includes/Parser.php

   1 <?php
   2
   3 // require_once('Tokenizer.php');
   4
   5 # PHP Parser
   6 #
   7 # Processes wiki markup
   8 #
   9 # There are two main entry points into the Parser class: parse() and preSaveTransform().
  10 # The parse() function produces HTML output, preSaveTransform() produces altered wiki markup.
  11 #
  12 # Globals used:
  13 #    objects:   $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
  14 #
  15 # NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  16 #
  17 #    settings:  $wgUseTex*, $wgUseCategoryMagic*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  18 #               $wgNamespacesWithSubpages, $wgLanguageCode, $wgAllowExternalImages*,
  19 #               $wgLocaltimezone
  20 #
  21 #      * only within ParserOptions
  22 #
  23 #
  24 #----------------------------------------
  25 #    Variable substitution O(N^2) attack
  26 #-----------------------------------------
  27 # Without countermeasures, it would be possible to attack the parser by saving a page
  28 # filled with a large number of inclusions of large pages. The size of the generated
  29 # page would be proportional to the square of the input size. Hence, we limit the number
  30 # of inclusions of any given page, thus bringing any attack back to O(N).
  31 #
  32
  33 define( "MAX_INCLUDE_REPEAT", 5 );
  34 define( "MAX_INCLUDE_SIZE", 1000000 ); // 1 Million
  35
  36 # Allowed values for $mOutputType
  37 define( "OT_HTML", 1 );
  38 define( "OT_WIKI", 2 );
  39 define( "OT_MSG", 3 );
  40
  41 # string parameter for extractTags which will cause it
  42 # to strip HTML comments in addition to regular
  43 # <XML>-style tags. This should not be anything we
  44 # may want to use in wikisyntax
  45 define( "STRIP_COMMENTS", "HTMLCommentStrip" );
  46
  47 # prefix for escaping, used in two functions at least
  48 define( "UNIQ_PREFIX", "NaodW29");
  49
  50 class Parser
  51 {
  52         # Persistent:
  53         var $mTagHooks;
  54
  55         # Cleared with clearState():
  56         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  57         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  58
  59         # Temporary:
  60         var $mOptions, $mTitle, $mOutputType,
  61             $mTemplates,        // cache of already loaded templates, avoids
  62                                 // multiple SQL queries for the same string
  63             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
  64                                 // in this path. Used for loop detection.
  65
  66         function Parser() {
  67                 $this->mTemplates = array();
  68                 $this->mTemplatePath = array();
  69                 $this->mTagHooks = array();
  70                 $this->clearState();
  71         }
  72
  73         function clearState() {
  74                 $this->mOutput = new ParserOutput;
  75                 $this->mAutonumber = 0;
  76                 $this->mLastSection = "";
  77                 $this->mDTopen = false;
  78                 $this->mVariables = false;
  79                 $this->mIncludeCount = array();
  80                 $this->mStripState = array();
  81                 $this->mArgStack = array();
  82                 $this->mInPre = false;
  83         }
  84
  85         # First pass--just handle <nowiki> sections, pass the rest off
  86         # to internalParse() which does all the real work.
  87         #
  88         # Returns a ParserOutput
  89         #
  90         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
  91                 global $wgUseTidy;
  92                 $fname = "Parser::parse";
  93                 wfProfileIn( $fname );
  94
  95                 if ( $clearState ) {
  96                         $this->clearState();
  97                 }
  98
  99                 $this->mOptions = $options;
 100                 $this->mTitle =& $title;
 101                 $this->mOutputType = OT_HTML;
 102
 103                 $stripState = NULL;
 104                 $text = $this->strip( $text, $this->mStripState );
 105                 $text = $this->internalParse( $text, $linestart );
 106                 $text = $this->unstrip( $text, $this->mStripState );
 107                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 108                 if(!$wgUseTidy) {
 109                         $fixtags = array(
 110                                 # french spaces, last one Guillemet-left
 111                                 # only if there is something before the space
 112                                 '/(.) (\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 113                                 # french spaces, Guillemet-right
 114                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 115                                 '/<hr *>/i' => '<hr />',
 116                                 '/<br *>/i' => '<br />',
 117                                 '/<center *>/i' => '<div class="center">',
 118                                 '/<\\/center *>/i' => '</div>',
 119                                 # Clean up spare ampersands; note that we probably ought to be
 120                                 # more careful about named entities.
 121                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 122                         );
 123                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 124                 } else {
 125                         $fixtags = array(
 126                                 # french spaces, last one Guillemet-left
 127                                 '/ (\\?|:|!|\\302\\273)/i' => '&nbsp;\\1',
 128                                 # french spaces, Guillemet-right
 129                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 130                                 '/([^> ]+(&#x30(1|3|9);)[^< ]*)/i' => '<span class="diacrit">\\1</span>',
 131                                 '/<center *>/i' => '<div class="center">',
 132                                 '/<\\/center *>/i' => '</div>'
 133                         );
 134                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 135                 }
 136                 # only once and last
 137                 $text = $this->doBlockLevels( $text, $linestart );
 138                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 139                 if($wgUseTidy) {
 140                         $text = $this->tidy($text);
 141                 }
 142                 $this->mOutput->setText( $text );
 143                 wfProfileOut( $fname );
 144                 return $this->mOutput;
 145         }
 146
 147         /* static */ function getRandomString() {
 148                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 149         }
 150
 151         # Replaces all occurrences of <$tag>content</$tag> in the text
 152         # with a random marker and returns the new text. the output parameter
 153         # $content will be an associative array filled with data on the form
 154         # $unique_marker => content.
 155
 156         # If $content is already set, the additional entries will be appended
 157
 158         # If $tag is set to STRIP_COMMENTS, the function will extract
 159         # <!-- HTML comments -->
 160
 161         /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ""){
 162                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 163                 if ( !$content ) {
 164                         $content = array( );
 165                 }
 166                 $n = 1;
 167                 $stripped = '';
 168
 169                 while ( '' != $text ) {
 170                         if($tag==STRIP_COMMENTS) {
 171                                 $p = preg_split( '/<!--/i', $text, 2 );
 172                         } else {
 173                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 174                         }
 175                         $stripped .= $p[0];
 176                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 177                                 $text = '';
 178                         } else {
 179                                 if($tag==STRIP_COMMENTS) {
 180                                         $q = preg_split( '/-->/i', $p[1], 2 );
 181                                 } else {
 182                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 183                                 }
 184                                 $marker = $rnd . sprintf('%08X', $n++);
 185                                 $content[$marker] = $q[0];
 186                                 $stripped .= $marker;
 187                                 $text = $q[1];
 188                         }
 189                 }
 190                 return $stripped;
 191         }
 192
 193         # Strips and renders <nowiki>, <pre>, <math>, <hiero>
 194         # If $render is set, performs necessary rendering operations on plugins
 195         # Returns the text, and fills an array with data needed in unstrip()
 196         # If the $state is already a valid strip state, it adds to the state
 197
 198         # When $stripcomments is set, HTML comments <!-- like this -->
 199         # will be stripped in addition to other tags. This is important
 200         # for section editing, where these comments cause confusion when
 201         # counting the sections in the wikisource
 202         function strip( $text, &$state, $stripcomments = false ) {
 203                 $render = ($this->mOutputType == OT_HTML);
 204                 $nowiki_content = array();
 205                 $math_content = array();
 206                 $pre_content = array();
 207                 $comment_content = array();
 208                 $ext_content = array();
 209
 210                 # Replace any instances of the placeholders
 211                 $uniq_prefix = UNIQ_PREFIX;
 212                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 213
 214
 215                 # nowiki
 216                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 217                 foreach( $nowiki_content as $marker => $content ){
 218                         if( $render ){
 219                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 220                         } else {
 221                                 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
 222                         }
 223                 }
 224
 225                 # math
 226                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 227                 foreach( $math_content as $marker => $content ){
 228                         if( $render ) {
 229                                 if( $this->mOptions->getUseTeX() ) {
 230                                         $math_content[$marker] = renderMath( $content );
 231                                 } else {
 232                                         $math_content[$marker] = "&lt;math&gt;$content&lt;math&gt;";
 233                                 }
 234                         } else {
 235                                 $math_content[$marker] = "<math>$content</math>";
 236                         }
 237                 }
 238
 239                 # pre
 240                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 241                 foreach( $pre_content as $marker => $content ){
 242                         if( $render ){
 243                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 244                         } else {
 245                                 $pre_content[$marker] = "<pre>$content</pre>";
 246                         }
 247                 }
 248
 249                 # Comments
 250                 if($stripcomments) {
 251                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 252                         foreach( $comment_content as $marker => $content ){
 253                                 $comment_content[$marker] = "<!--$content-->";
 254                         }
 255                 }
 256
 257                 # Extensions
 258                 foreach ( $this->mTagHooks as $tag => $callback ) {
 259                         $ext_contents[$tag] = array();
 260                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 261                         foreach( $ext_content[$tag] as $marker => $content ) {
 262                                 if ( $render ) {
 263                                         $ext_content[$tag][$marker] = $callback( $content );
 264                                 } else {
 265                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 266                                 }
 267                         }
 268                 }
 269
 270                 # Merge state with the pre-existing state, if there is one
 271                 if ( $state ) {
 272                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 273                         $state['math'] = $state['math'] + $math_content;
 274                         $state['pre'] = $state['pre'] + $pre_content;
 275                         $state['comment'] = $state['comment'] + $comment_content;
 276
 277                         foreach( $ext_content as $tag => $array ) {
 278                                 if ( array_key_exists( $tag, $state ) ) {
 279                                         $state[$tag] = $state[$tag] + $array;
 280                                 }
 281                         }
 282                 } else {
 283                         $state = array(
 284                           'nowiki' => $nowiki_content,
 285                           'math' => $math_content,
 286                           'pre' => $pre_content,
 287                           'comment' => $comment_content,
 288                         ) + $ext_content;
 289                 }
 290                 return $text;
 291         }
 292
 293         # always call unstripNoWiki() after this one
 294         function unstrip( $text, &$state ) {
 295                 # Must expand in reverse order, otherwise nested tags will be corrupted
 296                 $contentDict = end( $state );
 297                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 298                         if( key($state) != 'nowiki') {
 299                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 300                                         $text = str_replace( key( $contentDict ), $content, $text );
 301                                 }
 302                         }
 303                 }
 304
 305                 return $text;
 306         }
 307         # always call this after unstrip() to preserve the order
 308         function unstripNoWiki( $text, &$state ) {
 309                 # Must expand in reverse order, otherwise nested tags will be corrupted
 310                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 311                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 312                 }
 313
 314                 return $text;
 315         }
 316
 317         # Add an item to the strip state
 318         # Returns the unique tag which must be inserted into the stripped text
 319         # The tag will be replaced with the original text in unstrip()
 320
 321         function insertStripItem( $text, &$state ) {
 322                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 323                 if ( !$state ) {
 324                         $state = array(
 325                           'nowiki' => array(),
 326                           'math' => array(),
 327                           'pre' => array()
 328                         );
 329                 }
 330                 $state['item'][$rnd] = $text;
 331                 return $rnd;
 332         }
 333
 334         # categoryMagic
 335         # generate a list of subcategories and pages for a category
 336         # depending on wfMsg("usenewcategorypage") it either calls the new
 337         # or the old code. The new code will not work properly for some
 338         # languages due to sorting issues, so they might want to turn it
 339         # off.
 340         function categoryMagic() {
 341                 $msg = wfMsg('usenewcategorypage');
 342                 if ( '0' == @$msg[0] )
 343                 {
 344                         return $this->oldCategoryMagic();
 345                 } else {
 346                         return $this->newCategoryMagic();
 347                 }
 348         }
 349
 350         # This method generates the list of subcategories and pages for a category
 351         function oldCategoryMagic () {
 352                 global $wgLang , $wgUser ;
 353                 $fname = 'Parser::oldCategoryMagic';
 354
 355                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 356
 357                 $cns = Namespace::getCategory() ;
 358                 if ( $this->mTitle->getNamespace() != $cns ) return "" ; # This ain't a category page
 359
 360                 $r = "<br style=\"clear:both;\"/>\n";
 361
 362
 363                 $sk =& $wgUser->getSkin() ;
 364
 365                 $articles = array() ;
 366                 $children = array() ;
 367                 $data = array () ;
 368                 $id = $this->mTitle->getArticleID() ;
 369
 370                 # FIXME: add limits
 371                 $dbr =& wfGetDB( DB_SLAVE );
 372                 $cur = $dbr->tableName( 'cur' );
 373                 $categorylinks = $dbr->tableName( 'categorylinks' );
 374
 375                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 376                 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM $cur,$categorylinks " .
 377                         "WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 378                 $res = $dbr->query( $sql, $fname ) ;
 379                 while ( $x = $dbr->fetchObject ( $res ) ) $data[] = $x ;
 380
 381                 # For all pages that link to this category
 382                 foreach ( $data AS $x )
 383                 {
 384                         $t = $wgLang->getNsText ( $x->cur_namespace ) ;
 385                         if ( $t != "" ) $t .= ":" ;
 386                         $t .= $x->cur_title ;
 387
 388                         if ( $x->cur_namespace == $cns ) {
 389                                 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
 390                         } else {
 391                                 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
 392                         }
 393                 }
 394                 $dbr->freeResult ( $res ) ;
 395
 396                 # Showing subcategories
 397                 if ( count ( $children ) > 0 ) {
 398                         $r .= '<h2>'.wfMsg('subcategories')."</h2>\n" ;
 399                         $r .= implode ( ', ' , $children ) ;
 400                 }
 401
 402                 # Showing pages in this category
 403                 if ( count ( $articles ) > 0 ) {
 404                         $ti = $this->mTitle->getText() ;
 405                         $h =  wfMsg( 'category_header', $ti );
 406                         $r .= "<h2>{$h}</h2>\n" ;
 407                         $r .= implode ( ', ' , $articles ) ;
 408                 }
 409
 410                 return $r ;
 411         }
 412
 413
 414
 415         function newCategoryMagic () {
 416                 global $wgLang , $wgUser ;
 417                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 418
 419                 $cns = Namespace::getCategory() ;
 420                 if ( $this->mTitle->getNamespace() != $cns ) return '' ; # This ain't a category page
 421
 422                 $r = "<br style=\"clear:both;\"/>\n";
 423
 424
 425                 $sk =& $wgUser->getSkin() ;
 426
 427                 $articles = array() ;
 428                 $articles_start_char = array();
 429                 $children = array() ;
 430                 $children_start_char = array();
 431                 $data = array () ;
 432                 $id = $this->mTitle->getArticleID() ;
 433
 434                 # FIXME: add limits
 435                 $dbr =& wfGetDB( DB_SLAVE );
 436                 $cur = $dbr->tableName( 'cur' );
 437                 $categorylinks = $dbr->tableName( 'categorylinks' );
 438
 439                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 440                 $sql = "SELECT DISTINCT cur_title,cur_namespace,cl_sortkey FROM " .
 441                         "$cur,$categorylinks WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 442                 $res = $dbr->query ( $sql ) ;
 443                 while ( $x = $dbr->fetchObject ( $res ) )
 444                 {
 445                         $t = $ns = $wgLang->getNsText ( $x->cur_namespace ) ;
 446                         if ( $t != '' ) $t .= ':' ;
 447                         $t .= $x->cur_title ;
 448
 449                         if ( $x->cur_namespace == $cns ) {
 450                                 $ctitle = str_replace( '_',' ',$x->cur_title );
 451                                 array_push ( $children, $sk->makeKnownLink ( $t, $ctitle ) ) ; # Subcategory
 452
 453                                 // If there's a link from Category:A to Category:B, the sortkey of the resulting
 454                                 // entry in the categorylinks table is Category:A, not A, which it SHOULD be.
 455                                 // Workaround: If sortkey == "Category:".$title, than use $title for sorting,
 456                                 // else use sortkey...
 457                                 if ( ($ns.":".$ctitle) ==  $x->cl_sortkey ) {
 458                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cur_title ) );
 459                                 } else {
 460                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 461                                 }
 462                         } else {
 463                                 array_push ( $articles , $sk->makeKnownLink ( $t ) ) ; # Page in this category
 464                                 array_push ( $articles_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 465                         }
 466                 }
 467                 $dbr->freeResult ( $res ) ;
 468
 469                 $ti = $this->mTitle->getText() ;
 470
 471                 # Don't show subcategories section if there are none.
 472                 if ( count ( $children ) > 0 )
 473                 {
 474                         # Showing subcategories
 475                         $r .= '<h2>' . wfMsg( 'subcategories' ) . "</h2>\n"
 476                                 . wfMsg( 'subcategorycount', count( $children ) );
 477                         if ( count ( $children ) > 6 ) {
 478
 479                                 // divide list into three equal chunks
 480                                 $chunk = (int) (count ( $children ) / 3);
 481
 482                                 // get and display header
 483                                 $r .= '<table width="100%"><tr valign="top">';
 484
 485                                 $startChunk = 0;
 486                                 $endChunk = $chunk;
 487
 488                                 // loop through the chunks
 489                                 for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 490                                         $chunkIndex < 3;
 491                                         $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 492                                 {
 493
 494                                         $r .= '<td><ul>';
 495                                         // output all subcategories to category
 496                                         for ($index = $startChunk ;
 497                                                 $index < $endChunk && $index < count($children);
 498                                                 $index++ )
 499                                         {
 500                                                 // check for change of starting letter or begging of chunk
 501                                                 if ( ($children_start_char[$index] != $children_start_char[$index - 1])
 502                                                         || ($index == $startChunk) )
 503                                                 {
 504                                                         $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 505                                                 }
 506
 507                                                 $r .= "<li>{$children[$index]}</li>";
 508                                         }
 509                                         $r .= '</ul></td>';
 510
 511
 512                                 }
 513                                 $r .= '</tr></table>';
 514                         } else {
 515                                 // for short lists of subcategories to category.
 516
 517                                 $r .= "<h3>{$children_start_char[0]}</h3>\n";
 518                                 $r .= '<ul><li>'.$children[0].'</li>';
 519                                 for ($index = 1; $index < count($children); $index++ )
 520                                 {
 521                                         if ($children_start_char[$index] != $children_start_char[$index - 1])
 522                                         {
 523                                                 $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 524                                         }
 525
 526                                         $r .= "<li>{$children[$index]}</li>";
 527                                 }
 528                                 $r .= '</ul>';
 529                         }
 530                 } # END of if ( count($children) > 0 )
 531
 532                 $r .= '<h2>' . wfMsg( 'category_header', $ti ) . "</h2>\n" .
 533                         wfMsg( 'categoryarticlecount', count( $articles ) );
 534
 535                 # Showing articles in this category
 536                 if ( count ( $articles ) > 6) {
 537                         $ti = $this->mTitle->getText() ;
 538
 539                         // divide list into three equal chunks
 540                         $chunk = (int) (count ( $articles ) / 3);
 541
 542                         // get and display header
 543                         $r .= '<table width="100%"><tr valign="top">';
 544
 545                         // loop through the chunks
 546                         for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 547                                 $chunkIndex < 3;
 548                                 $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 549                         {
 550
 551                                 $r .= '<td><ul>';
 552
 553                                 // output all articles in category
 554                                 for ($index = $startChunk ;
 555                                         $index < $endChunk && $index < count($articles);
 556                                         $index++ )
 557                                 {
 558                                         // check for change of starting letter or begging of chunk
 559                                         if ( ($articles_start_char[$index] != $articles_start_char[$index - 1])
 560                                                 || ($index == $startChunk) )
 561                                         {
 562                                                 $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 563                                         }
 564
 565                                         $r .= "<li>{$articles[$index]}</li>";
 566                                 }
 567                                 $r .= '</ul></td>';
 568
 569
 570                         }
 571                         $r .= '</tr></table>';
 572                 } elseif ( count ( $articles )  > 0) {
 573                         // for short lists of articles in categories.
 574                         $ti = $this->mTitle->getText() ;
 575
 576                         $r .= '<h3>'.$articles_start_char[0]."</h3>\n";
 577                         $r .= '<ul><li>'.$articles[0].'</li>';
 578                         for ($index = 1; $index < count($articles); $index++ )
 579                         {
 580                                 if ($articles_start_char[$index] != $articles_start_char[$index - 1])
 581                                 {
 582                                         $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 583                                 }
 584
 585                                 $r .= "<li>{$articles[$index]}</li>";
 586                         }
 587                         $r .= '</ul>';
 588                 }
 589
 590
 591                 return $r ;
 592         }
 593
 594         # Return allowed HTML attributes
 595         function getHTMLattrs () {
 596                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 597                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 598                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 599                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 600                                 /* FONT */ 'type', 'start', 'value', 'compact',
 601                                 /* For various lists, mostly deprecated but safe */
 602                                 'summary', 'width', 'border', 'frame', 'rules',
 603                                 'cellspacing', 'cellpadding', 'valign', 'char',
 604                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 605                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 606                                 'id', 'class', 'name', 'style' /* For CSS */
 607                                 );
 608                 return $htmlattrs ;
 609         }
 610
 611         # Remove non approved attributes and javascript in css
 612         function fixTagAttributes ( $t ) {
 613                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 614                 $htmlattrs = $this->getHTMLattrs() ;
 615
 616                 # Strip non-approved attributes from the tag
 617                 $t = preg_replace(
 618                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 619                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 620                         $t);
 621                 # Strip javascript "expression" from stylesheets. Brute force approach:
 622                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 623
 624                 if( preg_match(
 625                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 626                         wfMungeToUtf8( $t ) ) )
 627                 {
 628                         $t='';
 629                 }
 630
 631                 return trim ( $t ) ;
 632         }
 633
 634         # interface with html tidy, used if $wgUseTidy = true
 635         function tidy ( $text ) {
 636                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 637                 global $wgInputEncoding, $wgOutputEncoding;
 638                 $fname = 'Parser::tidy';
 639                 wfProfileIn( $fname );
 640
 641                 $cleansource = '';
 642                 switch(strtoupper($wgOutputEncoding)) {
 643                         case 'ISO-8859-1':
 644                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 645                                 break;
 646                         case 'UTF-8':
 647                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 648                                 break;
 649                         default:
 650                                 $wgTidyOpts .= ' -raw';
 651                         }
 652
 653                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 654 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 655 '<head><title>test</title></head><body>'.$text.'</body></html>';
 656                 $descriptorspec = array(
 657                         0 => array('pipe', 'r'),
 658                         1 => array('pipe', 'w'),
 659                         2 => array('file', '/dev/null', 'a')
 660                 );
 661                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 662                 if (is_resource($process)) {
 663                         fwrite($pipes[0], $wrappedtext);
 664                         fclose($pipes[0]);
 665                         while (!feof($pipes[1])) {
 666                                 $cleansource .= fgets($pipes[1], 1024);
 667                         }
 668                         fclose($pipes[1]);
 669                         $return_value = proc_close($process);
 670                 }
 671
 672                 wfProfileOut( $fname );
 673
 674                 if( $cleansource == '' && $text != '') {
 675                         wfDebug( "Tidy error detected!\n" );
 676                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 677                 } else {
 678                         return $cleansource;
 679                 }
 680         }
 681
 682         # parse the wiki syntax used to render tables
 683         function doTableStuff ( $t ) {
 684                 $fname = 'Parser::doTableStuff';
 685                 wfProfileIn( $fname );
 686
 687                 $t = explode ( "\n" , $t ) ;
 688                 $td = array () ; # Is currently a td tag open?
 689                 $ltd = array () ; # Was it TD or TH?
 690                 $tr = array () ; # Is currently a tr tag open?
 691                 $ltr = array () ; # tr attributes
 692                 foreach ( $t AS $k => $x )
 693                 {
 694                         $x = trim ( $x ) ;
 695                         $fc = substr ( $x , 0 , 1 ) ;
 696                         if ( '{|' == substr ( $x , 0 , 2 ) )
 697                         {
 698                                 $t[$k] = "\n<table " . $this->fixTagAttributes ( substr ( $x , 2 ) ) . '>' ;
 699                                 array_push ( $td , false ) ;
 700                                 array_push ( $ltd , '' ) ;
 701                                 array_push ( $tr , false ) ;
 702                                 array_push ( $ltr , '' ) ;
 703                         }
 704                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 705                         else if ( '|}' == substr ( $x , 0 , 2 ) )
 706                         {
 707                                 $z = "</table>\n" ;
 708                                 $l = array_pop ( $ltd ) ;
 709                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 710                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 711                                 array_pop ( $ltr ) ;
 712                                 $t[$k] = $z ;
 713                         }
 714                         else if ( '|-' == substr ( $x , 0 , 2 ) ) # Allows for |---------------
 715                         {
 716                                 $x = substr ( $x , 1 ) ;
 717                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 718                                 $z = '' ;
 719                                 $l = array_pop ( $ltd ) ;
 720                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 721                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 722                                 array_pop ( $ltr ) ;
 723                                 $t[$k] = $z ;
 724                                 array_push ( $tr , false ) ;
 725                                 array_push ( $td , false ) ;
 726                                 array_push ( $ltd , '' ) ;
 727                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 728                         }
 729                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) # Caption
 730                         {
 731                                 if ( '|+' == substr ( $x , 0 , 2 ) )
 732                                 {
 733                                         $fc = '+' ;
 734                                         $x = substr ( $x , 1 ) ;
 735                                 }
 736                                 $after = substr ( $x , 1 ) ;
 737                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 738                                 $after = explode ( '||' , $after ) ;
 739                                 $t[$k] = '' ;
 740                                 foreach ( $after AS $theline )
 741                                 {
 742                                         $z = '' ;
 743                                         if ( $fc != '+' )
 744                                         {
 745                                                 $tra = array_pop ( $ltr ) ;
 746                                                 if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
 747                                                 array_push ( $tr , true ) ;
 748                                                 array_push ( $ltr , '' ) ;
 749                                         }
 750
 751                                         $l = array_pop ( $ltd ) ;
 752                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 753                                         if ( $fc == '|' ) $l = 'td' ;
 754                                         else if ( $fc == '!' ) $l = 'th' ;
 755                                         else if ( $fc == '+' ) $l = 'caption' ;
 756                                         else $l = '' ;
 757                                         array_push ( $ltd , $l ) ;
 758                                         $y = explode ( '|' , $theline , 2 ) ;
 759                                         if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 760                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 761                                         $t[$k] .= $y ;
 762                                         array_push ( $td , true ) ;
 763                                 }
 764                         }
 765                 }
 766
 767                 # Closing open td, tr && table
 768                 while ( count ( $td ) > 0 )
 769                 {
 770                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 771                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 772                         $t[] = '</table>' ;
 773                 }
 774
 775                 $t = implode ( "\n" , $t ) ;
 776                 #               $t = $this->removeHTMLtags( $t );
 777                 wfProfileOut( $fname );
 778                 return $t ;
 779         }
 780
 781         # Parses the text and adds the result to the strip state
 782         # Returns the strip tag
 783         function stripParse( $text, $newline, $args )
 784         {
 785                 $text = $this->strip( $text, $this->mStripState );
 786                 $text = $this->internalParse( $text, (bool)$newline, $args, false );
 787                 return $newline.$this->insertStripItem( $text, $this->mStripState );
 788         }
 789
 790         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 791                 $fname = 'Parser::internalParse';
 792                 wfProfileIn( $fname );
 793
 794                 $text = $this->removeHTMLtags( $text );
 795                 $text = $this->replaceVariables( $text, $args );
 796
 797                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 798
 799                 $text = $this->doHeadings( $text );
 800                 if($this->mOptions->getUseDynamicDates()) {
 801                         global $wgDateFormatter;
 802                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 803                 }
 804                 $text = $this->doAllQuotes( $text );
 805                 // $text = $this->doExponent( $text );
 806                 $text = $this->replaceExternalLinks( $text );
 807                 $text = $this->replaceInternalLinks ( $text );
 808                 $text = $this->replaceInternalLinks ( $text );
 809                 //$text = $this->doTokenizedParser ( $text );
 810                 $text = $this->doTableStuff ( $text ) ;
 811                 $text = $this->magicISBN( $text );
 812                 $text = $this->magicGEO( $text );
 813                 $text = $this->magicRFC( $text );
 814                 $text = $this->formatHeadings( $text, $isMain );
 815                 $sk =& $this->mOptions->getSkin();
 816                 $text = $sk->transformContent( $text );
 817
 818                 if ( $isMain && !isset ( $this->categoryMagicDone ) ) {
 819                         $text .= $this->categoryMagic () ;
 820                         $this->categoryMagicDone = true ;
 821                 }
 822
 823                 wfProfileOut( $fname );
 824                 return $text;
 825         }
 826
 827         # Parse ^^ tokens and return html
 828         /* private */ function doExponent ( $text )
 829         {
 830                 $fname = 'Parser::doExponent';
 831                 wfProfileIn( $fname);
 832                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 833                 wfProfileOut( $fname);
 834                 return $text;
 835         }
 836
 837     # Parse headers and return html
 838         /* private */ function doHeadings( $text ) {
 839                 $fname = 'Parser::doHeadings';
 840                 wfProfileIn( $fname );
 841                 for ( $i = 6; $i >= 1; --$i ) {
 842                         $h = substr( '======', 0, $i );
 843                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 844                           "<h{$i}>\\1</h{$i}>\\2", $text );
 845                 }
 846                 wfProfileOut( $fname );
 847                 return $text;
 848         }
 849
 850         /* private */ function doAllQuotes( $text ) {
 851                 $fname = 'Parser::doAllQuotes';
 852                 wfProfileIn( $fname );
 853                 $outtext = '';
 854                 $lines = explode( "\n", $text );
 855                 foreach ( $lines as $line ) {
 856                         $outtext .= $this->doQuotes ( '', $line, '' ) . "\n";
 857                 }
 858                 $outtext = substr($outtext, 0,-1);
 859                 wfProfileOut( $fname );
 860                 return $outtext;
 861         }
 862
 863         /* private */ function doQuotes( $pre, $text, $mode ) {
 864                 if ( preg_match( "/^(.*)''(.*)$/sU", $text, $m ) ) {
 865                         $m1_strong = ($m[1] == "") ? "" : "<strong>{$m[1]}</strong>";
 866                         $m1_em = ($m[1] == "") ? "" : "<em>{$m[1]}</em>";
 867                         if ( substr ($m[2], 0, 1) == '\'' ) {
 868                                 $m[2] = substr ($m[2], 1);
 869                                 if ($mode == 'em') {
 870                                         return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'emstrong' );
 871                                 } else if ($mode == 'strong') {
 872                                         return $m1_strong . $this->doQuotes ( '', $m[2], '' );
 873                                 } else if (($mode == 'emstrong') || ($mode == 'both')) {
 874                                         return $this->doQuotes ( '', $pre.$m1_strong.$m[2], 'em' );
 875                                 } else if ($mode == 'strongem') {
 876                                         return "<strong>{$pre}{$m1_em}</strong>" . $this->doQuotes ( '', $m[2], 'em' );
 877                                 } else {
 878                                         return $m[1] . $this->doQuotes ( '', $m[2], 'strong' );
 879                                 }
 880                         } else {
 881                                 if ($mode == 'strong') {
 882                                         return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'strongem' );
 883                                 } else if ($mode == 'em') {
 884                                         return $m1_em . $this->doQuotes ( '', $m[2], '' );
 885                                 } else if ($mode == 'emstrong') {
 886                                         return "<em>{$pre}{$m1_strong}</em>" . $this->doQuotes ( '', $m[2], 'strong' );
 887                                 } else if (($mode == 'strongem') || ($mode == 'both')) {
 888                                         return $this->doQuotes ( '', $pre.$m1_em.$m[2], 'strong' );
 889                                 } else {
 890                                         return $m[1] . $this->doQuotes ( '', $m[2], 'em' );
 891                                 }
 892                         }
 893                 } else {
 894                         $text_strong = ($text == '') ? '' : "<strong>{$text}</strong>";
 895                         $text_em = ($text == '') ? '' : "<em>{$text}</em>";
 896                         if ($mode == '') {
 897                                 return $pre . $text;
 898                         } else if ($mode == 'em') {
 899                                 return $pre . $text_em;
 900                         } else if ($mode == 'strong') {
 901                                 return $pre . $text_strong;
 902                         } else if ($mode == 'strongem') {
 903                                 return (($pre == '') && ($text == '')) ? '' : "<strong>{$pre}{$text_em}</strong>";
 904                         } else {
 905                                 return (($pre == '') && ($text == '')) ? '' : "<em>{$pre}{$text_strong}</em>";
 906                         }
 907                 }
 908         }
 909
 910         # Note: we have to do external links before the internal ones,
 911         # and otherwise take great care in the order of things here, so
 912         # that we don't end up interpreting some URLs twice.
 913
 914         /* private */ function replaceExternalLinks( $text ) {
 915                 $fname = 'Parser::replaceExternalLinks';
 916                 wfProfileIn( $fname );
 917                 $text = $this->subReplaceExternalLinks( $text, 'http', true );
 918                 $text = $this->subReplaceExternalLinks( $text, 'https', true );
 919                 $text = $this->subReplaceExternalLinks( $text, 'ftp', false );
 920                 $text = $this->subReplaceExternalLinks( $text, 'irc', false );
 921                 $text = $this->subReplaceExternalLinks( $text, 'gopher', false );
 922                 $text = $this->subReplaceExternalLinks( $text, 'news', false );
 923                 $text = $this->subReplaceExternalLinks( $text, 'mailto', false );
 924                 wfProfileOut( $fname );
 925                 return $text;
 926         }
 927
 928         /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber ) {
 929                 $unique = '4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3';
 930                 $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
 931
 932                 # this is  the list of separators that should be ignored if they
 933                 # are the last character of an URL but that should be included
 934                 # if they occur within the URL, e.g. "go to www.foo.com, where .."
 935                 # in this case, the last comma should not become part of the URL,
 936                 # but in "www.foo.com/123,2342,32.htm" it should.
 937                 $sep = ",;\.:";
 938                 $fnc = 'A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF';
 939                 $images = 'gif|png|jpg|jpeg';
 940
 941                 # PLEASE NOTE: The curly braces { } are not part of the regex,
 942                 # they are interpreted as part of the string (used to tell PHP
 943                 # that the content of the string should be inserted there).
 944                 $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
 945                   "((?i){$images})([^{$uc}]|$)/";
 946
 947                 $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
 948                 $sk =& $this->mOptions->getSkin();
 949
 950                 if ( $autonumber and $this->mOptions->getAllowExternalImages() ) { # Use img tags only for HTTP urls
 951                         $s = preg_replace( $e1, '\\1' . $sk->makeImage( "{$unique}:\\3" .
 952                           '/\\4.\\5', '\\4.\\5' ) . '\\6', $s );
 953                 }
 954                 $s = preg_replace( $e2, '\\1' . "<a href=\"{$unique}:\\3\"" .
 955                   $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
 956                   "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
 957                   '</a>\\5', $s );
 958                 $s = str_replace( $unique, $protocol, $s );
 959
 960                 $a = explode( "[{$protocol}:", " " . $s );
 961                 $s = array_shift( $a );
 962                 $s = substr( $s, 1 );
 963
 964                 # Regexp for URL in square brackets
 965                 $e1 = "/^([{$uc}{$sep}]+)\\](.*)\$/sD";
 966                 # Regexp for URL with link text in square brackets
 967                 $e2 = "/^([{$uc}{$sep}]+)\\s+([^\\]]+)\\](.*)\$/sD";
 968
 969                 foreach ( $a as $line ) {
 970
 971                         # CASE 1: Link in square brackets, e.g.
 972                         # some text [http://domain.tld/some.link] more text
 973                         if ( preg_match( $e1, $line, $m ) ) {
 974                                 $link = "{$protocol}:{$m[1]}";
 975                                 $trail = $m[2];
 976                                 if ( $autonumber ) { $text = "[" . ++$this->mAutonumber . "]"; }
 977                                 else { $text = wfEscapeHTML( $link ); }
 978                         }
 979
 980                         # CASE 2: Link with link text and text directly following it, e.g.
 981                         # This is a collection of [http://domain.tld/some.link link]s
 982                         else if ( preg_match( $e2, $line, $m ) ) {
 983                                 $link = "{$protocol}:{$m[1]}";
 984                                 $text = $m[2];
 985                                 $dtrail = '';
 986                                 $trail = $m[3];
 987                                 if ( preg_match( wfMsg ('linktrail'), $trail, $m2 ) ) {
 988                                         $dtrail = $m2[1];
 989                                         $trail = $m2[2];
 990                                 }
 991                         }
 992
 993                         # CASE 3: Nothing matches, just output the source text
 994                         else {
 995                                 $s .= "[{$protocol}:" . $line;
 996                                 continue;
 997                         }
 998
 999                         if( $link == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
1000                                 $paren = '';
1001                         } else {
1002                                 # Expand the URL for printable version
1003                                 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
1004                         }
1005                         $la = $sk->getExternalLinkAttributes( $link, $text );
1006                         $s .= "<a href='{$link}'{$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
1007
1008                 }
1009                 return $s;
1010         }
1011
1012
1013         /* private */ function replaceInternalLinks( $s ) {
1014                 global $wgLang, $wgLinkCache;
1015                 global $wgNamespacesWithSubpages, $wgLanguageCode;
1016                 static $fname = 'Parser::replaceInternalLinks' ;
1017                 wfProfileIn( $fname );
1018
1019                 wfProfileIn( $fname.'-setup' );
1020                 static $tc = FALSE;
1021                 # the % is needed to support urlencoded titles as well
1022                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1023                 $sk =& $this->mOptions->getSkin();
1024
1025                 $a = explode( '[[', ' ' . $s );
1026                 $s = array_shift( $a );
1027                 $s = substr( $s, 1 );
1028
1029                 # Match a link having the form [[namespace:link|alternate]]trail
1030                 static $e1 = FALSE;
1031                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1032                 # Match the end of a line for a word that's not followed by whitespace,
1033                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1034                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1035
1036                 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
1037                 # Special and Media are pseudo-namespaces; no pages actually exist in them
1038                 static $image = FALSE;
1039                 static $special = FALSE;
1040                 static $media = FALSE;
1041                 static $category = FALSE;
1042                 if ( !$image ) { $image = Namespace::getImage(); }
1043                 if ( !$special ) { $special = Namespace::getSpecial(); }
1044                 if ( !$media ) { $media = Namespace::getMedia(); }
1045                 if ( !$category ) { $category = Namespace::getCategory(); }
1046
1047                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1048
1049                 if ( $useLinkPrefixExtension ) {
1050                         if ( preg_match( $e2, $s, $m ) ) {
1051                                 $first_prefix = $m[2];
1052                                 $s = $m[1];
1053                         } else {
1054                                 $first_prefix = false;
1055                         }
1056                 } else {
1057                         $prefix = '';
1058                 }
1059
1060                 wfProfileOut( $fname.'-setup' );
1061
1062                 foreach ( $a as $line ) {
1063                         wfProfileIn( $fname.'-prefixhandling' );
1064                         if ( $useLinkPrefixExtension ) {
1065                                 if ( preg_match( $e2, $s, $m ) ) {
1066                                         $prefix = $m[2];
1067                                         $s = $m[1];
1068                                 } else {
1069                                         $prefix='';
1070                                 }
1071                                 # first link
1072                                 if($first_prefix) {
1073                                         $prefix = $first_prefix;
1074                                         $first_prefix = false;
1075                                 }
1076                         }
1077                         wfProfileOut( $fname.'-prefixhandling' );
1078
1079                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1080                                 $text = $m[2];
1081                                 # fix up urlencoded title texts
1082                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1083                                 $trail = $m[3];
1084                         } else { # Invalid form; output directly
1085                                 $s .= $prefix . '[[' . $line ;
1086                                 continue;
1087                         }
1088
1089                         /* Valid link forms:
1090                         Foobar -- normal
1091                         :Foobar -- override special treatment of prefix (images, language links)
1092                         /Foobar -- convert to CurrentPage/Foobar
1093                         /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1094                         */
1095                         $c = substr($m[1],0,1);
1096                         $noforce = ($c != ':');
1097                         if( $c == '/' ) { # subpage
1098                                 if(substr($m[1],-1,1)=='/') {                 # / at end means we don't want the slash to be shown
1099                                         $m[1]=substr($m[1],1,strlen($m[1])-2);
1100                                         $noslash=$m[1];
1101                                 } else {
1102                                         $noslash=substr($m[1],1);
1103                                 }
1104                                 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) { # subpages allowed here
1105                                         $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1106                                         if( '' == $text ) {
1107                                                 $text= $m[1];
1108                                         } # this might be changed for ugliness reasons
1109                                 } else {
1110                                         $link = $noslash; # no subpage allowed, use standard link
1111                                 }
1112                         } elseif( $noforce ) { # no subpage
1113                                 $link = $m[1];
1114                         } else {
1115                                 $link = substr( $m[1], 1 );
1116                         }
1117                         $wasblank = ( '' == $text );
1118                         if( $wasblank )
1119                         $text = $link;
1120
1121                         $nt = Title::newFromText( $link );
1122                         if( !$nt ) {
1123                                 $s .= $prefix . '[[' . $line;
1124                                 continue;
1125                         }
1126                         $ns = $nt->getNamespace();
1127                         $iw = $nt->getInterWiki();
1128                         if( $noforce ) {
1129                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1130                                         array_push( $this->mOutput->mLanguageLinks, $nt->getPrefixedText() );
1131                                         $tmp = $prefix . $trail ;
1132                                         $s .= (trim($tmp) == '')? '': $tmp;
1133                                         continue;
1134                                 }
1135                                 if ( $ns == $image ) {
1136                                         $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1137                                         $wgLinkCache->addImageLinkObj( $nt );
1138                                         continue;
1139                                 }
1140                                 if ( $ns == $category ) {
1141                                         $t = $nt->getText() ;
1142                                         $nnt = Title::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
1143
1144                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1145                                         $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1146                                         $wgLinkCache->resume();
1147
1148                                         $sortkey = $wasblank ? $this->mTitle->getPrefixedText() : $text;
1149                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1150                                         $this->mOutput->mCategoryLinks[] = $t ;
1151                                         $s .= $prefix . $trail ;
1152                                         continue;
1153                                 }
1154                         }
1155                         if( ( $nt->getPrefixedText() == $this->mTitle->getPrefixedText() ) &&
1156                         ( strpos( $link, '#' ) == FALSE ) ) {
1157                                 # Self-links are handled specially; generally de-link and change to bold.
1158                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1159                                 continue;
1160                         }
1161
1162                         if( $ns == $media ) {
1163                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1164                                 $wgLinkCache->addImageLinkObj( $nt );
1165                                 continue;
1166                         } elseif( $ns == $special ) {
1167                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1168                                 continue;
1169                         }
1170                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1171                 }
1172                 wfProfileOut( $fname );
1173                 return $s;
1174         }
1175
1176         # Some functions here used by doBlockLevels()
1177         #
1178         /* private */ function closeParagraph() {
1179                 $result = '';
1180                 if ( '' != $this->mLastSection ) {
1181                         $result = '</' . $this->mLastSection  . ">\n";
1182                 }
1183                 $this->mInPre = false;
1184                 $this->mLastSection = '';
1185                 return $result;
1186         }
1187         # getCommon() returns the length of the longest common substring
1188         # of both arguments, starting at the beginning of both.
1189         #
1190         /* private */ function getCommon( $st1, $st2 ) {
1191                 $fl = strlen( $st1 );
1192                 $shorter = strlen( $st2 );
1193                 if ( $fl < $shorter ) { $shorter = $fl; }
1194
1195                 for ( $i = 0; $i < $shorter; ++$i ) {
1196                         if ( $st1{$i} != $st2{$i} ) { break; }
1197                 }
1198                 return $i;
1199         }
1200         # These next three functions open, continue, and close the list
1201         # element appropriate to the prefix character passed into them.
1202         #
1203         /* private */ function openList( $char )
1204     {
1205                 $result = $this->closeParagraph();
1206
1207                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1208                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1209                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1210                 else if ( ';' == $char ) {
1211                         $result .= '<dl><dt>';
1212                         $this->mDTopen = true;
1213                 }
1214                 else { $result = '<!-- ERR 1 -->'; }
1215
1216                 return $result;
1217         }
1218
1219         /* private */ function nextItem( $char ) {
1220                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1221                 else if ( ':' == $char || ';' == $char ) {
1222                         $close = "</dd>";
1223                         if ( $this->mDTopen ) { $close = '</dt>'; }
1224                         if ( ';' == $char ) {
1225                                 $this->mDTopen = true;
1226                                 return $close . '<dt>';
1227                         } else {
1228                                 $this->mDTopen = false;
1229                                 return $close . '<dd>';
1230                         }
1231                 }
1232                 return '<!-- ERR 2 -->';
1233         }
1234
1235         /* private */function closeList( $char ) {
1236                 if ( '*' == $char ) { $text = '</li></ul>'; }
1237                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1238                 else if ( ':' == $char ) {
1239                         if ( $this->mDTopen ) {
1240                                 $this->mDTopen = false;
1241                                 $text = '</dt></dl>';
1242                         } else {
1243                                 $text = '</dd></dl>';
1244                         }
1245                 }
1246                 else {  return '<!-- ERR 3 -->'; }
1247                 return $text."\n";
1248         }
1249
1250         /* private */ function doBlockLevels( $text, $linestart ) {
1251                 $fname = 'Parser::doBlockLevels';
1252                 wfProfileIn( $fname );
1253
1254                 # Parsing through the text line by line.  The main thing
1255                 # happening here is handling of block-level elements p, pre,
1256                 # and making lists from lines starting with * # : etc.
1257                 #
1258                 $textLines = explode( "\n", $text );
1259
1260                 $lastPrefix = $output = $lastLine = '';
1261                 $this->mDTopen = $inBlockElem = false;
1262                 $prefixLength = 0;
1263                 $paragraphStack = false;
1264
1265                 if ( !$linestart ) {
1266                         $output .= array_shift( $textLines );
1267                 }
1268                 foreach ( $textLines as $oLine ) {
1269                         $lastPrefixLength = strlen( $lastPrefix );
1270                         $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1271                         $preOpenMatch = preg_match("/<pre/i", $oLine );
1272                         if ( !$this->mInPre ) {
1273                                 # Multiple prefixes may abut each other for nested lists.
1274                                 $prefixLength = strspn( $oLine, '*#:;' );
1275                                 $pref = substr( $oLine, 0, $prefixLength );
1276
1277                                 # eh?
1278                                 $pref2 = str_replace( ';', ':', $pref );
1279                                 $t = substr( $oLine, $prefixLength );
1280                                 $this->mInPre = !empty($preOpenMatch);
1281                         } else {
1282                                 # Don't interpret any other prefixes in preformatted text
1283                                 $prefixLength = 0;
1284                                 $pref = $pref2 = '';
1285                                 $t = $oLine;
1286                         }
1287
1288                         # List generation
1289                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1290                                 # Same as the last item, so no need to deal with nesting or opening stuff
1291                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1292                                 $paragraphStack = false;
1293
1294                                 if ( ";" == substr( $pref, -1 ) ) {
1295                                         # The one nasty exception: definition lists work like this:
1296                                         # ; title : definition text
1297                                         # So we check for : in the remainder text to split up the
1298                                         # title and definition, without b0rking links.
1299                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1300                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1301                                                 $term = $match[1];
1302                                                 $output .= $term . $this->nextItem( ':' );
1303                                                 $t = $match[2];
1304                                         }
1305                                 }
1306                         } elseif( $prefixLength || $lastPrefixLength ) {
1307                                 # Either open or close a level...
1308                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1309                                 $paragraphStack = false;
1310
1311                                 while( $commonPrefixLength < $lastPrefixLength ) {
1312                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1313                                         --$lastPrefixLength;
1314                                 }
1315                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1316                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1317                                 }
1318                                 while ( $prefixLength > $commonPrefixLength ) {
1319                                         $char = substr( $pref, $commonPrefixLength, 1 );
1320                                         $output .= $this->openList( $char );
1321
1322                                         if ( ';' == $char ) {
1323                                                 # FIXME: This is dupe of code above
1324                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1325                                                         $term = $match[1];
1326                                                         $output .= $term . $this->nextItem( ":" );
1327                                                         $t = $match[2];
1328                                                 }
1329                                         }
1330                                         ++$commonPrefixLength;
1331                                 }
1332                                 $lastPrefix = $pref2;
1333                         }
1334                         if( 0 == $prefixLength ) {
1335                                 # No prefix (not in list)--go to paragraph mode
1336                                 $uniq_prefix = UNIQ_PREFIX;
1337                                 // XXX: use a stack for nestable elements like span, table and div
1338                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1339                                 $closematch = preg_match(
1340                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1341                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1342                                 if ( $openmatch or $closematch ) {
1343                                         $paragraphStack = false;
1344                                         $output .= $this->closeParagraph();
1345                                         if($preOpenMatch and !$preCloseMatch) {
1346                                                 $this->mInPre = true;
1347                                         }
1348                                         if ( $closematch  ) {
1349                                                 $inBlockElem = false;
1350                                         } else {
1351                                                 $inBlockElem = true;
1352                                         }
1353                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1354                                         if ( " " == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1355                                                 // pre
1356                                                 if ($this->mLastSection != 'pre') {
1357                                                         $paragraphStack = false;
1358                                                         $output .= $this->closeParagraph().'<pre>';
1359                                                         $this->mLastSection = 'pre';
1360                                                 }
1361                                         } else {
1362                                                 // paragraph
1363                                                 if ( '' == trim($t) ) {
1364                                                         if ( $paragraphStack ) {
1365                                                                 $output .= $paragraphStack.'<br />';
1366                                                                 $paragraphStack = false;
1367                                                                 $this->mLastSection = 'p';
1368                                                         } else {
1369                                                                 if ($this->mLastSection != 'p' ) {
1370                                                                         $output .= $this->closeParagraph();
1371                                                                         $this->mLastSection = '';
1372                                                                         $paragraphStack = '<p>';
1373                                                                 } else {
1374                                                                         $paragraphStack = '</p><p>';
1375                                                                 }
1376                                                         }
1377                                                 } else {
1378                                                         if ( $paragraphStack ) {
1379                                                                 $output .= $paragraphStack;
1380                                                                 $paragraphStack = false;
1381                                                                 $this->mLastSection = 'p';
1382                                                         } else if ($this->mLastSection != 'p') {
1383                                                                 $output .= $this->closeParagraph().'<p>';
1384                                                                 $this->mLastSection = 'p';
1385                                                         }
1386                                                 }
1387                                         }
1388                                 }
1389                         }
1390                         if ($paragraphStack === false) {
1391                                 $output .= $t."\n";
1392                         }
1393                 }
1394                 while ( $prefixLength ) {
1395                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1396                         --$prefixLength;
1397                 }
1398                 if ( '' != $this->mLastSection ) {
1399                         $output .= '</' . $this->mLastSection . '>';
1400                         $this->mLastSection = '';
1401                 }
1402
1403                 wfProfileOut( $fname );
1404                 return $output;
1405         }
1406
1407         # Return value of a magic variable (like PAGENAME)
1408         function getVariableValue( $index ) {
1409                 global $wgLang, $wgSitename, $wgServer;
1410
1411                 switch ( $index ) {
1412                         case MAG_CURRENTMONTH:
1413                                 return date( 'm' );
1414                         case MAG_CURRENTMONTHNAME:
1415                                 return $wgLang->getMonthName( date('n') );
1416                         case MAG_CURRENTMONTHNAMEGEN:
1417                                 return $wgLang->getMonthNameGen( date('n') );
1418                         case MAG_CURRENTDAY:
1419                                 return date('j');
1420                         case MAG_PAGENAME:
1421                                 return $this->mTitle->getText();
1422                         case MAG_NAMESPACE:
1423                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1424                                 return $wgLang->getNsText($this->mTitle->getNamespace()); // Patch  by Dori
1425                         case MAG_CURRENTDAYNAME:
1426                                 return $wgLang->getWeekdayName( date('w')+1 );
1427                         case MAG_CURRENTYEAR:
1428                                 return date( 'Y' );
1429                         case MAG_CURRENTTIME:
1430                                 return $wgLang->time( wfTimestampNow(), false );
1431                         case MAG_NUMBEROFARTICLES:
1432                                 return wfNumberOfArticles();
1433                         case MAG_SITENAME:
1434                                 return $wgSitename;
1435                         case MAG_SERVER:
1436                                 return $wgServer;
1437                         default:
1438                                 return NULL;
1439                 }
1440         }
1441
1442         # initialise the magic variables (like CURRENTMONTHNAME)
1443         function initialiseVariables() {
1444                 global $wgVariableIDs;
1445                 $this->mVariables = array();
1446                 foreach ( $wgVariableIDs as $id ) {
1447                         $mw =& MagicWord::get( $id );
1448                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1449                 }
1450         }
1451
1452         /* private */ function replaceVariables( $text, $args = array() ) {
1453                 global $wgLang, $wgScript, $wgArticlePath;
1454
1455                 # Prevent too big inclusions
1456                 if(strlen($text)> MAX_INCLUDE_SIZE)
1457                    return $text;
1458
1459                 $fname = 'Parser::replaceVariables';
1460                 wfProfileIn( $fname );
1461
1462                 $bail = false;
1463                 $titleChars = Title::legalChars();
1464                 $nonBraceChars = str_replace( array( '{', '}' ), array( '', '' ), $titleChars );
1465
1466                 # This function is called recursively. To keep track of arguments we need a stack:
1467                 array_push( $this->mArgStack, $args );
1468
1469                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1470                 $GLOBALS['wgCurParser'] =& $this;
1471
1472
1473                 if ( $this->mOutputType == OT_HTML ) {
1474                         # Variable substitution
1475                         $text = preg_replace_callback( "/{{([$nonBraceChars]*?)}}/", 'wfVariableSubstitution', $text );
1476
1477                         # Argument substitution
1478                         $text = preg_replace_callback( "/(\\n?){{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1479                 }
1480                 # Template substitution
1481                 $regex = '/(\\n?){{(['.$nonBraceChars.']*)(\\|.*?|)}}/s';
1482                 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1483
1484                 array_pop( $this->mArgStack );
1485
1486                 wfProfileOut( $fname );
1487                 return $text;
1488         }
1489
1490         function variableSubstitution( $matches ) {
1491                 if ( !$this->mVariables ) {
1492                         $this->initialiseVariables();
1493                 }
1494                 if ( array_key_exists( $matches[1], $this->mVariables ) ) {
1495                         $text = $this->mVariables[$matches[1]];
1496                         $this->mOutput->mContainsOldMagic = true;
1497                 } else {
1498                         $text = $matches[0];
1499                 }
1500                 return $text;
1501         }
1502
1503         function braceSubstitution( $matches ) {
1504                 global $wgLinkCache, $wgLang;
1505                 $fname = 'Parser::braceSubstitution';
1506                 $found = false;
1507                 $nowiki = false;
1508                 $noparse = false;
1509
1510                 $title = NULL;
1511
1512                 # $newline is an optional newline character before the braces
1513                 # $part1 is the bit before the first |, and must contain only title characters
1514                 # $args is a list of arguments, starting from index 0, not including $part1
1515
1516                 $newline = $matches[1];
1517                 $part1 = $matches[2];
1518                 # If the third subpattern matched anything, it will start with |
1519                 if ( $matches[3] !== '' ) {
1520                         $args = explode( '|', substr( $matches[3], 1 ) );
1521                 } else {
1522                         $args = array();
1523                 }
1524                 $argc = count( $args );
1525
1526                 # {{{}}}
1527                 if ( strpos( $matches[0], '{{{' ) !== false ) {
1528                         $text = $matches[0];
1529                         $found = true;
1530                         $noparse = true;
1531                 }
1532
1533                 # SUBST
1534                 if ( !$found ) {
1535                         $mwSubst =& MagicWord::get( MAG_SUBST );
1536                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1537                                 if ( $this->mOutputType != OT_WIKI ) {
1538                                         # Invalid SUBST not replaced at PST time
1539                                         # Return without further processing
1540                                         $text = $matches[0];
1541                                         $found = true;
1542                                         $noparse= true;
1543                                 }
1544                         } elseif ( $this->mOutputType == OT_WIKI ) {
1545                                 # SUBST not found in PST pass, do nothing
1546                                 $text = $matches[0];
1547                                 $found = true;
1548                         }
1549                 }
1550
1551                 # MSG, MSGNW and INT
1552                 if ( !$found ) {
1553                         # Check for MSGNW:
1554                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1555                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1556                                 $nowiki = true;
1557                         } else {
1558                                 # Remove obsolete MSG:
1559                                 $mwMsg =& MagicWord::get( MAG_MSG );
1560                                 $mwMsg->matchStartAndRemove( $part1 );
1561                         }
1562
1563                         # Check if it is an internal message
1564                         $mwInt =& MagicWord::get( MAG_INT );
1565                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1566                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1567                                         $text = wfMsgReal( $part1, $args, true );
1568                                         $found = true;
1569                                 }
1570                         }
1571                 }
1572
1573                 # NS
1574                 if ( !$found ) {
1575                         # Check for NS: (namespace expansion)
1576                         $mwNs = MagicWord::get( MAG_NS );
1577                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1578                                 if ( intval( $part1 ) ) {
1579                                         $text = $wgLang->getNsText( intval( $part1 ) );
1580                                         $found = true;
1581                                 } else {
1582                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1583                                         if ( !is_null( $index ) ) {
1584                                                 $text = $wgLang->getNsText( $index );
1585                                                 $found = true;
1586                                         }
1587                                 }
1588                         }
1589                 }
1590
1591                 # LOCALURL and LOCALURLE
1592                 if ( !$found ) {
1593                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1594                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1595
1596                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1597                                 $func = 'getLocalURL';
1598                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1599                                 $func = 'escapeLocalURL';
1600                         } else {
1601                                 $func = '';
1602                         }
1603
1604                         if ( $func !== '' ) {
1605                                 $title = Title::newFromText( $part1 );
1606                                 if ( !is_null( $title ) ) {
1607                                         if ( $argc > 0 ) {
1608                                                 $text = $title->$func( $args[0] );
1609                                         } else {
1610                                                 $text = $title->$func();
1611                                         }
1612                                         $found = true;
1613                                 }
1614                         }
1615                 }
1616
1617                 # Internal variables
1618                 if ( !$this->mVariables ) {
1619                         $this->initialiseVariables();
1620                 }
1621                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1622                         $text = $this->mVariables[$part1];
1623                         $found = true;
1624                         $this->mOutput->mContainsOldMagic = true;
1625                 }
1626
1627                 # Template table test
1628
1629                 # Did we encounter this template already? If yes, it is in the cache
1630                 # and we need to check for loops.
1631                 if ( isset( $this->mTemplates[$part1] ) ) {
1632                         # Infinite loop test
1633                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1634                                 $noparse = true;
1635                                 $found = true;
1636                         }
1637                         # set $text to cached message.
1638                         $text = $this->mTemplates[$part1];
1639                         $found = true;
1640                 }
1641
1642                 # Load from database
1643                 if ( !$found ) {
1644                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1645                         if ( !is_null( $title ) && !$title->isExternal() ) {
1646                                 # Check for excessive inclusion
1647                                 $dbk = $title->getPrefixedDBkey();
1648                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1649                                         # This should never be reached.
1650                                         $article = new Article( $title );
1651                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1652                                         if ( $articleContent !== false ) {
1653                                                 $found = true;
1654                                                 $text = $articleContent;
1655
1656                                         }
1657                                 }
1658
1659                                 # If the title is valid but undisplayable, make a link to it
1660                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1661                                         $text = '[[' . $title->getPrefixedText() . ']]';
1662                                         $found = true;
1663                                 }
1664
1665                                 # Template cache array insertion
1666                                 $this->mTemplates[$part1] = $text;
1667                         }
1668                 }
1669
1670                 # Recursive parsing, escaping and link table handling
1671                 # Only for HTML output
1672                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1673                         $text = wfEscapeWikiText( $text );
1674                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1675                         # Clean up argument array
1676                         $assocArgs = array();
1677                         $index = 1;
1678                         foreach( $args as $arg ) {
1679                                 $eqpos = strpos( $arg, '=' );
1680                                 if ( $eqpos === false ) {
1681                                         $assocArgs[$index++] = $arg;
1682                                 } else {
1683                                         $name = trim( substr( $arg, 0, $eqpos ) );
1684                                         $value = trim( substr( $arg, $eqpos+1 ) );
1685                                         if ( $value === false ) {
1686                                                 $value = '';
1687                                         }
1688                                         if ( $name !== false ) {
1689                                                 $assocArgs[$name] = $value;
1690                                         }
1691                                 }
1692                         }
1693
1694                         # Do not enter included links in link table
1695                         if ( !is_null( $title ) ) {
1696                                 $wgLinkCache->suspend();
1697                         }
1698
1699                         # Add a new element to the templace recursion path
1700                         $this->mTemplatePath[$part1] = 1;
1701
1702                         # Run full parser on the included text
1703                         $text = $this->stripParse( $text, $newline, $assocArgs );
1704
1705                         # Resume the link cache and register the inclusion as a link
1706                         if ( !is_null( $title ) ) {
1707                                 $wgLinkCache->resume();
1708                                 $wgLinkCache->addLinkObj( $title );
1709                         }
1710                 }
1711                 # Empties the template path
1712                 $this->mTemplatePath = array();
1713
1714                 if ( !$found ) {
1715                         return $matches[0];
1716                 } else {
1717                         return $text;
1718                 }
1719         }
1720
1721         # Triple brace replacement -- used for template arguments
1722         function argSubstitution( $matches ) {
1723                 $newline = $matches[1];
1724                 $arg = trim( $matches[2] );
1725                 $text = $matches[0];
1726                 $inputArgs = end( $this->mArgStack );
1727
1728                 if ( array_key_exists( $arg, $inputArgs ) ) {
1729                         $text = $this->stripParse( $inputArgs[$arg], $newline, array() );
1730                 }
1731
1732                 return $text;
1733         }
1734
1735         # Returns true if the function is allowed to include this entity
1736         function incrementIncludeCount( $dbk ) {
1737                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1738                         $this->mIncludeCount[$dbk] = 0;
1739                 }
1740                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1741                         return true;
1742                 } else {
1743                         return false;
1744                 }
1745         }
1746
1747
1748         # Cleans up HTML, removes dangerous tags and attributes
1749         /* private */ function removeHTMLtags( $text ) {
1750                 global $wgUseTidy, $wgUserHtml;
1751                 $fname = 'Parser::removeHTMLtags';
1752                 wfProfileIn( $fname );
1753
1754                 if( $wgUserHtml ) {
1755                         $htmlpairs = array( # Tags that must be closed
1756                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1757                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1758                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
1759                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1760                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
1761                         );
1762                         $htmlsingle = array(
1763                                 'br', 'hr', 'li', 'dt', 'dd'
1764                         );
1765                         $htmlnest = array( # Tags that can be nested--??
1766                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1767                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
1768                         );
1769                         $tabletags = array( # Can only appear inside table
1770                                 'td', 'th', 'tr'
1771                         );
1772                 } else {
1773                         $htmlpairs = array();
1774                         $htmlsingle = array();
1775                         $htmlnest = array();
1776                         $tabletags = array();
1777                 }
1778
1779                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1780                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1781
1782                 $htmlattrs = $this->getHTMLattrs () ;
1783
1784                 # Remove HTML comments
1785                 $text = preg_replace( '/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU', '$2', $text );
1786
1787                 $bits = explode( '<', $text );
1788                 $text = array_shift( $bits );
1789                 if(!$wgUseTidy) {
1790                         $tagstack = array(); $tablestack = array();
1791                         foreach ( $bits as $x ) {
1792                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1793                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1794                                 $x, $regs );
1795                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1796                                 error_reporting( $prev );
1797
1798                                 $badtag = 0 ;
1799                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1800                                         # Check our stack
1801                                         if ( $slash ) {
1802                                                 # Closing a tag...
1803                                                 if ( ! in_array( $t, $htmlsingle ) &&
1804                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
1805                                                         @array_push( $tagstack, $ot );
1806                                                         $badtag = 1;
1807                                                 } else {
1808                                                         if ( $t == 'table' ) {
1809                                                                 $tagstack = array_pop( $tablestack );
1810                                                         }
1811                                                         $newparams = '';
1812                                                 }
1813                                         } else {
1814                                                 # Keep track for later
1815                                                 if ( in_array( $t, $tabletags ) &&
1816                                                 ! in_array( 'table', $tagstack ) ) {
1817                                                         $badtag = 1;
1818                                                 } else if ( in_array( $t, $tagstack ) &&
1819                                                 ! in_array ( $t , $htmlnest ) ) {
1820                                                         $badtag = 1 ;
1821                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1822                                                         if ( $t == 'table' ) {
1823                                                                 array_push( $tablestack, $tagstack );
1824                                                                 $tagstack = array();
1825                                                         }
1826                                                         array_push( $tagstack, $t );
1827                                                 }
1828                                                 # Strip non-approved attributes from the tag
1829                                                 $newparams = $this->fixTagAttributes($params);
1830
1831                                         }
1832                                         if ( ! $badtag ) {
1833                                                 $rest = str_replace( '>', '&gt;', $rest );
1834                                                 $text .= "<$slash$t $newparams$brace$rest";
1835                                                 continue;
1836                                         }
1837                                 }
1838                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1839                         }
1840                         # Close off any remaining tags
1841                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
1842                                 $text .= "</$t>\n";
1843                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
1844                         }
1845                 } else {
1846                         # this might be possible using tidy itself
1847                         foreach ( $bits as $x ) {
1848                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1849                                 $x, $regs );
1850                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1851                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1852                                         $newparams = $this->fixTagAttributes($params);
1853                                         $rest = str_replace( '>', '&gt;', $rest );
1854                                         $text .= "<$slash$t $newparams$brace$rest";
1855                                 } else {
1856                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1857                                 }
1858                         }
1859                 }
1860                 wfProfileOut( $fname );
1861                 return $text;
1862         }
1863
1864
1865 /*
1866  *
1867  * This function accomplishes several tasks:
1868  * 1) Auto-number headings if that option is enabled
1869  * 2) Add an [edit] link to sections for logged in users who have enabled the option
1870  * 3) Add a Table of contents on the top for users who have enabled the option
1871  * 4) Auto-anchor headings
1872  *
1873  * It loops through all headlines, collects the necessary data, then splits up the
1874  * string and re-inserts the newly formatted headlines.
1875  *
1876  */
1877
1878         /* private */ function formatHeadings( $text, $isMain=true ) {
1879                 global $wgInputEncoding, $wgMaxTocLevel;
1880
1881                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1882                 $doShowToc = $this->mOptions->getShowToc();
1883                 $forceTocHere = false;
1884                 if( !$this->mTitle->userCanEdit() ) {
1885                         $showEditLink = 0;
1886                         $rightClickHack = 0;
1887                 } else {
1888                         $showEditLink = $this->mOptions->getEditSection();
1889                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1890                 }
1891
1892                 # Inhibit editsection links if requested in the page
1893                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1894                 if( $esw->matchAndRemove( $text ) ) {
1895                         $showEditLink = 0;
1896                 }
1897                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1898                 # do not add TOC
1899                 $mw =& MagicWord::get( MAG_NOTOC );
1900                 if( $mw->matchAndRemove( $text ) ) {
1901                         $doShowToc = 0;
1902                 }
1903
1904                 # never add the TOC to the Main Page. This is an entry page that should not
1905                 # be more than 1-2 screens large anyway
1906                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
1907                         $doShowToc = 0;
1908                 }
1909
1910                 # Get all headlines for numbering them and adding funky stuff like [edit]
1911                 # links - this is for later, but we need the number of headlines right now
1912                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
1913
1914                 # if there are fewer than 4 headlines in the article, do not show TOC
1915                 if( $numMatches < 4 ) {
1916                         $doShowToc = 0;
1917                 }
1918
1919                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
1920                 # override above conditions and always show TOC at that place
1921                 $mw =& MagicWord::get( MAG_TOC );
1922                 if ($mw->match( $text ) ) {
1923                         $doShowToc = 1;
1924                         $forceTocHere = true;
1925                 } else {
1926                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1927                         # override above conditions and always show TOC above first header
1928                         $mw =& MagicWord::get( MAG_FORCETOC );
1929                         if ($mw->matchAndRemove( $text ) ) {
1930                                 $doShowToc = 1;
1931                         }
1932                 }
1933
1934
1935
1936                 # We need this to perform operations on the HTML
1937                 $sk =& $this->mOptions->getSkin();
1938
1939                 # headline counter
1940                 $headlineCount = 0;
1941
1942                 # Ugh .. the TOC should have neat indentation levels which can be
1943                 # passed to the skin functions. These are determined here
1944                 $toclevel = 0;
1945                 $toc = '';
1946                 $full = '';
1947                 $head = array();
1948                 $sublevelCount = array();
1949                 $level = 0;
1950                 $prevlevel = 0;
1951                 foreach( $matches[3] as $headline ) {
1952                         $numbering = '';
1953                         if( $level ) {
1954                                 $prevlevel = $level;
1955                         }
1956                         $level = $matches[1][$headlineCount];
1957                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
1958                                 # reset when we enter a new level
1959                                 $sublevelCount[$level] = 0;
1960                                 $toc .= $sk->tocIndent( $level - $prevlevel );
1961                                 $toclevel += $level - $prevlevel;
1962                         }
1963                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
1964                                 # reset when we step back a level
1965                                 $sublevelCount[$level+1]=0;
1966                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
1967                                 $toclevel -= $prevlevel - $level;
1968                         }
1969                         # count number of headlines for each level
1970                         @$sublevelCount[$level]++;
1971                         if( $doNumberHeadings || $doShowToc ) {
1972                                 $dot = 0;
1973                                 for( $i = 1; $i <= $level; $i++ ) {
1974                                         if( !empty( $sublevelCount[$i] ) ) {
1975                                                 if( $dot ) {
1976                                                         $numbering .= '.';
1977                                                 }
1978                                                 $numbering .= $sublevelCount[$i];
1979                                                 $dot = 1;
1980                                         }
1981                                 }
1982                         }
1983
1984                         # The canonized header is a version of the header text safe to use for links
1985                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
1986                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
1987                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
1988
1989                         # strip out HTML
1990                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
1991                         $tocline = trim( $canonized_headline );
1992                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
1993                         $replacearray = array(
1994                                 '%3A' => ':',
1995                                 '%' => '.'
1996                         );
1997                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
1998                         $refer[$headlineCount] = $canonized_headline;
1999
2000                         # count how many in assoc. array so we can track dupes in anchors
2001                         @$refers[$canonized_headline]++;
2002                         $refcount[$headlineCount]=$refers[$canonized_headline];
2003
2004                         # Prepend the number to the heading text
2005
2006                         if( $doNumberHeadings || $doShowToc ) {
2007                                 $tocline = $numbering . ' ' . $tocline;
2008
2009                                 # Don't number the heading if it is the only one (looks silly)
2010                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2011                                         # the two are different if the line contains a link
2012                                         $headline=$numbering . ' ' . $headline;
2013                                 }
2014                         }
2015
2016                         # Create the anchor for linking from the TOC to the section
2017                         $anchor = $canonized_headline;
2018                         if($refcount[$headlineCount] > 1 ) {
2019                                 $anchor .= '_' . $refcount[$headlineCount];
2020                         }
2021                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2022                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2023                         }
2024                         if( $showEditLink ) {
2025                                 if ( empty( $head[$headlineCount] ) ) {
2026                                         $head[$headlineCount] = '';
2027                                 }
2028                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
2029                         }
2030
2031                         # Add the edit section span
2032                         if( $rightClickHack ) {
2033                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
2034                         }
2035
2036                         # give headline the correct <h#> tag
2037                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
2038
2039                         $headlineCount++;
2040                 }
2041
2042                 if( $doShowToc ) {
2043                         $toclines = $headlineCount;
2044                         $toc .= $sk->tocUnindent( $toclevel );
2045                         $toc = $sk->tocTable( $toc );
2046                 }
2047
2048                 # split up and insert constructed headlines
2049
2050                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2051                 $i = 0;
2052
2053                 foreach( $blocks as $block ) {
2054                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2055                             # This is the [edit] link that appears for the top block of text when
2056                                 # section editing is enabled
2057
2058                                 # Disabled because it broke block formatting
2059                                 # For example, a bullet point in the top line
2060                                 # $full .= $sk->editSectionLink(0);
2061                         }
2062                         $full .= $block;
2063                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2064                         # Top anchor now in skin
2065                                 $full = $full.$toc;
2066                         }
2067
2068                         if( !empty( $head[$i] ) ) {
2069                                 $full .= $head[$i];
2070                         }
2071                         $i++;
2072                 }
2073                 if($forceTocHere) {
2074                         $mw =& MagicWord::get( MAG_TOC );
2075                         return $mw->replace( $toc, $full );
2076                 } else {
2077                         return $full;
2078                 }
2079         }
2080
2081         # Return an HTML link for the "ISBN 123456" text
2082         /* private */ function magicISBN( $text ) {
2083                 global $wgLang;
2084                 $fname = 'Parser::magicISBN';
2085                 wfProfileIn( $fname );
2086
2087                 $a = split( 'ISBN ', " $text" );
2088                 if ( count ( $a ) < 2 ) {
2089                         wfProfileOut( $fname );
2090                         return $text;
2091                 }
2092                 $text = substr( array_shift( $a ), 1);
2093                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2094
2095                 foreach ( $a as $x ) {
2096                         $isbn = $blank = '' ;
2097                         while ( ' ' == $x{0} ) {
2098                                 $blank .= ' ';
2099                                 $x = substr( $x, 1 );
2100                         }
2101                         while ( strstr( $valid, $x{0} ) != false ) {
2102                                 $isbn .= $x{0};
2103                                 $x = substr( $x, 1 );
2104                         }
2105                         $num = str_replace( '-', '', $isbn );
2106                         $num = str_replace( ' ', '', $num );
2107
2108                         if ( '' == $num ) {
2109                                 $text .= "ISBN $blank$x";
2110                         } else {
2111                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2112                                 $text .= '<a href="' .
2113                                 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
2114                                         "\" class=\"internal\">ISBN $isbn</a>";
2115                                 $text .= $x;
2116                         }
2117                 }
2118                 wfProfileOut( $fname );
2119                 return $text;
2120         }
2121
2122         # Return an HTML link for the "GEO ..." text
2123         /* private */ function magicGEO( $text ) {
2124                 global $wgLang, $wgUseGeoMode;
2125                 if ( !isset ( $wgUseGeoMode ) || !$wgUseGeoMode ) return $text ;
2126                 $fname = 'Parser::magicGEO';
2127                 wfProfileIn( $fname );
2128
2129                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2130                 $directions = array ( "N" => "North" , "S" => "South" , "E" => "East" , "W" => "West" ) ;
2131                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2132                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2133                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2134                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2135
2136                 $a = split( 'GEO ', " $text" );
2137                 if ( count ( $a ) < 2 ) {
2138                         wfProfileOut( $fname );
2139                         return $text;
2140                 }
2141                 $text = substr( array_shift( $a ), 1);
2142                 $valid = '0123456789.+-:';
2143
2144                 foreach ( $a as $x ) {
2145                         $geo = $blank = '' ;
2146                         while ( ' ' == $x{0} ) {
2147                                 $blank .= ' ';
2148                                 $x = substr( $x, 1 );
2149                         }
2150                         while ( strstr( $valid, $x{0} ) != false ) {
2151                                 $geo .= $x{0};
2152                                 $x = substr( $x, 1 );
2153                         }
2154                         $num = str_replace( '+', '', $geo );
2155                         $num = str_replace( ' ', '', $num );
2156
2157                         if ( '' == $num || count ( explode ( ":" , $num , 3 ) ) < 2 ) {
2158                                 $text .= "GEO $blank$x";
2159                         } else {
2160                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2161                                 $text .= '<a href="' .
2162                                 $titleObj->escapeLocalUrl( "coordinates={$num}" ) .
2163                                         "\" class=\"internal\">GEO $geo</a>";
2164                                 $text .= $x;
2165                         }
2166                 }
2167                 wfProfileOut( $fname );
2168                 return $text;
2169         }
2170
2171         # Return an HTML link for the "RFC 1234" text
2172         /* private */ function magicRFC( $text ) {
2173                 global $wgLang;
2174
2175                 $a = split( 'RFC ', ' '.$text );
2176                 if ( count ( $a ) < 2 ) return $text;
2177                 $text = substr( array_shift( $a ), 1);
2178                 $valid = '0123456789';
2179
2180                 foreach ( $a as $x ) {
2181                         $rfc = $blank = '' ;
2182                         while ( ' ' == $x{0} ) {
2183                                 $blank .= ' ';
2184                                 $x = substr( $x, 1 );
2185                         }
2186                         while ( strstr( $valid, $x{0} ) != false ) {
2187                                 $rfc .= $x{0};
2188                                 $x = substr( $x, 1 );
2189                         }
2190
2191                         if ( '' == $rfc ) {
2192                                 $text .= "RFC $blank$x";
2193                         } else {
2194                                 $url = wfmsg( 'rfcurl' );
2195                                 $url = str_replace( '$1', $rfc, $url);
2196                                 $sk =& $this->mOptions->getSkin();
2197                                 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
2198                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2199                         }
2200                 }
2201                 return $text;
2202         }
2203
2204         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2205                 $this->mOptions = $options;
2206                 $this->mTitle =& $title;
2207                 $this->mOutputType = OT_WIKI;
2208
2209                 if ( $clearState ) {
2210                         $this->clearState();
2211                 }
2212
2213                 $stripState = false;
2214                 $pairs = array(
2215                         "\r\n" => "\n",
2216                         );
2217                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2218                 // now with regexes
2219                 /*
2220                 $pairs = array(
2221                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2222                         "/<br *?>/i" => "<br />",
2223                 );
2224                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2225                 */
2226                 $text = $this->strip( $text, $stripState, false );
2227                 $text = $this->pstPass2( $text, $user );
2228                 $text = $this->unstrip( $text, $stripState );
2229                 $text = $this->unstripNoWiki( $text, $stripState );
2230                 return $text;
2231         }
2232
2233         /* private */ function pstPass2( $text, &$user ) {
2234                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2235
2236                 # Variable replacement
2237                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2238                 $text = $this->replaceVariables( $text );
2239
2240                 # Signatures
2241                 #
2242                 $n = $user->getName();
2243                 $k = $user->getOption( 'nickname' );
2244                 if ( '' == $k ) { $k = $n; }
2245                 if(isset($wgLocaltimezone)) {
2246                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2247                 }
2248                 /* Note: this is an ugly timezone hack for the European wikis */
2249                 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2250                   ' (' . date( 'T' ) . ')';
2251                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2252
2253                 $text = preg_replace( '/~~~~~/', $d, $text );
2254                 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText(
2255                   Namespace::getUser() ) . ":$n|$k]] $d", $text );
2256                 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText(
2257                   Namespace::getUser() ) . ":$n|$k]]", $text );
2258
2259                 # Context links: [[|name]] and [[name (context)|]]
2260                 #
2261                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2262                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2263                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2264                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2265
2266                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2267                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2268                 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/";          # [[namespace:page|]]
2269                 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2270                                                                                                                 # [[ns:page (cont)|]]
2271                 $context = "";
2272                 $t = $this->mTitle->getText();
2273                 if ( preg_match( $conpat, $t, $m ) ) {
2274                         $context = $m[2];
2275                 }
2276                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2277                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2278                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2279
2280                 if ( '' == $context ) {
2281                         $text = preg_replace( $p2, '[[\\1]]', $text );
2282                 } else {
2283                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2284                 }
2285
2286                 /*
2287                 $mw =& MagicWord::get( MAG_SUBST );
2288                 $wgCurParser = $this->fork();
2289                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2290                 $this->merge( $wgCurParser );
2291                 */
2292
2293                 # Trim trailing whitespace
2294                 # MAG_END (__END__) tag allows for trailing
2295                 # whitespace to be deliberately included
2296                 $text = rtrim( $text );
2297                 $mw =& MagicWord::get( MAG_END );
2298                 $mw->matchAndRemove( $text );
2299
2300                 return $text;
2301         }
2302
2303         # Set up some variables which are usually set up in parse()
2304         # so that an external function can call some class members with confidence
2305         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2306                 $this->mTitle =& $title;
2307                 $this->mOptions = $options;
2308                 $this->mOutputType = $outputType;
2309                 if ( $clearState ) {
2310                         $this->clearState();
2311                 }
2312         }
2313
2314         function transformMsg( $text, $options ) {
2315                 global $wgTitle;
2316                 static $executing = false;
2317
2318                 # Guard against infinite recursion
2319                 if ( $executing ) {
2320                         return $text;
2321                 }
2322                 $executing = true;
2323
2324                 $this->mTitle = $wgTitle;
2325                 $this->mOptions = $options;
2326                 $this->mOutputType = OT_MSG;
2327                 $this->clearState();
2328                 $text = $this->replaceVariables( $text );
2329
2330                 $executing = false;
2331                 return $text;
2332         }
2333
2334         # Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2335         # Callback will be called with the text within
2336         # Transform and return the text within
2337         function setHook( $tag, $callback ) {
2338                 $oldVal = @$this->mTagHooks[$tag];
2339                 $this->mTagHooks[$tag] = $callback;
2340                 return $oldVal;
2341         }
2342 }
2343
2344 class ParserOutput
2345 {
2346         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2347         var $mCacheTime; # Used in ParserCache
2348
2349         function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2350                 $containsOldMagic = false )
2351         {
2352                 $this->mText = $text;
2353                 $this->mLanguageLinks = $languageLinks;
2354                 $this->mCategoryLinks = $categoryLinks;
2355                 $this->mContainsOldMagic = $containsOldMagic;
2356                 $this->mCacheTime = "";
2357         }
2358
2359         function getText() { return $this->mText; }
2360         function getLanguageLinks() { return $this->mLanguageLinks; }
2361         function getCategoryLinks() { return $this->mCategoryLinks; }
2362         function getCacheTime() { return $this->mCacheTime; }
2363         function containsOldMagic() { return $this->mContainsOldMagic; }
2364         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2365         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2366         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2367         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2368         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2369
2370         function merge( $other ) {
2371                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2372                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2373                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2374         }
2375
2376 }
2377
2378 class ParserOptions
2379 {
2380         # All variables are private
2381         var $mUseTeX;                    # Use texvc to expand <math> tags
2382         var $mUseCategoryMagic;          # Treat [[Category:xxxx]] tags specially
2383         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2384         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2385         var $mAllowExternalImages;       # Allow external images inline
2386         var $mSkin;                      # Reference to the preferred skin
2387         var $mDateFormat;                # Date format index
2388         var $mEditSection;               # Create "edit section" links
2389         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2390         var $mNumberHeadings;            # Automatically number headings
2391         var $mShowToc;                   # Show table of contents
2392
2393         function getUseTeX()                        { return $this->mUseTeX; }
2394         function getUseCategoryMagic()              { return $this->mUseCategoryMagic; }
2395         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2396         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2397         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2398         function getSkin()                          { return $this->mSkin; }
2399         function getDateFormat()                    { return $this->mDateFormat; }
2400         function getEditSection()                   { return $this->mEditSection; }
2401         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
2402         function getNumberHeadings()                { return $this->mNumberHeadings; }
2403         function getShowToc()                       { return $this->mShowToc; }
2404
2405         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
2406         function setUseCategoryMagic( $x )          { return wfSetVar( $this->mUseCategoryMagic, $x ); }
2407         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
2408         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
2409         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
2410         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
2411         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
2412         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2413         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
2414         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
2415
2416     function setSkin( &$x ) { $this->mSkin =& $x; }
2417
2418         /* static */ function newFromUser( &$user ) {
2419                 $popts = new ParserOptions;
2420                 $popts->initialiseFromUser( $user );
2421                 return $popts;
2422         }
2423
2424         function initialiseFromUser( &$userInput ) {
2425                 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2426
2427                 if ( !$userInput ) {
2428                         $user = new User;
2429                         $user->setLoaded( true );
2430                 } else {
2431                         $user =& $userInput;
2432                 }
2433
2434                 $this->mUseTeX = $wgUseTeX;
2435                 $this->mUseCategoryMagic = $wgUseCategoryMagic;
2436                 $this->mUseDynamicDates = $wgUseDynamicDates;
2437                 $this->mInterwikiMagic = $wgInterwikiMagic;
2438                 $this->mAllowExternalImages = $wgAllowExternalImages;
2439                 $this->mSkin =& $user->getSkin();
2440                 $this->mDateFormat = $user->getOption( 'date' );
2441                 $this->mEditSection = $user->getOption( 'editsection' );
2442                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2443                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2444                 $this->mShowToc = $user->getOption( 'showtoc' );
2445         }
2446
2447
2448 }
2449
2450 # Regex callbacks, used in Parser::replaceVariables
2451 function wfBraceSubstitution( $matches )
2452 {
2453         global $wgCurParser;
2454         return $wgCurParser->braceSubstitution( $matches );
2455 }
2456
2457 function wfArgSubstitution( $matches )
2458 {
2459         global $wgCurParser;
2460         return $wgCurParser->argSubstitution( $matches );
2461 }
2462
2463 function wfVariableSubstitution( $matches )
2464 {
2465         global $wgCurParser;
2466         return $wgCurParser->variableSubstitution( $matches );
2467 }
2468
2469 ?>