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