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         $namespace = $wgImportTargetNamespace;
  34         $frompage = '';
  35         $history = true;
  36
  37         if( $wgRequest->wasPosted() && $wgRequest->getVal( 'action' ) == 'submit') {
  38                 $isUpload = false;
  39                 $namespace = $wgRequest->getIntOrNull( 'namespace' );
  40
  41                 switch( $wgRequest->getVal( "source" ) ) {
  42                 case "upload":
  43                         $isUpload = true;
  44                         if( $wgUser->isAllowed( 'importupload' ) ) {
  45                                 $source = ImportStreamSource::newFromUpload( "xmlimport" );
  46                         } else {
  47                                 return $wgOut->permissionRequired( 'importupload' );
  48                         }
  49                         break;
  50                 case "interwiki":
  51                         $interwiki = $wgRequest->getVal( 'interwiki' );
  52                         $history = $wgRequest->getCheck( 'interwikiHistory' );
  53                         $frompage = $wgRequest->getText( "frompage" );
  54                         $source = ImportStreamSource::newFromInterwiki(
  55                                 $interwiki,
  56                                 $frompage,
  57                                 $history );
  58                         break;
  59                 default:
  60                         $source = new WikiErrorMsg( "importunknownsource" );
  61                 }
  62
  63                 if( WikiError::isError( $source ) ) {
  64                         $wgOut->addWikiText( wfEscapeWikiText( $source->getMessage() ) );
  65                 } else {
  66                         $wgOut->addWikiText( wfMsg( "importstart" ) );
  67
  68                         $importer = new WikiImporter( $source );
  69                         if( !is_null( $namespace ) ) {
  70                                 $importer->setTargetNamespace( $namespace );
  71                         }
  72                         $reporter = new ImportReporter( $importer, $isUpload, $interwiki );
  73
  74                         $reporter->open();
  75                         $result = $importer->doImport();
  76                         $reporter->close();
  77
  78                         if( WikiError::isError( $result ) ) {
  79                                 $wgOut->addWikiText( wfMsg( "importfailed",
  80                                         wfEscapeWikiText( $result->getMessage() ) ) );
  81                         } else {
  82                                 # Success!
  83                                 $wgOut->addWikiText( wfMsg( "importsuccess" ) );
  84                         }
  85                 }
  86         }
  87
  88         $action = $wgTitle->getLocalUrl( 'action=submit' );
  89
  90         if( $wgUser->isAllowed( 'importupload' ) ) {
  91                 $wgOut->addWikiText( wfMsg( "importtext" ) );
  92                 $wgOut->addHTML( "
  93 <fieldset>
  94         <legend>" . wfMsgHtml('upload') . "</legend>
  95         <form enctype='multipart/form-data' method='post' action=\"$action\">
  96                 <input type='hidden' name='action' value='submit' />
  97                 <input type='hidden' name='source' value='upload' />
  98                 <input type='file' name='xmlimport' value='' size='30' />
  99                 <input type='submit' value=\"" . wfMsgHtml( "uploadbtn" ) . "\" />
 100         </form>
 101 </fieldset>
 102 " );
 103         } else {
 104                 if( empty( $wgImportSources ) ) {
 105                         $wgOut->addWikiText( wfMsg( 'importnosources' ) );
 106                 }
 107         }
 108
 109         if( !empty( $wgImportSources ) ) {
 110                 $wgOut->addHTML( "
 111 <fieldset>
 112         <legend>" . wfMsgHtml('importinterwiki') . "</legend>
 113         <form method='post' action=\"$action\">" .
 114                 $wgOut->parse( wfMsg( 'import-interwiki-text' ) ) . "
 115                 <input type='hidden' name='action' value='submit' />
 116                 <input type='hidden' name='source' value='interwiki' />
 117                 <table>
 118                         <tr>
 119                                 <td>
 120                                         <select name='interwiki'>" );
 121                 foreach( $wgImportSources as $prefix ) {
 122                         $iw = htmlspecialchars( $prefix );
 123                         $selected = ($interwiki === $prefix) ? ' selected="selected"' : '';
 124                         $wgOut->addHTML( "<option value=\"$iw\"$selected>$iw</option>\n" );
 125                 }
 126                 $wgOut->addHTML( "
 127                                         </select>
 128                                 </td>
 129                                 <td>" .
 130                                         wfInput( 'frompage', 50, $frompage ) .
 131                                 "</td>
 132                         </tr>
 133                         <tr>
 134                                 <td></td>
 135                                 <td>" .
 136                                         wfCheckLabel( wfMsg( 'import-interwiki-history' ),
 137                                                 'interwikiHistory', 'interwikiHistory', $history ) .
 138                                 "</td>
 139                         </tr>
 140                         <tr>
 141                                 <td></td>
 142                                 <td>
 143                                         " . wfMsgHtml( 'import-interwiki-namespace' ) . " " .
 144                                                 HTMLnamespaceselector( $namespace, '' ) . "
 145                                 </td>
 146                         </tr>
 147                         <tr>
 148                                 <td></td>
 149                                 <td>" .
 150                                         wfSubmitButton( wfMsg( 'import-interwiki-submit' ) ) .
 151                                 "</td>
 152                         </tr>
 153                 </table>
 154         </form>
 155 </fieldset>
 156 " );
 157         }
 158 }
 159
 160 /**
 161  * Reporting callback
 162  * @addtogroup SpecialPage
 163  */
 164 class ImportReporter {
 165         function __construct( $importer, $upload, $interwiki ) {
 166                 $importer->setPageOutCallback( array( $this, 'reportPage' ) );
 167                 $this->mPageCount = 0;
 168                 $this->mIsUpload = $upload;
 169                 $this->mInterwiki = $interwiki;
 170         }
 171
 172         function open() {
 173                 global $wgOut;
 174                 $wgOut->addHtml( "<ul>\n" );
 175         }
 176
 177         function reportPage( $title, $origTitle, $revisionCount, $successCount ) {
 178                 global $wgOut, $wgUser, $wgLang, $wgContLang;
 179
 180                 $skin = $wgUser->getSkin();
 181
 182                 $this->mPageCount++;
 183
 184                 $localCount = $wgLang->formatNum( $successCount );
 185                 $contentCount = $wgContLang->formatNum( $successCount );
 186
 187                 $wgOut->addHtml( "<li>" . $skin->makeKnownLinkObj( $title ) .
 188                         " " .
 189                         wfMsgExt( 'import-revision-count', array( 'parsemag', 'escape' ), $localCount ) .
 190                         "</li>\n" );
 191
 192                 if( $successCount > 0 ) {
 193                         $log = new LogPage( 'import' );
 194                         if( $this->mIsUpload ) {
 195                                 $detail = wfMsgExt( 'import-logentry-upload-detail', array( 'content', 'parsemag' ),
 196                                         $contentCount );
 197                                 $log->addEntry( 'upload', $title, $detail );
 198                         } else {
 199                                 $interwiki = '[[:' . $this->mInterwiki . ':' .
 200                                         $origTitle->getPrefixedText() . ']]';
 201                                 $detail = wfMsgExt( 'import-logentry-interwiki-detail', array( 'content', 'parsemag' ),
 202                                         $contentCount, $interwiki );
 203                                 $log->addEntry( 'interwiki', $title, $detail );
 204                         }
 205
 206                         $comment = $detail; // quick
 207                         $dbw = wfGetDB( DB_MASTER );
 208                         $nullRevision = Revision::newNullRevision(
 209                                 $dbw, $title->getArticleId(), $comment, true );
 210                         $nullRevision->insertOn( $dbw );
 211                         # Update page record
 212                         $article = new Article( $title );
 213                         $article->updateRevisionOn( $dbw, $nullRevision );
 214                 }
 215         }
 216
 217         function close() {
 218                 global $wgOut;
 219                 if( $this->mPageCount == 0 ) {
 220                         $wgOut->addHtml( "<li>" . wfMsgHtml( 'importnopages' ) . "</li>\n" );
 221                 }
 222                 $wgOut->addHtml( "</ul>\n" );
 223         }
 224 }
 225
 226 /**
 227  *
 228  * @addtogroup SpecialPage
 229  */
 230 class WikiRevision {
 231         var $title = null;
 232         var $id = 0;
 233         var $timestamp = "20010115000000";
 234         var $user = 0;
 235         var $user_text = "";
 236         var $text = "";
 237         var $comment = "";
 238         var $minor = false;
 239
 240         function setTitle( $title ) {
 241                 if( is_object( $title ) ) {
 242                         $this->title = $title;
 243                 } elseif( is_null( $title ) ) {
 244                         throw new MWException( "WikiRevision given a null title in import. You may need to adjust \$wgLegalTitleChars." );
 245                 } else {
 246                         throw new MWException( "WikiRevision given non-object title in import." );
 247                 }
 248         }
 249
 250         function setID( $id ) {
 251                 $this->id = $id;
 252         }
 253
 254         function setTimestamp( $ts ) {
 255                 # 2003-08-05T18:30:02Z
 256                 $this->timestamp = wfTimestamp( TS_MW, $ts );
 257         }
 258
 259         function setUsername( $user ) {
 260                 $this->user_text = $user;
 261         }
 262
 263         function setUserIP( $ip ) {
 264                 $this->user_text = $ip;
 265         }
 266
 267         function setText( $text ) {
 268                 $this->text = $text;
 269         }
 270
 271         function setComment( $text ) {
 272                 $this->comment = $text;
 273         }
 274
 275         function setMinor( $minor ) {
 276                 $this->minor = (bool)$minor;
 277         }
 278
 279         function getTitle() {
 280                 return $this->title;
 281         }
 282
 283         function getID() {
 284                 return $this->id;
 285         }
 286
 287         function getTimestamp() {
 288                 return $this->timestamp;
 289         }
 290
 291         function getUser() {
 292                 return $this->user_text;
 293         }
 294
 295         function getText() {
 296                 return $this->text;
 297         }
 298
 299         function getComment() {
 300                 return $this->comment;
 301         }
 302
 303         function getMinor() {
 304                 return $this->minor;
 305         }
 306
 307         function importOldRevision() {
 308                 $dbw = wfGetDB( DB_MASTER );
 309
 310                 # Sneak a single revision into place
 311                 $user = User::newFromName( $this->getUser() );
 312                 if( $user ) {
 313                         $userId = intval( $user->getId() );
 314                         $userText = $user->getName();
 315                 } else {
 316                         $userId = 0;
 317                         $userText = $this->getUser();
 318                 }
 319
 320                 // avoid memory leak...?
 321                 $linkCache =& LinkCache::singleton();
 322                 $linkCache->clear();
 323
 324                 $article = new Article( $this->title );
 325                 $pageId = $article->getId();
 326                 if( $pageId == 0 ) {
 327                         # must create the page...
 328                         $pageId = $article->insertOn( $dbw );
 329                         $created = true;
 330                 } else {
 331                         $created = false;
 332
 333                         $prior = Revision::loadFromTimestamp( $dbw, $this->title, $this->timestamp );
 334                         if( !is_null( $prior ) ) {
 335                                 // FIXME: this could fail slightly for multiple matches :P
 336                                 wfDebug( __METHOD__ . ": skipping existing revision for [[" .
 337                                         $this->title->getPrefixedText() . "]], timestamp " .
 338                                         $this->timestamp . "\n" );
 339                                 return false;
 340                         }
 341                 }
 342
 343                 # FIXME: Use original rev_id optionally
 344                 # FIXME: blah blah blah
 345
 346                 #if( $numrows > 0 ) {
 347                 #       return wfMsg( "importhistoryconflict" );
 348                 #}
 349
 350                 # Insert the row
 351                 $revision = new Revision( array(
 352                         'page'       => $pageId,
 353                         'text'       => $this->getText(),
 354                         'comment'    => $this->getComment(),
 355                         'user'       => $userId,
 356                         'user_text'  => $userText,
 357                         'timestamp'  => $this->timestamp,
 358                         'minor_edit' => $this->minor,
 359                         ) );
 360                 $revId = $revision->insertOn( $dbw );
 361                 $changed = $article->updateIfNewerOn( $dbw, $revision );
 362
 363                 if( $created ) {
 364                         wfDebug( __METHOD__ . ": running onArticleCreate\n" );
 365                         Article::onArticleCreate( $this->title );
 366
 367                         wfDebug( __METHOD__ . ": running create updates\n" );
 368                         $article->createUpdates( $revision );
 369
 370                 } elseif( $changed ) {
 371                         wfDebug( __METHOD__ . ": running onArticleEdit\n" );
 372                         Article::onArticleEdit( $this->title );
 373
 374                         wfDebug( __METHOD__ . ": running edit updates\n" );
 375                         $article->editUpdates(
 376                                 $this->getText(),
 377                                 $this->getComment(),
 378                                 $this->minor,
 379                                 $this->timestamp,
 380                                 $revId );
 381                 }
 382
 383                 return true;
 384         }
 385
 386 }
 387
 388 /**
 389  * implements Special:Import
 390  * @addtogroup SpecialPage
 391  */
 392 class WikiImporter {
 393         var $mSource = null;
 394         var $mPageCallback = null;
 395         var $mPageOutCallback = null;
 396         var $mRevisionCallback = null;
 397         var $mTargetNamespace = null;
 398         var $lastfield;
 399
 400         function WikiImporter( $source ) {
 401                 $this->setRevisionCallback( array( &$this, "importRevision" ) );
 402                 $this->mSource = $source;
 403         }
 404
 405         function throwXmlError( $err ) {
 406                 $this->debug( "FAILURE: $err" );
 407                 wfDebug( "WikiImporter XML error: $err\n" );
 408         }
 409
 410         # --------------
 411
 412         function doImport() {
 413                 if( empty( $this->mSource ) ) {
 414                         return new WikiErrorMsg( "importnotext" );
 415                 }
 416
 417                 $parser = xml_parser_create( "UTF-8" );
 418
 419                 # case folding violates XML standard, turn it off
 420                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 421
 422                 xml_set_object( $parser, $this );
 423                 xml_set_element_handler( $parser, "in_start", "" );
 424
 425                 $offset = 0; // for context extraction on error reporting
 426                 do {
 427                         $chunk = $this->mSource->readChunk();
 428                         if( !xml_parse( $parser, $chunk, $this->mSource->atEnd() ) ) {
 429                                 wfDebug( "WikiImporter::doImport encountered XML parsing error\n" );
 430                                 return new WikiXmlError( $parser, 'XML import parse failure', $chunk, $offset );
 431                         }
 432                         $offset += strlen( $chunk );
 433                 } while( $chunk !== false && !$this->mSource->atEnd() );
 434                 xml_parser_free( $parser );
 435
 436                 return true;
 437         }
 438
 439         function debug( $data ) {
 440                 #wfDebug( "IMPORT: $data\n" );
 441         }
 442
 443         function notice( $data ) {
 444                 global $wgCommandLineMode;
 445                 if( $wgCommandLineMode ) {
 446                         print "$data\n";
 447                 } else {
 448                         global $wgOut;
 449                         $wgOut->addHTML( "<li>" . htmlspecialchars( $data ) . "</li>\n" );
 450                 }
 451         }
 452
 453         /**
 454          * Sets the action to perform as each new page in the stream is reached.
 455          * @param callable $callback
 456          * @return callable
 457          */
 458         function setPageCallback( $callback ) {
 459                 $previous = $this->mPageCallback;
 460                 $this->mPageCallback = $callback;
 461                 return $previous;
 462         }
 463
 464         /**
 465          * Sets the action to perform as each page in the stream is completed.
 466          * Callback accepts the page title (as a Title object), a second object
 467          * with the original title form (in case it's been overridden into a
 468          * local namespace), and a count of revisions.
 469          *
 470          * @param callable $callback
 471          * @return callable
 472          */
 473         function setPageOutCallback( $callback ) {
 474                 $previous = $this->mPageOutCallback;
 475                 $this->mPageOutCallback = $callback;
 476                 return $previous;
 477         }
 478
 479         /**
 480          * Sets the action to perform as each page revision is reached.
 481          * @param callable $callback
 482          * @return callable
 483          */
 484         function setRevisionCallback( $callback ) {
 485                 $previous = $this->mRevisionCallback;
 486                 $this->mRevisionCallback = $callback;
 487                 return $previous;
 488         }
 489
 490         /**
 491          * Set a target namespace to override the defaults
 492          */
 493         function setTargetNamespace( $namespace ) {
 494                 if( is_null( $namespace ) ) {
 495                         // Don't override namespaces
 496                         $this->mTargetNamespace = null;
 497                 } elseif( $namespace >= 0 ) {
 498                         // FIXME: Check for validity
 499                         $this->mTargetNamespace = intval( $namespace );
 500                 } else {
 501                         return false;
 502                 }
 503         }
 504
 505         /**
 506          * Default per-revision callback, performs the import.
 507          * @param WikiRevision $revision
 508          * @private
 509          */
 510         function importRevision( &$revision ) {
 511                 $dbw = wfGetDB( DB_MASTER );
 512                 return $dbw->deadlockLoop( array( &$revision, 'importOldRevision' ) );
 513         }
 514
 515         /**
 516          * Alternate per-revision callback, for debugging.
 517          * @param WikiRevision $revision
 518          * @private
 519          */
 520         function debugRevisionHandler( &$revision ) {
 521                 $this->debug( "Got revision:" );
 522                 if( is_object( $revision->title ) ) {
 523                         $this->debug( "-- Title: " . $revision->title->getPrefixedText() );
 524                 } else {
 525                         $this->debug( "-- Title: <invalid>" );
 526                 }
 527                 $this->debug( "-- User: " . $revision->user_text );
 528                 $this->debug( "-- Timestamp: " . $revision->timestamp );
 529                 $this->debug( "-- Comment: " . $revision->comment );
 530                 $this->debug( "-- Text: " . $revision->text );
 531         }
 532
 533         /**
 534          * Notify the callback function when a new <page> is reached.
 535          * @param Title $title
 536          * @private
 537          */
 538         function pageCallback( $title ) {
 539                 if( is_callable( $this->mPageCallback ) ) {
 540                         call_user_func( $this->mPageCallback, $title );
 541                 }
 542         }
 543
 544         /**
 545          * Notify the callback function when a </page> is closed.
 546          * @param Title $title
 547          * @param Title $origTitle
 548          * @param int $revisionCount
 549          * @param int $successCount number of revisions for which callback returned true
 550          * @private
 551          */
 552         function pageOutCallback( $title, $origTitle, $revisionCount, $successCount ) {
 553                 if( is_callable( $this->mPageOutCallback ) ) {
 554                         call_user_func( $this->mPageOutCallback, $title, $origTitle,
 555                                 $revisionCount, $successCount );
 556                 }
 557         }
 558
 559
 560         # XML parser callbacks from here out -- beware!
 561         function donothing( $parser, $x, $y="" ) {
 562                 #$this->debug( "donothing" );
 563         }
 564
 565         function in_start( $parser, $name, $attribs ) {
 566                 $this->debug( "in_start $name" );
 567                 if( $name != "mediawiki" ) {
 568                         return $this->throwXMLerror( "Expected <mediawiki>, got <$name>" );
 569                 }
 570                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 571         }
 572
 573         function in_mediawiki( $parser, $name, $attribs ) {
 574                 $this->debug( "in_mediawiki $name" );
 575                 if( $name == 'siteinfo' ) {
 576                         xml_set_element_handler( $parser, "in_siteinfo", "out_siteinfo" );
 577                 } elseif( $name == 'page' ) {
 578                         $this->workRevisionCount = 0;
 579                         $this->workSuccessCount = 0;
 580                         xml_set_element_handler( $parser, "in_page", "out_page" );
 581                 } else {
 582                         return $this->throwXMLerror( "Expected <page>, got <$name>" );
 583                 }
 584         }
 585         function out_mediawiki( $parser, $name ) {
 586                 $this->debug( "out_mediawiki $name" );
 587                 if( $name != "mediawiki" ) {
 588                         return $this->throwXMLerror( "Expected </mediawiki>, got </$name>" );
 589                 }
 590                 xml_set_element_handler( $parser, "donothing", "donothing" );
 591         }
 592
 593
 594         function in_siteinfo( $parser, $name, $attribs ) {
 595                 // no-ops for now
 596                 $this->debug( "in_siteinfo $name" );
 597                 switch( $name ) {
 598                 case "sitename":
 599                 case "base":
 600                 case "generator":
 601                 case "case":
 602                 case "namespaces":
 603                 case "namespace":
 604                         break;
 605                 default:
 606                         return $this->throwXMLerror( "Element <$name> not allowed in <siteinfo>." );
 607                 }
 608         }
 609
 610         function out_siteinfo( $parser, $name ) {
 611                 if( $name == "siteinfo" ) {
 612                         xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 613                 }
 614         }
 615
 616
 617         function in_page( $parser, $name, $attribs ) {
 618                 $this->debug( "in_page $name" );
 619                 switch( $name ) {
 620                 case "id":
 621                 case "title":
 622                 case "restrictions":
 623                         $this->appendfield = $name;
 624                         $this->appenddata = "";
 625                         $this->parenttag = "page";
 626                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 627                         xml_set_character_data_handler( $parser, "char_append" );
 628                         break;
 629                 case "revision":
 630                         if( is_object( $this->pageTitle ) ) {
 631                                 $this->workRevision = new WikiRevision;
 632                                 $this->workRevision->setTitle( $this->pageTitle );
 633                                 $this->workRevisionCount++;
 634                         } else {
 635                                 // Skipping items due to invalid page title
 636                                 $this->workRevision = null;
 637                         }
 638                         xml_set_element_handler( $parser, "in_revision", "out_revision" );
 639                         break;
 640                 default:
 641                         return $this->throwXMLerror( "Element <$name> not allowed in a <page>." );
 642                 }
 643         }
 644
 645         function out_page( $parser, $name ) {
 646                 $this->debug( "out_page $name" );
 647                 if( $name != "page" ) {
 648                         return $this->throwXMLerror( "Expected </page>, got </$name>" );
 649                 }
 650                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 651
 652                 $this->pageOutCallback( $this->pageTitle, $this->origTitle,
 653                         $this->workRevisionCount, $this->workSuccessCount );
 654
 655                 $this->workTitle = null;
 656                 $this->workRevision = null;
 657                 $this->workRevisionCount = 0;
 658                 $this->workSuccessCount = 0;
 659                 $this->pageTitle = null;
 660                 $this->origTitle = null;
 661         }
 662
 663         function in_nothing( $parser, $name, $attribs ) {
 664                 $this->debug( "in_nothing $name" );
 665                 return $this->throwXMLerror( "No child elements allowed here; got <$name>" );
 666         }
 667         function char_append( $parser, $data ) {
 668                 $this->debug( "char_append '$data'" );
 669                 $this->appenddata .= $data;
 670         }
 671         function out_append( $parser, $name ) {
 672                 $this->debug( "out_append $name" );
 673                 if( $name != $this->appendfield ) {
 674                         return $this->throwXMLerror( "Expected </{$this->appendfield}>, got </$name>" );
 675                 }
 676                 xml_set_element_handler( $parser, "in_$this->parenttag", "out_$this->parenttag" );
 677                 xml_set_character_data_handler( $parser, "donothing" );
 678
 679                 switch( $this->appendfield ) {
 680                 case "title":
 681                         $this->workTitle = $this->appenddata;
 682                         $this->origTitle = Title::newFromText( $this->workTitle );
 683                         if( !is_null( $this->mTargetNamespace ) && !is_null( $this->origTitle ) ) {
 684                                 $this->pageTitle = Title::makeTitle( $this->mTargetNamespace,
 685                                         $this->origTitle->getDbKey() );
 686                         } else {
 687                                 $this->pageTitle = Title::newFromText( $this->workTitle );
 688                         }
 689                         if( is_null( $this->pageTitle ) ) {
 690                                 // Invalid page title? Ignore the page
 691                                 $this->notice( "Skipping invalid page title '$this->workTitle'" );
 692                         } else {
 693                                 $this->pageCallback( $this->workTitle );
 694                         }
 695                         break;
 696                 case "id":
 697                         if ( $this->parenttag == 'revision' ) {
 698                                 if( $this->workRevision )
 699                                         $this->workRevision->setID( $this->appenddata );
 700                         }
 701                         break;
 702                 case "text":
 703                         if( $this->workRevision )
 704                                 $this->workRevision->setText( $this->appenddata );
 705                         break;
 706                 case "username":
 707                         if( $this->workRevision )
 708                                 $this->workRevision->setUsername( $this->appenddata );
 709                         break;
 710                 case "ip":
 711                         if( $this->workRevision )
 712                                 $this->workRevision->setUserIP( $this->appenddata );
 713                         break;
 714                 case "timestamp":
 715                         if( $this->workRevision )
 716                                 $this->workRevision->setTimestamp( $this->appenddata );
 717                         break;
 718                 case "comment":
 719                         if( $this->workRevision )
 720                                 $this->workRevision->setComment( $this->appenddata );
 721                         break;
 722                 case "minor":
 723                         if( $this->workRevision )
 724                                 $this->workRevision->setMinor( true );
 725                         break;
 726                 default:
 727                         $this->debug( "Bad append: {$this->appendfield}" );
 728                 }
 729                 $this->appendfield = "";
 730                 $this->appenddata = "";
 731         }
 732
 733         function in_revision( $parser, $name, $attribs ) {
 734                 $this->debug( "in_revision $name" );
 735                 switch( $name ) {
 736                 case "id":
 737                 case "timestamp":
 738                 case "comment":
 739                 case "minor":
 740                 case "text":
 741                         $this->parenttag = "revision";
 742                         $this->appendfield = $name;
 743                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 744                         xml_set_character_data_handler( $parser, "char_append" );
 745                         break;
 746                 case "contributor":
 747                         xml_set_element_handler( $parser, "in_contributor", "out_contributor" );
 748                         break;
 749                 default:
 750                         return $this->throwXMLerror( "Element <$name> not allowed in a <revision>." );
 751                 }
 752         }
 753
 754         function out_revision( $parser, $name ) {
 755                 $this->debug( "out_revision $name" );
 756                 if( $name != "revision" ) {
 757                         return $this->throwXMLerror( "Expected </revision>, got </$name>" );
 758                 }
 759                 xml_set_element_handler( $parser, "in_page", "out_page" );
 760
 761                 if( $this->workRevision ) {
 762                         $ok = call_user_func_array( $this->mRevisionCallback,
 763                                 array( &$this->workRevision, &$this ) );
 764                         if( $ok ) {
 765                                 $this->workSuccessCount++;
 766                         }
 767                 }
 768         }
 769
 770         function in_contributor( $parser, $name, $attribs ) {
 771                 $this->debug( "in_contributor $name" );
 772                 switch( $name ) {
 773                 case "username":
 774                 case "ip":
 775                 case "id":
 776                         $this->parenttag = "contributor";
 777                         $this->appendfield = $name;
 778                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 779                         xml_set_character_data_handler( $parser, "char_append" );
 780                         break;
 781                 default:
 782                         $this->throwXMLerror( "Invalid tag <$name> in <contributor>" );
 783                 }
 784         }
 785
 786         function out_contributor( $parser, $name ) {
 787                 $this->debug( "out_contributor $name" );
 788                 if( $name != "contributor" ) {
 789                         return $this->throwXMLerror( "Expected </contributor>, got </$name>" );
 790                 }
 791                 xml_set_element_handler( $parser, "in_revision", "out_revision" );
 792         }
 793
 794 }
 795
 796 /**
 797  * @todo document (e.g. one-sentence class description).
 798  * @addtogroup SpecialPage
 799  */
 800 class ImportStringSource {
 801         function ImportStringSource( $string ) {
 802                 $this->mString = $string;
 803                 $this->mRead = false;
 804         }
 805
 806         function atEnd() {
 807                 return $this->mRead;
 808         }
 809
 810         function readChunk() {
 811                 if( $this->atEnd() ) {
 812                         return false;
 813                 } else {
 814                         $this->mRead = true;
 815                         return $this->mString;
 816                 }
 817         }
 818 }
 819
 820 /**
 821  * @todo document (e.g. one-sentence class description).
 822  * @addtogroup SpecialPage
 823  */
 824 class ImportStreamSource {
 825         function ImportStreamSource( $handle ) {
 826                 $this->mHandle = $handle;
 827         }
 828
 829         function atEnd() {
 830                 return feof( $this->mHandle );
 831         }
 832
 833         function readChunk() {
 834                 return fread( $this->mHandle, 32768 );
 835         }
 836
 837         static function newFromFile( $filename ) {
 838                 $file = @fopen( $filename, 'rt' );
 839                 if( !$file ) {
 840                         return new WikiErrorMsg( "importcantopen" );
 841                 }
 842                 return new ImportStreamSource( $file );
 843         }
 844
 845         static function newFromUpload( $fieldname = "xmlimport" ) {
 846                 $upload =& $_FILES[$fieldname];
 847
 848                 if( !isset( $upload ) || !$upload['name'] ) {
 849                         return new WikiErrorMsg( 'importnofile' );
 850                 }
 851                 if( !empty( $upload['error'] ) ) {
 852                         switch($upload['error']){
 853                                 case 1: # The uploaded file exceeds the upload_max_filesize directive in php.ini.
 854                                         return new WikiErrorMsg( 'importuploaderrorsize' );
 855                                 case 2: # The uploaded file exceeds the MAX_FILE_SIZE directive that was specified in the HTML form.
 856                                         return new WikiErrorMsg( 'importuploaderrorsize' );
 857                                 case 3: # The uploaded file was only partially uploaded
 858                                         return new WikiErrorMsg( 'importuploaderrorpartial' );
 859                             case 6: #Missing a temporary folder. Introduced in PHP 4.3.10 and PHP 5.0.3.
 860                                 return new WikiErrorMsg( 'importuploaderrortemp' );
 861                             # case else: # Currently impossible
 862                         }
 863
 864                 }
 865                 $fname = $upload['tmp_name'];
 866                 if( is_uploaded_file( $fname ) ) {
 867                         return ImportStreamSource::newFromFile( $fname );
 868                 } else {
 869                         return new WikiErrorMsg( 'importnofile' );
 870                 }
 871         }
 872
 873         function newFromURL( $url, $method = 'GET' ) {
 874                 wfDebug( __METHOD__ . ": opening $url\n" );
 875                 # Use the standard HTTP fetch function; it times out
 876                 # quicker and sorts out user-agent problems which might
 877                 # otherwise prevent importing from large sites, such
 878                 # as the Wikimedia cluster, etc.
 879                 $data = Http::request( $method, $url );
 880                 if( $data !== false ) {
 881                         $file = tmpfile();
 882                         fwrite( $file, $data );
 883                         fflush( $file );
 884                         fseek( $file, 0 );
 885                         return new ImportStreamSource( $file );
 886                 } else {
 887                         return new WikiErrorMsg( 'importcantopen' );
 888                 }
 889         }
 890
 891         public static function newFromInterwiki( $interwiki, $page, $history=false ) {
 892                 $link = Title::newFromText( "$interwiki:Special:Export/$page" );
 893                 if( is_null( $link ) || $link->getInterwiki() == '' ) {
 894                         return new WikiErrorMsg( 'importbadinterwiki' );
 895                 } else {
 896                         $params = $history ? 'history=1' : '';
 897                         $url = $link->getFullUrl( $params );
 898                         # For interwikis, use POST to avoid redirects.
 899                         return ImportStreamSource::newFromURL( $url, "POST" );
 900                 }
 901         }
 902 }
 903
 904
 905