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