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                 $html_content = array();
 205                 $nowiki_content = array();
 206                 $math_content = array();
 207                 $pre_content = array();
 208                 $comment_content = array();
 209                 $ext_content = array();
 210
 211                 # Replace any instances of the placeholders
 212                 $uniq_prefix = UNIQ_PREFIX;
 213                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 214
 215                 # html
 216                 global $wgRawHtml;
 217                 if( $wgRawHtml ) {
 218                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 219                         foreach( $html_content as $marker => $content ) {
 220                                 if ($render ) {
 221                                         # Raw and unchecked for validity.
 222                                         $html_content[$marker] = $content;
 223                                 } else {
 224                                         $html_content[$marker] = "<html>$content</html>";
 225                                 }
 226                         }
 227                 }
 228
 229                 # nowiki
 230                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 231                 foreach( $nowiki_content as $marker => $content ) {
 232                         if( $render ){
 233                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 234                         } else {
 235                                 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
 236                         }
 237                 }
 238
 239                 # math
 240                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 241                 foreach( $math_content as $marker => $content ){
 242                         if( $render ) {
 243                                 if( $this->mOptions->getUseTeX() ) {
 244                                         $math_content[$marker] = renderMath( $content );
 245                                 } else {
 246                                         $math_content[$marker] = "&lt;math&gt;$content&lt;math&gt;";
 247                                 }
 248                         } else {
 249                                 $math_content[$marker] = "<math>$content</math>";
 250                         }
 251                 }
 252
 253                 # pre
 254                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 255                 foreach( $pre_content as $marker => $content ){
 256                         if( $render ){
 257                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 258                         } else {
 259                                 $pre_content[$marker] = "<pre>$content</pre>";
 260                         }
 261                 }
 262
 263                 # Comments
 264                 if($stripcomments) {
 265                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 266                         foreach( $comment_content as $marker => $content ){
 267                                 $comment_content[$marker] = "<!--$content-->";
 268                         }
 269                 }
 270
 271                 # Extensions
 272                 foreach ( $this->mTagHooks as $tag => $callback ) {
 273                         $ext_contents[$tag] = array();
 274                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 275                         foreach( $ext_content[$tag] as $marker => $content ) {
 276                                 if ( $render ) {
 277                                         $ext_content[$tag][$marker] = $callback( $content );
 278                                 } else {
 279                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 280                                 }
 281                         }
 282                 }
 283
 284                 # Merge state with the pre-existing state, if there is one
 285                 if ( $state ) {
 286                         $state['html'] = $state['html'] + $html_content;
 287                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 288                         $state['math'] = $state['math'] + $math_content;
 289                         $state['pre'] = $state['pre'] + $pre_content;
 290                         $state['comment'] = $state['comment'] + $comment_content;
 291
 292                         foreach( $ext_content as $tag => $array ) {
 293                                 if ( array_key_exists( $tag, $state ) ) {
 294                                         $state[$tag] = $state[$tag] + $array;
 295                                 }
 296                         }
 297                 } else {
 298                         $state = array(
 299                           'html' => $html_content,
 300                           'nowiki' => $nowiki_content,
 301                           'math' => $math_content,
 302                           'pre' => $pre_content,
 303                           'comment' => $comment_content,
 304                         ) + $ext_content;
 305                 }
 306                 return $text;
 307         }
 308
 309         # always call unstripNoWiki() after this one
 310         function unstrip( $text, &$state ) {
 311                 # Must expand in reverse order, otherwise nested tags will be corrupted
 312                 $contentDict = end( $state );
 313                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 314                         if( key($state) != 'nowiki' && key($state) != 'html') {
 315                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 316                                         $text = str_replace( key( $contentDict ), $content, $text );
 317                                 }
 318                         }
 319                 }
 320
 321                 return $text;
 322         }
 323         # always call this after unstrip() to preserve the order
 324         function unstripNoWiki( $text, &$state ) {
 325                 # Must expand in reverse order, otherwise nested tags will be corrupted
 326                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 327                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 328                 }
 329
 330                 global $wgRawHtml;
 331                 if ($wgRawHtml) {
 332                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 333                                 $text = str_replace( key( $state['html'] ), $content, $text );
 334                         }
 335                 }
 336
 337                 return $text;
 338         }
 339
 340         # Add an item to the strip state
 341         # Returns the unique tag which must be inserted into the stripped text
 342         # The tag will be replaced with the original text in unstrip()
 343
 344         function insertStripItem( $text, &$state ) {
 345                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 346                 if ( !$state ) {
 347                         $state = array(
 348                           'html' => array(),
 349                           'nowiki' => array(),
 350                           'math' => array(),
 351                           'pre' => array()
 352                         );
 353                 }
 354                 $state['item'][$rnd] = $text;
 355                 return $rnd;
 356         }
 357
 358         # categoryMagic
 359         # generate a list of subcategories and pages for a category
 360         # depending on wfMsg("usenewcategorypage") it either calls the new
 361         # or the old code. The new code will not work properly for some
 362         # languages due to sorting issues, so they might want to turn it
 363         # off.
 364         function categoryMagic() {
 365                 $msg = wfMsg('usenewcategorypage');
 366                 if ( '0' == @$msg[0] )
 367                 {
 368                         return $this->oldCategoryMagic();
 369                 } else {
 370                         return $this->newCategoryMagic();
 371                 }
 372         }
 373
 374         # This method generates the list of subcategories and pages for a category
 375         function oldCategoryMagic () {
 376                 global $wgLang , $wgUser ;
 377                 $fname = 'Parser::oldCategoryMagic';
 378
 379                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 380
 381                 $cns = Namespace::getCategory() ;
 382                 if ( $this->mTitle->getNamespace() != $cns ) return "" ; # This ain't a category page
 383
 384                 $r = "<br style=\"clear:both;\"/>\n";
 385
 386
 387                 $sk =& $wgUser->getSkin() ;
 388
 389                 $articles = array() ;
 390                 $children = array() ;
 391                 $data = array () ;
 392                 $id = $this->mTitle->getArticleID() ;
 393
 394                 # FIXME: add limits
 395                 $dbr =& wfGetDB( DB_SLAVE );
 396                 $cur = $dbr->tableName( 'cur' );
 397                 $categorylinks = $dbr->tableName( 'categorylinks' );
 398
 399                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 400                 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM $cur,$categorylinks " .
 401                         "WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 402                 $res = $dbr->query( $sql, $fname ) ;
 403                 while ( $x = $dbr->fetchObject ( $res ) ) $data[] = $x ;
 404
 405                 # For all pages that link to this category
 406                 foreach ( $data AS $x )
 407                 {
 408                         $t = $wgLang->getNsText ( $x->cur_namespace ) ;
 409                         if ( $t != "" ) $t .= ":" ;
 410                         $t .= $x->cur_title ;
 411
 412                         if ( $x->cur_namespace == $cns ) {
 413                                 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
 414                         } else {
 415                                 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
 416                         }
 417                 }
 418                 $dbr->freeResult ( $res ) ;
 419
 420                 # Showing subcategories
 421                 if ( count ( $children ) > 0 ) {
 422                         $r .= '<h2>'.wfMsg('subcategories')."</h2>\n" ;
 423                         $r .= implode ( ', ' , $children ) ;
 424                 }
 425
 426                 # Showing pages in this category
 427                 if ( count ( $articles ) > 0 ) {
 428                         $ti = $this->mTitle->getText() ;
 429                         $h =  wfMsg( 'category_header', $ti );
 430                         $r .= "<h2>{$h}</h2>\n" ;
 431                         $r .= implode ( ', ' , $articles ) ;
 432                 }
 433
 434                 return $r ;
 435         }
 436
 437
 438
 439         function newCategoryMagic () {
 440                 global $wgLang , $wgUser ;
 441                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 442
 443                 $cns = Namespace::getCategory() ;
 444                 if ( $this->mTitle->getNamespace() != $cns ) return '' ; # This ain't a category page
 445
 446                 $r = "<br style=\"clear:both;\"/>\n";
 447
 448
 449                 $sk =& $wgUser->getSkin() ;
 450
 451                 $articles = array() ;
 452                 $articles_start_char = array();
 453                 $children = array() ;
 454                 $children_start_char = array();
 455                 $data = array () ;
 456                 $id = $this->mTitle->getArticleID() ;
 457
 458                 # FIXME: add limits
 459                 $dbr =& wfGetDB( DB_SLAVE );
 460                 $cur = $dbr->tableName( 'cur' );
 461                 $categorylinks = $dbr->tableName( 'categorylinks' );
 462
 463                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 464                 $sql = "SELECT DISTINCT cur_title,cur_namespace,cl_sortkey FROM " .
 465                         "$cur,$categorylinks WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 466                 $res = $dbr->query ( $sql ) ;
 467                 while ( $x = $dbr->fetchObject ( $res ) )
 468                 {
 469                         $t = $ns = $wgLang->getNsText ( $x->cur_namespace ) ;
 470                         if ( $t != '' ) $t .= ':' ;
 471                         $t .= $x->cur_title ;
 472
 473                         if ( $x->cur_namespace == $cns ) {
 474                                 $ctitle = str_replace( '_',' ',$x->cur_title );
 475                                 array_push ( $children, $sk->makeKnownLink ( $t, $ctitle ) ) ; # Subcategory
 476
 477                                 // If there's a link from Category:A to Category:B, the sortkey of the resulting
 478                                 // entry in the categorylinks table is Category:A, not A, which it SHOULD be.
 479                                 // Workaround: If sortkey == "Category:".$title, than use $title for sorting,
 480                                 // else use sortkey...
 481                                 if ( ($ns.":".$ctitle) ==  $x->cl_sortkey ) {
 482                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cur_title ) );
 483                                 } else {
 484                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 485                                 }
 486                         } else {
 487                                 array_push ( $articles , $sk->makeKnownLink ( $t ) ) ; # Page in this category
 488                                 array_push ( $articles_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 489                         }
 490                 }
 491                 $dbr->freeResult ( $res ) ;
 492
 493                 $ti = $this->mTitle->getText() ;
 494
 495                 # Don't show subcategories section if there are none.
 496                 if ( count ( $children ) > 0 )
 497                 {
 498                         # Showing subcategories
 499                         $r .= '<h2>' . wfMsg( 'subcategories' ) . "</h2>\n"
 500                                 . wfMsg( 'subcategorycount', count( $children ) );
 501                         if ( count ( $children ) > 6 ) {
 502
 503                                 // divide list into three equal chunks
 504                                 $chunk = (int) (count ( $children ) / 3);
 505
 506                                 // get and display header
 507                                 $r .= '<table width="100%"><tr valign="top">';
 508
 509                                 $startChunk = 0;
 510                                 $endChunk = $chunk;
 511
 512                                 // loop through the chunks
 513                                 for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 514                                         $chunkIndex < 3;
 515                                         $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 516                                 {
 517
 518                                         $r .= '<td><ul>';
 519                                         // output all subcategories to category
 520                                         for ($index = $startChunk ;
 521                                                 $index < $endChunk && $index < count($children);
 522                                                 $index++ )
 523                                         {
 524                                                 // check for change of starting letter or begging of chunk
 525                                                 if ( ($children_start_char[$index] != $children_start_char[$index - 1])
 526                                                         || ($index == $startChunk) )
 527                                                 {
 528                                                         $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 529                                                 }
 530
 531                                                 $r .= "<li>{$children[$index]}</li>";
 532                                         }
 533                                         $r .= '</ul></td>';
 534
 535
 536                                 }
 537                                 $r .= '</tr></table>';
 538                         } else {
 539                                 // for short lists of subcategories to category.
 540
 541                                 $r .= "<h3>{$children_start_char[0]}</h3>\n";
 542                                 $r .= '<ul><li>'.$children[0].'</li>';
 543                                 for ($index = 1; $index < count($children); $index++ )
 544                                 {
 545                                         if ($children_start_char[$index] != $children_start_char[$index - 1])
 546                                         {
 547                                                 $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 548                                         }
 549
 550                                         $r .= "<li>{$children[$index]}</li>";
 551                                 }
 552                                 $r .= '</ul>';
 553                         }
 554                 } # END of if ( count($children) > 0 )
 555
 556                 $r .= '<h2>' . wfMsg( 'category_header', $ti ) . "</h2>\n" .
 557                         wfMsg( 'categoryarticlecount', count( $articles ) );
 558
 559                 # Showing articles in this category
 560                 if ( count ( $articles ) > 6) {
 561                         $ti = $this->mTitle->getText() ;
 562
 563                         // divide list into three equal chunks
 564                         $chunk = (int) (count ( $articles ) / 3);
 565
 566                         // get and display header
 567                         $r .= '<table width="100%"><tr valign="top">';
 568
 569                         // loop through the chunks
 570                         for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 571                                 $chunkIndex < 3;
 572                                 $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 573                         {
 574
 575                                 $r .= '<td><ul>';
 576
 577                                 // output all articles in category
 578                                 for ($index = $startChunk ;
 579                                         $index < $endChunk && $index < count($articles);
 580                                         $index++ )
 581                                 {
 582                                         // check for change of starting letter or begging of chunk
 583                                         if ( ($articles_start_char[$index] != $articles_start_char[$index - 1])
 584                                                 || ($index == $startChunk) )
 585                                         {
 586                                                 $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 587                                         }
 588
 589                                         $r .= "<li>{$articles[$index]}</li>";
 590                                 }
 591                                 $r .= '</ul></td>';
 592
 593
 594                         }
 595                         $r .= '</tr></table>';
 596                 } elseif ( count ( $articles )  > 0) {
 597                         // for short lists of articles in categories.
 598                         $ti = $this->mTitle->getText() ;
 599
 600                         $r .= '<h3>'.$articles_start_char[0]."</h3>\n";
 601                         $r .= '<ul><li>'.$articles[0].'</li>';
 602                         for ($index = 1; $index < count($articles); $index++ )
 603                         {
 604                                 if ($articles_start_char[$index] != $articles_start_char[$index - 1])
 605                                 {
 606                                         $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 607                                 }
 608
 609                                 $r .= "<li>{$articles[$index]}</li>";
 610                         }
 611                         $r .= '</ul>';
 612                 }
 613
 614
 615                 return $r ;
 616         }
 617
 618         # Return allowed HTML attributes
 619         function getHTMLattrs () {
 620                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 621                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 622                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 623                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 624                                 /* FONT */ 'type', 'start', 'value', 'compact',
 625                                 /* For various lists, mostly deprecated but safe */
 626                                 'summary', 'width', 'border', 'frame', 'rules',
 627                                 'cellspacing', 'cellpadding', 'valign', 'char',
 628                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 629                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 630                                 'id', 'class', 'name', 'style' /* For CSS */
 631                                 );
 632                 return $htmlattrs ;
 633         }
 634
 635         # Remove non approved attributes and javascript in css
 636         function fixTagAttributes ( $t ) {
 637                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 638                 $htmlattrs = $this->getHTMLattrs() ;
 639
 640                 # Strip non-approved attributes from the tag
 641                 $t = preg_replace(
 642                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 643                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 644                         $t);
 645                 # Strip javascript "expression" from stylesheets. Brute force approach:
 646                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 647
 648                 if( preg_match(
 649                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 650                         wfMungeToUtf8( $t ) ) )
 651                 {
 652                         $t='';
 653                 }
 654
 655                 return trim ( $t ) ;
 656         }
 657
 658         # interface with html tidy, used if $wgUseTidy = true
 659         function tidy ( $text ) {
 660                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 661                 global $wgInputEncoding, $wgOutputEncoding;
 662                 $fname = 'Parser::tidy';
 663                 wfProfileIn( $fname );
 664
 665                 $cleansource = '';
 666                 switch(strtoupper($wgOutputEncoding)) {
 667                         case 'ISO-8859-1':
 668                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 669                                 break;
 670                         case 'UTF-8':
 671                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 672                                 break;
 673                         default:
 674                                 $wgTidyOpts .= ' -raw';
 675                         }
 676
 677                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 678 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 679 '<head><title>test</title></head><body>'.$text.'</body></html>';
 680                 $descriptorspec = array(
 681                         0 => array('pipe', 'r'),
 682                         1 => array('pipe', 'w'),
 683                         2 => array('file', '/dev/null', 'a')
 684                 );
 685                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 686                 if (is_resource($process)) {
 687                         fwrite($pipes[0], $wrappedtext);
 688                         fclose($pipes[0]);
 689                         while (!feof($pipes[1])) {
 690                                 $cleansource .= fgets($pipes[1], 1024);
 691                         }
 692                         fclose($pipes[1]);
 693                         $return_value = proc_close($process);
 694                 }
 695
 696                 wfProfileOut( $fname );
 697
 698                 if( $cleansource == '' && $text != '') {
 699                         wfDebug( "Tidy error detected!\n" );
 700                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 701                 } else {
 702                         return $cleansource;
 703                 }
 704         }
 705
 706         # parse the wiki syntax used to render tables
 707         function doTableStuff ( $t ) {
 708                 $fname = 'Parser::doTableStuff';
 709                 wfProfileIn( $fname );
 710
 711                 $t = explode ( "\n" , $t ) ;
 712                 $td = array () ; # Is currently a td tag open?
 713                 $ltd = array () ; # Was it TD or TH?
 714                 $tr = array () ; # Is currently a tr tag open?
 715                 $ltr = array () ; # tr attributes
 716                 $indent_level = 0; # indent level of the table
 717                 foreach ( $t AS $k => $x )
 718                 {
 719                         $x = trim ( $x ) ;
 720                         $fc = substr ( $x , 0 , 1 ) ;
 721                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) )
 722                         {
 723                                 $indent_level = strlen( $matches[1] );
 724                                 $t[$k] = "\n" .
 725                                         str_repeat( "<dl><dd>", $indent_level ) .
 726                                         "<table " . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 727                                 array_push ( $td , false ) ;
 728                                 array_push ( $ltd , '' ) ;
 729                                 array_push ( $tr , false ) ;
 730                                 array_push ( $ltr , '' ) ;
 731                         }
 732                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 733                         else if ( '|}' == substr ( $x , 0 , 2 ) )
 734                         {
 735                                 $z = "</table>\n" ;
 736                                 $l = array_pop ( $ltd ) ;
 737                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 738                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 739                                 array_pop ( $ltr ) ;
 740                                 $t[$k] = $z . str_repeat( "</dd></dl>", $indent_level );
 741                         }
 742                         else if ( '|-' == substr ( $x , 0 , 2 ) ) # Allows for |---------------
 743                         {
 744                                 $x = substr ( $x , 1 ) ;
 745                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 746                                 $z = '' ;
 747                                 $l = array_pop ( $ltd ) ;
 748                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 749                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 750                                 array_pop ( $ltr ) ;
 751                                 $t[$k] = $z ;
 752                                 array_push ( $tr , false ) ;
 753                                 array_push ( $td , false ) ;
 754                                 array_push ( $ltd , '' ) ;
 755                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 756                         }
 757                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) # Caption
 758                         {
 759                                 if ( '|+' == substr ( $x , 0 , 2 ) )
 760                                 {
 761                                         $fc = '+' ;
 762                                         $x = substr ( $x , 1 ) ;
 763                                 }
 764                                 $after = substr ( $x , 1 ) ;
 765                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 766                                 $after = explode ( '||' , $after ) ;
 767                                 $t[$k] = '' ;
 768                                 foreach ( $after AS $theline )
 769                                 {
 770                                         $z = '' ;
 771                                         if ( $fc != '+' )
 772                                         {
 773                                                 $tra = array_pop ( $ltr ) ;
 774                                                 if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
 775                                                 array_push ( $tr , true ) ;
 776                                                 array_push ( $ltr , '' ) ;
 777                                         }
 778
 779                                         $l = array_pop ( $ltd ) ;
 780                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 781                                         if ( $fc == '|' ) $l = 'td' ;
 782                                         else if ( $fc == '!' ) $l = 'th' ;
 783                                         else if ( $fc == '+' ) $l = 'caption' ;
 784                                         else $l = '' ;
 785                                         array_push ( $ltd , $l ) ;
 786                                         $y = explode ( '|' , $theline , 2 ) ;
 787                                         if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 788                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 789                                         $t[$k] .= $y ;
 790                                         array_push ( $td , true ) ;
 791                                 }
 792                         }
 793                 }
 794
 795                 # Closing open td, tr && table
 796                 while ( count ( $td ) > 0 )
 797                 {
 798                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 799                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 800                         $t[] = '</table>' ;
 801                 }
 802
 803                 $t = implode ( "\n" , $t ) ;
 804                 #               $t = $this->removeHTMLtags( $t );
 805                 wfProfileOut( $fname );
 806                 return $t ;
 807         }
 808
 809         # Parses the text and adds the result to the strip state
 810         # Returns the strip tag
 811         function stripParse( $text, $newline, $args )
 812         {
 813                 $text = $this->strip( $text, $this->mStripState );
 814                 $text = $this->internalParse( $text, (bool)$newline, $args, false );
 815                 return $newline.$this->insertStripItem( $text, $this->mStripState );
 816         }
 817
 818         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 819                 $fname = 'Parser::internalParse';
 820                 wfProfileIn( $fname );
 821
 822                 $text = $this->removeHTMLtags( $text );
 823                 $text = $this->replaceVariables( $text, $args );
 824
 825                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 826
 827                 $text = $this->doHeadings( $text );
 828                 if($this->mOptions->getUseDynamicDates()) {
 829                         global $wgDateFormatter;
 830                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 831                 }
 832                 $text = $this->doAllQuotes( $text );
 833                 // $text = $this->doExponent( $text );
 834                 $text = $this->replaceExternalLinks( $text );
 835                 $text = $this->doMagicLinks( $text );
 836                 $text = $this->replaceInternalLinks ( $text );
 837                 $text = $this->replaceInternalLinks ( $text );
 838                 //$text = $this->doTokenizedParser ( $text );
 839                 $text = $this->doTableStuff( $text );
 840                 $text = $this->formatHeadings( $text, $isMain );
 841                 $sk =& $this->mOptions->getSkin();
 842                 $text = $sk->transformContent( $text );
 843
 844                 if ( $isMain && !isset ( $this->categoryMagicDone ) ) {
 845                         $text .= $this->categoryMagic () ;
 846                         $this->categoryMagicDone = true ;
 847                 }
 848
 849                 wfProfileOut( $fname );
 850                 return $text;
 851         }
 852
 853         /* private */ function &doMagicLinks( &$text ) {
 854                 $text = $this->magicISBN( $text );
 855                 $text = $this->magicGEO( $text );
 856                 $text = $this->magicRFC( $text );
 857                 return $text;
 858         }
 859
 860         # Parse ^^ tokens and return html
 861         /* private */ function doExponent ( $text )
 862         {
 863                 $fname = 'Parser::doExponent';
 864                 wfProfileIn( $fname);
 865                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 866                 wfProfileOut( $fname);
 867                 return $text;
 868         }
 869
 870     # Parse headers and return html
 871         /* private */ function doHeadings( $text ) {
 872                 $fname = 'Parser::doHeadings';
 873                 wfProfileIn( $fname );
 874                 for ( $i = 6; $i >= 1; --$i ) {
 875                         $h = substr( '======', 0, $i );
 876                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 877                           "<h{$i}>\\1</h{$i}>\\2", $text );
 878                 }
 879                 wfProfileOut( $fname );
 880                 return $text;
 881         }
 882
 883         /* private */ function doAllQuotes( $text ) {
 884                 $fname = 'Parser::doAllQuotes';
 885                 wfProfileIn( $fname );
 886                 $outtext = '';
 887                 $lines = explode( "\n", $text );
 888                 foreach ( $lines as $line ) {
 889                         $outtext .= $this->doQuotes ( '', $line, '' ) . "\n";
 890                 }
 891                 $outtext = substr($outtext, 0,-1);
 892                 wfProfileOut( $fname );
 893                 return $outtext;
 894         }
 895
 896         /* private */ function doQuotes( $pre, $text, $mode ) {
 897                 if ( preg_match( "/^(.*)''(.*)$/sU", $text, $m ) ) {
 898                         $m1_strong = ($m[1] == "") ? "" : "<strong>{$m[1]}</strong>";
 899                         $m1_em = ($m[1] == "") ? "" : "<em>{$m[1]}</em>";
 900                         if ( substr ($m[2], 0, 1) == '\'' ) {
 901                                 $m[2] = substr ($m[2], 1);
 902                                 if ($mode == 'em') {
 903                                         return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'emstrong' );
 904                                 } else if ($mode == 'strong') {
 905                                         return $m1_strong . $this->doQuotes ( '', $m[2], '' );
 906                                 } else if (($mode == 'emstrong') || ($mode == 'both')) {
 907                                         return $this->doQuotes ( '', $pre.$m1_strong.$m[2], 'em' );
 908                                 } else if ($mode == 'strongem') {
 909                                         return "<strong>{$pre}{$m1_em}</strong>" . $this->doQuotes ( '', $m[2], 'em' );
 910                                 } else {
 911                                         return $m[1] . $this->doQuotes ( '', $m[2], 'strong' );
 912                                 }
 913                         } else {
 914                                 if ($mode == 'strong') {
 915                                         return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'strongem' );
 916                                 } else if ($mode == 'em') {
 917                                         return $m1_em . $this->doQuotes ( '', $m[2], '' );
 918                                 } else if ($mode == 'emstrong') {
 919                                         return "<em>{$pre}{$m1_strong}</em>" . $this->doQuotes ( '', $m[2], 'strong' );
 920                                 } else if (($mode == 'strongem') || ($mode == 'both')) {
 921                                         return $this->doQuotes ( '', $pre.$m1_em.$m[2], 'strong' );
 922                                 } else {
 923                                         return $m[1] . $this->doQuotes ( '', $m[2], 'em' );
 924                                 }
 925                         }
 926                 } else {
 927                         $text_strong = ($text == '') ? '' : "<strong>{$text}</strong>";
 928                         $text_em = ($text == '') ? '' : "<em>{$text}</em>";
 929                         if ($mode == '') {
 930                                 return $pre . $text;
 931                         } else if ($mode == 'em') {
 932                                 return $pre . $text_em;
 933                         } else if ($mode == 'strong') {
 934                                 return $pre . $text_strong;
 935                         } else if ($mode == 'strongem') {
 936                                 return (($pre == '') && ($text == '')) ? '' : "<strong>{$pre}{$text_em}</strong>";
 937                         } else {
 938                                 return (($pre == '') && ($text == '')) ? '' : "<em>{$pre}{$text_strong}</em>";
 939                         }
 940                 }
 941         }
 942
 943         # Note: we have to do external links before the internal ones,
 944         # and otherwise take great care in the order of things here, so
 945         # that we don't end up interpreting some URLs twice.
 946
 947         /* private */ function replaceExternalLinks( $text ) {
 948                 $fname = 'Parser::replaceExternalLinks';
 949                 wfProfileIn( $fname );
 950                 $text = $this->subReplaceExternalLinks( $text, 'http', true );
 951                 $text = $this->subReplaceExternalLinks( $text, 'https', true );
 952                 $text = $this->subReplaceExternalLinks( $text, 'ftp', false );
 953                 $text = $this->subReplaceExternalLinks( $text, 'irc', false );
 954                 $text = $this->subReplaceExternalLinks( $text, 'gopher', false );
 955                 $text = $this->subReplaceExternalLinks( $text, 'news', false );
 956                 $text = $this->subReplaceExternalLinks( $text, 'mailto', false );
 957                 wfProfileOut( $fname );
 958                 return $text;
 959         }
 960
 961         /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber ) {
 962                 $unique = '4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3';
 963                 $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
 964
 965                 # this is  the list of separators that should be ignored if they
 966                 # are the last character of an URL but that should be included
 967                 # if they occur within the URL, e.g. "go to www.foo.com, where .."
 968                 # in this case, the last comma should not become part of the URL,
 969                 # but in "www.foo.com/123,2342,32.htm" it should.
 970                 $sep = ",;\.:";
 971                 $fnc = 'A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF';
 972                 $images = 'gif|png|jpg|jpeg';
 973
 974                 # PLEASE NOTE: The curly braces { } are not part of the regex,
 975                 # they are interpreted as part of the string (used to tell PHP
 976                 # that the content of the string should be inserted there).
 977                 $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
 978                   "((?i){$images})([^{$uc}]|$)/";
 979
 980                 $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
 981                 $sk =& $this->mOptions->getSkin();
 982
 983                 if ( $autonumber and $this->mOptions->getAllowExternalImages() ) { # Use img tags only for HTTP urls
 984                         $s = preg_replace( $e1, '\\1' . $sk->makeImage( "{$unique}:\\3" .
 985                           '/\\4.\\5', '\\4.\\5' ) . '\\6', $s );
 986                 }
 987                 $s = preg_replace( $e2, '\\1' . "<a href=\"{$unique}:\\3\"" .
 988                   $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
 989                   "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
 990                   '</a>\\5', $s );
 991                 $s = str_replace( $unique, $protocol, $s );
 992
 993                 $a = explode( "[{$protocol}:", " " . $s );
 994                 $s = array_shift( $a );
 995                 $s = substr( $s, 1 );
 996
 997                 # Regexp for URL in square brackets
 998                 $e1 = "/^([{$uc}{$sep}]+)\\](.*)\$/sD";
 999                 # Regexp for URL with link text in square brackets
1000                 $e2 = "/^([{$uc}{$sep}]+)\\s+([^\\]]+)\\](.*)\$/sD";
1001
1002                 foreach ( $a as $line ) {
1003
1004                         # CASE 1: Link in square brackets, e.g.
1005                         # some text [http://domain.tld/some.link] more text
1006                         if ( preg_match( $e1, $line, $m ) ) {
1007                                 $link = "{$protocol}:{$m[1]}";
1008                                 $trail = $m[2];
1009                                 if ( $autonumber ) { $text = "[" . ++$this->mAutonumber . "]"; }
1010                                 else { $text = wfEscapeHTML( $link ); }
1011                         }
1012
1013                         # CASE 2: Link with link text and text directly following it, e.g.
1014                         # This is a collection of [http://domain.tld/some.link link]s
1015                         else if ( preg_match( $e2, $line, $m ) ) {
1016                                 $link = "{$protocol}:{$m[1]}";
1017                                 $text = $m[2];
1018                                 $dtrail = '';
1019                                 $trail = $m[3];
1020                                 if ( preg_match( wfMsg ('linktrail'), $trail, $m2 ) ) {
1021                                         $dtrail = $m2[1];
1022                                         $trail = $m2[2];
1023                                 }
1024                         }
1025
1026                         # CASE 3: Nothing matches, just output the source text
1027                         else {
1028                                 $s .= "[{$protocol}:" . $line;
1029                                 continue;
1030                         }
1031
1032                         if( $link == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
1033                                 $paren = '';
1034                         } else {
1035                                 # Expand the URL for printable version
1036                                 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
1037                         }
1038                         $la = $sk->getExternalLinkAttributes( $link, $text );
1039                         $s .= "<a href='{$link}'{$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
1040
1041                 }
1042                 return $s;
1043         }
1044
1045
1046         /* private */ function replaceInternalLinks( $s ) {
1047                 global $wgLang, $wgLinkCache;
1048                 global $wgNamespacesWithSubpages, $wgLanguageCode;
1049                 static $fname = 'Parser::replaceInternalLinks' ;
1050                 wfProfileIn( $fname );
1051
1052                 wfProfileIn( $fname.'-setup' );
1053                 static $tc = FALSE;
1054                 # the % is needed to support urlencoded titles as well
1055                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1056                 $sk =& $this->mOptions->getSkin();
1057
1058                 $isRedirect = false ;
1059                 if ( trim ( substr ( $s , 0 , 10 ) ) == '#REDIRECT' ) $isRedirect = true ;
1060                 $a = explode( '[[', ' ' . $s );
1061                 $s = array_shift( $a );
1062                 $s = substr( $s, 1 );
1063
1064                 # Match a link having the form [[namespace:link|alternate]]trail
1065                 static $e1 = FALSE;
1066                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1067                 # Match the end of a line for a word that's not followed by whitespace,
1068                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1069                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1070
1071                 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
1072                 # Special and Media are pseudo-namespaces; no pages actually exist in them
1073                 static $image = FALSE;
1074                 static $special = FALSE;
1075                 static $media = FALSE;
1076                 static $category = FALSE;
1077                 if ( !$image ) { $image = Namespace::getImage(); }
1078                 if ( !$special ) { $special = Namespace::getSpecial(); }
1079                 if ( !$media ) { $media = Namespace::getMedia(); }
1080                 if ( !$category ) { $category = Namespace::getCategory(); }
1081
1082                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1083
1084                 if ( $useLinkPrefixExtension ) {
1085                         if ( preg_match( $e2, $s, $m ) ) {
1086                                 $first_prefix = $m[2];
1087                                 $s = $m[1];
1088                         } else {
1089                                 $first_prefix = false;
1090                         }
1091                 } else {
1092                         $prefix = '';
1093                 }
1094
1095                 wfProfileOut( $fname.'-setup' );
1096
1097                 foreach ( $a as $line ) {
1098                         wfProfileIn( $fname.'-prefixhandling' );
1099                         if ( $useLinkPrefixExtension ) {
1100                                 if ( preg_match( $e2, $s, $m ) ) {
1101                                         $prefix = $m[2];
1102                                         $s = $m[1];
1103                                 } else {
1104                                         $prefix='';
1105                                 }
1106                                 # first link
1107                                 if($first_prefix) {
1108                                         $prefix = $first_prefix;
1109                                         $first_prefix = false;
1110                                 }
1111                         }
1112                         wfProfileOut( $fname.'-prefixhandling' );
1113
1114                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1115                                 $text = $m[2];
1116                                 # fix up urlencoded title texts
1117                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1118                                 $trail = $m[3];
1119                         } else { # Invalid form; output directly
1120                                 $s .= $prefix . '[[' . $line ;
1121                                 continue;
1122                         }
1123
1124                         /* Valid link forms:
1125                         Foobar -- normal
1126                         :Foobar -- override special treatment of prefix (images, language links)
1127                         /Foobar -- convert to CurrentPage/Foobar
1128                         /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1129                         */
1130                         $c = substr($m[1],0,1);
1131                         $noforce = ($c != ':');
1132                         if( $c == '/' ) { # subpage
1133                                 if(substr($m[1],-1,1)=='/') {                 # / at end means we don't want the slash to be shown
1134                                         $m[1]=substr($m[1],1,strlen($m[1])-2);
1135                                         $noslash=$m[1];
1136                                 } else {
1137                                         $noslash=substr($m[1],1);
1138                                 }
1139                                 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) { # subpages allowed here
1140                                         $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1141                                         if( '' == $text ) {
1142                                                 $text= $m[1];
1143                                         } # this might be changed for ugliness reasons
1144                                 } else {
1145                                         $link = $noslash; # no subpage allowed, use standard link
1146                                 }
1147                         } elseif( $noforce ) { # no subpage
1148                                 $link = $m[1];
1149                         } else {
1150                                 $link = substr( $m[1], 1 );
1151                         }
1152                         $wasblank = ( '' == $text );
1153                         if( $wasblank )
1154                         $text = $link;
1155
1156                         $nt = Title::newFromText( $link );
1157                         if( !$nt ) {
1158                                 $s .= $prefix . '[[' . $line;
1159                                 continue;
1160                         }
1161                         $ns = $nt->getNamespace();
1162                         $iw = $nt->getInterWiki();
1163                         if( $noforce ) {
1164                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1165                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1166                                         $tmp = $prefix . $trail ;
1167                                         $s .= (trim($tmp) == '')? '': $tmp;
1168                                         continue;
1169                                 }
1170                                 if ( $ns == $image ) {
1171                                         $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1172                                         $wgLinkCache->addImageLinkObj( $nt );
1173                                         continue;
1174                                 }
1175                                 if ( $ns == $category && !$isRedirect ) {
1176                                         $t = $nt->getText() ;
1177                                         $nnt = Title::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
1178
1179                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1180                                         $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1181                                         $wgLinkCache->resume();
1182
1183                                         $sortkey = $wasblank ? $this->mTitle->getPrefixedText() : $text;
1184                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1185                                         $this->mOutput->mCategoryLinks[] = $t ;
1186                                         $s .= $prefix . $trail ;
1187                                         continue;
1188                                 }
1189                         }
1190                         if( ( $nt->getPrefixedText() == $this->mTitle->getPrefixedText() ) &&
1191                         ( strpos( $link, '#' ) == FALSE ) ) {
1192                                 # Self-links are handled specially; generally de-link and change to bold.
1193                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1194                                 continue;
1195                         }
1196
1197                         if( $ns == $media ) {
1198                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1199                                 $wgLinkCache->addImageLinkObj( $nt );
1200                                 continue;
1201                         } elseif( $ns == $special ) {
1202                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1203                                 continue;
1204                         }
1205                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1206                 }
1207                 wfProfileOut( $fname );
1208                 return $s;
1209         }
1210
1211         # Some functions here used by doBlockLevels()
1212         #
1213         /* private */ function closeParagraph() {
1214                 $result = '';
1215                 if ( '' != $this->mLastSection ) {
1216                         $result = '</' . $this->mLastSection  . ">\n";
1217                 }
1218                 $this->mInPre = false;
1219                 $this->mLastSection = '';
1220                 return $result;
1221         }
1222         # getCommon() returns the length of the longest common substring
1223         # of both arguments, starting at the beginning of both.
1224         #
1225         /* private */ function getCommon( $st1, $st2 ) {
1226                 $fl = strlen( $st1 );
1227                 $shorter = strlen( $st2 );
1228                 if ( $fl < $shorter ) { $shorter = $fl; }
1229
1230                 for ( $i = 0; $i < $shorter; ++$i ) {
1231                         if ( $st1{$i} != $st2{$i} ) { break; }
1232                 }
1233                 return $i;
1234         }
1235         # These next three functions open, continue, and close the list
1236         # element appropriate to the prefix character passed into them.
1237         #
1238         /* private */ function openList( $char )
1239     {
1240                 $result = $this->closeParagraph();
1241
1242                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1243                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1244                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1245                 else if ( ';' == $char ) {
1246                         $result .= '<dl><dt>';
1247                         $this->mDTopen = true;
1248                 }
1249                 else { $result = '<!-- ERR 1 -->'; }
1250
1251                 return $result;
1252         }
1253
1254         /* private */ function nextItem( $char ) {
1255                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1256                 else if ( ':' == $char || ';' == $char ) {
1257                         $close = "</dd>";
1258                         if ( $this->mDTopen ) { $close = '</dt>'; }
1259                         if ( ';' == $char ) {
1260                                 $this->mDTopen = true;
1261                                 return $close . '<dt>';
1262                         } else {
1263                                 $this->mDTopen = false;
1264                                 return $close . '<dd>';
1265                         }
1266                 }
1267                 return '<!-- ERR 2 -->';
1268         }
1269
1270         /* private */function closeList( $char ) {
1271                 if ( '*' == $char ) { $text = '</li></ul>'; }
1272                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1273                 else if ( ':' == $char ) {
1274                         if ( $this->mDTopen ) {
1275                                 $this->mDTopen = false;
1276                                 $text = '</dt></dl>';
1277                         } else {
1278                                 $text = '</dd></dl>';
1279                         }
1280                 }
1281                 else {  return '<!-- ERR 3 -->'; }
1282                 return $text."\n";
1283         }
1284
1285         /* private */ function doBlockLevels( $text, $linestart ) {
1286                 $fname = 'Parser::doBlockLevels';
1287                 wfProfileIn( $fname );
1288
1289                 # Parsing through the text line by line.  The main thing
1290                 # happening here is handling of block-level elements p, pre,
1291                 # and making lists from lines starting with * # : etc.
1292                 #
1293                 $textLines = explode( "\n", $text );
1294
1295                 $lastPrefix = $output = $lastLine = '';
1296                 $this->mDTopen = $inBlockElem = false;
1297                 $prefixLength = 0;
1298                 $paragraphStack = false;
1299
1300                 if ( !$linestart ) {
1301                         $output .= array_shift( $textLines );
1302                 }
1303                 foreach ( $textLines as $oLine ) {
1304                         $lastPrefixLength = strlen( $lastPrefix );
1305                         $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1306                         $preOpenMatch = preg_match("/<pre/i", $oLine );
1307                         if ( !$this->mInPre ) {
1308                                 # Multiple prefixes may abut each other for nested lists.
1309                                 $prefixLength = strspn( $oLine, '*#:;' );
1310                                 $pref = substr( $oLine, 0, $prefixLength );
1311
1312                                 # eh?
1313                                 $pref2 = str_replace( ';', ':', $pref );
1314                                 $t = substr( $oLine, $prefixLength );
1315                                 $this->mInPre = !empty($preOpenMatch);
1316                         } else {
1317                                 # Don't interpret any other prefixes in preformatted text
1318                                 $prefixLength = 0;
1319                                 $pref = $pref2 = '';
1320                                 $t = $oLine;
1321                         }
1322
1323                         # List generation
1324                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1325                                 # Same as the last item, so no need to deal with nesting or opening stuff
1326                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1327                                 $paragraphStack = false;
1328
1329                                 if ( ";" == substr( $pref, -1 ) ) {
1330                                         # The one nasty exception: definition lists work like this:
1331                                         # ; title : definition text
1332                                         # So we check for : in the remainder text to split up the
1333                                         # title and definition, without b0rking links.
1334                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1335                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1336                                                 $term = $match[1];
1337                                                 $output .= $term . $this->nextItem( ':' );
1338                                                 $t = $match[2];
1339                                         }
1340                                 }
1341                         } elseif( $prefixLength || $lastPrefixLength ) {
1342                                 # Either open or close a level...
1343                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1344                                 $paragraphStack = false;
1345
1346                                 while( $commonPrefixLength < $lastPrefixLength ) {
1347                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1348                                         --$lastPrefixLength;
1349                                 }
1350                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1351                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1352                                 }
1353                                 while ( $prefixLength > $commonPrefixLength ) {
1354                                         $char = substr( $pref, $commonPrefixLength, 1 );
1355                                         $output .= $this->openList( $char );
1356
1357                                         if ( ';' == $char ) {
1358                                                 # FIXME: This is dupe of code above
1359                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1360                                                         $term = $match[1];
1361                                                         $output .= $term . $this->nextItem( ":" );
1362                                                         $t = $match[2];
1363                                                 }
1364                                         }
1365                                         ++$commonPrefixLength;
1366                                 }
1367                                 $lastPrefix = $pref2;
1368                         }
1369                         if( 0 == $prefixLength ) {
1370                                 # No prefix (not in list)--go to paragraph mode
1371                                 $uniq_prefix = UNIQ_PREFIX;
1372                                 // XXX: use a stack for nestable elements like span, table and div
1373                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1374                                 $closematch = preg_match(
1375                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1376                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1377                                 if ( $openmatch or $closematch ) {
1378                                         $paragraphStack = false;
1379                                         $output .= $this->closeParagraph();
1380                                         if($preOpenMatch and !$preCloseMatch) {
1381                                                 $this->mInPre = true;
1382                                         }
1383                                         if ( $closematch  ) {
1384                                                 $inBlockElem = false;
1385                                         } else {
1386                                                 $inBlockElem = true;
1387                                         }
1388                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1389                                         if ( " " == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1390                                                 // pre
1391                                                 if ($this->mLastSection != 'pre') {
1392                                                         $paragraphStack = false;
1393                                                         $output .= $this->closeParagraph().'<pre>';
1394                                                         $this->mLastSection = 'pre';
1395                                                 }
1396                                         } else {
1397                                                 // paragraph
1398                                                 if ( '' == trim($t) ) {
1399                                                         if ( $paragraphStack ) {
1400                                                                 $output .= $paragraphStack.'<br />';
1401                                                                 $paragraphStack = false;
1402                                                                 $this->mLastSection = 'p';
1403                                                         } else {
1404                                                                 if ($this->mLastSection != 'p' ) {
1405                                                                         $output .= $this->closeParagraph();
1406                                                                         $this->mLastSection = '';
1407                                                                         $paragraphStack = '<p>';
1408                                                                 } else {
1409                                                                         $paragraphStack = '</p><p>';
1410                                                                 }
1411                                                         }
1412                                                 } else {
1413                                                         if ( $paragraphStack ) {
1414                                                                 $output .= $paragraphStack;
1415                                                                 $paragraphStack = false;
1416                                                                 $this->mLastSection = 'p';
1417                                                         } else if ($this->mLastSection != 'p') {
1418                                                                 $output .= $this->closeParagraph().'<p>';
1419                                                                 $this->mLastSection = 'p';
1420                                                         }
1421                                                 }
1422                                         }
1423                                 }
1424                         }
1425                         if ($paragraphStack === false) {
1426                                 $output .= $t."\n";
1427                         }
1428                 }
1429                 while ( $prefixLength ) {
1430                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1431                         --$prefixLength;
1432                 }
1433                 if ( '' != $this->mLastSection ) {
1434                         $output .= '</' . $this->mLastSection . '>';
1435                         $this->mLastSection = '';
1436                 }
1437
1438                 wfProfileOut( $fname );
1439                 return $output;
1440         }
1441
1442         # Return value of a magic variable (like PAGENAME)
1443         function getVariableValue( $index ) {
1444                 global $wgLang, $wgSitename, $wgServer;
1445
1446                 switch ( $index ) {
1447                         case MAG_CURRENTMONTH:
1448                                 return $wgLang->formatNum( date( 'm' ) );
1449                         case MAG_CURRENTMONTHNAME:
1450                                 return $wgLang->getMonthName( date('n') );
1451                         case MAG_CURRENTMONTHNAMEGEN:
1452                                 return $wgLang->getMonthNameGen( date('n') );
1453                         case MAG_CURRENTDAY:
1454                                 return $wgLang->formatNum( date('j') );
1455                         case MAG_PAGENAME:
1456                                 return $this->mTitle->getText();
1457                         case MAG_NAMESPACE:
1458                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1459                                 return $wgLang->getNsText($this->mTitle->getNamespace()); // Patch  by Dori
1460                         case MAG_CURRENTDAYNAME:
1461                                 return $wgLang->getWeekdayName( date('w')+1 );
1462                         case MAG_CURRENTYEAR:
1463                                 return $wgLang->formatNum( date( 'Y' ) );
1464                         case MAG_CURRENTTIME:
1465                                 return $wgLang->time( wfTimestampNow(), false );
1466                         case MAG_NUMBEROFARTICLES:
1467                                 return $wgLang->formatNum( wfNumberOfArticles() );
1468                         case MAG_SITENAME:
1469                                 return $wgSitename;
1470                         case MAG_SERVER:
1471                                 return $wgServer;
1472                         default:
1473                                 return NULL;
1474                 }
1475         }
1476
1477         # initialise the magic variables (like CURRENTMONTHNAME)
1478         function initialiseVariables() {
1479                 global $wgVariableIDs;
1480                 $this->mVariables = array();
1481                 foreach ( $wgVariableIDs as $id ) {
1482                         $mw =& MagicWord::get( $id );
1483                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1484                 }
1485         }
1486
1487         /* private */ function replaceVariables( $text, $args = array() ) {
1488                 global $wgLang, $wgScript, $wgArticlePath;
1489
1490                 # Prevent too big inclusions
1491                 if(strlen($text)> MAX_INCLUDE_SIZE)
1492                    return $text;
1493
1494                 $fname = 'Parser::replaceVariables';
1495                 wfProfileIn( $fname );
1496
1497                 $bail = false;
1498                 $titleChars = Title::legalChars();
1499                 $nonBraceChars = str_replace( array( '{', '}' ), array( '', '' ), $titleChars );
1500
1501                 # This function is called recursively. To keep track of arguments we need a stack:
1502                 array_push( $this->mArgStack, $args );
1503
1504                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1505                 $GLOBALS['wgCurParser'] =& $this;
1506
1507
1508                 if ( $this->mOutputType == OT_HTML ) {
1509                         # Variable substitution
1510                         $text = preg_replace_callback( "/{{([$nonBraceChars]*?)}}/", 'wfVariableSubstitution', $text );
1511
1512                         # Argument substitution
1513                         $text = preg_replace_callback( "/(\\n?){{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1514                 }
1515                 # Template substitution
1516                 $regex = '/(\\n?){{(['.$nonBraceChars.']*)(\\|.*?|)}}/s';
1517                 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1518
1519                 array_pop( $this->mArgStack );
1520
1521                 wfProfileOut( $fname );
1522                 return $text;
1523         }
1524
1525         function variableSubstitution( $matches ) {
1526                 if ( !$this->mVariables ) {
1527                         $this->initialiseVariables();
1528                 }
1529                 if ( array_key_exists( $matches[1], $this->mVariables ) ) {
1530                         $text = $this->mVariables[$matches[1]];
1531                         $this->mOutput->mContainsOldMagic = true;
1532                 } else {
1533                         $text = $matches[0];
1534                 }
1535                 return $text;
1536         }
1537
1538         # Split template arguments
1539         function getTemplateArgs( $argsString ) {
1540                 if ( $argsString === '' ) {
1541                         return array();
1542                 }
1543
1544                 $args = explode( '|', substr( $argsString, 1 ) );
1545
1546                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1547                 # merged with the next arg because the '|' character between belongs
1548                 # to the link syntax and not the template parameter syntax.
1549                 $argc = count($args);
1550                 $i = 0;
1551                 for ( $i = 0; $i < $argc-1; $i++ ) {
1552                         if ( substr_count ( $args[$i], "[[" ) != substr_count ( $args[$i], "]]" ) ) {
1553                                 $args[$i] .= "|".$args[$i+1];
1554                                 array_splice($args, $i+1, 1);
1555                                 $i--;
1556                                 $argc--;
1557                         }
1558                 }
1559
1560                 return $args;
1561         }
1562
1563         function braceSubstitution( $matches ) {
1564                 global $wgLinkCache, $wgLang;
1565                 $fname = 'Parser::braceSubstitution';
1566                 $found = false;
1567                 $nowiki = false;
1568                 $noparse = false;
1569
1570                 $title = NULL;
1571
1572                 # $newline is an optional newline character before the braces
1573                 # $part1 is the bit before the first |, and must contain only title characters
1574                 # $args is a list of arguments, starting from index 0, not including $part1
1575
1576                 $newline = $matches[1];
1577                 $part1 = $matches[2];
1578                 # If the third subpattern matched anything, it will start with |
1579
1580                 $args = $this->getTemplateArgs($matches[3]);
1581                 $argc = count( $args );
1582
1583                 # {{{}}}
1584                 if ( strpos( $matches[0], '{{{' ) !== false ) {
1585                         $text = $matches[0];
1586                         $found = true;
1587                         $noparse = true;
1588                 }
1589
1590                 # SUBST
1591                 if ( !$found ) {
1592                         $mwSubst =& MagicWord::get( MAG_SUBST );
1593                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1594                                 if ( $this->mOutputType != OT_WIKI ) {
1595                                         # Invalid SUBST not replaced at PST time
1596                                         # Return without further processing
1597                                         $text = $matches[0];
1598                                         $found = true;
1599                                         $noparse= true;
1600                                 }
1601                         } elseif ( $this->mOutputType == OT_WIKI ) {
1602                                 # SUBST not found in PST pass, do nothing
1603                                 $text = $matches[0];
1604                                 $found = true;
1605                         }
1606                 }
1607
1608                 # MSG, MSGNW and INT
1609                 if ( !$found ) {
1610                         # Check for MSGNW:
1611                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1612                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1613                                 $nowiki = true;
1614                         } else {
1615                                 # Remove obsolete MSG:
1616                                 $mwMsg =& MagicWord::get( MAG_MSG );
1617                                 $mwMsg->matchStartAndRemove( $part1 );
1618                         }
1619
1620                         # Check if it is an internal message
1621                         $mwInt =& MagicWord::get( MAG_INT );
1622                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1623                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1624                                         $text = wfMsgReal( $part1, $args, true );
1625                                         $found = true;
1626                                 }
1627                         }
1628                 }
1629
1630                 # NS
1631                 if ( !$found ) {
1632                         # Check for NS: (namespace expansion)
1633                         $mwNs = MagicWord::get( MAG_NS );
1634                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1635                                 if ( intval( $part1 ) ) {
1636                                         $text = $wgLang->getNsText( intval( $part1 ) );
1637                                         $found = true;
1638                                 } else {
1639                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1640                                         if ( !is_null( $index ) ) {
1641                                                 $text = $wgLang->getNsText( $index );
1642                                                 $found = true;
1643                                         }
1644                                 }
1645                         }
1646                 }
1647
1648                 # LOCALURL and LOCALURLE
1649                 if ( !$found ) {
1650                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1651                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1652
1653                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1654                                 $func = 'getLocalURL';
1655                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1656                                 $func = 'escapeLocalURL';
1657                         } else {
1658                                 $func = '';
1659                         }
1660
1661                         if ( $func !== '' ) {
1662                                 $title = Title::newFromText( $part1 );
1663                                 if ( !is_null( $title ) ) {
1664                                         if ( $argc > 0 ) {
1665                                                 $text = $title->$func( $args[0] );
1666                                         } else {
1667                                                 $text = $title->$func();
1668                                         }
1669                                         $found = true;
1670                                 }
1671                         }
1672                 }
1673
1674                 # Internal variables
1675                 if ( !$this->mVariables ) {
1676                         $this->initialiseVariables();
1677                 }
1678                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1679                         $text = $this->mVariables[$part1];
1680                         $found = true;
1681                         $this->mOutput->mContainsOldMagic = true;
1682                 }
1683
1684                 # Template table test
1685
1686                 # Did we encounter this template already? If yes, it is in the cache
1687                 # and we need to check for loops.
1688                 if ( isset( $this->mTemplates[$part1] ) ) {
1689                         # Infinite loop test
1690                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1691                                 $noparse = true;
1692                                 $found = true;
1693                         }
1694                         # set $text to cached message.
1695                         $text = $this->mTemplates[$part1];
1696                         $found = true;
1697                 }
1698
1699                 # Load from database
1700                 if ( !$found ) {
1701                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1702                         if ( !is_null( $title ) && !$title->isExternal() ) {
1703                                 # Check for excessive inclusion
1704                                 $dbk = $title->getPrefixedDBkey();
1705                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1706                                         # This should never be reached.
1707                                         $article = new Article( $title );
1708                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1709                                         if ( $articleContent !== false ) {
1710                                                 $found = true;
1711                                                 $text = $articleContent;
1712
1713                                         }
1714                                 }
1715
1716                                 # If the title is valid but undisplayable, make a link to it
1717                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1718                                         $text = '[[' . $title->getPrefixedText() . ']]';
1719                                         $found = true;
1720                                 }
1721
1722                                 # Template cache array insertion
1723                                 $this->mTemplates[$part1] = $text;
1724                         }
1725                 }
1726
1727                 # Recursive parsing, escaping and link table handling
1728                 # Only for HTML output
1729                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1730                         $text = wfEscapeWikiText( $text );
1731                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1732                         # Clean up argument array
1733                         $assocArgs = array();
1734                         $index = 1;
1735                         foreach( $args as $arg ) {
1736                                 $eqpos = strpos( $arg, '=' );
1737                                 if ( $eqpos === false ) {
1738                                         $assocArgs[$index++] = $arg;
1739                                 } else {
1740                                         $name = trim( substr( $arg, 0, $eqpos ) );
1741                                         $value = trim( substr( $arg, $eqpos+1 ) );
1742                                         if ( $value === false ) {
1743                                                 $value = '';
1744                                         }
1745                                         if ( $name !== false ) {
1746                                                 $assocArgs[$name] = $value;
1747                                         }
1748                                 }
1749                         }
1750
1751                         # Do not enter included links in link table
1752                         if ( !is_null( $title ) ) {
1753                                 $wgLinkCache->suspend();
1754                         }
1755
1756                         # Add a new element to the templace recursion path
1757                         $this->mTemplatePath[$part1] = 1;
1758
1759                         # Run full parser on the included text
1760                         $text = $this->stripParse( $text, $newline, $assocArgs );
1761
1762                         # Resume the link cache and register the inclusion as a link
1763                         if ( !is_null( $title ) ) {
1764                                 $wgLinkCache->resume();
1765                                 $wgLinkCache->addLinkObj( $title );
1766                         }
1767                 }
1768                 # Empties the template path
1769                 $this->mTemplatePath = array();
1770
1771                 if ( !$found ) {
1772                         return $matches[0];
1773                 } else {
1774                         return $text;
1775                 }
1776         }
1777
1778         # Triple brace replacement -- used for template arguments
1779         function argSubstitution( $matches ) {
1780                 $newline = $matches[1];
1781                 $arg = trim( $matches[2] );
1782                 $text = $matches[0];
1783                 $inputArgs = end( $this->mArgStack );
1784
1785                 if ( array_key_exists( $arg, $inputArgs ) ) {
1786                         $text = $this->stripParse( $inputArgs[$arg], $newline, array() );
1787                 }
1788
1789                 return $text;
1790         }
1791
1792         # Returns true if the function is allowed to include this entity
1793         function incrementIncludeCount( $dbk ) {
1794                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1795                         $this->mIncludeCount[$dbk] = 0;
1796                 }
1797                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1798                         return true;
1799                 } else {
1800                         return false;
1801                 }
1802         }
1803
1804
1805         # Cleans up HTML, removes dangerous tags and attributes
1806         /* private */ function removeHTMLtags( $text ) {
1807                 global $wgUseTidy, $wgUserHtml;
1808                 $fname = 'Parser::removeHTMLtags';
1809                 wfProfileIn( $fname );
1810
1811                 if( $wgUserHtml ) {
1812                         $htmlpairs = array( # Tags that must be closed
1813                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1814                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1815                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
1816                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1817                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
1818                         );
1819                         $htmlsingle = array(
1820                                 'br', 'hr', 'li', 'dt', 'dd'
1821                         );
1822                         $htmlnest = array( # Tags that can be nested--??
1823                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1824                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
1825                         );
1826                         $tabletags = array( # Can only appear inside table
1827                                 'td', 'th', 'tr'
1828                         );
1829                 } else {
1830                         $htmlpairs = array();
1831                         $htmlsingle = array();
1832                         $htmlnest = array();
1833                         $tabletags = array();
1834                 }
1835
1836                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1837                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1838
1839                 $htmlattrs = $this->getHTMLattrs () ;
1840
1841                 # Remove HTML comments
1842                 $text = preg_replace( '/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU', '$2', $text );
1843
1844                 $bits = explode( '<', $text );
1845                 $text = array_shift( $bits );
1846                 if(!$wgUseTidy) {
1847                         $tagstack = array(); $tablestack = array();
1848                         foreach ( $bits as $x ) {
1849                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1850                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1851                                 $x, $regs );
1852                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1853                                 error_reporting( $prev );
1854
1855                                 $badtag = 0 ;
1856                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1857                                         # Check our stack
1858                                         if ( $slash ) {
1859                                                 # Closing a tag...
1860                                                 if ( ! in_array( $t, $htmlsingle ) &&
1861                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
1862                                                         @array_push( $tagstack, $ot );
1863                                                         $badtag = 1;
1864                                                 } else {
1865                                                         if ( $t == 'table' ) {
1866                                                                 $tagstack = array_pop( $tablestack );
1867                                                         }
1868                                                         $newparams = '';
1869                                                 }
1870                                         } else {
1871                                                 # Keep track for later
1872                                                 if ( in_array( $t, $tabletags ) &&
1873                                                 ! in_array( 'table', $tagstack ) ) {
1874                                                         $badtag = 1;
1875                                                 } else if ( in_array( $t, $tagstack ) &&
1876                                                 ! in_array ( $t , $htmlnest ) ) {
1877                                                         $badtag = 1 ;
1878                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1879                                                         if ( $t == 'table' ) {
1880                                                                 array_push( $tablestack, $tagstack );
1881                                                                 $tagstack = array();
1882                                                         }
1883                                                         array_push( $tagstack, $t );
1884                                                 }
1885                                                 # Strip non-approved attributes from the tag
1886                                                 $newparams = $this->fixTagAttributes($params);
1887
1888                                         }
1889                                         if ( ! $badtag ) {
1890                                                 $rest = str_replace( '>', '&gt;', $rest );
1891                                                 $text .= "<$slash$t $newparams$brace$rest";
1892                                                 continue;
1893                                         }
1894                                 }
1895                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1896                         }
1897                         # Close off any remaining tags
1898                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
1899                                 $text .= "</$t>\n";
1900                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
1901                         }
1902                 } else {
1903                         # this might be possible using tidy itself
1904                         foreach ( $bits as $x ) {
1905                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1906                                 $x, $regs );
1907                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1908                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1909                                         $newparams = $this->fixTagAttributes($params);
1910                                         $rest = str_replace( '>', '&gt;', $rest );
1911                                         $text .= "<$slash$t $newparams$brace$rest";
1912                                 } else {
1913                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1914                                 }
1915                         }
1916                 }
1917                 wfProfileOut( $fname );
1918                 return $text;
1919         }
1920
1921
1922 /*
1923  *
1924  * This function accomplishes several tasks:
1925  * 1) Auto-number headings if that option is enabled
1926  * 2) Add an [edit] link to sections for logged in users who have enabled the option
1927  * 3) Add a Table of contents on the top for users who have enabled the option
1928  * 4) Auto-anchor headings
1929  *
1930  * It loops through all headlines, collects the necessary data, then splits up the
1931  * string and re-inserts the newly formatted headlines.
1932  *
1933  */
1934
1935         /* private */ function formatHeadings( $text, $isMain=true ) {
1936                 global $wgInputEncoding, $wgMaxTocLevel;
1937
1938                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1939                 $doShowToc = $this->mOptions->getShowToc();
1940                 $forceTocHere = false;
1941                 if( !$this->mTitle->userCanEdit() ) {
1942                         $showEditLink = 0;
1943                         $rightClickHack = 0;
1944                 } else {
1945                         $showEditLink = $this->mOptions->getEditSection();
1946                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1947                 }
1948
1949                 # Inhibit editsection links if requested in the page
1950                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1951                 if( $esw->matchAndRemove( $text ) ) {
1952                         $showEditLink = 0;
1953                 }
1954                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1955                 # do not add TOC
1956                 $mw =& MagicWord::get( MAG_NOTOC );
1957                 if( $mw->matchAndRemove( $text ) ) {
1958                         $doShowToc = 0;
1959                 }
1960
1961                 # never add the TOC to the Main Page. This is an entry page that should not
1962                 # be more than 1-2 screens large anyway
1963                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
1964                         $doShowToc = 0;
1965                 }
1966
1967                 # Get all headlines for numbering them and adding funky stuff like [edit]
1968                 # links - this is for later, but we need the number of headlines right now
1969                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
1970
1971                 # if there are fewer than 4 headlines in the article, do not show TOC
1972                 if( $numMatches < 4 ) {
1973                         $doShowToc = 0;
1974                 }
1975
1976                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
1977                 # override above conditions and always show TOC at that place
1978                 $mw =& MagicWord::get( MAG_TOC );
1979                 if ($mw->match( $text ) ) {
1980                         $doShowToc = 1;
1981                         $forceTocHere = true;
1982                 } else {
1983                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1984                         # override above conditions and always show TOC above first header
1985                         $mw =& MagicWord::get( MAG_FORCETOC );
1986                         if ($mw->matchAndRemove( $text ) ) {
1987                                 $doShowToc = 1;
1988                         }
1989                 }
1990
1991
1992
1993                 # We need this to perform operations on the HTML
1994                 $sk =& $this->mOptions->getSkin();
1995
1996                 # headline counter
1997                 $headlineCount = 0;
1998
1999                 # Ugh .. the TOC should have neat indentation levels which can be
2000                 # passed to the skin functions. These are determined here
2001                 $toclevel = 0;
2002                 $toc = '';
2003                 $full = '';
2004                 $head = array();
2005                 $sublevelCount = array();
2006                 $level = 0;
2007                 $prevlevel = 0;
2008                 foreach( $matches[3] as $headline ) {
2009                         $numbering = '';
2010                         if( $level ) {
2011                                 $prevlevel = $level;
2012                         }
2013                         $level = $matches[1][$headlineCount];
2014                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2015                                 # reset when we enter a new level
2016                                 $sublevelCount[$level] = 0;
2017                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2018                                 $toclevel += $level - $prevlevel;
2019                         }
2020                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2021                                 # reset when we step back a level
2022                                 $sublevelCount[$level+1]=0;
2023                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2024                                 $toclevel -= $prevlevel - $level;
2025                         }
2026                         # count number of headlines for each level
2027                         @$sublevelCount[$level]++;
2028                         if( $doNumberHeadings || $doShowToc ) {
2029                                 $dot = 0;
2030                                 for( $i = 1; $i <= $level; $i++ ) {
2031                                         if( !empty( $sublevelCount[$i] ) ) {
2032                                                 if( $dot ) {
2033                                                         $numbering .= '.';
2034                                                 }
2035                                                 $numbering .= $sublevelCount[$i];
2036                                                 $dot = 1;
2037                                         }
2038                                 }
2039                         }
2040
2041                         # The canonized header is a version of the header text safe to use for links
2042                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2043                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2044                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2045
2046                         # strip out HTML
2047                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2048                         $tocline = trim( $canonized_headline );
2049                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2050                         $replacearray = array(
2051                                 '%3A' => ':',
2052                                 '%' => '.'
2053                         );
2054                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2055                         $refer[$headlineCount] = $canonized_headline;
2056
2057                         # count how many in assoc. array so we can track dupes in anchors
2058                         @$refers[$canonized_headline]++;
2059                         $refcount[$headlineCount]=$refers[$canonized_headline];
2060
2061                         # Prepend the number to the heading text
2062
2063                         if( $doNumberHeadings || $doShowToc ) {
2064                                 $tocline = $numbering . ' ' . $tocline;
2065
2066                                 # Don't number the heading if it is the only one (looks silly)
2067                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2068                                         # the two are different if the line contains a link
2069                                         $headline=$numbering . ' ' . $headline;
2070                                 }
2071                         }
2072
2073                         # Create the anchor for linking from the TOC to the section
2074                         $anchor = $canonized_headline;
2075                         if($refcount[$headlineCount] > 1 ) {
2076                                 $anchor .= '_' . $refcount[$headlineCount];
2077                         }
2078                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2079                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2080                         }
2081                         if( $showEditLink ) {
2082                                 if ( empty( $head[$headlineCount] ) ) {
2083                                         $head[$headlineCount] = '';
2084                                 }
2085                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
2086                         }
2087
2088                         # Add the edit section span
2089                         if( $rightClickHack ) {
2090                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
2091                         }
2092
2093                         # give headline the correct <h#> tag
2094                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
2095
2096                         $headlineCount++;
2097                 }
2098
2099                 if( $doShowToc ) {
2100                         $toclines = $headlineCount;
2101                         $toc .= $sk->tocUnindent( $toclevel );
2102                         $toc = $sk->tocTable( $toc );
2103                 }
2104
2105                 # split up and insert constructed headlines
2106
2107                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2108                 $i = 0;
2109
2110                 foreach( $blocks as $block ) {
2111                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2112                             # This is the [edit] link that appears for the top block of text when
2113                                 # section editing is enabled
2114
2115                                 # Disabled because it broke block formatting
2116                                 # For example, a bullet point in the top line
2117                                 # $full .= $sk->editSectionLink(0);
2118                         }
2119                         $full .= $block;
2120                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2121                         # Top anchor now in skin
2122                                 $full = $full.$toc;
2123                         }
2124
2125                         if( !empty( $head[$i] ) ) {
2126                                 $full .= $head[$i];
2127                         }
2128                         $i++;
2129                 }
2130                 if($forceTocHere) {
2131                         $mw =& MagicWord::get( MAG_TOC );
2132                         return $mw->replace( $toc, $full );
2133                 } else {
2134                         return $full;
2135                 }
2136         }
2137
2138         # Return an HTML link for the "ISBN 123456" text
2139         /* private */ function magicISBN( $text ) {
2140                 global $wgLang;
2141                 $fname = 'Parser::magicISBN';
2142                 wfProfileIn( $fname );
2143
2144                 $a = split( 'ISBN ', " $text" );
2145                 if ( count ( $a ) < 2 ) {
2146                         wfProfileOut( $fname );
2147                         return $text;
2148                 }
2149                 $text = substr( array_shift( $a ), 1);
2150                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2151
2152                 foreach ( $a as $x ) {
2153                         $isbn = $blank = '' ;
2154                         while ( ' ' == $x{0} ) {
2155                                 $blank .= ' ';
2156                                 $x = substr( $x, 1 );
2157                         }
2158                         while ( strstr( $valid, $x{0} ) != false ) {
2159                                 $isbn .= $x{0};
2160                                 $x = substr( $x, 1 );
2161                         }
2162                         $num = str_replace( '-', '', $isbn );
2163                         $num = str_replace( ' ', '', $num );
2164
2165                         if ( '' == $num ) {
2166                                 $text .= "ISBN $blank$x";
2167                         } else {
2168                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2169                                 $text .= '<a href="' .
2170                                 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
2171                                         "\" class=\"internal\">ISBN $isbn</a>";
2172                                 $text .= $x;
2173                         }
2174                 }
2175                 wfProfileOut( $fname );
2176                 return $text;
2177         }
2178
2179         # Return an HTML link for the "GEO ..." text
2180         /* private */ function magicGEO( $text ) {
2181                 global $wgLang, $wgUseGeoMode;
2182                 if ( !isset ( $wgUseGeoMode ) || !$wgUseGeoMode ) return $text ;
2183                 $fname = 'Parser::magicGEO';
2184                 wfProfileIn( $fname );
2185
2186                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2187                 $directions = array ( "N" => "North" , "S" => "South" , "E" => "East" , "W" => "West" ) ;
2188                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2189                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2190                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2191                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2192
2193                 $a = split( 'GEO ', " $text" );
2194                 if ( count ( $a ) < 2 ) {
2195                         wfProfileOut( $fname );
2196                         return $text;
2197                 }
2198                 $text = substr( array_shift( $a ), 1);
2199                 $valid = '0123456789.+-:';
2200
2201                 foreach ( $a as $x ) {
2202                         $geo = $blank = '' ;
2203                         while ( ' ' == $x{0} ) {
2204                                 $blank .= ' ';
2205                                 $x = substr( $x, 1 );
2206                         }
2207                         while ( strstr( $valid, $x{0} ) != false ) {
2208                                 $geo .= $x{0};
2209                                 $x = substr( $x, 1 );
2210                         }
2211                         $num = str_replace( '+', '', $geo );
2212                         $num = str_replace( ' ', '', $num );
2213
2214                         if ( '' == $num || count ( explode ( ":" , $num , 3 ) ) < 2 ) {
2215                                 $text .= "GEO $blank$x";
2216                         } else {
2217                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2218                                 $text .= '<a href="' .
2219                                 $titleObj->escapeLocalUrl( "coordinates={$num}" ) .
2220                                         "\" class=\"internal\">GEO $geo</a>";
2221                                 $text .= $x;
2222                         }
2223                 }
2224                 wfProfileOut( $fname );
2225                 return $text;
2226         }
2227
2228         # Return an HTML link for the "RFC 1234" text
2229         /* private */ function magicRFC( $text ) {
2230                 global $wgLang;
2231
2232                 $a = split( 'RFC ', ' '.$text );
2233                 if ( count ( $a ) < 2 ) return $text;
2234                 $text = substr( array_shift( $a ), 1);
2235                 $valid = '0123456789';
2236
2237                 foreach ( $a as $x ) {
2238                         $rfc = $blank = '' ;
2239                         while ( ' ' == $x{0} ) {
2240                                 $blank .= ' ';
2241                                 $x = substr( $x, 1 );
2242                         }
2243                         while ( strstr( $valid, $x{0} ) != false ) {
2244                                 $rfc .= $x{0};
2245                                 $x = substr( $x, 1 );
2246                         }
2247
2248                         if ( '' == $rfc ) {
2249                                 $text .= "RFC $blank$x";
2250                         } else {
2251                                 $url = wfmsg( 'rfcurl' );
2252                                 $url = str_replace( '$1', $rfc, $url);
2253                                 $sk =& $this->mOptions->getSkin();
2254                                 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
2255                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2256                         }
2257                 }
2258                 return $text;
2259         }
2260
2261         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2262                 $this->mOptions = $options;
2263                 $this->mTitle =& $title;
2264                 $this->mOutputType = OT_WIKI;
2265
2266                 if ( $clearState ) {
2267                         $this->clearState();
2268                 }
2269
2270                 $stripState = false;
2271                 $pairs = array(
2272                         "\r\n" => "\n",
2273                         );
2274                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2275                 // now with regexes
2276                 /*
2277                 $pairs = array(
2278                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2279                         "/<br *?>/i" => "<br />",
2280                 );
2281                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2282                 */
2283                 $text = $this->strip( $text, $stripState, false );
2284                 $text = $this->pstPass2( $text, $user );
2285                 $text = $this->unstrip( $text, $stripState );
2286                 $text = $this->unstripNoWiki( $text, $stripState );
2287                 return $text;
2288         }
2289
2290         /* private */ function pstPass2( $text, &$user ) {
2291                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2292
2293                 # Variable replacement
2294                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2295                 $text = $this->replaceVariables( $text );
2296
2297                 # Signatures
2298                 #
2299                 $n = $user->getName();
2300                 $k = $user->getOption( 'nickname' );
2301                 if ( '' == $k ) { $k = $n; }
2302                 if(isset($wgLocaltimezone)) {
2303                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2304                 }
2305                 /* Note: this is an ugly timezone hack for the European wikis */
2306                 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2307                   ' (' . date( 'T' ) . ')';
2308                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2309
2310                 $text = preg_replace( '/~~~~~/', $d, $text );
2311                 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText(
2312                   Namespace::getUser() ) . ":$n|$k]] $d", $text );
2313                 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText(
2314                   Namespace::getUser() ) . ":$n|$k]]", $text );
2315
2316                 # Context links: [[|name]] and [[name (context)|]]
2317                 #
2318                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2319                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2320                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2321                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2322
2323                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2324                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2325                 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/";          # [[namespace:page|]]
2326                 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2327                                                                                                                 # [[ns:page (cont)|]]
2328                 $context = "";
2329                 $t = $this->mTitle->getText();
2330                 if ( preg_match( $conpat, $t, $m ) ) {
2331                         $context = $m[2];
2332                 }
2333                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2334                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2335                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2336
2337                 if ( '' == $context ) {
2338                         $text = preg_replace( $p2, '[[\\1]]', $text );
2339                 } else {
2340                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2341                 }
2342
2343                 /*
2344                 $mw =& MagicWord::get( MAG_SUBST );
2345                 $wgCurParser = $this->fork();
2346                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2347                 $this->merge( $wgCurParser );
2348                 */
2349
2350                 # Trim trailing whitespace
2351                 # MAG_END (__END__) tag allows for trailing
2352                 # whitespace to be deliberately included
2353                 $text = rtrim( $text );
2354                 $mw =& MagicWord::get( MAG_END );
2355                 $mw->matchAndRemove( $text );
2356
2357                 return $text;
2358         }
2359
2360         # Set up some variables which are usually set up in parse()
2361         # so that an external function can call some class members with confidence
2362         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2363                 $this->mTitle =& $title;
2364                 $this->mOptions = $options;
2365                 $this->mOutputType = $outputType;
2366                 if ( $clearState ) {
2367                         $this->clearState();
2368                 }
2369         }
2370
2371         function transformMsg( $text, $options ) {
2372                 global $wgTitle;
2373                 static $executing = false;
2374
2375                 # Guard against infinite recursion
2376                 if ( $executing ) {
2377                         return $text;
2378                 }
2379                 $executing = true;
2380
2381                 $this->mTitle = $wgTitle;
2382                 $this->mOptions = $options;
2383                 $this->mOutputType = OT_MSG;
2384                 $this->clearState();
2385                 $text = $this->replaceVariables( $text );
2386
2387                 $executing = false;
2388                 return $text;
2389         }
2390
2391         # Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2392         # Callback will be called with the text within
2393         # Transform and return the text within
2394         function setHook( $tag, $callback ) {
2395                 $oldVal = @$this->mTagHooks[$tag];
2396                 $this->mTagHooks[$tag] = $callback;
2397                 return $oldVal;
2398         }
2399 }
2400
2401 class ParserOutput
2402 {
2403         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2404         var $mCacheTime; # Used in ParserCache
2405
2406         function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2407                 $containsOldMagic = false )
2408         {
2409                 $this->mText = $text;
2410                 $this->mLanguageLinks = $languageLinks;
2411                 $this->mCategoryLinks = $categoryLinks;
2412                 $this->mContainsOldMagic = $containsOldMagic;
2413                 $this->mCacheTime = "";
2414         }
2415
2416         function getText() { return $this->mText; }
2417         function getLanguageLinks() { return $this->mLanguageLinks; }
2418         function getCategoryLinks() { return $this->mCategoryLinks; }
2419         function getCacheTime() { return $this->mCacheTime; }
2420         function containsOldMagic() { return $this->mContainsOldMagic; }
2421         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2422         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2423         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2424         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2425         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2426
2427         function merge( $other ) {
2428                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2429                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2430                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2431         }
2432
2433 }
2434
2435 class ParserOptions
2436 {
2437         # All variables are private
2438         var $mUseTeX;                    # Use texvc to expand <math> tags
2439         var $mUseCategoryMagic;          # Treat [[Category:xxxx]] tags specially
2440         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2441         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2442         var $mAllowExternalImages;       # Allow external images inline
2443         var $mSkin;                      # Reference to the preferred skin
2444         var $mDateFormat;                # Date format index
2445         var $mEditSection;               # Create "edit section" links
2446         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2447         var $mNumberHeadings;            # Automatically number headings
2448         var $mShowToc;                   # Show table of contents
2449
2450         function getUseTeX()                        { return $this->mUseTeX; }
2451         function getUseCategoryMagic()              { return $this->mUseCategoryMagic; }
2452         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2453         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2454         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2455         function getSkin()                          { return $this->mSkin; }
2456         function getDateFormat()                    { return $this->mDateFormat; }
2457         function getEditSection()                   { return $this->mEditSection; }
2458         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
2459         function getNumberHeadings()                { return $this->mNumberHeadings; }
2460         function getShowToc()                       { return $this->mShowToc; }
2461
2462         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
2463         function setUseCategoryMagic( $x )          { return wfSetVar( $this->mUseCategoryMagic, $x ); }
2464         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
2465         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
2466         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
2467         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
2468         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
2469         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2470         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
2471         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
2472
2473     function setSkin( &$x ) { $this->mSkin =& $x; }
2474
2475         /* static */ function newFromUser( &$user ) {
2476                 $popts = new ParserOptions;
2477                 $popts->initialiseFromUser( $user );
2478                 return $popts;
2479         }
2480
2481         function initialiseFromUser( &$userInput ) {
2482                 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2483
2484                 if ( !$userInput ) {
2485                         $user = new User;
2486                         $user->setLoaded( true );
2487                 } else {
2488                         $user =& $userInput;
2489                 }
2490
2491                 $this->mUseTeX = $wgUseTeX;
2492                 $this->mUseCategoryMagic = $wgUseCategoryMagic;
2493                 $this->mUseDynamicDates = $wgUseDynamicDates;
2494                 $this->mInterwikiMagic = $wgInterwikiMagic;
2495                 $this->mAllowExternalImages = $wgAllowExternalImages;
2496                 $this->mSkin =& $user->getSkin();
2497                 $this->mDateFormat = $user->getOption( 'date' );
2498                 $this->mEditSection = $user->getOption( 'editsection' );
2499                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2500                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2501                 $this->mShowToc = $user->getOption( 'showtoc' );
2502         }
2503
2504
2505 }
2506
2507 # Regex callbacks, used in Parser::replaceVariables
2508 function wfBraceSubstitution( $matches )
2509 {
2510         global $wgCurParser;
2511         return $wgCurParser->braceSubstitution( $matches );
2512 }
2513
2514 function wfArgSubstitution( $matches )
2515 {
2516         global $wgCurParser;
2517         return $wgCurParser->argSubstitution( $matches );
2518 }
2519
2520 function wfVariableSubstitution( $matches )
2521 {
2522         global $wgCurParser;
2523         return $wgCurParser->variableSubstitution( $matches );
2524 }
2525
2526 ?>