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