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