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