includes/Parser.php

   1 <?php
   2
   3 require_once('Tokenizer.php');
   4
   5 if( $GLOBALS['wgUseWikiHiero'] ){
   6         require_once('extensions/wikihiero/wikihiero.php');
   7 }
   8 if( $GLOBALS['wgUseTimeline'] ){
   9         require_once('extensions/timeline/Timeline.php');
  10 }
  11
  12 # PHP Parser
  13 #
  14 # Processes wiki markup
  15 #
  16 # There are two main entry points into the Parser class: parse() and preSaveTransform().
  17 # The parse() function produces HTML output, preSaveTransform() produces altered wiki markup.
  18 #
  19 # Globals used:
  20 #    objects:   $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
  21 #
  22 # NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  23 #
  24 #    settings:  $wgUseTex*, $wgUseCategoryMagic*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  25 #               $wgNamespacesWithSubpages, $wgLanguageCode, $wgAllowExternalImages*,
  26 #               $wgLocaltimezone
  27 #
  28 #      * only within ParserOptions
  29 #
  30 #
  31 #----------------------------------------
  32 #    Variable substitution O(N^2) attack
  33 #-----------------------------------------
  34 # Without countermeasures, it would be possible to attack the parser by saving a page
  35 # filled with a large number of inclusions of large pages. The size of the generated
  36 # page would be proportional to the square of the input size. Hence, we limit the number
  37 # of inclusions of any given page, thus bringing any attack back to O(N).
  38 #
  39
  40 define( "MAX_INCLUDE_REPEAT", 5 );
  41
  42 # Allowed values for $mOutputType
  43 define( "OT_HTML", 1 );
  44 define( "OT_WIKI", 2 );
  45 define( "OT_MSG", 3 );
  46
  47 # string parameter for extractTags which will cause it
  48 # to strip HTML comments in addition to regular
  49 # <XML>-style tags. This should not be anything we
  50 # may want to use in wikisyntax
  51 define( "STRIP_COMMENTS", "HTMLCommentStrip" );
  52
  53 # prefix for escaping, used in two functions at least
  54 define( "UNIQ_PREFIX", "NaodW29");
  55
  56 class Parser
  57 {
  58         # Cleared with clearState():
  59         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  60         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  61
  62         # Temporary:
  63         var $mOptions, $mTitle, $mOutputType;
  64
  65         function Parser()
  66         {
  67                 $this->clearState();
  68         }
  69
  70         function clearState()
  71         {
  72                 $this->mOutput = new ParserOutput;
  73                 $this->mAutonumber = 0;
  74                 $this->mLastSection = "";
  75                 $this->mDTopen = false;
  76                 $this->mVariables = false;
  77                 $this->mIncludeCount = array();
  78                 $this->mStripState = array();
  79                 $this->mArgStack = array();
  80         }
  81
  82         # First pass--just handle <nowiki> sections, pass the rest off
  83         # to internalParse() which does all the real work.
  84         #
  85         # Returns a ParserOutput
  86         #
  87         function parse( $text, &$title, $options, $linestart = true, $clearState = true )
  88         {
  89                 global $wgUseTidy;
  90                 $fname = "Parser::parse";
  91                 wfProfileIn( $fname );
  92
  93                 if ( $clearState ) {
  94                         $this->clearState();
  95                 }
  96
  97                 $this->mOptions = $options;
  98                 $this->mTitle =& $title;
  99                 $this->mOutputType = OT_HTML;
 100
 101                 $stripState = NULL;
 102                 $text = $this->strip( $text, $this->mStripState );
 103                 $text = $this->internalParse( $text, $linestart );
 104                 $text = $this->unstrip( $text, $this->mStripState );
 105                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 106                 if(!$wgUseTidy) {
 107                         $fixtags = array(
 108                                 "/<hr *>/i" => '<hr/>',
 109                                 "/<br *>/i" => '<br/>',
 110                                 "/<center *>/i"=>'<div class="center">',
 111                                 "/<\\/center *>/i" => '</div>',
 112                                 # Clean up spare ampersands; note that we probably ought to be
 113                                 # more careful about named entities.
 114                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 115                         );
 116                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 117                 } else {
 118                         $fixtags = array(
 119                                 "/<center *>/i"=>'<div class="center">',
 120                                 "/<\\/center *>/i" => '</div>'
 121                         );
 122                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 123                 }
 124                 # only once and last
 125                 $text = $this->doBlockLevels( $text, $linestart );
 126                 if($wgUseTidy) {
 127                         $text = $this->tidy($text);
 128                 }
 129                 $this->mOutput->setText( $text );
 130                 wfProfileOut( $fname );
 131                 return $this->mOutput;
 132         }
 133
 134         /* static */ function getRandomString()
 135         {
 136                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 137         }
 138
 139         # Replaces all occurrences of <$tag>content</$tag> in the text
 140         # with a random marker and returns the new text. the output parameter
 141         # $content will be an associative array filled with data on the form
 142         # $unique_marker => content.
 143
 144         # If $content is already set, the additional entries will be appended
 145
 146         # If $tag is set to STRIP_COMMENTS, the function will extract
 147         # <!-- HTML comments -->
 148
 149         /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ""){
 150                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 151                 if ( !$content ) {
 152                         $content = array( );
 153                 }
 154                 $n = 1;
 155                 $stripped = "";
 156
 157                 while ( "" != $text ) {
 158                         if($tag==STRIP_COMMENTS) {
 159                                 $p = preg_split( "/<!--/i", $text, 2 );
 160                         } else {
 161                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 162                         }
 163                         $stripped .= $p[0];
 164                         if ( ( count( $p ) < 2 ) || ( "" == $p[1] ) ) {
 165                                 $text = "";
 166                         } else {
 167                                 if($tag==STRIP_COMMENTS) {
 168                                         $q = preg_split( "/-->/i", $p[1], 2 );
 169                                 } else {
 170                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 171                                 }
 172                                 $marker = $rnd . sprintf("%08X", $n++);
 173                                 $content[$marker] = $q[0];
 174                                 $stripped .= $marker;
 175                                 $text = $q[1];
 176                         }
 177                 }
 178                 return $stripped;
 179         }
 180
 181         # Strips and renders <nowiki>, <pre>, <math>, <hiero>
 182         # If $render is set, performs necessary rendering operations on plugins
 183         # Returns the text, and fills an array with data needed in unstrip()
 184         # If the $state is already a valid strip state, it adds to the state
 185
 186         # When $stripcomments is set, HTML comments <!-- like this -->
 187         # will be stripped in addition to other tags. This is important
 188         # for section editing, where these comments cause confusion when
 189         # counting the sections in the wikisource
 190         function strip( $text, &$state, $stripcomments = false )
 191         {
 192                 $render = ($this->mOutputType == OT_HTML);
 193                 $nowiki_content = array();
 194                 $hiero_content = array();
 195                 $math_content = array();
 196                 $pre_content = array();
 197                 $comment_content = array();
 198
 199                 # Replace any instances of the placeholders
 200                 $uniq_prefix = UNIQ_PREFIX;
 201                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 202
 203                 $text = Parser::extractTags("nowiki", $text, $nowiki_content, $uniq_prefix);
 204                 foreach( $nowiki_content as $marker => $content ){
 205                         if( $render ){
 206                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 207                         } else {
 208                                 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
 209                         }
 210                 }
 211
 212                 $text = Parser::extractTags("hiero", $text, $hiero_content, $uniq_prefix);
 213                 foreach( $hiero_content as $marker => $content ){
 214                         if( $render && $GLOBALS['wgUseWikiHiero']){
 215                                 $hiero_content[$marker] = WikiHiero( $content, WH_MODE_HTML);
 216                         } else {
 217                                 $hiero_content[$marker] = "<hiero>$content</hiero>";
 218                         }
 219                 }
 220
 221                 $text = Parser::extractTags("math", $text, $math_content, $uniq_prefix);
 222                 foreach( $math_content as $marker => $content ){
 223                         if( $render ) {
 224                                 if( $this->mOptions->getUseTeX() ) {
 225                                         $math_content[$marker] = renderMath( $content );
 226                                 } else {
 227                                         $math_content[$marker] = "&lt;math&gt;$content&lt;math&gt;";
 228                                 }
 229                         } else {
 230                                 $math_content[$marker] = "<math>$content</math>";
 231                         }
 232                 }
 233
 234                 $text = Parser::extractTags("pre", $text, $pre_content, $uniq_prefix);
 235                 foreach( $pre_content as $marker => $content ){
 236                         if( $render ){
 237                                 $pre_content[$marker] = "<pre>" . wfEscapeHTMLTagsOnly( $content ) . "</pre>";
 238                         } else {
 239                                 $pre_content[$marker] = "<pre>$content</pre>";
 240                         }
 241                 }
 242                 if($stripcomments) {
 243                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 244                         foreach( $comment_content as $marker => $content ){
 245                                 $comment_content[$marker] = "<!--$content-->";
 246                         }
 247                 }
 248
 249                 # Merge state with the pre-existing state, if there is one
 250                 if ( $state ) {
 251                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 252                         $state['hiero'] = $state['hiero'] + $hiero_content;
 253                         $state['math'] = $state['math'] + $math_content;
 254                         $state['pre'] = $state['pre'] + $pre_content;
 255                         $state['comment'] = $state['comment'] + $comment_content;
 256                 } else {
 257                         $state = array(
 258                           'nowiki' => $nowiki_content,
 259                           'hiero' => $hiero_content,
 260                           'math' => $math_content,
 261                           'pre' => $pre_content,
 262                           'comment' => $comment_content
 263                         );
 264                 }
 265                 return $text;
 266         }
 267
 268         function unstrip( $text, &$state )
 269         {
 270                 # Must expand in reverse order, otherwise nested tags will be corrupted
 271                 $contentDict = end( $state );
 272                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 273                         for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 274                                 $text = str_replace( key( $contentDict ), $content, $text );
 275                         }
 276                 }
 277
 278                 return $text;
 279         }
 280
 281         # Add an item to the strip state
 282         # Returns the unique tag which must be inserted into the stripped text
 283         # The tag will be replaced with the original text in unstrip()
 284
 285         function insertStripItem( $text, &$state )
 286         {
 287                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 288                 if ( !$state ) {
 289                         $state = array(
 290                           'nowiki' => array(),
 291                           'hiero' => array(),
 292                           'math' => array(),
 293                           'pre' => array()
 294                         );
 295                 }
 296                 $state['item'][$rnd] = $text;
 297                 return $rnd;
 298         }
 299
 300         # This method generates the list of subcategories and pages for a category
 301         function categoryMagic ()
 302         {
 303                 global $wgLang , $wgUser ;
 304                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 305
 306                 $cns = Namespace::getCategory() ;
 307                 if ( $this->mTitle->getNamespace() != $cns ) return "" ; # This ain't a category page
 308
 309                 $r = "<br style=\"clear:both;\"/>\n";
 310
 311
 312                 $sk =& $wgUser->getSkin() ;
 313
 314                 $articles = array() ;
 315                 $children = array() ;
 316                 $data = array () ;
 317                 $id = $this->mTitle->getArticleID() ;
 318
 319                 # FIXME: add limits
 320                 $t = wfStrencode( $this->mTitle->getDBKey() );
 321                 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM cur,categorylinks WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 322                 $res = wfQuery ( $sql, DB_READ ) ;
 323                 while ( $x = wfFetchObject ( $res ) ) $data[] = $x ;
 324
 325                 # For all pages that link to this category
 326                 foreach ( $data AS $x )
 327                 {
 328                         $t = $wgLang->getNsText ( $x->cur_namespace ) ;
 329                         if ( $t != "" ) $t .= ":" ;
 330                         $t .= $x->cur_title ;
 331
 332                         if ( $x->cur_namespace == $cns ) {
 333                                 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
 334                         } else {
 335                                 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
 336                         }
 337                 }
 338                 wfFreeResult ( $res ) ;
 339
 340                 # Showing subcategories
 341                 if ( count ( $children ) > 0 ) {
 342                         $r .= "<h2>".wfMsg("subcategories")."</h2>\n" ;
 343                         $r .= implode ( ", " , $children ) ;
 344                 }
 345
 346                 # Showing pages in this category
 347                 if ( count ( $articles ) > 0 ) {
 348                         $ti = $this->mTitle->getText() ;
 349                         $h =  wfMsg( "category_header", $ti );
 350                         $r .= "<h2>{$h}</h2>\n" ;
 351                         $r .= implode ( ", " , $articles ) ;
 352                 }
 353
 354
 355                 return $r ;
 356         }
 357
 358         function getHTMLattrs ()
 359         {
 360                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 361                                 "title", "align", "lang", "dir", "width", "height",
 362                                 "bgcolor", "clear", /* BR */ "noshade", /* HR */
 363                                 "cite", /* BLOCKQUOTE, Q */ "size", "face", "color",
 364                                 /* FONT */ "type", "start", "value", "compact",
 365                                 /* For various lists, mostly deprecated but safe */
 366                                 "summary", "width", "border", "frame", "rules",
 367                                 "cellspacing", "cellpadding", "valign", "char",
 368                                 "charoff", "colgroup", "col", "span", "abbr", "axis",
 369                                 "headers", "scope", "rowspan", "colspan", /* Tables */
 370                                 "id", "class", "name", "style" /* For CSS */
 371                                 );
 372                 return $htmlattrs ;
 373         }
 374
 375         function fixTagAttributes ( $t )
 376         {
 377                 if ( trim ( $t ) == "" ) return "" ; # Saves runtime ;-)
 378                 $htmlattrs = $this->getHTMLattrs() ;
 379
 380                 # Strip non-approved attributes from the tag
 381                 $t = preg_replace(
 382                         "/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e",
 383                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 384                         $t);
 385                 # Strip javascript "expression" from stylesheets. Brute force approach:
 386                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 387
 388                 if( preg_match(
 389                         "/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is",
 390                         wfMungeToUtf8( $t ) ) )
 391                 {
 392                         $t="";
 393                 }
 394
 395                 return trim ( $t ) ;
 396         }
 397
 398         /* interface with html tidy, used if $wgUseTidy = true */
 399         function tidy ( $text ) {
 400                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 401                 global $wgInputEncoding, $wgOutputEncoding;
 402                 $cleansource = '';
 403                 switch(strtoupper($wgOutputEncoding)) {
 404                         case 'ISO-8859-1':
 405                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 406                                 break;
 407                         case 'UTF-8':
 408                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 409                                 break;
 410                         default:
 411                                 $wgTidyOpts .= ' -raw';
 412                         }
 413
 414                 $text = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 415 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 416 '<head><title>test</title></head><body>'.$text.'</body></html>';
 417                 $descriptorspec = array(
 418                         0 => array("pipe", "r"),
 419                         1 => array("pipe", "w"),
 420                         2 => array("file", "/dev/null", "a")
 421                 );
 422                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 423                 if (is_resource($process)) {
 424                         fwrite($pipes[0], $text);
 425                         fclose($pipes[0]);
 426                         while (!feof($pipes[1])) {
 427                                 $cleansource .= fgets($pipes[1], 1024);
 428                         }
 429                         fclose($pipes[1]);
 430                         $return_value = proc_close($process);
 431                 }
 432                 if( $cleansource == '' && $text != '') {
 433                         return '<h2>'.wfMsg('seriousxhtmlerrors').'</h2><pre>'.htmlspecialchars($text).'</pre>';
 434                 } else {
 435                         return $cleansource;
 436                 }
 437         }
 438
 439         function doTableStuff ( $t )
 440         {
 441                 $t = explode ( "\n" , $t ) ;
 442                 $td = array () ; # Is currently a td tag open?
 443                         $ltd = array () ; # Was it TD or TH?
 444                         $tr = array () ; # Is currently a tr tag open?
 445                         $ltr = array () ; # tr attributes
 446                         foreach ( $t AS $k => $x )
 447                         {
 448                                 $x = trim ( $x ) ;
 449                                 $fc = substr ( $x , 0 , 1 ) ;
 450                                 if ( "{|" == substr ( $x , 0 , 2 ) )
 451                                 {
 452                                         $t[$k] = "\n<table " . $this->fixTagAttributes ( substr ( $x , 3 ) ) . ">" ;
 453                                         array_push ( $td , false ) ;
 454                                         array_push ( $ltd , "" ) ;
 455                                         array_push ( $tr , false ) ;
 456                                         array_push ( $ltr , "" ) ;
 457                                 }
 458                                 else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 459                                 else if ( "|}" == substr ( $x , 0 , 2 ) )
 460                                 {
 461                                         $z = "</table>\n" ;
 462                                         $l = array_pop ( $ltd ) ;
 463                                         if ( array_pop ( $tr ) ) $z = "</tr>" . $z ;
 464                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 465                                         array_pop ( $ltr ) ;
 466                                         $t[$k] = $z ;
 467                                 }
 468                                 /*      else if ( "|_" == substr ( $x , 0 , 2 ) ) # Caption
 469                                                 {
 470                                                 $z = trim ( substr ( $x , 2 ) ) ;
 471                                                 $t[$k] = "<caption>{$z}</caption>\n" ;
 472                                                 }*/
 473                                 else if ( "|-" == substr ( $x , 0 , 2 ) ) # Allows for |---------------
 474                                 {
 475                                         $x = substr ( $x , 1 ) ;
 476                                         while ( $x != "" && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 477                                         $z = "" ;
 478                                         $l = array_pop ( $ltd ) ;
 479                                         if ( array_pop ( $tr ) ) $z = "</tr>" . $z ;
 480                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 481                                         array_pop ( $ltr ) ;
 482                                         $t[$k] = $z ;
 483                                         array_push ( $tr , false ) ;
 484                                         array_push ( $td , false ) ;
 485                                         array_push ( $ltd , "" ) ;
 486                                         array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 487                                 }
 488                                 else if ( "|" == $fc || "!" == $fc || "|+" == substr ( $x , 0 , 2 ) ) # Caption
 489                                 {
 490                                         if ( "|+" == substr ( $x , 0 , 2 ) )
 491                                         {
 492                                                 $fc = "+" ;
 493                                                 $x = substr ( $x , 1 ) ;
 494                                         }
 495                                         $after = substr ( $x , 1 ) ;
 496                                         if ( $fc == "!" ) $after = str_replace ( "!!" , "||" , $after ) ;
 497                                         $after = explode ( "||" , $after ) ;
 498                                         $t[$k] = "" ;
 499                                         foreach ( $after AS $theline )
 500                                         {
 501                                                 $z = "" ;
 502                                                 if ( $fc != "+" )
 503                                                 {
 504                                                         $tra = array_pop ( $ltr ) ;
 505                                                         if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
 506                                                         array_push ( $tr , true ) ;
 507                                                         array_push ( $ltr , "" ) ;
 508                                                 }
 509
 510                                                 $l = array_pop ( $ltd ) ;
 511                                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 512                                                 if ( $fc == "|" ) $l = "td" ;
 513                                                 else if ( $fc == "!" ) $l = "th" ;
 514                                                 else if ( $fc == "+" ) $l = "caption" ;
 515                                                 else $l = "" ;
 516                                                 array_push ( $ltd , $l ) ;
 517                                                 $y = explode ( "|" , $theline , 2 ) ;
 518                                                 if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 519                                                 else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 520                                                 $t[$k] .= $y ;
 521                                                 array_push ( $td , true ) ;
 522                                         }
 523                                 }
 524                         }
 525
 526                 # Closing open td, tr && table
 527                 while ( count ( $td ) > 0 )
 528                 {
 529                         if ( array_pop ( $td ) ) $t[] = "</td>" ;
 530                         if ( array_pop ( $tr ) ) $t[] = "</tr>" ;
 531                         $t[] = "</table>" ;
 532                 }
 533
 534                 $t = implode ( "\n" , $t ) ;
 535                 #               $t = $this->removeHTMLtags( $t );
 536                 return $t ;
 537         }
 538
 539         # Parses the text and adds the result to the strip state
 540         # Returns the strip tag
 541         function stripParse( $text, $linestart, $args )
 542         {
 543                 $text = $this->strip( $text, $this->mStripState );
 544                 $text = $this->internalParse( $text, $linestart, $args, false );
 545                 if( $linestart ) {
 546                         $text = "\n" . $text;
 547                 }
 548                 return $this->insertStripItem( $text, $this->mStripState );
 549         }
 550
 551         function internalParse( $text, $linestart, $args = array(), $isMain=true )
 552         {
 553                 $fname = "Parser::internalParse";
 554                 wfProfileIn( $fname );
 555
 556                 $text = $this->removeHTMLtags( $text );
 557                 $text = $this->replaceVariables( $text, $args );
 558
 559                 # $text = preg_replace( "/(^|\n)-----*/", "\\1<hr>", $text );
 560
 561                 $text = $this->doHeadings( $text );
 562                 if($this->mOptions->getUseDynamicDates()) {
 563                         global $wgDateFormatter;
 564                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 565                 }
 566                 $text = $this->replaceExternalLinks( $text );
 567                 $text = $this->doTokenizedParser ( $text );
 568                 $text = $this->doTableStuff ( $text ) ;
 569                 $text = $this->formatHeadings( $text, $isMain );
 570                 $sk =& $this->mOptions->getSkin();
 571                 $text = $sk->transformContent( $text );
 572
 573                 if ( !isset ( $this->categoryMagicDone ) ) {
 574                         $text .= $this->categoryMagic () ;
 575                         $this->categoryMagicDone = true ;
 576                 }
 577
 578                 wfProfileOut( $fname );
 579                 return $text;
 580         }
 581
 582
 583         /* private */ function doHeadings( $text )
 584         {
 585                 for ( $i = 6; $i >= 1; --$i ) {
 586                         $h = substr( "======", 0, $i );
 587                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 588                           "<h{$i}>\\1</h{$i}>\\2", $text );
 589                 }
 590                 return $text;
 591         }
 592
 593         # Note: we have to do external links before the internal ones,
 594         # and otherwise take great care in the order of things here, so
 595         # that we don't end up interpreting some URLs twice.
 596
 597         /* private */ function replaceExternalLinks( $text )
 598         {
 599                 $fname = "Parser::replaceExternalLinks";
 600                 wfProfileIn( $fname );
 601                 $text = $this->subReplaceExternalLinks( $text, "http", true );
 602                 $text = $this->subReplaceExternalLinks( $text, "https", true );
 603                 $text = $this->subReplaceExternalLinks( $text, "ftp", false );
 604                 $text = $this->subReplaceExternalLinks( $text, "irc", false );
 605                 $text = $this->subReplaceExternalLinks( $text, "gopher", false );
 606                 $text = $this->subReplaceExternalLinks( $text, "news", false );
 607                 $text = $this->subReplaceExternalLinks( $text, "mailto", false );
 608                 wfProfileOut( $fname );
 609                 return $text;
 610         }
 611
 612         /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber )
 613         {
 614                 $unique = "4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3";
 615                 $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
 616
 617                 # this is  the list of separators that should be ignored if they
 618                 # are the last character of an URL but that should be included
 619                 # if they occur within the URL, e.g. "go to www.foo.com, where .."
 620                 # in this case, the last comma should not become part of the URL,
 621                 # but in "www.foo.com/123,2342,32.htm" it should.
 622                 $sep = ",;\.:";
 623                 $fnc = "A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF";
 624                 $images = "gif|png|jpg|jpeg";
 625
 626                 # PLEASE NOTE: The curly braces { } are not part of the regex,
 627                 # they are interpreted as part of the string (used to tell PHP
 628                 # that the content of the string should be inserted there).
 629                 $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
 630                   "((?i){$images})([^{$uc}]|$)/";
 631
 632                 $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
 633                 $sk =& $this->mOptions->getSkin();
 634
 635                 if ( $autonumber and $this->mOptions->getAllowExternalImages() ) { # Use img tags only for HTTP urls
 636                         $s = preg_replace( $e1, "\\1" . $sk->makeImage( "{$unique}:\\3" .
 637                           "/\\4.\\5", "\\4.\\5" ) . "\\6", $s );
 638                 }
 639                 $s = preg_replace( $e2, "\\1" . "<a href=\"{$unique}:\\3\"" .
 640                   $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
 641                   "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
 642                   "</a>\\5", $s );
 643                 $s = str_replace( $unique, $protocol, $s );
 644
 645                 $a = explode( "[{$protocol}:", " " . $s );
 646                 $s = array_shift( $a );
 647                 $s = substr( $s, 1 );
 648
 649                 $e1 = "/^([{$uc}"."{$sep}]+)](.*)\$/sD";
 650                 $e2 = "/^([{$uc}"."{$sep}]+)\\s+([^\\]]+)](.*)\$/sD";
 651
 652                 foreach ( $a as $line ) {
 653                         if ( preg_match( $e1, $line, $m ) ) {
 654                                 $link = "{$protocol}:{$m[1]}";
 655                                 $trail = $m[2];
 656                                 if ( $autonumber ) { $text = "[" . ++$this->mAutonumber . "]"; }
 657                                 else { $text = wfEscapeHTML( $link ); }
 658                         } else if ( preg_match( $e2, $line, $m ) ) {
 659                                 $link = "{$protocol}:{$m[1]}";
 660                                 $text = $m[2];
 661                                 $trail = $m[3];
 662                         } else {
 663                                 $s .= "[{$protocol}:" . $line;
 664                                 continue;
 665                         }
 666                         if( $link == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
 667                                 $paren = "";
 668                         } else {
 669                                 # Expand the URL for printable version
 670                                 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
 671                         }
 672                         $la = $sk->getExternalLinkAttributes( $link, $text );
 673                         $s .= "<a href='{$link}'{$la}>{$text}</a>{$paren}{$trail}";
 674
 675                 }
 676                 return $s;
 677         }
 678
 679         /* private */ function handle4Quotes( &$state, $token )
 680         {
 681                 /* This one makes some assumptions.
 682                  * '''Caesar''''s army  => <strong>Caesar</strong>'s army
 683                  * ''''Caesar'''' was a roman emperor => '<strong>Caesar</strong>' was a roman emperor
 684                  * These assumptions might be wrong, but any other assumption might be wrong, too.
 685                  * So here we go */
 686                 if ( $state["strong"] !== false ) {
 687                         return $this->handle3Quotes( $state, $token ) . "'";
 688                 } else {
 689                         return "'" . $this->handle3Quotes( $state, $token );
 690                 }
 691         }
 692
 693
 694         /* private */ function handle3Quotes( &$state, $token )
 695         {
 696                 if ( $state["strong"] !== false ) {
 697                         if ( $state["em"] !== false && $state["em"] > $state["strong"] )
 698                         {
 699                                 # ''' lala ''lala '''
 700                                 $s = "</em></strong><em>";
 701                         } else {
 702                                 $s = "</strong>";
 703                         }
 704                         $state["strong"] = FALSE;
 705                 } else {
 706                         $s = "<strong>";
 707                         $state["strong"] = $token["pos"];
 708                 }
 709                 return $s;
 710         }
 711
 712         /* private */ function handle2Quotes( &$state, $token )
 713         {
 714                 if ( $state["em"] !== false ) {
 715                         if ( $state["strong"] !== false && $state["strong"] > $state["em"] )
 716                         {
 717                                 # ''lala'''lala'' ....'''
 718                                 $s = "</strong></em><strong>";
 719                         } else {
 720                                 $s = "</em>";
 721                         }
 722                         $state["em"] = FALSE;
 723                 } else {
 724                         $s = "<em>";
 725                         $state["em"] = $token["pos"];
 726
 727                 }
 728                 return $s;
 729         }
 730
 731         /* private */ function handle5Quotes( &$state, $token )
 732         {
 733                 $s = "";
 734                 if ( $state["em"] !== false && $state["strong"] !== false ) {
 735                         if ( $state["em"] < $state["strong"] ) {
 736                                 $s .= "</strong></em>";
 737                         } else {
 738                                 $s .= "</em></strong>";
 739                         }
 740                         $state["strong"] = $state["em"] = FALSE;
 741                 } elseif ( $state["em"] !== false ) {
 742                         $s .= "</em><strong>";
 743                         $state["em"] = FALSE;
 744                         $state["strong"] = $token["pos"];
 745                 } elseif ( $state["strong"] !== false ) {
 746                         $s .= "</strong><em>";
 747                         $state["strong"] = FALSE;
 748                         $state["em"] = $token["pos"];
 749                 } else { # not $em and not $strong
 750                         $s .= "<strong><em>";
 751                         $state["strong"] = $state["em"] = $token["pos"];
 752                 }
 753                 return $s;
 754         }
 755
 756         /* private */ function doTokenizedParser( $str )
 757         {
 758                 global $wgLang; # for language specific parser hook
 759                 global $wgUploadDirectory, $wgUseTimeline;
 760
 761                 $tokenizer=Tokenizer::newFromString( $str );
 762                 $tokenStack = array();
 763
 764                 $s="";
 765                 $state["em"]      = FALSE;
 766                 $state["strong"]  = FALSE;
 767                 $tagIsOpen = FALSE;
 768                 $threeopen = false;
 769
 770                 # The tokenizer splits the text into tokens and returns them one by one.
 771                 # Every call to the tokenizer returns a new token.
 772                 while ( $token = $tokenizer->nextToken() )
 773                 {
 774                         switch ( $token["type"] )
 775                         {
 776                                 case "text":
 777                                         # simple text with no further markup
 778                                         $txt = $token["text"];
 779                                         break;
 780                                 case "blank":
 781                                         # Text that contains blanks that have to be converted to
 782                                         # non-breakable spaces for French.
 783                                         # U+202F NARROW NO-BREAK SPACE might be a better choice, but
 784                                         # browser support for Unicode spacing is poor.
 785                                         $txt = str_replace( " ", "&nbsp;", $token["text"] );
 786                                         break;
 787                                 case "[[[":
 788                                         # remember the tag opened with 3 [
 789                                         $threeopen = true;
 790                                 case "[[":
 791                                         # link opening tag.
 792                                         # FIXME : Treat orphaned open tags (stack not empty when text is over)
 793                                         $tagIsOpen = TRUE;
 794                                         array_push( $tokenStack, $token );
 795                                         $txt="";
 796                                         break;
 797
 798                                 case "]]]":
 799                                 case "]]":
 800                                         # link close tag.
 801                                         # get text from stack, glue it together, and call the code to handle a
 802                                         # link
 803
 804                                         if ( count( $tokenStack ) == 0 )
 805                                         {
 806                                                 # stack empty. Found a ]] without an opening [[
 807                                                 $txt = "]]";
 808                                         } else {
 809                                                 $linkText = "";
 810                                                 $lastToken = array_pop( $tokenStack );
 811                                                 while ( !(($lastToken["type"] == "[[[") or ($lastToken["type"] == "[[")) )
 812                                                 {
 813                                                         if( !empty( $lastToken["text"] ) ) {
 814                                                                 $linkText = $lastToken["text"] . $linkText;
 815                                                         }
 816                                                         $lastToken = array_pop( $tokenStack );
 817                                                 }
 818
 819                                                 $txt = $linkText ."]]";
 820
 821                                                 if( isset( $lastToken["text"] ) ) {
 822                                                         $prefix = $lastToken["text"];
 823                                                 } else {
 824                                                         $prefix = "";
 825                                                 }
 826                                                 $nextToken = $tokenizer->previewToken();
 827                                                 if ( $nextToken["type"] == "text" )
 828                                                 {
 829                                                         # Preview just looks at it. Now we have to fetch it.
 830                                                         $nextToken = $tokenizer->nextToken();
 831                                                         $txt .= $nextToken["text"];
 832                                                 }
 833                                                 $txt = $this->handleInternalLink( $this->unstrip($txt,$this->mStripState), $prefix );
 834
 835                                                 # did the tag start with 3 [ ?
 836                                                 if($threeopen) {
 837                                                         # show the first as text
 838                                                         $txt = "[".$txt;
 839                                                         $threeopen=false;
 840                                                 }
 841
 842                                         }
 843                                         $tagIsOpen = (count( $tokenStack ) != 0);
 844                                         break;
 845                                 case "----":
 846                                         $txt = "\n<hr />\n";
 847                                         break;
 848                                 case "'''":
 849                                         # This and the four next ones handle quotes
 850                                         $txt = $this->handle3Quotes( $state, $token );
 851                                         break;
 852                                 case "''":
 853                                         $txt = $this->handle2Quotes( $state, $token );
 854                                         break;
 855                                 case "'''''":
 856                                         $txt = $this->handle5Quotes( $state, $token );
 857                                         break;
 858                                 case "''''":
 859                                         $txt = $this->handle4Quotes( $state, $token );
 860                                         break;
 861                                 case "":
 862                                         # empty token
 863                                         $txt="";
 864                                         break;
 865                                 case "h":
 866                                         #heading- used to close all unbalanced bold or em tags in this section
 867                                         $txt = '';
 868                                         if( $state['em'] !== false and
 869                                         ( $state['strong'] === false or $state['em'] > $state['strong'] ) )
 870                                         {
 871                                                 $s .= '</em>';
 872                                                 $state['em'] = false;
 873                                         }
 874                                         if ( $state['strong'] !== false ) $txt .= '</strong>';
 875                                         if ( $state['em'] !== false ) $txt .= '</em>';
 876                                         $state['strong'] = $state['em'] = false;
 877                                         break;
 878                                 case "RFC ":
 879                                         if ( $tagIsOpen ) {
 880                                                 $txt = "RFC ";
 881                                         } else {
 882                                                 $txt = $this->doMagicRFC( $tokenizer );
 883                                         }
 884                                         break;
 885                                 case "ISBN ":
 886                                         if ( $tagIsOpen ) {
 887                                                 $txt = "ISBN ";
 888                                         } else {
 889                                                 $txt = $this->doMagicISBN( $tokenizer );
 890                                         }
 891                                         break;
 892                                 case "<timeline>":
 893                                         if ( $wgUseTimeline &&
 894                                              "" != ( $timelinesrc = $tokenizer->readAllUntil("&lt;/timeline&gt;") ) )
 895                                         {
 896                                                 $txt = renderTimeline( $timelinesrc );
 897                                         } else {
 898                                                 $txt=$token["text"];
 899                                         }
 900                                         break;
 901                                 default:
 902                                         # Call language specific Hook.
 903                                         $txt = $wgLang->processToken( $token, $tokenStack );
 904                                         if ( NULL == $txt ) {
 905                                                 # An unkown token. Highlight.
 906                                                 $txt = "<font color=\"#FF0000\"><b>".$token["type"]."</b></font>";
 907                                                 $txt .= "<font color=\"#FFFF00\"><b>".$token["text"]."</b></font>";
 908                                         }
 909                                         break;
 910                         }
 911                         # If we're parsing the interior of a link, don't append the interior to $s,
 912                         # but push it to the stack so it can be processed when a ]] token is found.
 913                         if ( $tagIsOpen  && $txt != "" ) {
 914                                 $token["type"] = "text";
 915                                 $token["text"] = $txt;
 916                                 array_push( $tokenStack, $token );
 917                         } else {
 918                                 $s .= $txt;
 919                         }
 920                 } #end while
 921
 922                 # make 100% sure all strong and em tags are closed
 923                 # doBlockLevels often messes the last bit up though, but invalid nesting is better than unclosed tags
 924                 # tidy solves this though
 925                 if( $state['em'] !== false and
 926                 ( $state['strong'] === false or $state['em'] > $state['strong'] ) )
 927                 {
 928                         $s .= '</em>';
 929                         $state['em'] = false;
 930                 }
 931                 if ( $state['strong'] !== false ) $s .= '</strong>';
 932                 if ( $state['em'] !== false ) $s .= '</em>';
 933
 934                 if ( count( $tokenStack ) != 0 )
 935                 {
 936                         # still objects on stack. opened [[ tag without closing ]] tag.
 937                         $txt = "";
 938                         while ( $lastToken = array_pop( $tokenStack ) )
 939                         {
 940                                 if ( $lastToken["type"] == "text" )
 941                                 {
 942                                         $txt = $lastToken["text"] . $txt;
 943                                 } else {
 944                                         $txt = $lastToken["type"] . $txt;
 945                                 }
 946                         }
 947                         $s .= $txt;
 948                 }
 949                 return $s;
 950         }
 951
 952         /* private */ function handleInternalLink( $line, $prefix )
 953         {
 954                 global $wgLang, $wgLinkCache;
 955                 global $wgNamespacesWithSubpages, $wgLanguageCode;
 956                 static $fname = "Parser::handleInternalLink" ;
 957                 wfProfileIn( $fname );
 958
 959                 wfProfileIn( "$fname-setup" );
 960                 static $tc = FALSE;
 961                 if ( !$tc ) { $tc = Title::legalChars() . "#"; }
 962                 $sk =& $this->mOptions->getSkin();
 963
 964                 # Match a link having the form [[namespace:link|alternate]]trail
 965                 static $e1 = FALSE;
 966                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
 967                 # Match the end of a line for a word that's not followed by whitespace,
 968                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
 969                 #$e2 = "/^(.*)\\b(\\w+)\$/suD";
 970                 #$e2 = "/^(.*\\s)(\\S+)\$/suD";
 971                 static $e2 = '/^(.*\s)([a-zA-Z\x80-\xff]+)$/sD';
 972
 973
 974                 # Special and Media are pseudo-namespaces; no pages actually exist in them
 975                 static $image = FALSE;
 976                 static $special = FALSE;
 977                 static $media = FALSE;
 978                 static $category = FALSE;
 979                 if ( !$image ) { $image = Namespace::getImage(); }
 980                 if ( !$special ) { $special = Namespace::getSpecial(); }
 981                 if ( !$media ) { $media = Namespace::getMedia(); }
 982                 if ( !$category ) { $category = Namespace::getCategory(); }
 983
 984                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
 985
 986                 wfProfileOut( "$fname-setup" );
 987                 $s = "";
 988
 989                 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
 990                         $text = $m[2];
 991                         $trail = $m[3];
 992                 } else { # Invalid form; output directly
 993                         $s .= $prefix . "[[" . $line ;
 994                         return $s;
 995                 }
 996
 997                 /* Valid link forms:
 998                 Foobar -- normal
 999                 :Foobar -- override special treatment of prefix (images, language links)
1000                 /Foobar -- convert to CurrentPage/Foobar
1001                 /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1002                 */
1003                 $c = substr($m[1],0,1);
1004                 $noforce = ($c != ":");
1005                 if( $c == "/" ) { # subpage
1006                         if(substr($m[1],-1,1)=="/") {                 # / at end means we don't want the slash to be shown
1007                                 $m[1]=substr($m[1],1,strlen($m[1])-2);
1008                                 $noslash=$m[1];
1009                         } else {
1010                                 $noslash=substr($m[1],1);
1011                         }
1012                         if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) { # subpages allowed here
1013                                 $link = $this->mTitle->getPrefixedText(). "/" . trim($noslash);
1014                                 if( "" == $text ) {
1015                                         $text= $m[1];
1016                                 } # this might be changed for ugliness reasons
1017                         } else {
1018                                 $link = $noslash; # no subpage allowed, use standard link
1019                         }
1020                 } elseif( $noforce ) { # no subpage
1021                         $link = $m[1];
1022                 } else {
1023                         $link = substr( $m[1], 1 );
1024                 }
1025                 $wasblank = ( "" == $text );
1026                 if( $wasblank )
1027                         $text = $link;
1028
1029                 $nt = Title::newFromText( $link );
1030                 if( !$nt ) {
1031                         $s .= $prefix . "[[" . $line;
1032                         return $s;
1033                 }
1034                 $ns = $nt->getNamespace();
1035                 $iw = $nt->getInterWiki();
1036                 if( $noforce ) {
1037                         if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1038                                 array_push( $this->mOutput->mLanguageLinks, $nt->getPrefixedText() );
1039                                 $s .= $prefix . $trail ;
1040                                 return (trim($s) == '')? '': $s;
1041                         }
1042                         if( $ns == $image ) {
1043                                 $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1044                                 $wgLinkCache->addImageLinkObj( $nt );
1045                                 return $s;
1046                         }
1047                         if ( $ns == $category ) {
1048                                 $t = $nt->getText() ;
1049                                 $nnt = Title::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
1050
1051                                 $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1052                                 $t = $sk->makeLinkObj( $nnt, $t, "", "" , $prefix );
1053                                 $wgLinkCache->resume();
1054
1055                                 $sortkey = $wasblank ? $this->mTitle->getPrefixedText() : $text;
1056                                 $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1057                                 $this->mOutput->mCategoryLinks[] = $t ;
1058                                 $s .= $prefix . $trail ;
1059                                 return $s ;
1060                         }
1061                 }
1062                 if( ( $nt->getPrefixedText() == $this->mTitle->getPrefixedText() ) &&
1063                     ( strpos( $link, "#" ) == FALSE ) ) {
1064                         # Self-links are handled specially; generally de-link and change to bold.
1065                         $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, "", $trail );
1066                         return $s;
1067                 }
1068
1069                 if( $ns == $media ) {
1070                         $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1071                         $wgLinkCache->addImageLinkObj( $nt );
1072                         return $s;
1073                 } elseif( $ns == $special ) {
1074                         $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, "", $trail );
1075                         return $s;
1076                 }
1077                 $s .= $sk->makeLinkObj( $nt, $text, "", $trail , $prefix );
1078
1079                 wfProfileOut( $fname );
1080                 return $s;
1081         }
1082
1083         # Some functions here used by doBlockLevels()
1084         #
1085         /* private */ function closeParagraph()
1086         {
1087                 $result = "";
1088                 if ( '' != $this->mLastSection ) {
1089                         $result = "</" . $this->mLastSection  . ">\n";
1090                 }
1091                 $this->mInPre = false;
1092                 $this->mLastSection = "";
1093                 return $result;
1094         }
1095         # getCommon() returns the length of the longest common substring
1096         # of both arguments, starting at the beginning of both.
1097         #
1098         /* private */ function getCommon( $st1, $st2 )
1099         {
1100                 $fl = strlen( $st1 );
1101                 $shorter = strlen( $st2 );
1102                 if ( $fl < $shorter ) { $shorter = $fl; }
1103
1104                 for ( $i = 0; $i < $shorter; ++$i ) {
1105                         if ( $st1{$i} != $st2{$i} ) { break; }
1106                 }
1107                 return $i;
1108         }
1109         # These next three functions open, continue, and close the list
1110         # element appropriate to the prefix character passed into them.
1111         #
1112         /* private */ function openList( $char )
1113     {
1114                 $result = $this->closeParagraph();
1115
1116                 if ( "*" == $char ) { $result .= "<ul><li>"; }
1117                 else if ( "#" == $char ) { $result .= "<ol><li>"; }
1118                 else if ( ":" == $char ) { $result .= "<dl><dd>"; }
1119                 else if ( ";" == $char ) {
1120                         $result .= "<dl><dt>";
1121                         $this->mDTopen = true;
1122                 }
1123                 else { $result = "<!-- ERR 1 -->"; }
1124
1125                 return $result;
1126         }
1127
1128         /* private */ function nextItem( $char )
1129         {
1130                 if ( "*" == $char || "#" == $char ) { return "</li><li>"; }
1131                 else if ( ":" == $char || ";" == $char ) {
1132                         $close = "</dd>";
1133                         if ( $this->mDTopen ) { $close = "</dt>"; }
1134                         if ( ";" == $char ) {
1135                                 $this->mDTopen = true;
1136                                 return $close . "<dt>";
1137                         } else {
1138                                 $this->mDTopen = false;
1139                                 return $close . "<dd>";
1140                         }
1141                 }
1142                 return "<!-- ERR 2 -->";
1143         }
1144
1145         /* private */function closeList( $char )
1146         {
1147                 if ( "*" == $char ) { $text = "</li></ul>"; }
1148                 else if ( "#" == $char ) { $text = "</li></ol>"; }
1149                 else if ( ":" == $char ) {
1150                         if ( $this->mDTopen ) {
1151                                 $this->mDTopen = false;
1152                                 $text = "</dt></dl>";
1153                         } else {
1154                                 $text = "</dd></dl>";
1155                         }
1156                 }
1157                 else {  return "<!-- ERR 3 -->"; }
1158                 return $text."\n";
1159         }
1160
1161         /* private */ function doBlockLevels( $text, $linestart ) {
1162                 $fname = "Parser::doBlockLevels";
1163                 wfProfileIn( $fname );
1164
1165                 # Parsing through the text line by line.  The main thing
1166                 # happening here is handling of block-level elements p, pre,
1167                 # and making lists from lines starting with * # : etc.
1168                 #
1169                 $textLines = explode( "\n", $text );
1170
1171                 $lastPrefix = $output = $lastLine = '';
1172                 $this->mDTopen = $inBlockElem = false;
1173                 $prefixLength = 0;
1174                 $paragraphStack = false;
1175
1176                 if ( !$linestart ) {
1177                         $output .= array_shift( $textLines );
1178                 }
1179                 foreach ( $textLines as $oLine ) {
1180                         $lastPrefixLength = strlen( $lastPrefix );
1181                         $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1182                         $preOpenMatch = preg_match("/<pre/i", $oLine );
1183                         if (!$this->mInPre) {
1184                                 $this->mInPre = !empty($preOpenMatch);
1185                         }
1186                         if ( !$this->mInPre ) {
1187                                 # Multiple prefixes may abut each other for nested lists.
1188                                 $prefixLength = strspn( $oLine, "*#:;" );
1189                                 $pref = substr( $oLine, 0, $prefixLength );
1190
1191                                 # eh?
1192                                 $pref2 = str_replace( ";", ":", $pref );
1193                                 $t = substr( $oLine, $prefixLength );
1194                         } else {
1195                                 # Don't interpret any other prefixes in preformatted text
1196                                 $prefixLength = 0;
1197                                 $pref = $pref2 = '';
1198                                 $t = $oLine;
1199                         }
1200
1201                         # List generation
1202                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1203                                 # Same as the last item, so no need to deal with nesting or opening stuff
1204                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1205                                 $paragraphStack = false;
1206
1207                                 if ( ";" == substr( $pref, -1 ) ) {
1208                                         # The one nasty exception: definition lists work like this:
1209                                         # ; title : definition text
1210                                         # So we check for : in the remainder text to split up the
1211                                         # title and definition, without b0rking links.
1212                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1213                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1214                                                 $term = $match[1];
1215                                                 $output .= $term . $this->nextItem( ":" );
1216                                                 $t = $match[2];
1217                                         }
1218                                 }
1219                         } elseif( $prefixLength || $lastPrefixLength ) {
1220                                 # Either open or close a level...
1221                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1222                                 $paragraphStack = false;
1223
1224                                 while( $commonPrefixLength < $lastPrefixLength ) {
1225                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1226                                         --$lastPrefixLength;
1227                                 }
1228                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1229                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1230                                 }
1231                                 while ( $prefixLength > $commonPrefixLength ) {
1232                                         $char = substr( $pref, $commonPrefixLength, 1 );
1233                                         $output .= $this->openList( $char );
1234
1235                                         if ( ";" == $char ) {
1236                                                 # FIXME: This is dupe of code above
1237                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1238                                                         $term = $match[1];
1239                                                         $output .= $term . $this->nextItem( ":" );
1240                                                         $t = $match[2];
1241                                                 }
1242                                         }
1243                                         ++$commonPrefixLength;
1244                                 }
1245                                 $lastPrefix = $pref2;
1246                         }
1247                         if( 0 == $prefixLength ) {
1248                                 # No prefix (not in list)--go to paragraph mode
1249                                 $uniq_prefix = UNIQ_PREFIX;
1250                                 // XXX: use a stack for nestable elements like span, table and div
1251                                 $openmatch = preg_match("/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<td|<p|<ul|<li)/i", $t );
1252                                 $closematch = preg_match(
1253                                         "/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|".
1254                                         "<div|<\\/div|<hr|<\\/td|<\\/pre|<\\/p|".$uniq_prefix."-pre|<\\/li|<\\/ul)/i", $t );
1255                                 if ( $openmatch or $closematch ) {
1256                                         $paragraphStack = false;
1257                                         $output .= $this->closeParagraph();
1258                                         if($preOpenMatch and !$preCloseMatch) {
1259                                                 $this->mInPre = true;
1260                                         }
1261                                         if ( $closematch  ) {
1262                                                 $inBlockElem = false;
1263                                         } else {
1264                                                 $inBlockElem = true;
1265                                         }
1266                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1267                                         if ( " " == $t{0} and trim($t) != '' ) {
1268                                                 // pre
1269                                                 if ($this->mLastSection != 'pre') {
1270                                                         $paragraphStack = false;
1271                                                         $output .= $this->closeParagraph().'<pre>';
1272                                                         $this->mLastSection = 'pre';
1273                                                 }
1274                                         } else {
1275                                                 // paragraph
1276                                                 if ( '' == trim($t) ) {
1277                                                         if ( $paragraphStack ) {
1278                                                                 $output .= $paragraphStack.'<br/>';
1279                                                                 $paragraphStack = false;
1280                                                                 $this->mLastSection = 'p';
1281                                                         } else {
1282                                                                 if ($this->mLastSection != 'p' ) {
1283                                                                         $output .= $this->closeParagraph();
1284                                                                         $this->mLastSection = '';
1285                                                                         $paragraphStack = "<p>";
1286                                                                 } else {
1287                                                                         $paragraphStack = '</p><p>';
1288                                                                 }
1289                                                         }
1290                                                 } else {
1291                                                         if ( $paragraphStack ) {
1292                                                                 $output .= $paragraphStack;
1293                                                                 $paragraphStack = false;
1294                                                                 $this->mLastSection = 'p';
1295                                                         } else if ($this->mLastSection != 'p') {
1296                                                                 $output .= $this->closeParagraph().'<p>';
1297                                                                 $this->mLastSection = 'p';
1298                                                         }
1299                                                 }
1300                                         }
1301                                 }
1302                         }
1303                         if ($paragraphStack === false) {
1304                                 $output .= $t."\n";
1305                         }
1306                 }
1307                 while ( $prefixLength ) {
1308                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1309                         --$prefixLength;
1310                 }
1311                 if ( "" != $this->mLastSection ) {
1312                         $output .= "</" . $this->mLastSection . ">";
1313                         $this->mLastSection = "";
1314                 }
1315
1316                 wfProfileOut( $fname );
1317                 return $output;
1318         }
1319
1320         function getVariableValue( $index ) {
1321                 global $wgLang, $wgSitename, $wgServer;
1322
1323                 switch ( $index ) {
1324                         case MAG_CURRENTMONTH:
1325                                 return date( "m" );
1326                         case MAG_CURRENTMONTHNAME:
1327                                 return $wgLang->getMonthName( date("n") );
1328                         case MAG_CURRENTMONTHNAMEGEN:
1329                                 return $wgLang->getMonthNameGen( date("n") );
1330                         case MAG_CURRENTDAY:
1331                                 return date("j");
1332                         case MAG_PAGENAME:
1333                                 return $this->mTitle->getText();
1334                         case MAG_NAMESPACE:
1335                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1336                                 return $wgLang->getNsText($this->mTitle->getNamespace()); // Patch  by Dori
1337                         case MAG_CURRENTDAYNAME:
1338                                 return $wgLang->getWeekdayName( date("w")+1 );
1339                         case MAG_CURRENTYEAR:
1340                                 return date( "Y" );
1341                         case MAG_CURRENTTIME:
1342                                 return $wgLang->time( wfTimestampNow(), false );
1343                         case MAG_NUMBEROFARTICLES:
1344                                 return wfNumberOfArticles();
1345                         case MAG_SITENAME:
1346                                 return $wgSitename;
1347                         case MAG_SERVER:
1348                                 return $wgServer;
1349                         default:
1350                                 return NULL;
1351                 }
1352         }
1353
1354         function initialiseVariables()
1355         {
1356                 global $wgVariableIDs;
1357                 $this->mVariables = array();
1358                 foreach ( $wgVariableIDs as $id ) {
1359                         $mw =& MagicWord::get( $id );
1360                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1361                 }
1362         }
1363
1364         /* private */ function replaceVariables( $text, $args = array() )
1365         {
1366                 global $wgLang, $wgScript, $wgArticlePath;
1367
1368                 $fname = "Parser::replaceVariables";
1369                 wfProfileIn( $fname );
1370
1371                 $bail = false;
1372                 if ( !$this->mVariables ) {
1373                         $this->initialiseVariables();
1374                 }
1375                 $titleChars = Title::legalChars();
1376
1377                 # This function is called recursively. To keep track of arguments we need a stack:
1378                 array_push( $this->mArgStack, $args );
1379
1380                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1381                 $GLOBALS['wgCurParser'] =& $this;
1382
1383                 # Argument substitution
1384                 if ( $this->mOutputType == OT_HTML ) {
1385                         $text = preg_replace_callback( "/(\\n?){{{([$titleChars]*?)}}}/", "wfArgSubstitution", $text );
1386                 }
1387
1388                 # Double brace substitution
1389                 $regex = "/(\\n?){{([$titleChars]*?)(\\|.*?|)}}/s";
1390                 $text = preg_replace_callback( $regex, "wfBraceSubstitution", $text );
1391
1392                 array_pop( $this->mArgStack );
1393
1394                 return $text;
1395         }
1396
1397         function braceSubstitution( $matches )
1398         {
1399                 global $wgLinkCache, $wgLang;
1400                 $fname = "Parser::braceSubstitution";
1401                 $found = false;
1402                 $nowiki = false;
1403                 $noparse = false;
1404
1405                 $title = NULL;
1406
1407                 # $newline is an optional newline character before the braces
1408                 # $part1 is the bit before the first |, and must contain only title characters
1409                 # $args is a list of arguments, starting from index 0, not including $part1
1410
1411                 $newline = $matches[1];
1412                 $part1 = $matches[2];
1413                 # If the third subpattern matched anything, it will start with |
1414                 if ( $matches[3] !== "" ) {
1415                         $args = explode( "|", substr( $matches[3], 1 ) );
1416                 } else {
1417                         $args = array();
1418                 }
1419                 $argc = count( $args );
1420
1421                 # {{{}}}
1422                 if ( strpos( $matches[0], "{{{" ) !== false ) {
1423                         $text = $matches[0];
1424                         $found = true;
1425                         $noparse = true;
1426                 }
1427
1428                 # SUBST
1429                 if ( !$found ) {
1430                         $mwSubst =& MagicWord::get( MAG_SUBST );
1431                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1432                                 if ( $this->mOutputType != OT_WIKI ) {
1433                                         # Invalid SUBST not replaced at PST time
1434                                         # Return without further processing
1435                                         $text = $matches[0];
1436                                         $found = true;
1437                                         $noparse= true;
1438                                 }
1439                         } elseif ( $this->mOutputType == OT_WIKI ) {
1440                                 # SUBST not found in PST pass, do nothing
1441                                 $text = $matches[0];
1442                                 $found = true;
1443                         }
1444                 }
1445
1446                 # MSG, MSGNW and INT
1447                 if ( !$found ) {
1448                         # Check for MSGNW:
1449                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1450                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1451                                 $nowiki = true;
1452                         } else {
1453                                 # Remove obsolete MSG:
1454                                 $mwMsg =& MagicWord::get( MAG_MSG );
1455                                 $mwMsg->matchStartAndRemove( $part1 );
1456                         }
1457
1458                         # Check if it is an internal message
1459                         $mwInt =& MagicWord::get( MAG_INT );
1460                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1461                                 if ( $this->incrementIncludeCount( "int:$part1" ) ) {
1462                                         $text = wfMsgReal( $part1, $args, true );
1463                                         $found = true;
1464                                 }
1465                         }
1466                 }
1467
1468                 # NS
1469                 if ( !$found ) {
1470                         # Check for NS: (namespace expansion)
1471                         $mwNs = MagicWord::get( MAG_NS );
1472                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1473                                 if ( intval( $part1 ) ) {
1474                                         $text = $wgLang->getNsText( intval( $part1 ) );
1475                                         $found = true;
1476                                 } else {
1477                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1478                                         if ( !is_null( $index ) ) {
1479                                                 $text = $wgLang->getNsText( $index );
1480                                                 $found = true;
1481                                         }
1482                                 }
1483                         }
1484                 }
1485
1486                 # LOCALURL and LOCALURLE
1487                 if ( !$found ) {
1488                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1489                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1490
1491                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1492                                 $func = 'getLocalURL';
1493                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1494                                 $func = 'escapeLocalURL';
1495                         } else {
1496                                 $func = '';
1497                         }
1498
1499                         if ( $func !== '' ) {
1500                                 $title = Title::newFromText( $part1 );
1501                                 if ( !is_null( $title ) ) {
1502                                         if ( $argc > 0 ) {
1503                                                 $text = $title->$func( $args[0] );
1504                                         } else {
1505                                                 $text = $title->$func();
1506                                         }
1507                                         $found = true;
1508                                 }
1509                         }
1510                 }
1511
1512                 # Internal variables
1513                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1514                         $text = $this->mVariables[$part1];
1515                         $found = true;
1516                         $this->mOutput->mContainsOldMagic = true;
1517                 }
1518 /*
1519                 # Arguments input from the caller
1520                 $inputArgs = end( $this->mArgStack );
1521                 if ( !$found && array_key_exists( $part1, $inputArgs ) ) {
1522                         $text = $inputArgs[$part1];
1523                         $found = true;
1524                 }
1525 */
1526                 # Load from database
1527                 if ( !$found ) {
1528                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1529                         if ( !is_null( $title ) && !$title->isExternal() ) {
1530                                 # Check for excessive inclusion
1531                                 $dbk = $title->getPrefixedDBkey();
1532                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1533                                         $article = new Article( $title );
1534                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1535                                         if ( $articleContent !== false ) {
1536                                                 $found = true;
1537                                                 $text = $articleContent;
1538
1539                                         }
1540                                 }
1541
1542                                 # If the title is valid but undisplayable, make a link to it
1543                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1544                                         $text = "[[" . $title->getPrefixedText() . "]]";
1545                                         $found = true;
1546                                 }
1547                         }
1548                 }
1549
1550                 # Recursive parsing, escaping and link table handling
1551                 # Only for HTML output
1552                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1553                         $text = wfEscapeWikiText( $text );
1554                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1555                         # Clean up argument array
1556                         $assocArgs = array();
1557                         $index = 1;
1558                         foreach( $args as $arg ) {
1559                                 $eqpos = strpos( $arg, "=" );
1560                                 if ( $eqpos === false ) {
1561                                         $assocArgs[$index++] = $arg;
1562                                 } else {
1563                                         $name = trim( substr( $arg, 0, $eqpos ) );
1564                                         $value = trim( substr( $arg, $eqpos+1 ) );
1565                                         if ( $value === false ) {
1566                                                 $value = "";
1567                                         }
1568                                         if ( $name !== false ) {
1569                                                 $assocArgs[$name] = $value;
1570                                         }
1571                                 }
1572                         }
1573
1574                         # Do not enter included links in link table
1575                         if ( !is_null( $title ) ) {
1576                                 $wgLinkCache->suspend();
1577                         }
1578
1579                         # Run full parser on the included text
1580                         $text = $this->stripParse( $text, (bool)$newline, $assocArgs );
1581
1582                         # Resume the link cache and register the inclusion as a link
1583                         if ( !is_null( $title ) ) {
1584                                 $wgLinkCache->resume();
1585                                 $wgLinkCache->addLinkObj( $title );
1586                         }
1587                 }
1588
1589                 if ( !$found ) {
1590                         return $matches[0];
1591                 } else {
1592                         return $text;
1593                 }
1594         }
1595
1596         # Triple brace replacement -- used for template arguments
1597         function argSubstitution( $matches )
1598         {
1599                 $newline = $matches[1];
1600                 $arg = trim( $matches[2] );
1601                 $text = $matches[0];
1602                 $inputArgs = end( $this->mArgStack );
1603
1604                 if ( array_key_exists( $arg, $inputArgs ) ) {
1605                         $text = $this->stripParse( $inputArgs[$arg], (bool)$newline, array() );
1606                 }
1607
1608                 return $text;
1609         }
1610
1611         # Returns true if the function is allowed to include this entity
1612         function incrementIncludeCount( $dbk )
1613         {
1614                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1615                         $this->mIncludeCount[$dbk] = 0;
1616                 }
1617                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1618                         return true;
1619                 } else {
1620                         return false;
1621                 }
1622         }
1623
1624
1625         # Cleans up HTML, removes dangerous tags and attributes
1626         /* private */ function removeHTMLtags( $text )
1627         {
1628                 global $wgUseTidy, $wgUserHtml;
1629                 $fname = "Parser::removeHTMLtags";
1630                 wfProfileIn( $fname );
1631
1632                 if( $wgUserHtml ) {
1633                         $htmlpairs = array( # Tags that must be closed
1634                                 "b", "del", "i", "ins", "u", "font", "big", "small", "sub", "sup", "h1",
1635                                 "h2", "h3", "h4", "h5", "h6", "cite", "code", "em", "s",
1636                                 "strike", "strong", "tt", "var", "div", "center",
1637                                 "blockquote", "ol", "ul", "dl", "table", "caption", "pre",
1638                                 "ruby", "rt" , "rb" , "rp", "p"
1639                         );
1640                         $htmlsingle = array(
1641                                 "br", "hr", "li", "dt", "dd"
1642                         );
1643                         $htmlnest = array( # Tags that can be nested--??
1644                                 "table", "tr", "td", "th", "div", "blockquote", "ol", "ul",
1645                                 "dl", "font", "big", "small", "sub", "sup"
1646                         );
1647                         $tabletags = array( # Can only appear inside table
1648                                 "td", "th", "tr"
1649                         );
1650                 } else {
1651                         $htmlpairs = array();
1652                         $htmlsingle = array();
1653                         $htmlnest = array();
1654                         $tabletags = array();
1655                 }
1656
1657                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1658                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1659
1660                 $htmlattrs = $this->getHTMLattrs () ;
1661
1662                 # Remove HTML comments
1663                 $text = preg_replace( "/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU", "$2", $text );
1664
1665                 $bits = explode( "<", $text );
1666                 $text = array_shift( $bits );
1667                 if(!$wgUseTidy) {
1668                         $tagstack = array(); $tablestack = array();
1669                         foreach ( $bits as $x ) {
1670                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1671                                 preg_match( "/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/",
1672                                 $x, $regs );
1673                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1674                                 error_reporting( $prev );
1675
1676                                 $badtag = 0 ;
1677                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1678                                         # Check our stack
1679                                         if ( $slash ) {
1680                                                 # Closing a tag...
1681                                                 if ( ! in_array( $t, $htmlsingle ) &&
1682                                                 ( count($tagstack) && $ot = array_pop( $tagstack ) ) != $t ) {
1683                                                         if(!empty($ot)) array_push( $tagstack, $ot );
1684                                                         $badtag = 1;
1685                                                 } else {
1686                                                         if ( $t == "table" ) {
1687                                                                 $tagstack = array_pop( $tablestack );
1688                                                         }
1689                                                         $newparams = "";
1690                                                 }
1691                                         } else {
1692                                                 # Keep track for later
1693                                                 if ( in_array( $t, $tabletags ) &&
1694                                                 ! in_array( "table", $tagstack ) ) {
1695                                                         $badtag = 1;
1696                                                 } else if ( in_array( $t, $tagstack ) &&
1697                                                 ! in_array ( $t , $htmlnest ) ) {
1698                                                         $badtag = 1 ;
1699                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1700                                                         if ( $t == "table" ) {
1701                                                                 array_push( $tablestack, $tagstack );
1702                                                                 $tagstack = array();
1703                                                         }
1704                                                         array_push( $tagstack, $t );
1705                                                 }
1706                                                 # Strip non-approved attributes from the tag
1707                                                 $newparams = $this->fixTagAttributes($params);
1708
1709                                         }
1710                                         if ( ! $badtag ) {
1711                                                 $rest = str_replace( ">", "&gt;", $rest );
1712                                                 $text .= "<$slash$t $newparams$brace$rest";
1713                                                 continue;
1714                                         }
1715                                 }
1716                                 $text .= "&lt;" . str_replace( ">", "&gt;", $x);
1717                         }
1718                         # Close off any remaining tags
1719                         while ( $t = array_pop( $tagstack ) ) {
1720                                 $text .= "</$t>\n";
1721                                 if ( $t == "table" ) { $tagstack = array_pop( $tablestack ); }
1722                         }
1723                 } else {
1724                         # this might be possible using tidy itself
1725                         foreach ( $bits as $x ) {
1726                                 preg_match( "/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/",
1727                                 $x, $regs );
1728                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1729                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1730                                         $newparams = $this->fixTagAttributes($params);
1731                                         $rest = str_replace( ">", "&gt;", $rest );
1732                                         $text .= "<$slash$t $newparams$brace$rest";
1733                                 } else {
1734                                         $text .= "&lt;" . str_replace( ">", "&gt;", $x);
1735                                 }
1736                         }
1737                 }
1738                 wfProfileOut( $fname );
1739                 return $text;
1740         }
1741
1742
1743 /*
1744  *
1745  * This function accomplishes several tasks:
1746  * 1) Auto-number headings if that option is enabled
1747  * 2) Add an [edit] link to sections for logged in users who have enabled the option
1748  * 3) Add a Table of contents on the top for users who have enabled the option
1749  * 4) Auto-anchor headings
1750  *
1751  * It loops through all headlines, collects the necessary data, then splits up the
1752  * string and re-inserts the newly formatted headlines.
1753  *
1754  */
1755
1756         /* private */ function formatHeadings( $text, $isMain=true )
1757         {
1758                 global $wgInputEncoding;
1759
1760                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1761                 $doShowToc = $this->mOptions->getShowToc();
1762                 if( !$this->mTitle->userCanEdit() ) {
1763                         $showEditLink = 0;
1764                         $rightClickHack = 0;
1765                 } else {
1766                         $showEditLink = $this->mOptions->getEditSection();
1767                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1768                 }
1769
1770                 # Inhibit editsection links if requested in the page
1771                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1772                 if( $esw->matchAndRemove( $text ) ) {
1773                         $showEditLink = 0;
1774                 }
1775                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1776                 # do not add TOC
1777                 $mw =& MagicWord::get( MAG_NOTOC );
1778                 if( $mw->matchAndRemove( $text ) ) {
1779                         $doShowToc = 0;
1780                 }
1781
1782                 # never add the TOC to the Main Page. This is an entry page that should not
1783                 # be more than 1-2 screens large anyway
1784                 if( $this->mTitle->getPrefixedText() == wfMsg("mainpage") ) {
1785                         $doShowToc = 0;
1786                 }
1787
1788                 # Get all headlines for numbering them and adding funky stuff like [edit]
1789                 # links - this is for later, but we need the number of headlines right now
1790                 $numMatches = preg_match_all( "/<H([1-6])(.*?" . ">)(.*?)<\/H[1-6]>/i", $text, $matches );
1791
1792                 # if there are fewer than 4 headlines in the article, do not show TOC
1793                 if( $numMatches < 4 ) {
1794                         $doShowToc = 0;
1795                 }
1796
1797                 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1798                 # override above conditions and always show TOC
1799                 $mw =& MagicWord::get( MAG_FORCETOC );
1800                 if ($mw->matchAndRemove( $text ) ) {
1801                         $doShowToc = 1;
1802                 }
1803
1804
1805                 # We need this to perform operations on the HTML
1806                 $sk =& $this->mOptions->getSkin();
1807
1808                 # headline counter
1809                 $headlineCount = 0;
1810
1811                 # Ugh .. the TOC should have neat indentation levels which can be
1812                 # passed to the skin functions. These are determined here
1813                 $toclevel = 0;
1814                 $toc = "";
1815                 $full = "";
1816                 $head = array();
1817                 $sublevelCount = array();
1818                 $level = 0;
1819                 $prevlevel = 0;
1820                 foreach( $matches[3] as $headline ) {
1821                         $numbering = "";
1822                         if( $level ) {
1823                                 $prevlevel = $level;
1824                         }
1825                         $level = $matches[1][$headlineCount];
1826                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
1827                                 # reset when we enter a new level
1828                                 $sublevelCount[$level] = 0;
1829                                 $toc .= $sk->tocIndent( $level - $prevlevel );
1830                                 $toclevel += $level - $prevlevel;
1831                         }
1832                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
1833                                 # reset when we step back a level
1834                                 $sublevelCount[$level+1]=0;
1835                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
1836                                 $toclevel -= $prevlevel - $level;
1837                         }
1838                         # count number of headlines for each level
1839                         @$sublevelCount[$level]++;
1840                         if( $doNumberHeadings || $doShowToc ) {
1841                                 $dot = 0;
1842                                 for( $i = 1; $i <= $level; $i++ ) {
1843                                         if( !empty( $sublevelCount[$i] ) ) {
1844                                                 if( $dot ) {
1845                                                         $numbering .= ".";
1846                                                 }
1847                                                 $numbering .= $sublevelCount[$i];
1848                                                 $dot = 1;
1849                                         }
1850                                 }
1851                         }
1852
1853                         # The canonized header is a version of the header text safe to use for links
1854                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
1855                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
1856
1857                         # strip out HTML
1858                         $canonized_headline = preg_replace( "/<.*?" . ">/","",$canonized_headline );
1859                         $tocline = trim( $canonized_headline );
1860                         $canonized_headline = preg_replace("/[ \\?&\\/<>\\(\\)\\[\\]=,+']+/", '_', urlencode( do_html_entity_decode( $tocline, ENT_COMPAT, $wgInputEncoding ) ) );
1861                         # strip out urlencoded &nbsp; (inserted for french spaces, e.g. first space in 'something : something')
1862                         $canonized_headline = str_replace('%C2%A0','_', $canonized_headline);
1863                         $refer[$headlineCount] = $canonized_headline;
1864
1865                         # count how many in assoc. array so we can track dupes in anchors
1866                         @$refers[$canonized_headline]++;
1867                         $refcount[$headlineCount]=$refers[$canonized_headline];
1868
1869                         # Prepend the number to the heading text
1870
1871                         if( $doNumberHeadings || $doShowToc ) {
1872                                 $tocline = $numbering . " " . $tocline;
1873
1874                                 # Don't number the heading if it is the only one (looks silly)
1875                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
1876                                         # the two are different if the line contains a link
1877                                         $headline=$numbering . " " . $headline;
1878                                 }
1879                         }
1880
1881                         # Create the anchor for linking from the TOC to the section
1882                         $anchor = $canonized_headline;
1883                         if($refcount[$headlineCount] > 1 ) {
1884                                 $anchor .= "_" . $refcount[$headlineCount];
1885                         }
1886                         if( $doShowToc ) {
1887                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
1888                         }
1889                         if( $showEditLink ) {
1890                                 if ( empty( $head[$headlineCount] ) ) {
1891                                         $head[$headlineCount] = "";
1892                                 }
1893                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
1894                         }
1895
1896                         # Add the edit section span
1897                         if( $rightClickHack ) {
1898                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
1899                         }
1900
1901                         # give headline the correct <h#> tag
1902                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
1903
1904                         $headlineCount++;
1905                 }
1906
1907                 if( $doShowToc ) {
1908                         $toclines = $headlineCount;
1909                         $toc .= $sk->tocUnindent( $toclevel );
1910                         $toc = $sk->tocTable( $toc );
1911                 }
1912
1913                 # split up and insert constructed headlines
1914
1915                 $blocks = preg_split( "/<H[1-6].*?" . ">.*?<\/H[1-6]>/i", $text );
1916                 $i = 0;
1917
1918                 foreach( $blocks as $block ) {
1919                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
1920                             # This is the [edit] link that appears for the top block of text when
1921                                 # section editing is enabled
1922
1923                                 # Disabled because it broke block formatting
1924                                 # For example, a bullet point in the top line
1925                                 # $full .= $sk->editSectionLink(0);
1926                         }
1927                         $full .= $block;
1928                         if( $doShowToc && !$i && $isMain) {
1929                         # Top anchor now in skin
1930                                 $full = $full.$toc;
1931                         }
1932
1933                         if( !empty( $head[$i] ) ) {
1934                                 $full .= $head[$i];
1935                         }
1936                         $i++;
1937                 }
1938
1939                 return $full;
1940         }
1941
1942         /* private */ function doMagicISBN( &$tokenizer )
1943         {
1944                 global $wgLang;
1945
1946                 # Check whether next token is a text token
1947                 # If yes, fetch it and convert the text into a
1948                 # Special::BookSources link
1949                 $token = $tokenizer->previewToken();
1950                 while ( $token["type"] == "" )
1951                 {
1952                         $tokenizer->nextToken();
1953                         $token = $tokenizer->previewToken();
1954                 }
1955                 if ( $token["type"] == "text" )
1956                 {
1957                         $token = $tokenizer->nextToken();
1958                         $x = $token["text"];
1959                         $valid = "0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ";
1960
1961                         $isbn = $blank = "" ;
1962                         while ( " " == $x{0} ) {
1963                                 $blank .= " ";
1964                                 $x = substr( $x, 1 );
1965                         }
1966                         while ( strstr( $valid, $x{0} ) != false ) {
1967                                 $isbn .= $x{0};
1968                                 $x = substr( $x, 1 );
1969                         }
1970                         $num = str_replace( "-", "", $isbn );
1971                         $num = str_replace( " ", "", $num );
1972
1973                         if ( "" == $num ) {
1974                                 $text = "ISBN $blank$x";
1975                         } else {
1976                                 $titleObj = Title::makeTitle( NS_SPECIAL, "Booksources" );
1977                                 $text = "<a href=\"" .
1978                                 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
1979                                         "\" class=\"internal\">ISBN $isbn</a>";
1980                                 $text .= $x;
1981                         }
1982                 } else {
1983                         $text = "ISBN ";
1984                 }
1985                 return $text;
1986         }
1987         /* private */ function doMagicRFC( &$tokenizer )
1988         {
1989                 global $wgLang;
1990
1991                 # Check whether next token is a text token
1992                 # If yes, fetch it and convert the text into a
1993                 # link to an RFC source
1994                 $token = $tokenizer->previewToken();
1995                 while ( $token["type"] == "" )
1996                 {
1997                         $tokenizer->nextToken();
1998                         $token = $tokenizer->previewToken();
1999                 }
2000                 if ( $token["type"] == "text" )
2001                 {
2002                         $token = $tokenizer->nextToken();
2003                         $x = $token["text"];
2004                         $valid = "0123456789";
2005
2006                         $rfc = $blank = "" ;
2007                         while ( " " == $x{0} ) {
2008                                 $blank .= " ";
2009                                 $x = substr( $x, 1 );
2010                         }
2011                         while ( strstr( $valid, $x{0} ) != false ) {
2012                                 $rfc .= $x{0};
2013                                 $x = substr( $x, 1 );
2014                         }
2015
2016                         if ( "" == $rfc ) {
2017                                 $text .= "RFC $blank$x";
2018                         } else {
2019                                 $url = wfmsg( "rfcurl" );
2020                                 $url = str_replace( "$1", $rfc, $url);
2021                                 $sk =& $this->mOptions->getSkin();
2022                                 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
2023                                 $text = "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2024                         }
2025                 } else {
2026                         $text = "RFC ";
2027                 }
2028                 return $text;
2029         }
2030
2031         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true )
2032         {
2033                 $this->mOptions = $options;
2034                 $this->mTitle =& $title;
2035                 $this->mOutputType = OT_WIKI;
2036
2037                 if ( $clearState ) {
2038                         $this->clearState();
2039                 }
2040
2041                 $stripState = false;
2042                 $pairs = array(
2043                         "\r\n" => "\n",
2044                         );
2045                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2046                 // now with regexes
2047                 $pairs = array(
2048                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2049                         "/<br *?>/i" => "<br/>",
2050                 );
2051                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2052                 $text = $this->strip( $text, $stripState, false );
2053                 $text = $this->pstPass2( $text, $user );
2054                 $text = $this->unstrip( $text, $stripState );
2055                 return $text;
2056         }
2057
2058         /* private */ function pstPass2( $text, &$user )
2059         {
2060                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2061
2062                 # Variable replacement
2063                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2064                 $text = $this->replaceVariables( $text );
2065
2066                 # Signatures
2067                 #
2068                 $n = $user->getName();
2069                 $k = $user->getOption( "nickname" );
2070                 if ( "" == $k ) { $k = $n; }
2071                 if(isset($wgLocaltimezone)) {
2072                         $oldtz = getenv("TZ"); putenv("TZ=$wgLocaltimezone");
2073                 }
2074                 /* Note: this is an ugly timezone hack for the European wikis */
2075                 $d = $wgLang->timeanddate( date( "YmdHis" ), false ) .
2076                   " (" . date( "T" ) . ")";
2077                 if(isset($wgLocaltimezone)) putenv("TZ=$oldtz");
2078
2079                 $text = preg_replace( "/~~~~~/", $d, $text );
2080                 $text = preg_replace( "/~~~~/", "[[" . $wgLang->getNsText(
2081                   Namespace::getUser() ) . ":$n|$k]] $d", $text );
2082                 $text = preg_replace( "/~~~/", "[[" . $wgLang->getNsText(
2083                   Namespace::getUser() ) . ":$n|$k]]", $text );
2084
2085                 # Context links: [[|name]] and [[name (context)|]]
2086                 #
2087                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2088                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2089                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2090                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2091
2092                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2093                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2094                 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/";          # [[namespace:page|]]
2095                 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2096                                                                                                                 # [[ns:page (cont)|]]
2097                 $context = "";
2098                 $t = $this->mTitle->getText();
2099                 if ( preg_match( $conpat, $t, $m ) ) {
2100                         $context = $m[2];
2101                 }
2102                 $text = preg_replace( $p4, "[[\\1:\\2 (\\3)|\\2]]", $text );
2103                 $text = preg_replace( $p1, "[[\\1 (\\2)|\\1]]", $text );
2104                 $text = preg_replace( $p3, "[[\\1:\\2|\\2]]", $text );
2105
2106                 if ( "" == $context ) {
2107                         $text = preg_replace( $p2, "[[\\1]]", $text );
2108                 } else {
2109                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2110                 }
2111
2112                 /*
2113                 $mw =& MagicWord::get( MAG_SUBST );
2114                 $wgCurParser = $this->fork();
2115                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2116                 $this->merge( $wgCurParser );
2117                 */
2118
2119                 # Trim trailing whitespace
2120                 # MAG_END (__END__) tag allows for trailing
2121                 # whitespace to be deliberately included
2122                 $text = rtrim( $text );
2123                 $mw =& MagicWord::get( MAG_END );
2124                 $mw->matchAndRemove( $text );
2125
2126                 return $text;
2127         }
2128
2129         # Set up some variables which are usually set up in parse()
2130         # so that an external function can call some class members with confidence
2131         function startExternalParse( &$title, $options, $outputType, $clearState = true )
2132         {
2133                 $this->mTitle =& $title;
2134                 $this->mOptions = $options;
2135                 $this->mOutputType = $outputType;
2136                 if ( $clearState ) {
2137                         $this->clearState();
2138                 }
2139         }
2140
2141         function transformMsg( $text, $options ) {
2142                 global $wgTitle;
2143                 static $executing = false;
2144
2145                 # Guard against infinite recursion
2146                 if ( $executing ) {
2147                         return $text;
2148                 }
2149                 $executing = true;
2150
2151                 $this->mTitle = $wgTitle;
2152                 $this->mOptions = $options;
2153                 $this->mOutputType = OT_MSG;
2154                 $this->clearState();
2155                 $text = $this->replaceVariables( $text );
2156
2157                 $executing = false;
2158                 return $text;
2159         }
2160 }
2161
2162 class ParserOutput
2163 {
2164         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2165
2166         function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2167                 $containsOldMagic = false )
2168         {
2169                 $this->mText = $text;
2170                 $this->mLanguageLinks = $languageLinks;
2171                 $this->mCategoryLinks = $categoryLinks;
2172                 $this->mContainsOldMagic = $containsOldMagic;
2173         }
2174
2175         function getText() { return $this->mText; }
2176         function getLanguageLinks() { return $this->mLanguageLinks; }
2177         function getCategoryLinks() { return $this->mCategoryLinks; }
2178         function containsOldMagic() { return $this->mContainsOldMagic; }
2179         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2180         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2181         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2182         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2183
2184         function merge( $other ) {
2185                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2186                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2187                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2188         }
2189
2190 }
2191
2192 class ParserOptions
2193 {
2194         # All variables are private
2195         var $mUseTeX;                    # Use texvc to expand <math> tags
2196         var $mUseCategoryMagic;          # Treat [[Category:xxxx]] tags specially
2197         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2198         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2199         var $mAllowExternalImages;       # Allow external images inline
2200         var $mSkin;                      # Reference to the preferred skin
2201         var $mDateFormat;                # Date format index
2202         var $mEditSection;               # Create "edit section" links
2203         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2204         var $mNumberHeadings;            # Automatically number headings
2205         var $mShowToc;                   # Show table of contents
2206
2207         function getUseTeX() { return $this->mUseTeX; }
2208         function getUseCategoryMagic() { return $this->mUseCategoryMagic; }
2209         function getUseDynamicDates() { return $this->mUseDynamicDates; }
2210         function getInterwikiMagic() { return $this->mInterwikiMagic; }
2211         function getAllowExternalImages() { return $this->mAllowExternalImages; }
2212         function getSkin() { return $this->mSkin; }
2213         function getDateFormat() { return $this->mDateFormat; }
2214         function getEditSection() { return $this->mEditSection; }
2215         function getEditSectionOnRightClick() { return $this->mEditSectionOnRightClick; }
2216         function getNumberHeadings() { return $this->mNumberHeadings; }
2217         function getShowToc() { return $this->mShowToc; }
2218
2219         function setUseTeX( $x ) { return wfSetVar( $this->mUseTeX, $x ); }
2220         function setUseCategoryMagic( $x ) { return wfSetVar( $this->mUseCategoryMagic, $x ); }
2221         function setUseDynamicDates( $x ) { return wfSetVar( $this->mUseDynamicDates, $x ); }
2222         function setInterwikiMagic( $x ) { return wfSetVar( $this->mInterwikiMagic, $x ); }
2223         function setAllowExternalImages( $x ) { return wfSetVar( $this->mAllowExternalImages, $x ); }
2224         function setSkin( $x ) { return wfSetRef( $this->mSkin, $x ); }
2225         function setDateFormat( $x ) { return wfSetVar( $this->mDateFormat, $x ); }
2226         function setEditSection( $x ) { return wfSetVar( $this->mEditSection, $x ); }
2227         function setEditSectionOnRightClick( $x ) { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2228         function setNumberHeadings( $x ) { return wfSetVar( $this->mNumberHeadings, $x ); }
2229         function setShowToc( $x ) { return wfSetVar( $this->mShowToc, $x ); }
2230
2231         /* static */ function newFromUser( &$user )
2232         {
2233                 $popts = new ParserOptions;
2234                 $popts->initialiseFromUser( $user );
2235                 return $popts;
2236         }
2237
2238         function initialiseFromUser( &$userInput )
2239         {
2240                 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2241
2242                 if ( !$userInput ) {
2243                         $user = new User;
2244                         $user->setLoaded( true );
2245                 } else {
2246                         $user =& $userInput;
2247                 }
2248
2249                 $this->mUseTeX = $wgUseTeX;
2250                 $this->mUseCategoryMagic = $wgUseCategoryMagic;
2251                 $this->mUseDynamicDates = $wgUseDynamicDates;
2252                 $this->mInterwikiMagic = $wgInterwikiMagic;
2253                 $this->mAllowExternalImages = $wgAllowExternalImages;
2254                 $this->mSkin =& $user->getSkin();
2255                 $this->mDateFormat = $user->getOption( "date" );
2256                 $this->mEditSection = $user->getOption( "editsection" );
2257                 $this->mEditSectionOnRightClick = $user->getOption( "editsectiononrightclick" );
2258                 $this->mNumberHeadings = $user->getOption( "numberheadings" );
2259                 $this->mShowToc = $user->getOption( "showtoc" );
2260         }
2261
2262
2263 }
2264
2265 # Regex callbacks, used in Parser::replaceVariables
2266 function wfBraceSubstitution( $matches )
2267 {
2268         global $wgCurParser;
2269         return $wgCurParser->braceSubstitution( $matches );
2270 }
2271
2272 function wfArgSubstitution( $matches )
2273 {
2274         global $wgCurParser;
2275         return $wgCurParser->argSubstitution( $matches );
2276 }
2277
2278 ?>