includes/SpecialImport.php

   1 <?php
   2 # Copyright (C) 2003 Brion Vibber <brion@pobox.com>
   3 # http://www.mediawiki.org/
   4 #
   5 # This program is free software; you can redistribute it and/or modify
   6 # it under the terms of the GNU General Public License as published by
   7 # the Free Software Foundation; either version 2 of the License, or
   8 # (at your option) any later version.
   9 #
  10 # This program is distributed in the hope that it will be useful,
  11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13 # GNU General Public License for more details.
  14 #
  15 # You should have received a copy of the GNU General Public License along
  16 # with this program; if not, write to the Free Software Foundation, Inc.,
  17 # 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  18 # http://www.gnu.org/copyleft/gpl.html
  19
  20 /**
  21  *
  22  * @package MediaWiki
  23  * @subpackage SpecialPage
  24  */
  25
  26 /**
  27  * Constructor
  28  */
  29 function wfSpecialImport( $page = '' ) {
  30         global $wgOut, $wgLang, $wgRequest, $wgTitle;
  31         global $wgImportSources;
  32
  33         ###
  34         $wgOut->addWikiText( "Special:Import is not ready for this beta release, sorry." );
  35         return;
  36         ###
  37
  38         if( $wgRequest->wasPosted() && $wgRequest->getVal( 'action' ) == 'submit') {
  39                 $importer = new WikiImporter();
  40
  41                 switch( $wgRequest->getVal( "source" ) ) {
  42                 case "upload":
  43                         $ok = $importer->setupFromUpload( "xmlimport" );
  44                         break;
  45                 case "interwiki":
  46                         $ok = $importer->setupFromInterwiki(
  47                                 $wgRequest->getVal( "interwiki" ),
  48                                 $wgRequest->getText( "frompage" ) );
  49                         break;
  50                 default:
  51                         $ok = false;
  52                 }
  53
  54                 if( $ok ) {
  55                         $importer->setRevisionHandler( "wfImportOldRevision" );
  56                         if( $importer->doImport() ) {
  57                                 # Success!
  58                                 $wgOut->addHTML( "<p>" . wfMsg( "importsuccess" ) . "</p>" );
  59                         } else {
  60                                 $wgOut->addHTML( "<p>" . wfMsg( "importfailed",
  61                                         htmlspecialchars( $importer->getError() ) ) . "</p>" );
  62                         }
  63                 } else {
  64                         $wgOut->addWikiText( htmlspecialchars( $importer->getError() ) );
  65                 }
  66         }
  67
  68         $wgOut->addWikiText( "<p>" . wfMsg( "importtext" ) . "</p>" );
  69         $action = $wgTitle->escapeLocalUrl();
  70         $wgOut->addHTML( "
  71 <fieldset>
  72         <legend>Upload XML</legend>
  73         <form enctype='multipart/form-data' method='post' action=\"$action\">
  74                 <input type='hidden' name='action' value='submit' />
  75                 <input type='hidden' name='source' value='upload' />
  76                 <input type='hidden' name='MAX_FILE_SIZE' value='200000' />
  77                 <input type='file' name='xmlimport' value='' size='30' />
  78                 <input type='submit' value='" . htmlspecialchars( wfMsg( "uploadbtn" ) ) . "'/>
  79         </form>
  80 </fieldset>
  81 " );
  82
  83         if( !empty( $wgImportSources ) ) {
  84                 $wgOut->addHTML( "
  85 <fieldset>
  86         <legend>Interwiki import</legend>
  87         <form method='post' action=\"$action\">
  88                 <input type='hidden' name='action' value='submit' />
  89                 <input type='hidden' name='source' value='interwiki' />
  90                 <select name='interwiki'>
  91 " );
  92                 foreach( $wgImportSources as $interwiki ) {
  93                         $iw = htmlspecialchars( $interwiki );
  94                         $wgOut->addHTML( "<option value=\"$iw\">$iw</option>\n" );
  95                 }
  96                 $wgOut->addHTML( "
  97                 </select>
  98                 <input name='frompage' />
  99                 <input type='submit' />
 100         </form>
 101 </fieldset>
 102 " );
 103         }
 104 }
 105
 106 function wfImportOldRevision( &$revision ) {
 107         $dbw =& wfGetDB( DB_MASTER );
 108         $dbw->deadlockLoop( array( &$revision, 'importOldRevision' ) );
 109 }
 110
 111 /**
 112  *
 113  * @package MediaWiki
 114  * @subpackage SpecialPage
 115  */
 116 class WikiRevision {
 117         var $title = NULL;
 118         var $timestamp = "20010115000000";
 119         var $user = 0;
 120         var $user_text = "";
 121         var $text = "";
 122         var $comment = "";
 123
 124         function setTitle( $text ) {
 125                 $text = $this->fixEncoding( $text );
 126                 $this->title = Title::newFromText( $text );
 127         }
 128
 129         function setTimestamp( $ts ) {
 130                 # 2003-08-05T18:30:02Z
 131                 $this->timestamp = preg_replace( '/^(....)-(..)-(..)T(..):(..):(..)Z$/', '$1$2$3$4$5$6', $ts );
 132         }
 133
 134         function setUsername( $user ) {
 135                 $this->user_text = $this->fixEncoding( $user );
 136         }
 137
 138         function setUserIP( $ip ) {
 139                 $this->user_text = $this->fixEncoding( $ip );
 140         }
 141
 142         function setText( $text ) {
 143                 $this->text = $this->fixEncoding( $text );
 144         }
 145
 146         function setComment( $text ) {
 147                 $this->comment = $this->fixEncoding( $text );
 148         }
 149
 150         function fixEncoding( $data ) {
 151                 global $wgLang, $wgInputEncoding;
 152
 153                 if( strcasecmp( $wgInputEncoding, "utf-8" ) == 0 ) {
 154                         return $data;
 155                 } else {
 156                         return $wgLang->iconv( "utf-8", $wgInputEncoding, $data );
 157                 }
 158         }
 159
 160         function getTitle() {
 161                 return $this->title;
 162         }
 163
 164         function getTimestamp() {
 165                 return $this->timestamp;
 166         }
 167
 168         function getUser() {
 169                 return $this->user_text;
 170         }
 171
 172         function getText() {
 173                 return $this->text;
 174         }
 175
 176         function getComment() {
 177                 return $this->comment;
 178         }
 179 }
 180
 181 /**
 182  *
 183  * @package MediaWiki
 184  * @subpackage SpecialPage
 185  */
 186 class WikiImporter {
 187         var $mSource = NULL;
 188         var $mError = "";
 189         var $mXmlError = XML_ERROR_NONE;
 190         var $mRevisionHandler = NULL;
 191         var $lastfield;
 192
 193         function WikiImporter() {
 194                 $this->setRevisionHandler( array( &$this, "defaultRevisionHandler" ) );
 195         }
 196
 197         function setError( $err ) {
 198                 $this->mError = $err;
 199                 return false;
 200         }
 201
 202         function getError() {
 203                 if( $this->mXmlError == XML_ERROR_NONE ) {
 204                         return $this->mError;
 205                 } else {
 206                         return xml_error_string( $this->mXmlError );
 207                 }
 208         }
 209
 210         function throwXmlError( $err ) {
 211                 $this->debug( "FAILURE: $err" );
 212         }
 213
 214         function setupFromFile( $filename ) {
 215                 $this->mSource = file_get_contents( $filename );
 216                 return true;
 217         }
 218
 219         function setupFromUpload( $fieldname = "xmlimport" ) {
 220                 global $wgOut;
 221
 222                 $upload =& $_FILES[$fieldname];
 223
 224                 if( !isset( $upload ) ) {
 225                         return $this->setError( wfMsg( "importnofile" ) );
 226                 }
 227                 if( !empty( $upload['error'] ) ) {
 228                         return $this->setError( wfMsg( "importuploaderror", $upload['error'] ) );
 229                 }
 230                 $fname = $upload['tmp_name'];
 231                 if( is_uploaded_file( $fname ) ) {
 232                         return $this->setupFromFile( $fname );
 233                 } else {
 234                         return $this->setError( wfMsg( "importnofile" ) );
 235                 }
 236         }
 237
 238         function setupFromURL( $url ) {
 239                 # fopen-wrappers are normally turned off for security.
 240                 ini_set( "allow_url_fopen", true );
 241                 $ret = $this->setupFromFile( $url );
 242                 ini_set( "allow_url_fopen", false );
 243                 return $ret;
 244         }
 245
 246         function setupFromInterwiki( $interwiki, $page ) {
 247                 $base = Title::getInterwikiLink( $interwiki );
 248                 if( empty( $base ) ) {
 249                         return false;
 250                 } else {
 251                         $import = wfUrlencode( "Special:Export/$page" );
 252                         $url = str_replace( "$1", $import, $base );
 253                         $this->notice( "Importing from $url" );
 254                         return $this->setupFromURL( $url );
 255                 }
 256         }
 257
 258         # --------------
 259
 260         function doImport() {
 261                 if( empty( $this->mSource ) ) {
 262                         return $this->setError( wfMsg( "importnotext" ) );
 263                 }
 264
 265                 $parser = xml_parser_create( "UTF-8" );
 266
 267                 # case folding violates XML standard, turn it off
 268                 xml_parser_set_option( $parser, XML_OPTION_CASE_FOLDING, false );
 269
 270                 xml_set_object( $parser, &$this );
 271                 xml_set_element_handler( $parser, "in_start", "" );
 272
 273                 if( !xml_parse( $parser, $this->mSource, true ) ) {
 274                         # return error message
 275                         $this->mXmlError = xml_get_error_code( $parser );
 276                         xml_parser_free( $parser );
 277                         return false;
 278                 }
 279                 xml_parser_free( $parser );
 280
 281                 return true;
 282         }
 283
 284         function debug( $data ) {
 285                 global $wgOut;
 286                 # $this->notice( "DEBUG: $data\n" );
 287         }
 288
 289         function notice( $data ) {
 290                 global $wgCommandLineMode;
 291                 if( $wgCommandLineMode ) {
 292                         print "$data\n";
 293                 } else {
 294                         global $wgOut;
 295                         $wgOut->addHTML( "<li>$data</li>\n" );
 296                 }
 297         }
 298
 299         function setRevisionHandler( $functionref ) {
 300                 $this->mRevisionHandler = $functionref;
 301         }
 302
 303         function defaultRevisionHandler( &$revision ) {
 304                 $this->debug( "Got revision:" );
 305                 if( is_object( $revision->title ) ) {
 306                         $this->debug( "-- Title: " . $revision->title->getPrefixedText() );
 307                 } else {
 308                         $this->debug( "-- Title: <invalid>" );
 309                 }
 310                 $this->debug( "-- User: " . $revision->user_text );
 311                 $this->debug( "-- Timestamp: " . $revision->timestamp );
 312                 $this->debug( "-- Comment: " . $revision->comment );
 313                 $this->debug( "-- Text: " . $revision->text );
 314         }
 315
 316
 317
 318         # XML parser callbacks from here out -- beware!
 319         function donothing( $parser, $x, $y="" ) {
 320                 #$this->debug( "donothing" );
 321         }
 322
 323         function in_start( $parser, $name, $attribs ) {
 324                 $this->debug( "in_start $name" );
 325                 if( $name != "mediawiki" ) {
 326                         return $this->throwXMLerror( "Expected <mediawiki>, got <$name>" );
 327                 }
 328                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 329         }
 330
 331         function in_mediawiki( $parser, $name, $attribs ) {
 332                 $this->debug( "in_mediawiki $name" );
 333                 if( $name != "page" ) {
 334                         return $this->throwXMLerror( "Expected <page>, got <$name>" );
 335                 }
 336                 xml_set_element_handler( $parser, "in_page", "out_page" );
 337         }
 338         function out_mediawiki( $parser, $name ) {
 339                 $this->debug( "out_mediawiki $name" );
 340                 if( $name != "mediawiki" ) {
 341                         return $this->throwXMLerror( "Expected </mediawiki>, got </$name>" );
 342                 }
 343                 xml_set_element_handler( $parser, "donothing", "donothing" );
 344         }
 345
 346         function in_page( $parser, $name, $attribs ) {
 347                 $this->debug( "in_page $name" );
 348                 switch( $name ) {
 349                 case "id":
 350                 case "title":
 351                 case "restrictions":
 352                         $this->appendfield = $name;
 353                         $this->appenddata = "";
 354                         $this->parenttag = "page";
 355                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 356                         xml_set_character_data_handler( $parser, "char_append" );
 357                         break;
 358                 case "revision":
 359                         $this->workRevision = new WikiRevision;
 360                         $this->workRevision->setTitle( $this->workTitle );
 361                         xml_set_element_handler( $parser, "in_revision", "out_revision" );
 362                         break;
 363                 default:
 364                         return $this->throwXMLerror( "Element <$name> not allowed in a <page>." );
 365                 }
 366         }
 367
 368         function out_page( $parser, $name ) {
 369                 $this->debug( "out_page $name" );
 370                 if( $name != "page" ) {
 371                         return $this->throwXMLerror( "Expected </page>, got </$name>" );
 372                 }
 373                 xml_set_element_handler( $parser, "in_mediawiki", "out_mediawiki" );
 374
 375                 $this->workTitle = NULL;
 376                 $this->workRevision = NULL;
 377         }
 378
 379         function in_nothing( $parser, $name, $attribs ) {
 380                 $this->debug( "in_nothing $name" );
 381                 return $this->throwXMLerror( "No child elements allowed here; got <$name>" );
 382         }
 383         function char_append( $parser, $data ) {
 384                 $this->debug( "char_append '$data'" );
 385                 $this->appenddata .= $data;
 386         }
 387         function out_append( $parser, $name ) {
 388                 $this->debug( "out_append $name" );
 389                 if( $name != $this->appendfield ) {
 390                         return $this->throwXMLerror( "Expected </{$this->appendfield}>, got </$name>" );
 391                 }
 392                 xml_set_element_handler( $parser, "in_$this->parenttag", "out_$this->parenttag" );
 393                 xml_set_character_data_handler( $parser, "donothing" );
 394                 switch( $this->appendfield ) {
 395                 case "title":
 396                         $this->workTitle = $this->appenddata;
 397                         break;
 398                 case "text":
 399                         $this->workRevision->setText( $this->appenddata );
 400                         break;
 401                 case "username":
 402                         $this->workRevision->setUsername( $this->appenddata );
 403                         break;
 404                 case "ip":
 405                         $this->workRevision->setUserIP( $this->appenddata );
 406                         break;
 407                 case "timestamp":
 408                         $this->workRevision->setTimestamp( $this->appenddata );
 409                         break;
 410                 case "comment":
 411                         $this->workRevision->setComment( $this->appenddata );
 412                         break;
 413                 default;
 414                         $this->debug( "Bad append: {$this->appendfield}" );
 415                 }
 416                 $this->appendfield = "";
 417                 $this->appenddata = "";
 418         }
 419
 420         function in_revision( $parser, $name, $attribs ) {
 421                 $this->debug( "in_revision $name" );
 422                 switch( $name ) {
 423                 case "id":
 424                 case "timestamp":
 425                 case "comment":
 426                 case "text":
 427                         $this->parenttag = "revision";
 428                         $this->appendfield = $name;
 429                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 430                         xml_set_character_data_handler( $parser, "char_append" );
 431                         break;
 432                 case "contributor":
 433                         xml_set_element_handler( $parser, "in_contributor", "out_contributor" );
 434                         break;
 435                 default:
 436                         return $this->throwXMLerror( "Element <$name> not allowed in a <revision>." );
 437                 }
 438         }
 439
 440         function out_revision( $parser, $name ) {
 441                 $this->debug( "out_revision $name" );
 442                 if( $name != "revision" ) {
 443                         return $this->throwXMLerror( "Expected </revision>, got </$name>" );
 444                 }
 445                 xml_set_element_handler( $parser, "in_page", "out_page" );
 446
 447                 $out = call_user_func( $this->mRevisionHandler, &$this->workRevision, &$this );
 448                 if( !empty( $out ) ) {
 449                         global $wgOut;
 450                         $wgOut->addHTML( "<li>" . $out . "</li>\n" );
 451                 }
 452         }
 453
 454         function in_contributor( $parser, $name, $attribs ) {
 455                 $this->debug( "in_contributor $name" );
 456                 switch( $name ) {
 457                 case "username":
 458                 case "ip":
 459                         $this->parenttag = "contributor";
 460                         $this->appendfield = $name;
 461                         xml_set_element_handler( $parser, "in_nothing", "out_append" );
 462                         xml_set_character_data_handler( $parser, "char_append" );
 463                         break;
 464                 default:
 465                         $this->throwXMLerror( "Invalid tag <$name> in <contributor>" );
 466                 }
 467         }
 468
 469         function out_contributor( $parser, $name ) {
 470                 $this->debug( "out_contributor $name" );
 471                 if( $name != "contributor" ) {
 472                         return $this->throwXMLerror( "Expected </contributor>, got </$name>" );
 473                 }
 474                 xml_set_element_handler( $parser, "in_revision", "out_revision" );
 475         }
 476
 477         function importOldRevision() {
 478                 $fname = "WikiImporter::importOldRevision";
 479                 $dbw =& wfGetDB( DB_MASTER );
 480
 481                 # Sneak a single revision into place
 482                 $user = User::newFromName( $this->getUser() );
 483
 484                 $res = $dbw->select( 'old', 1,
 485                         $this->title->oldCond() + array( 'old_timestamp' => $this->timestamp ),
 486                         $fname, 'FOR UPDATE'
 487                 );
 488
 489                 $numrows = $dbw->numRows( $res );
 490                 $dbw->freeResult( $res );
 491                 if( $numrows > 0 ) {
 492                         return wfMsg( "importhistoryconflict" );
 493                 }
 494
 495                 # Insert the row
 496                 $oldIgnore = $dbw->ignoreErrors( true );
 497                 $success = $dbw->insert( 'old',
 498                         array(
 499                                 'old_namespace' => intval( $this->title->getNamespace() ),
 500                                 'old_title' => $this->title->getDBkey(),
 501                                 'old_text' => $this->getText(),
 502                                 'old_comment' => $this->getComment(),
 503                                 'old_user' => intval( $user->getId() ),
 504                                 'old_user_text' => $user->getName(),
 505                                 'old_timestamp' => $this->timestamp,
 506                                 'inverse_timestamp' => wfInvertTimestamp( $this->timestamp ),
 507                                 'old_minor_edit' => 0,
 508                                 'old_flags' => ''
 509                         ), $fname
 510                 );
 511
 512                 return wfMsg( "ok" );
 513         }
 514 }
 515
 516
 517 ?>