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