploadbtn' ) ) .
 110                         Xml::closeElement( 'form' ) .
 111                         Xml::closeElement( 'fieldset' ) );
 112         } else {
 113                 if( empty( $wgImportSources ) ) {
 114                         $wgOut->addWikiText( wfMsg( 'importnosources' ) );
 115                 }
 116         }
 117
 118         if( !empty( $wgImportSources ) ) {
 119                 $wgOut->addHTML(
 120                         Xml::openElement( 'fieldset' ) .
 121                         Xml::element( 'legend', null, wfMsg( 'importinterwiki' ) ) .
 122                         Xml::openElement( 'form', array( 'method' => 'post', 'action' => $action ) ) .
 123                         wfMsgExt( 'import-interwiki-text', array( 'parse' ) ) .
 124                         Xml::hidden( 'action', 'submit' ) .
 125                         Xml::hidden( 'source', 'interwiki' ) .
 126                         Xml::openElement( 'table' ) .
 127                         "<tr>
 128                                 <td>" .
 129                                         Xml::openElement( 'select', array( 'name' => 'interwiki' ) ) );
 130                 foreach( $wgImportSources as $prefix ) {
 131                         $iw = htmlspecialchars( $prefix );
 132                         $selected = ($interwiki === $prefix) ? ' selected="selected"' : '';
 133                         $wgOut->addHTML( Xml::option( $iw, $iw, $selected ) );
 134                 }
 135                 $wgOut->addHTML(
 136                                         Xml::closeElement( 'select' ) .
 137                                 "</td>
 138                                 <td>" .
 139                                         Xml::input( 'frompage', 50, $frompage ) .
 140                                 "</td>
 141                         </tr>
 142                         <tr>
 143                                 <td></td>
 144                                 <td>" .
 145                                         Xml::checkLabel( wfMsg( 'import-interwiki-history' ), 'interwikiHistory', 'interwikiHistory', $history ) .
 146                                 "</td>
 147                         </tr>
 148                         <tr>
 149                                 <td></td>
 150                                 <td>" .
 151                                         Xml::inputLabel( wfMsg( 'import-articlename' ), 'articlename', 'articlename', 50, $articleName ) .
 152                                 "</td>
 153                         </tr>
 154                         <tr>
 155                                 <td></td>
 156                                 <td>" .
 157                                         Xml::label( wfMsg( 'import-interwiki-namespace' ), 'namespace' ) .
 158                                         Xml::namespaceSelector( $namespace, '' ) .
 159                                 "</td>
 160                         </tr>
 161                         <tr>
 162                                 <td></td>
 163                                 <td>" .
 164                                         Xml::submitButton( wfMsg( 'import-interwiki-submit' ) ) .
 165                                 "</td>
 166                         </tr>" .
 167                         Xml::closeElement( 'table' ).
 168                         Xml::closeElement( 'form' ) .
 169                         Xml::closeElement( 'fieldset' ) );
 170         }
 171 }
 172
 173 /**
 174  * Reporting callback
 175  * @addtogroup SpecialPage
 176  */
 177 class ImportReporter {
 178         function __construct( $importer, $upload, $interwiki, $frompage ) {
 179                 $importer->setPageOutCallback( array( $this, 'reportPage' ) );
 180                 $this->mPageCount = 0;
 181                 $this->mIsUpload = $upload;
 182                 $this->mInterwiki = $interwiki;
 183                 $this->mFrompage = $frompage;
 184         }
 185
 186         function open() {
 187                 global $wgOut;
 188                 $wgOut->addHtml( "<ul>\n" );
 189         }
 190
 191         function reportPage( $title, $origTitle, $revisionCount, $successCount ) {
 192                 global $wgOut, $wgUser, $wgLang, $wgContLang;
 193
 194                 $skin = $wgUser->getSkin();
 195
 196                 $this->mPageCount++;
 197
 198                 $localCount = $wgLang->formatNum( $successCount );
 199                 $contentCount = $wgContLang->formatNum( $successCount );
 200
 201                 if( $successCount > 0 ) {
 202                         $wgOut->addHtml( "<li>" . $skin->makeKnownLinkObj( $title ) .
 203                                 " " .
 204                                 wfMsgExt( 'import-revision-count', array( 'parsemag', 'escape' ), $localCount ) .
 205                                 "</li>\n" );
 206
 207                         $log = new LogPage( 'import' );
 208                         if( $this->mIsUpload ) {
 209                                 $detail = wfMsgForContent( 'import-logentry-upload-detail',
 210                                         $contentCount );
 211                                 $log->addEntry( 'upload', $title, $detail );
 212                         } else {
 213                                 // Show the source article name in log
 214                                 $origin = $this->mFrompage != $origTitle->getPrefixedText()
 215                                         ? $this->mFrompage
 216                                         : $origTitle->getPrefixedText();
 217                                 $interwiki = '[[:' . $this->mInterwiki . ':' . $origin . ']]';
 218                                 $detail = wfMsgForContent( 'import-logentry-interwiki-detail',
 219                                         $contentCount, $interwiki );
 220                                 $log->addEntry( 'interwiki', $title, $detail );
 221                         }
 222
 223                         $comment = $detail; // quick
 224                         $dbw = wfGetDB( DB_MASTER );
 225                         $nullRevision = Revision::newNullRevision(
 226                                 $dbw, $title->getArticleId(), $comment, true );
 227                         $nullRevision->insertOn( $dbw );
 228                 } else {
 229                         $wgOut->addHtml( '<li>' . wfMsgExt( 'import-nonewrevisions', array( 'parse', ) ) . '</li>' );
 230                 }
 231
 232         }
 233
 234         function close() {
 235                 global $wgOut;
 236                 if( $this->mPageCount == 0 ) {
 237                         $wgOut->addHtml( "</ul>\n" );
 238                         return new WikiErrorMsg( "importnopages" );
 239                 }
 240                 $wgOut->addHtml( "</ul>\n" );
 241
 242                 return $this->mPageCount;
 243         }
 244 }
 245
 246 /**
 247  *
 248  * @addtogroup SpecialPage
 249  */
 250 class WikiRevision {
 251         var $title = null;
 252         var $id = 0;
 253         var $timestamp = "20010115000000";
 254         var $user = 0;
 255         var $user_text = "";
 256         var $text = "";
 257         var $comment = "";
 258         var $minor = false;
 259
 260         function setTitle( $title ) {
 261                 if( is_object( $title ) ) {
 262                         $this->title = $title;
 263                 } elseif( is_null( $title ) ) {
 264                         throw new MWException( "WikiRevision given a null title in import. You may need to adjust \$wgLegalTitleChars." );
 265                 } else {
 266                         throw new MWException( "WikiRevision given non-object title in import." );
 267                 }
 268         }
 269
 270         function setID( $id ) {
 271                 $this->id = $id;
 272         }
 273
 274         function setTimestamp( $ts ) {
 275                 # 2003-08-05T18:30:02Z
 276                 $this->timestamp = wfTimestamp( TS_MW, $ts );
 277         }
 278
 279         function setUsername( $user ) {
 280                 $this->user_text = $user;
 281         }
 282
 283         function setUserIP( $ip ) {
 284                 $this->user_text = $ip;
 285         }
 286
 287         function setText( $text ) {
 288                 $this->text = $text;
 289         }
 290
 291         function setComment( $text ) {
 292                 $this->comment = $text;
 293         }
 294
 295         function setMinor( $minor ) {
 296                 $this->minor = (bool)$minor;
 297         }
 298
 299         function getTitle() {
 300                 return $this->title;
 301         }
 302
 303         function getID() {
 304                 return $this->id;
 305         }
 306
 307         function getTimestamp() {
 308                 return $this->timestamp;
 309         }
 310
 311         function getUser() {
 312                 return $this->user_text;
 313         }
 314
 315         function getText() {
 316                 return $this->text;
 317         }
 318
 319         function getComment() {
 320                 return $this->comment;
 321         }
 322
 323         function getMinor() {
 324                 return $this->minor;
 325         }
 326
 327         function importOldRevision() {
 328                 $dbw = wfGetDB( DB_MASTER );
 329
 330                 # Sneak a single revision into place
 331                 $user = User::newFromName( $this->getUser() );
 332                 if( $user ) {
 333                         $userId = intval( $user->getId() );
 334                         $userText = $user->getName();
 335                 } else {
 336                         $userId = 0;
 337                         $userText = $this->getUser();
 338                 }
 339
 340                 // avoid memory leak...?
 341                 $linkCache =& LinkCache::singleton();
 342                 $linkCache->clear();
 343
 344                 $article = new Article( $this->title );
 345                 $pageId = $article->getId();
 346                 if( $pageId == 0 ) {
 347                         # must create the page...
 348                         $pageId = $article->insertOn( $dbw );
 349                         $created = true;
 350                 } else {
 351                         $created = false;
 352
 353                         $prior = Revision::loadFromTimestamp( $dbw, $this->title, $this->timestamp );
 354                         if( !is_null( $prior ) ) {
 355                                 // FIXME: this could fail slightly for multiple matches :P
 356                                 wfDebug( __METHOD__ . ": skipping existing revision for [[" .
 357                                         $this->title->getPrefixedText() . "]], timestamp " .
 358                                         $this->timestamp . "\n" );
 359                                 return false;
 360                         }
 361                 }
 362
 363                 # FIXME: Use original rev_id optionally
 364                 # FIXME: blah blah blah
 365
 366                 #if( $numrows > 0 ) {
 367                 #       return wfMsg( "importhistoryconflict" );
 368                 #}
 369
 370                 # Insert the row
 371                 $revision = new Revision( array(
 372                         'page'       => $pageId,
 373                         'text'       => $this->getText(),
 374                         'comment'    => $this->getComment(),
 375                         'user'       => $userId,
 376                         'user_text'  => $userText,
 377                         'timestamp'  => $this->timestamp,
 378                         'minor_edit' => $this->minor,
 379                         ) );
 380                 $revId = $revision->insertOn( $dbw );
 381                 $changed = $article->updateIfNewerOn( $dbw, $revision );
 382
 383                 if( $created ) {
 384                         wfDebug( __METHOD__ . ": running onArticleCreate\n" );
 385                         Article::onArticleCreate( $this->title );
 386
 387                         wfDebug( __METHOD__ . ": running create updates\n" );
 388                         $article->createUpdates( $revision );
 389
 390                 } elseif( $changed ) {
 391                         wfDebug( __METHOD__ . ": running onArticleEdit\n" );
 392                         Article::onArticleEdit( $this->title );
 393
 394                         wfDebug( __METHOD__ . ": running edit updates\n" );
 395                         $article->editUpdates(
 396                                 $this->getText(),
 397                                 $this->getComment(),
 398                                 $this->minor,
 399                                 $this->timestamp,
 400                                 $revId );
 401                 }
 402
 403                 return true;
 404         }
 405
 406 }
 407
 408 /**
 409  * implements Special:Import
 410  * @addtogroup SpecialPage
 411  */
 412 class WikiImporter {
 413         var $mSource = null;
 414         var $mPageCallback = null;
 415         var $mPageOutCallback = null;
 416         var $mRevisionCallback = null;
 417         var $mTargetNamespace = null;
 418         var $mTargetArticleName = null;
 419         var $lastfield;
 420
 421         function WikiImporter( $source ) {
 422                 $this->setRevisionCallback( array( &$this, "importRevision" ) );
 423                 $this->mSource = $source;
 424         }
 425
 426         function throwXmlError( $err ) {
 427                 $this->debug( "FAILURE: $err" );
 428                 wfDebug( "WikiImporter XML error: $err\n" );
 429         }
 430
 431         # --------------
 432
 433         function doImport() {
 434                 if( empty( $this->mSource ) ) {
 435                         return new WikiErrorMsg( "importnotext" );
 436                 }
 437
 438                 $parser = xml_parser_create( "UTF-8" );
 439
 440                 # case folding violates XML standard, turn it off
 441                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 442
 443                 xml_set_object( $parser, $this );
 444                 xml_set_element_handler( $parser, "in_start", "" );
 445
 446                 $offset = 0; // for context extraction on error reporting
 447                 do {
 448                         $chunk = $this->mSource->readChunk();
 449                         if( !xml_parse( $parser, $chunk, $this->mSource->atEnd() ) ) {
 450                                 wfDebug( "WikiImporter::doImport encountered XML parsing error\n" );
 451                                 return new WikiXmlError( $parser, wfMsgHtml( 'import-parse-failure' ), $chunk, $offset );
 452                         }
 453                         $offset += strlen( $chunk );
 454                 } while( $chunk !== false && !$this->mSource->atEnd() );
 455                 xml_parser_free( $parser );
 456
 457                 return true;
 458         }
 459
 460         function debug( $data ) {
 461                 #wfDebug( "IMPORT: $data\n" );
 462         }
 463
 464         function notice( $data ) {
 465                 global $wgCommandLineMode;
 466                 if( $wgCommandLineMode ) {
 467                         print "$data\n";
 468                 } else {
 469                         global $wgOut;
 470                         $wgOut->addHTML( "<li>" . htmlspecialchars( $data ) . "</li>\n" );
 471                 }
 472         }
 473
 474         /**
 475          * Sets the action to perform as each new page in the stream is reached.
 476          * @param callable $callback
 477          * @return callable
 478          */
 479         function setPageCallback( $callback ) {
 480                 $previous = $this->mPageCallback;
 481                 $this->mPageCallback = $callback;
 482                 return $previous;
 483         }
 484
 485         /**
 486          * Sets the action to perform as each page in the stream is completed.
 487          * Callback accepts the page title (as a Title object), a second object
 488          * with the original title form (in case it's been overridden into a
 489          * local namespace), and a count of revisions.
 490          *
 491          * @param callable $callback
 492          * @return callable
 493          */
 494         function setPageOutCallback( $callback ) {
 495                 $previous = $this->mPageOutCallback;
 496                 $this->mPageOutCallback = $callback;
 497                 return $previous;
 498         }
 499
 500         /**
 501          * Sets the action to perform as each page revision is reached.
 502          * @param callable $callback
 503          * @return callable
 504          */
 505         function setRevisionCallback( $callback ) {
 506                 $previous = $this->mRevisionCallback;
 507                 $this->mRevisionCallback = $callback;
 508                 return $previous;
 509         }
 510
 511         /**
 512          * Set a target namespace to override the defaults
 513          */
 514         function setTargetNamespace( $namespace ) {
 515                 if( is_null( $namespace ) ) {
 516                         // Don't override namespaces
 517                         $this->mTargetNamespace = null;
 518                 } elseif( $namespace >= 0 ) {
 519                         // FIXME: Check for validity
 520                         $this->mTargetNamespace = intval( $namespace );
 521                 } else {
 522                         return false;
 523                 }
 524         }
 525
 526         /**
 527          * Set a target articlename to override the defaults
 528          */
 529         function setTargetArticleName( $articleName ) {
 530                 $this->mTargetArticleName = $articleName;
 531         }
 532
 533         /**
 534          * Default per-revision callback, performs the import.
 535          * @param WikiRevision $revision
 536          * @private
 537          */
 538         function importRevision( &$revision ) {
 539                 $dbw = wfGetDB( DB_MASTER );
 540                 return $dbw->deadlockLoop( array( &$revision, 'importOldRevision' ) );
 541         }
 542
 543         /**
 544          * Alternate per-revision callback, for debugging.
 545          * @param WikiRevision $revision
 546          * @private
 547          */
 548         function debugRevisionHandler( &$revision ) {
 549                 $this->debug( "Got revision:" );
 550                 if( is_object( $revision->title ) ) {
 551                         $this->debug( "-- Title: " . $revision->title->getPrefixedText() );
 552                 } else {
 553                         $this->debug( "-- Title: <invalid>" );
 554                 }
 555                 $this->debug( "-- User: " . $revision->user_text );
 556                 $this->debug( "-- Timestamp: " . $revision->timestamp );
 557                 $this->debug( "-- Comment: " . $revision->comment );
 558                 $this->debug( "-- Text: " . $revision->text );
 559         }
 560
 561         /**
 562          * Notify the callback function when a new <page> is reached.
 563          * @param Title $title
 564          * @private
 565          */
 566         function pageCallback( $title ) {
 567                 if( is_callable( $this->mPageCallback ) ) {
 568                         call_user_func( $this->mPageCallback, $title );
 569                 }
 570         }
 571
 572         /**
 573          * Notify the callback function when a </page> is closed.
 574          * @param Title $title
 575          * @param Title $origTitle
 576          * @param int $revisionCount
 577          * @param int $successCount number of revisions for which callback returned true
 578          * @private
 579          */
 580         function pageOutCallback( $title, $origTitle, $revisionCount, $successCount ) {
 581                 if( is_callable( $this->mPageOutCallback ) ) {
 582                         call_user_func( $this->mPageOutCallback, $title, $origTitle,
 583                                 $revisionCount, $successCount );
 584                 }
 585         }
 586
 587
 588         # XML parser callbacks from here out -- beware!
 589         function donothing( $parser, $x, $y="" ) {
 590                 #$this->debug( "donothing" );
 591         }
 592
 593         function in_start( $parser, $name, $attribs ) {
 594                 $this->debug( "in_start $name" );
 595                 if( $name != "mediawiki" ) {
 596                         return $this->throwXMLerror( "Expected <mediawiki>, got <$name>" );
 597                 }
 598                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 599         }
 600
 601         function in_mediawiki( $parser, $name, $attribs ) {
 602                 $this->debug( "in_mediawiki $name" );
 603                 if( $name == 'siteinfo' ) {
 604                         xml_set_element_handler( $parser, "in_siteinfo", "out_siteinfo" );
 605                 } elseif( $name == 'page' ) {
 606                         $this->workRevisionCount = 0;
 607                         $this->workSuccessCount = 0;
 608                         xml_set_element_handler( $parser, "in_page", "out_page" );
 609                 } else {
 610                         return $this->throwXMLerror( "Expected <page>, got <$name>" );
 611                 }
 612         }
 613         function out_mediawiki( $parser, $name ) {
 614                 $this->debug( "out_mediawiki $name" );
 615                 if( $name != "mediawiki" ) {
 616                         return $this->throwXMLerror( "Expected </mediawiki>, got </$name>" );
 617                 }
 618                 xml_set_element_handler( $parser, "donothing", "donothing" );
 619         }
 620
 621
 622         function in_siteinfo( $parser, $name, $attribs ) {
 623                 // no-ops for now
 624                 $this->debug( "in_siteinfo $name" );
 625                 switch( $name ) {
 626                 case "sitename":
 627                 case "base":
 628                 case "generator":
 629                 case "case":
 630                 case "namespaces":
 631                 case "namespace":
 632                         break;
 633                 default:
 634                         return $this->throwXMLerror( "Element <$name> not allowed in <siteinfo>." );
 635                 }
 636         }
 637
 638         function out_siteinfo( $parser, $name ) {
 639                 if( $name == "siteinfo" ) {
 640                         xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 641                 }
 642         }
 643
 644
 645         function in_page( $parser, $name, $attribs ) {
 646                 $this->debug( "in_page $name" );
 647                 switch( $name ) {
 648                 case "id":
 649                 case "title":
 650                 case "restrictions":
 651                         $this->appendfield = $name;
 652                         $this->appenddata = "";
 653                         $this->parenttag = "page";
 654                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 655                         xml_set_character_data_handler( $parser, "char_append" );
 656                         break;
 657                 case "revision":
 658                         if( is_object( $this->pageTitle ) ) {
 659                                 $this->workRevision = new WikiRevision;
 660                                 $this->workRevision->setTitle( $this->pageTitle );
 661                                 $this->workRevisionCount++;
 662                         } else {
 663                                 // Skipping items due to invalid page title
 664                                 $this->workRevision = null;
 665                         }
 666                         xml_set_element_handler( $parser, "in_revision", "out_revision" );
 667                         break;
 668                 default:
 669                         return $this->throwXMLerror( "Element <$name> not allowed in a <page>." );
 670                 }
 671         }
 672
 673         function out_page( $parser, $name ) {
 674                 $this->debug( "out_page $name" );
 675                 if( $name != "page" ) {
 676                         return $this->throwXMLerror( "Expected </page>, got </$name>" );
 677                 }
 678                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 679
 680                 $this->pageOutCallback( $this->pageTitle, $this->origTitle,
 681                         $this->workRevisionCount, $this->workSuccessCount );
 682
 683                 $this->workTitle = null;
 684                 $this->workRevision = null;
 685                 $this->workRevisionCount = 0;
 686                 $this->workSuccessCount = 0;
 687                 $this->pageTitle = null;
 688                 $this->origTitle = null;
 689         }
 690
 691         function in_nothing( $parser, $name, $attribs ) {
 692                 $this->debug( "in_nothing $name" );
 693                 return $this->throwXMLerror( "No child elements allowed here; got <$name>" );
 694         }
 695         function char_append( $parser, $data ) {
 696                 $this->debug( "char_append '$data'" );
 697                 $this->appenddata .= $data;
 698         }
 699         function out_append( $parser, $name ) {
 700                 $this->debug( "out_append $name" );
 701                 if( $name != $this->appendfield ) {
 702                         return $this->throwXMLerror( "Expected </{$this->appendfield}>, got </$name>" );
 703                 }
 704                 xml_set_element_handler( $parser, "in_$this->parenttag", "out_$this->parenttag" );
 705                 xml_set_character_data_handler( $parser, "donothing" );
 706
 707                 switch( $this->appendfield ) {
 708                 case "title":
 709                         $this->workTitle = $this->appenddata;
 710
 711                         if( $this->mTargetArticleName != '' ) {
 712                                 // Import article with a different target article name
 713                                 $this->workTitle = $this->mTargetArticleName;
 714                         }
 715                         $this->origTitle = Title::newFromText( $this->workTitle );
 716                         if( !is_null( $this->mTargetNamespace ) && !is_null( $this->origTitle ) ) {
 717                                 $this->pageTitle = Title::makeTitle( $this->mTargetNamespace,
 718                                         $this->origTitle->getDbKey() );
 719                         } else {
 720                                 $this->pageTitle = Title::newFromText( $this->workTitle );
 721                         }
 722                         if( is_null( $this->pageTitle ) ) {
 723                                 // Invalid page title? Ignore the page
 724                                 $this->notice( "Skipping invalid page title '$this->workTitle'" );
 725                         } else {
 726                                 $this->pageCallback( $this->workTitle );
 727                         }
 728                         break;
 729                 case "id":
 730                         if ( $this->parenttag == 'revision' ) {
 731                                 if( $this->workRevision )
 732                                         $this->workRevision->setID( $this->appenddata );
 733                         }
 734                         break;
 735                 case "text":
 736                         if( $this->workRevision )
 737                                 $this->workRevision->setText( $this->appenddata );
 738                         break;
 739                 case "username":
 740                         if( $this->workRevision )
 741                                 $this->workRevision->setUsername( $this->appenddata );
 742                         break;
 743                 case "ip":
 744                         if( $this->workRevision )
 745                                 $this->workRevision->setUserIP( $this->appenddata );
 746                         break;
 747                 case "timestamp":
 748                         if( $this->workRevision )
 749                                 $this->workRevision->setTimestamp( $this->appenddata );
 750                         break;
 751                 case "comment":
 752                         if( $this->workRevision )
 753                                 $this->workRevision->setComment( $this->appenddata );
 754                         break;
 755                 case "minor":
 756                         if( $this->workRevision )
 757                                 $this->workRevision->setMinor( true );
 758                         break;
 759                 default:
 760                         $this->debug( "Bad append: {$this->appendfield}" );
 761                 }
 762                 $this->appendfield = "";
 763                 $this->appenddata = "";
 764         }
 765
 766         function in_revision( $parser, $name, $attribs ) {
 767                 $this->debug( "in_revision $name" );
 768                 switch( $name ) {
 769                 case "id":
 770                 case "timestamp":
 771                 case "comment":
 772                 case "minor":
 773                 case "text":
 774                         $this->parenttag = "revision";
 775                         $this->appendfield = $name;
 776                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 777                         xml_set_character_data_handler( $parser, "char_append" );
 778                         break;
 779                 case "contributor":
 780                         xml_set_element_handler( $parser, "in_contributor", "out_contributor" );
 781                         break;
 782                 default:
 783                         return $this->throwXMLerror( "Element <$name> not allowed in a <revision>." );
 784                 }
 785         }
 786
 787         function out_revision( $parser, $name ) {
 788                 $this->debug( "out_revision $name" );
 789                 if( $name != "revision" ) {
 790                         return $this->throwXMLerror( "Expected </revision>, got </$name>" );
 791                 }
 792                 xml_set_element_handler( $parser, "in_page", "out_page" );
 793
 794                 if( $this->workRevision ) {
 795                         $ok = call_user_func_array( $this->mRevisionCallback,
 796                                 array( &$this->workRevision, &$this ) );
 797                         if( $ok ) {
 798                                 $this->workSuccessCount++;
 799                         }
 800                 }
 801         }
 802
 803         function in_contributor( $parser, $name, $attribs ) {
 804                 $this->debug( "in_contributor $name" );
 805                 switch( $name ) {
 806                 case "username":
 807                 case "ip":
 808                 case "id":
 809                         $this->parenttag = "contributor";
 810                         $this->appendfield = $name;
 811                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 812                         xml_set_character_data_handler( $parser, "char_append" );
 813                         break;
 814                 default:
 815                         $this->throwXMLerror( "Invalid tag <$name> in <contributor>" );
 816                 }
 817         }
 818
 819         function out_contributor( $parser, $name ) {
 820                 $this->debug( "out_contributor $name" );
 821                 if( $name != "contributor" ) {
 822                         return $this->throwXMLerror( "Expected </contributor>, got </$name>" );
 823                 }
 824                 xml_set_element_handler( $parser, "in_revision", "out_revision" );
 825         }
 826
 827 }
 828
 829 /**
 830  * @todo document (e.g. one-sentence class description).
 831  * @addtogroup SpecialPage
 832  */
 833 class ImportStringSource {
 834         function ImportStringSource( $string ) {
 835                 $this->mString = $string;
 836                 $this->mRead = false;
 837         }
 838
 839         function atEnd() {
 840                 return $this->mRead;
 841         }
 842
 843         function readChunk() {
 844                 if( $this->atEnd() ) {
 845                         return false;
 846                 } else {
 847                         $this->mRead = true;
 848                         return $this->mString;
 849                 }
 850         }
 851 }
 852
 853 /**
 854  * @todo document (e.g. one-sentence class description).
 855  * @addtogroup SpecialPage
 856  */
 857 class ImportStreamSource {
 858         function ImportStreamSource( $handle ) {
 859                 $this->mHandle = $handle;
 860         }
 861
 862         function atEnd() {
 863                 return feof( $this->mHandle );
 864         }
 865
 866         function readChunk() {
 867                 return fread( $this->mHandle, 32768 );
 868         }
 869
 870         static function newFromFile( $filename ) {
 871                 $file = @fopen( $filename, 'rt' );
 872                 if( !$file ) {
 873                         return new WikiErrorMsg( "importcantopen" );
 874                 }
 875                 return new ImportStreamSource( $file );
 876         }
 877
 878         static function newFromUpload( $fieldname = "xmlimport" ) {
 879                 $upload =& $_FILES[$fieldname];
 880
 881                 if( !isset( $upload ) || !$upload['name'] ) {
 882                         return new WikiErrorMsg( 'importnofile' );
 883                 }
 884                 if( !empty( $upload['error'] ) ) {
 885                         return new WikiErrorMsg( 'importuploaderror', $upload['error'] );
 886                 }
 887                 $fname = $upload['tmp_name'];
 888                 if( is_uploaded_file( $fname ) ) {
 889                         return ImportStreamSource::newFromFile( $fname );
 890                 } else {
 891                         return new WikiErrorMsg( 'importnofile' );
 892                 }
 893         }
 894
 895         function newFromURL( $url, $method = 'GET' ) {
 896                 wfDebug( __METHOD__ . ": opening $url\n" );
 897                 # Use the standard HTTP fetch function; it times out
 898                 # quicker and sorts out user-agent problems which might
 899                 # otherwise prevent importing from large sites, such
 900                 # as the Wikimedia cluster, etc.
 901                 $data = Http::request( $method, $url );
 902                 if( $data !== false ) {
 903                         $file = tmpfile();
 904                         fwrite( $file, $data );
 905                         fflush( $file );
 906                         fseek( $file, 0 );
 907                         return new ImportStreamSource( $file );
 908                 } else {
 909                         return new WikiErrorMsg( 'importcantopen' );
 910                 }
 911         }
 912
 913         public static function newFromInterwiki( $interwiki, $page, $history=false ) {
 914                 if( $page == '' ) {
 915                         return new WikiErrorMsg( 'import-noarticle' );
 916                 }
 917                 $link = Title::newFromText( "$interwiki:Special:Export/$page" );
 918                 if( is_null( $link ) || $link->getInterwiki() == '' ) {
 919                         return new WikiErrorMsg( 'importbadinterwiki' );
 920                 } else {
 921                         $params = $history ? 'history=1' : '';
 922                         $url = $link->getFullUrl( $params );
 923                         # For interwikis, use POST to avoid redirects.
 924                         return ImportStreamSource::newFromURL( $url, "POST" );
 925                 }
 926         }
 927 }
 928
 929
 930