includes/Parser.php

   1 <?php
   2 /**
   3  * File for Parser and related classes
   4  *
   5  * @package MediaWiki
   6  */
   7
   8 /** */
   9 require_once( 'Sanitizer.php' );
  10
  11 /**
  12  * Update this version number when the ParserOutput format
  13  * changes in an incompatible way, so the parser cache
  14  * can automatically discard old data.
  15  */
  16 define( 'MW_PARSER_VERSION', '1.5.0' );
  17
  18 /**
  19  * Variable substitution O(N^2) attack
  20  *
  21  * Without countermeasures, it would be possible to attack the parser by saving
  22  * a page filled with a large number of inclusions of large pages. The size of
  23  * the generated page would be proportional to the square of the input size.
  24  * Hence, we limit the number of inclusions of any given page, thus bringing any
  25  * attack back to O(N).
  26  */
  27
  28 define( 'MAX_INCLUDE_REPEAT', 100 );
  29 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  30
  31 define( 'RLH_FOR_UPDATE', 1 );
  32
  33 # Allowed values for $mOutputType
  34 define( 'OT_HTML', 1 );
  35 define( 'OT_WIKI', 2 );
  36 define( 'OT_MSG' , 3 );
  37
  38 # string parameter for extractTags which will cause it
  39 # to strip HTML comments in addition to regular
  40 # <XML>-style tags. This should not be anything we
  41 # may want to use in wikisyntax
  42 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  43
  44 # prefix for escaping, used in two functions at least
  45 define( 'UNIQ_PREFIX', 'NaodW29');
  46
  47 # Constants needed for external link processing
  48 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  49 define( 'HTTP_PROTOCOLS', 'http|https' );
  50 # Everything except bracket, space, or control characters
  51 define( 'EXT_LINK_URL_CLASS', '[^]<>"\\x00-\\x20\\x7F]' );
  52 # Including space
  53 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  54 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  55 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  56 define( 'EXT_LINK_BRACKETED',  '/\[(\b('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  57 define( 'EXT_IMAGE_REGEX',
  58         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  59         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  60         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  61 );
  62
  63 /**
  64  * PHP Parser
  65  *
  66  * Processes wiki markup
  67  *
  68  * <pre>
  69  * There are three main entry points into the Parser class:
  70  * parse()
  71  *   produces HTML output
  72  * preSaveTransform().
  73  *   produces altered wiki markup.
  74  * transformMsg()
  75  *   performs brace substitution on MediaWiki messages
  76  *
  77  * Globals used:
  78  *    objects:   $wgLang, $wgDateFormatter, $wgLinkCache
  79  *
  80  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  81  *
  82  * settings:
  83  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  84  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  85  *  $wgLocaltimezone
  86  *
  87  *  * only within ParserOptions
  88  * </pre>
  89  *
  90  * @package MediaWiki
  91  */
  92 class Parser
  93 {
  94         /**#@+
  95          * @access private
  96          */
  97         # Persistent:
  98         var $mTagHooks;
  99
 100         # Cleared with clearState():
 101         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
 102         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
 103
 104         # Temporary:
 105         var $mOptions, $mTitle, $mOutputType,
 106             $mTemplates,        // cache of already loaded templates, avoids
 107                                 // multiple SQL queries for the same string
 108             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 109                                 // in this path. Used for loop detection.
 110
 111         /**#@-*/
 112
 113         /**
 114          * Constructor
 115          *
 116          * @access public
 117          */
 118         function Parser() {
 119                 $this->mTemplates = array();
 120                 $this->mTemplatePath = array();
 121                 $this->mTagHooks = array();
 122                 $this->clearState();
 123         }
 124
 125         /**
 126          * Clear Parser state
 127          *
 128          * @access private
 129          */
 130         function clearState() {
 131                 $this->mOutput = new ParserOutput;
 132                 $this->mAutonumber = 0;
 133                 $this->mLastSection = '';
 134                 $this->mDTopen = false;
 135                 $this->mVariables = false;
 136                 $this->mIncludeCount = array();
 137                 $this->mStripState = array();
 138                 $this->mArgStack = array();
 139                 $this->mInPre = false;
 140         }
 141
 142         /**
 143          * First pass--just handle <nowiki> sections, pass the rest off
 144          * to internalParse() which does all the real work.
 145          *
 146          * @access private
 147          * @param string $text Text we want to parse
 148          * @param Title &$title A title object
 149          * @param array $options
 150          * @param boolean $linestart
 151          * @param boolean $clearState
 152          * @return ParserOutput a ParserOutput
 153          */
 154         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 155                 global $wgUseTidy, $wgContLang;
 156                 $fname = 'Parser::parse';
 157                 wfProfileIn( $fname );
 158
 159                 if ( $clearState ) {
 160                         $this->clearState();
 161                 }
 162
 163                 $this->mOptions = $options;
 164                 $this->mTitle =& $title;
 165                 $this->mOutputType = OT_HTML;
 166
 167                 $stripState = NULL;
 168                 global $fnord; $fnord = 1;
 169                 //$text = $this->strip( $text, $this->mStripState );
 170                 // VOODOO MAGIC FIX! Sometimes the above segfaults in PHP5.
 171                 $x =& $this->mStripState;
 172                 $text = $this->strip( $text, $x );
 173
 174                 $text = $this->internalParse( $text, $linestart );
 175
 176                 $dashReplace = array(
 177                                                  '/ - /' => "&nbsp;&ndash; ", # N dash
 178                                                  '/(?<=[0-9])-(?=[0-9])/' => "&ndash;", # N dash between numbers
 179                                                  '/ -- /' => "&nbsp;&mdash; " # M dash
 180                                                  );
 181                 $text = preg_replace( array_keys($dashReplace), array_values($dashReplace), $text );
 182
 183
 184                 $text = $this->unstrip( $text, $this->mStripState );
 185                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 186                 global $wgUseTidy;
 187                 if(!$wgUseTidy) {
 188                         $fixtags = array(
 189                                 # french spaces, last one Guillemet-left
 190                                 # only if there is something before the space
 191                                 '/(.) (?=\\?|:|;|!|\\302\\273)/' => '\\1&nbsp;\\2',
 192                                 # french spaces, Guillemet-right
 193                                 '/(\\302\\253) /' => '\\1&nbsp;',
 194                                 '/<hr *>/i' => '<hr />',
 195                                 '/<br *>/i' => '<br />',
 196                                 '/<center *>/i' => '<div class="center">',
 197                                 '/<\\/center *>/i' => '</div>',
 198                         );
 199                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 200                         $text = Sanitizer::normalizeCharReferences( $text );
 201                 } else {
 202                         $fixtags = array(
 203                                 # french spaces, last one Guillemet-left
 204                                 '/ (\\?|:|;|!|\\302\\273)/' => '&nbsp;\\1',
 205                                 # french spaces, Guillemet-right
 206                                 '/(\\302\\253) /' => '\\1&nbsp;',
 207                                 '/<center *>/i' => '<div class="center">',
 208                                 '/<\\/center *>/i' => '</div>'
 209                         );
 210                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 211                 }
 212                 # only once and last
 213                 $text = $this->doBlockLevels( $text, $linestart );
 214
 215                 $this->replaceLinkHolders( $text );
 216                 $text = $wgContLang->convert($text);
 217
 218                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 219                 if ($wgUseTidy) {
 220                         $text = Parser::tidy($text);
 221                 }
 222
 223                 $this->mOutput->setText( $text );
 224                 wfProfileOut( $fname );
 225                 return $this->mOutput;
 226         }
 227
 228         /**
 229          * Get a random string
 230          *
 231          * @access private
 232          * @static
 233          */
 234         function getRandomString() {
 235                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 236         }
 237
 238         /**
 239          * Replaces all occurrences of <$tag>content</$tag> in the text
 240          * with a random marker and returns the new text. the output parameter
 241          * $content will be an associative array filled with data on the form
 242          * $unique_marker => content.
 243          *
 244          * If $content is already set, the additional entries will be appended
 245          * If $tag is set to STRIP_COMMENTS, the function will extract
 246          * <!-- HTML comments -->
 247          *
 248          * @access private
 249          * @static
 250          */
 251         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 252                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 253                 if ( !$content ) {
 254                         $content = array( );
 255                 }
 256                 $n = 1;
 257                 $stripped = '';
 258
 259                 while ( '' != $text ) {
 260                         if($tag==STRIP_COMMENTS) {
 261                                 $p = preg_split( '/<!--/', $text, 2 );
 262                         } else {
 263                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 264                         }
 265                         $stripped .= $p[0];
 266                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 267                                 $text = '';
 268                         } else {
 269                                 if($tag==STRIP_COMMENTS) {
 270                                         $q = preg_split( '/-->/i', $p[1], 2 );
 271                                 } else {
 272                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 273                                 }
 274                                 $marker = $rnd . sprintf('%08X', $n++);
 275                                 $content[$marker] = $q[0];
 276                                 $stripped .= $marker;
 277                                 $text = $q[1];
 278                         }
 279                 }
 280                 return $stripped;
 281         }
 282
 283         /**
 284          * Strips and renders nowiki, pre, math, hiero
 285          * If $render is set, performs necessary rendering operations on plugins
 286          * Returns the text, and fills an array with data needed in unstrip()
 287          * If the $state is already a valid strip state, it adds to the state
 288          *
 289          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 290          *  will be stripped in addition to other tags. This is important
 291          *  for section editing, where these comments cause confusion when
 292          *  counting the sections in the wikisource
 293          *
 294          * @access private
 295          */
 296         function strip( $text, &$state, $stripcomments = false ) {
 297                 $render = ($this->mOutputType == OT_HTML);
 298                 $html_content = array();
 299                 $nowiki_content = array();
 300                 $math_content = array();
 301                 $pre_content = array();
 302                 $comment_content = array();
 303                 $ext_content = array();
 304                 $gallery_content = array();
 305
 306                 # Replace any instances of the placeholders
 307                 $uniq_prefix = UNIQ_PREFIX;
 308                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 309
 310                 # html
 311                 global $wgRawHtml, $wgWhitelistEdit;
 312                 if( $wgRawHtml && $wgWhitelistEdit ) {
 313                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 314                         foreach( $html_content as $marker => $content ) {
 315                                 if ($render ) {
 316                                         # Raw and unchecked for validity.
 317                                         $html_content[$marker] = $content;
 318                                 } else {
 319                                         $html_content[$marker] = '<html>'.$content.'</html>';
 320                                 }
 321                         }
 322                 }
 323
 324                 # nowiki
 325                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 326                 foreach( $nowiki_content as $marker => $content ) {
 327                         if( $render ){
 328                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 329                         } else {
 330                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 331                         }
 332                 }
 333
 334                 # math
 335                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 336                 foreach( $math_content as $marker => $content ){
 337                         if( $render ) {
 338                                 if( $this->mOptions->getUseTeX() ) {
 339                                         $math_content[$marker] = renderMath( $content );
 340                                 } else {
 341                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 342                                 }
 343                         } else {
 344                                 $math_content[$marker] = '<math>'.$content.'</math>';
 345                         }
 346                 }
 347
 348                 # pre
 349                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 350                 foreach( $pre_content as $marker => $content ){
 351                         if( $render ){
 352                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 353                         } else {
 354                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 355                         }
 356                 }
 357
 358                 # gallery
 359                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 360                 foreach( $gallery_content as $marker => $content ) {
 361                         require_once( 'ImageGallery.php' );
 362                         if ( $render ) {
 363                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 364                         } else {
 365                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 366                         }
 367                 }
 368
 369                 # Comments
 370                 if($stripcomments) {
 371                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 372                         foreach( $comment_content as $marker => $content ){
 373                                 $comment_content[$marker] = '<!--'.$content.'-->';
 374                         }
 375                 }
 376
 377                 # Extensions
 378                 foreach ( $this->mTagHooks as $tag => $callback ) {
 379                         $ext_contents[$tag] = array();
 380                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 381                         foreach( $ext_content[$tag] as $marker => $content ) {
 382                                 if ( $render ) {
 383                                         $ext_content[$tag][$marker] = $callback( $content );
 384                                 } else {
 385                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 386                                 }
 387                         }
 388                 }
 389
 390                 # Merge state with the pre-existing state, if there is one
 391                 if ( $state ) {
 392                         $state['html'] = $state['html'] + $html_content;
 393                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 394                         $state['math'] = $state['math'] + $math_content;
 395                         $state['pre'] = $state['pre'] + $pre_content;
 396                         $state['comment'] = $state['comment'] + $comment_content;
 397                         $state['gallery'] = $state['gallery'] + $gallery_content;
 398
 399                         foreach( $ext_content as $tag => $array ) {
 400                                 if ( array_key_exists( $tag, $state ) ) {
 401                                         $state[$tag] = $state[$tag] + $array;
 402                                 }
 403                         }
 404                 } else {
 405                         $state = array(
 406                           'html' => $html_content,
 407                           'nowiki' => $nowiki_content,
 408                           'math' => $math_content,
 409                           'pre' => $pre_content,
 410                           'comment' => $comment_content,
 411                           'gallery' => $gallery_content,
 412                         ) + $ext_content;
 413                 }
 414                 return $text;
 415         }
 416
 417         /**
 418          * restores pre, math, and hiero removed by strip()
 419          *
 420          * always call unstripNoWiki() after this one
 421          * @access private
 422          */
 423         function unstrip( $text, &$state ) {
 424                 # Must expand in reverse order, otherwise nested tags will be corrupted
 425                 $contentDict = end( $state );
 426                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 427                         if( key($state) != 'nowiki' && key($state) != 'html') {
 428                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 429                                         $text = str_replace( key( $contentDict ), $content, $text );
 430                                 }
 431                         }
 432                 }
 433
 434                 return $text;
 435         }
 436
 437         /**
 438          * always call this after unstrip() to preserve the order
 439          *
 440          * @access private
 441          */
 442         function unstripNoWiki( $text, &$state ) {
 443                 # Must expand in reverse order, otherwise nested tags will be corrupted
 444                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 445                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 446                 }
 447
 448                 global $wgRawHtml;
 449                 if ($wgRawHtml) {
 450                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 451                                 $text = str_replace( key( $state['html'] ), $content, $text );
 452                         }
 453                 }
 454
 455                 return $text;
 456         }
 457
 458         /**
 459          * Add an item to the strip state
 460          * Returns the unique tag which must be inserted into the stripped text
 461          * The tag will be replaced with the original text in unstrip()
 462          *
 463          * @access private
 464          */
 465         function insertStripItem( $text, &$state ) {
 466                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 467                 if ( !$state ) {
 468                         $state = array(
 469                           'html' => array(),
 470                           'nowiki' => array(),
 471                           'math' => array(),
 472                           'pre' => array()
 473                         );
 474                 }
 475                 $state['item'][$rnd] = $text;
 476                 return $rnd;
 477         }
 478
 479         /**
 480          * interface with html tidy, used if $wgUseTidy = true
 481          *
 482          * @access public
 483          * @static
 484          */
 485         function tidy ( $text ) {
 486                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 487                 global $wgInputEncoding, $wgOutputEncoding;
 488                 $fname = 'Parser::tidy';
 489                 wfProfileIn( $fname );
 490
 491                 $cleansource = '';
 492                 $opts = '';
 493                 switch(strtoupper($wgOutputEncoding)) {
 494                         case 'ISO-8859-1':
 495                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 496                                 break;
 497                         case 'UTF-8':
 498                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 499                                 break;
 500                         default:
 501                                 $opts .= ' -raw';
 502                         }
 503
 504                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 505 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 506 '<head><title>test</title></head><body>'.$text.'</body></html>';
 507                 $descriptorspec = array(
 508                         0 => array('pipe', 'r'),
 509                         1 => array('pipe', 'w'),
 510                         2 => array('file', '/dev/null', 'a')
 511                 );
 512                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 513                 if (is_resource($process)) {
 514                         fwrite($pipes[0], $wrappedtext);
 515                         fclose($pipes[0]);
 516                         while (!feof($pipes[1])) {
 517                                 $cleansource .= fgets($pipes[1], 1024);
 518                         }
 519                         fclose($pipes[1]);
 520                         $return_value = proc_close($process);
 521                 }
 522
 523                 wfProfileOut( $fname );
 524
 525                 if( $cleansource == '' && $text != '') {
 526                         wfDebug( "Tidy error detected!\n" );
 527                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 528                 } else {
 529                         return $cleansource;
 530                 }
 531         }
 532
 533         /**
 534          * parse the wiki syntax used to render tables
 535          *
 536          * @access private
 537          */
 538         function doTableStuff ( $t ) {
 539                 $fname = 'Parser::doTableStuff';
 540                 wfProfileIn( $fname );
 541
 542                 $t = explode ( "\n" , $t ) ;
 543                 $td = array () ; # Is currently a td tag open?
 544                 $ltd = array () ; # Was it TD or TH?
 545                 $tr = array () ; # Is currently a tr tag open?
 546                 $ltr = array () ; # tr attributes
 547                 $indent_level = 0; # indent level of the table
 548                 foreach ( $t AS $k => $x )
 549                 {
 550                         $x = trim ( $x ) ;
 551                         $fc = substr ( $x , 0 , 1 ) ;
 552                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 553                                 $indent_level = strlen( $matches[1] );
 554                                 $t[$k] = "\n" .
 555                                         str_repeat( '<dl><dd>', $indent_level ) .
 556                                         '<table' . Sanitizer::fixTagAttributes ( $matches[2], 'table' ) . '>' ;
 557                                 array_push ( $td , false ) ;
 558                                 array_push ( $ltd , '' ) ;
 559                                 array_push ( $tr , false ) ;
 560                                 array_push ( $ltr , '' ) ;
 561                         }
 562                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 563                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 564                                 $z = "</table>" . substr ( $x , 2) . "\n";
 565                                 $l = array_pop ( $ltd ) ;
 566                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 567                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 568                                 array_pop ( $ltr ) ;
 569                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 570                         }
 571                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 572                                 $x = substr ( $x , 1 ) ;
 573                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 574                                 $z = '' ;
 575                                 $l = array_pop ( $ltd ) ;
 576                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 577                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 578                                 array_pop ( $ltr ) ;
 579                                 $t[$k] = $z ;
 580                                 array_push ( $tr , false ) ;
 581                                 array_push ( $td , false ) ;
 582                                 array_push ( $ltd , '' ) ;
 583                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $x, 'tr' ) ) ;
 584                         }
 585                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 586                                 # $x is a table row
 587                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 588                                         $fc = '+' ;
 589                                         $x = substr ( $x , 1 ) ;
 590                                 }
 591                                 $after = substr ( $x , 1 ) ;
 592                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 593                                 $after = explode ( '||' , $after ) ;
 594                                 $t[$k] = '' ;
 595
 596                                 # Loop through each table cell
 597                                 foreach ( $after AS $theline )
 598                                 {
 599                                         $z = '' ;
 600                                         if ( $fc != '+' )
 601                                         {
 602                                                 $tra = array_pop ( $ltr ) ;
 603                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 604                                                 array_push ( $tr , true ) ;
 605                                                 array_push ( $ltr , '' ) ;
 606                                         }
 607
 608                                         $l = array_pop ( $ltd ) ;
 609                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 610                                         if ( $fc == '|' ) $l = 'td' ;
 611                                         else if ( $fc == '!' ) $l = 'th' ;
 612                                         else if ( $fc == '+' ) $l = 'caption' ;
 613                                         else $l = '' ;
 614                                         array_push ( $ltd , $l ) ;
 615
 616                                         # Cell parameters
 617                                         $y = explode ( '|' , $theline , 2 ) ;
 618                                         # Note that a '|' inside an invalid link should not
 619                                         # be mistaken as delimiting cell parameters
 620                                         if ( strpos( $y[0], '[[' ) !== false ) {
 621                                                 $y = array ($theline);
 622                                         }
 623                                         if ( count ( $y ) == 1 )
 624                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 625                                         else $y = $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($y[0], $l).">{$y[1]}" ;
 626                                         $t[$k] .= $y ;
 627                                         array_push ( $td , true ) ;
 628                                 }
 629                         }
 630                 }
 631
 632                 # Closing open td, tr && table
 633                 while ( count ( $td ) > 0 )
 634                 {
 635                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 636                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 637                         $t[] = '</table>' ;
 638                 }
 639
 640                 $t = implode ( "\n" , $t ) ;
 641                 #               $t = Sanitizer::removeHTMLtags( $t );
 642                 wfProfileOut( $fname );
 643                 return $t ;
 644         }
 645
 646         /**
 647          * Helper function for parse() that transforms wiki markup into
 648          * HTML. Only called for $mOutputType == OT_HTML.
 649          *
 650          * @access private
 651          */
 652         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 653                 global $wgContLang;
 654
 655                 $fname = 'Parser::internalParse';
 656                 wfProfileIn( $fname );
 657
 658                 $text = Sanitizer::removeHTMLtags( $text );
 659                 $text = $this->replaceVariables( $text, $args );
 660
 661                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 662
 663                 $text = $this->doHeadings( $text );
 664                 if($this->mOptions->getUseDynamicDates()) {
 665                         global $wgDateFormatter;
 666                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 667                 }
 668                 $text = $this->doAllQuotes( $text );
 669                 $text = $this->replaceInternalLinks( $text );
 670                 $text = $this->replaceExternalLinks( $text );
 671
 672                 # replaceInternalLinks may sometimes leave behind
 673                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 674                 $text = str_replace("http-noparse://","http://",$text);
 675
 676                 $text = $this->doMagicLinks( $text );
 677                 $text = $this->doTableStuff( $text );
 678                 $text = $this->formatHeadings( $text, $isMain );
 679
 680                 wfProfileOut( $fname );
 681                 return $text;
 682         }
 683
 684         /**
 685          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 686          * magic external links.
 687          *
 688          * @access private
 689          */
 690         function &doMagicLinks( &$text ) {
 691                 $text = $this->magicISBN( $text );
 692                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 693                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 694                 return $text;
 695         }
 696
 697         /**
 698          * Parse ^^ tokens and return html
 699          *
 700          * @access private
 701          */
 702         function doExponent( $text ) {
 703                 $fname = 'Parser::doExponent';
 704                 wfProfileIn( $fname );
 705                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 706                 wfProfileOut( $fname );
 707                 return $text;
 708         }
 709
 710         /**
 711          * Parse headers and return html
 712          *
 713          * @access private
 714          */
 715         function doHeadings( $text ) {
 716                 $fname = 'Parser::doHeadings';
 717                 wfProfileIn( $fname );
 718                 for ( $i = 6; $i >= 1; --$i ) {
 719                         $h = substr( '======', 0, $i );
 720                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 721                           "<h{$i}>\\1</h{$i}>\\2", $text );
 722                 }
 723                 wfProfileOut( $fname );
 724                 return $text;
 725         }
 726
 727         /**
 728          * Replace single quotes with HTML markup
 729          * @access private
 730          * @return string the altered text
 731          */
 732         function doAllQuotes( $text ) {
 733                 $fname = 'Parser::doAllQuotes';
 734                 wfProfileIn( $fname );
 735                 $outtext = '';
 736                 $lines = explode( "\n", $text );
 737                 foreach ( $lines as $line ) {
 738                         $outtext .= $this->doQuotes ( $line ) . "\n";
 739                 }
 740                 $outtext = substr($outtext, 0,-1);
 741                 wfProfileOut( $fname );
 742                 return $outtext;
 743         }
 744
 745         /**
 746          * Helper function for doAllQuotes()
 747          * @access private
 748          */
 749         function doQuotes( $text ) {
 750                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 751                 if ( count( $arr ) == 1 )
 752                         return $text;
 753                 else
 754                 {
 755                         # First, do some preliminary work. This may shift some apostrophes from
 756                         # being mark-up to being text. It also counts the number of occurrences
 757                         # of bold and italics mark-ups.
 758                         $i = 0;
 759                         $numbold = 0;
 760                         $numitalics = 0;
 761                         foreach ( $arr as $r )
 762                         {
 763                                 if ( ( $i % 2 ) == 1 )
 764                                 {
 765                                         # If there are ever four apostrophes, assume the first is supposed to
 766                                         # be text, and the remaining three constitute mark-up for bold text.
 767                                         if ( strlen( $arr[$i] ) == 4 )
 768                                         {
 769                                                 $arr[$i-1] .= "'";
 770                                                 $arr[$i] = "'''";
 771                                         }
 772                                         # If there are more than 5 apostrophes in a row, assume they're all
 773                                         # text except for the last 5.
 774                                         else if ( strlen( $arr[$i] ) > 5 )
 775                                         {
 776                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 777                                                 $arr[$i] = "'''''";
 778                                         }
 779                                         # Count the number of occurrences of bold and italics mark-ups.
 780                                         # We are not counting sequences of five apostrophes.
 781                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 782                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 783                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 784                                 }
 785                                 $i++;
 786                         }
 787
 788                         # If there is an odd number of both bold and italics, it is likely
 789                         # that one of the bold ones was meant to be an apostrophe followed
 790                         # by italics. Which one we cannot know for certain, but it is more
 791                         # likely to be one that has a single-letter word before it.
 792                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 793                         {
 794                                 $i = 0;
 795                                 $firstsingleletterword = -1;
 796                                 $firstmultiletterword = -1;
 797                                 $firstspace = -1;
 798                                 foreach ( $arr as $r )
 799                                 {
 800                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 801                                         {
 802                                                 $x1 = substr ($arr[$i-1], -1);
 803                                                 $x2 = substr ($arr[$i-1], -2, 1);
 804                                                 if ($x1 == ' ') {
 805                                                         if ($firstspace == -1) $firstspace = $i;
 806                                                 } else if ($x2 == ' ') {
 807                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 808                                                 } else {
 809                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 810                                                 }
 811                                         }
 812                                         $i++;
 813                                 }
 814
 815                                 # If there is a single-letter word, use it!
 816                                 if ($firstsingleletterword > -1)
 817                                 {
 818                                         $arr [ $firstsingleletterword ] = "''";
 819                                         $arr [ $firstsingleletterword-1 ] .= "'";
 820                                 }
 821                                 # If not, but there's a multi-letter word, use that one.
 822                                 else if ($firstmultiletterword > -1)
 823                                 {
 824                                         $arr [ $firstmultiletterword ] = "''";
 825                                         $arr [ $firstmultiletterword-1 ] .= "'";
 826                                 }
 827                                 # ... otherwise use the first one that has neither.
 828                                 # (notice that it is possible for all three to be -1 if, for example,
 829                                 # there is only one pentuple-apostrophe in the line)
 830                                 else if ($firstspace > -1)
 831                                 {
 832                                         $arr [ $firstspace ] = "''";
 833                                         $arr [ $firstspace-1 ] .= "'";
 834                                 }
 835                         }
 836
 837                         # Now let's actually convert our apostrophic mush to HTML!
 838                         $output = '';
 839                         $buffer = '';
 840                         $state = '';
 841                         $i = 0;
 842                         foreach ($arr as $r)
 843                         {
 844                                 if (($i % 2) == 0)
 845                                 {
 846                                         if ($state == 'both')
 847                                                 $buffer .= $r;
 848                                         else
 849                                                 $output .= $r;
 850                                 }
 851                                 else
 852                                 {
 853                                         if (strlen ($r) == 2)
 854                                         {
 855                                                 if ($state == 'i')
 856                                                 { $output .= '</i>'; $state = ''; }
 857                                                 else if ($state == 'bi')
 858                                                 { $output .= '</i>'; $state = 'b'; }
 859                                                 else if ($state == 'ib')
 860                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 861                                                 else if ($state == 'both')
 862                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 863                                                 else # $state can be 'b' or ''
 864                                                 { $output .= '<i>'; $state .= 'i'; }
 865                                         }
 866                                         else if (strlen ($r) == 3)
 867                                         {
 868                                                 if ($state == 'b')
 869                                                 { $output .= '</b>'; $state = ''; }
 870                                                 else if ($state == 'bi')
 871                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 872                                                 else if ($state == 'ib')
 873                                                 { $output .= '</b>'; $state = 'i'; }
 874                                                 else if ($state == 'both')
 875                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 876                                                 else # $state can be 'i' or ''
 877                                                 { $output .= '<b>'; $state .= 'b'; }
 878                                         }
 879                                         else if (strlen ($r) == 5)
 880                                         {
 881                                                 if ($state == 'b')
 882                                                 { $output .= '</b><i>'; $state = 'i'; }
 883                                                 else if ($state == 'i')
 884                                                 { $output .= '</i><b>'; $state = 'b'; }
 885                                                 else if ($state == 'bi')
 886                                                 { $output .= '</i></b>'; $state = ''; }
 887                                                 else if ($state == 'ib')
 888                                                 { $output .= '</b></i>'; $state = ''; }
 889                                                 else if ($state == 'both')
 890                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 891                                                 else # ($state == '')
 892                                                 { $buffer = ''; $state = 'both'; }
 893                                         }
 894                                 }
 895                                 $i++;
 896                         }
 897                         # Now close all remaining tags.  Notice that the order is important.
 898                         if ($state == 'b' || $state == 'ib')
 899                                 $output .= '</b>';
 900                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 901                                 $output .= '</i>';
 902                         if ($state == 'bi')
 903                                 $output .= '</b>';
 904                         if ($state == 'both')
 905                                 $output .= '<b><i>'.$buffer.'</i></b>';
 906                         return $output;
 907                 }
 908         }
 909
 910         /**
 911          * Replace external links
 912          *
 913          * Note: this is all very hackish and the order of execution matters a lot.
 914          * Make sure to run maintenance/parserTests.php if you change this code.
 915          *
 916          * @access private
 917          */
 918         function replaceExternalLinks( $text ) {
 919                 $fname = 'Parser::replaceExternalLinks';
 920                 wfProfileIn( $fname );
 921
 922                 $sk =& $this->mOptions->getSkin();
 923                 global $wgContLang;
 924                 $linktrail = $wgContLang->linkTrail();
 925
 926                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 927
 928                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 929
 930                 $i = 0;
 931                 while ( $i<count( $bits ) ) {
 932                         $url = $bits[$i++];
 933                         $protocol = $bits[$i++];
 934                         $text = $bits[$i++];
 935                         $trail = $bits[$i++];
 936
 937                         # The characters '<' and '>' (which were escaped by
 938                         # removeHTMLtags()) should not be included in
 939                         # URLs, per RFC 2396.
 940                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 941                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 942                                 $url = substr($url, 0, $m2[0][1]);
 943                         }
 944
 945                         # If the link text is an image URL, replace it with an <img> tag
 946                         # This happened by accident in the original parser, but some people used it extensively
 947                         $img = $this->maybeMakeImageLink( $text );
 948                         if ( $img !== false ) {
 949                                 $text = $img;
 950                         }
 951
 952                         $dtrail = '';
 953
 954                         # Set linktype for CSS - if URL==text, link is essentially free
 955                         $linktype = ($text == $url) ? 'free' : 'text';
 956
 957                         # No link text, e.g. [http://domain.tld/some.link]
 958                         if ( $text == '' ) {
 959                                 # Autonumber if allowed
 960                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 961                                         $text = '[' . ++$this->mAutonumber . ']';
 962                                         $linktype = 'autonumber';
 963                                 } else {
 964                                         # Otherwise just use the URL
 965                                         $text = htmlspecialchars( $url );
 966                                         $linktype = 'free';
 967                                 }
 968                         } else {
 969                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 970                                 # Check for trail
 971                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 972                                         $dtrail = $m2[1];
 973                                         $trail = $m2[2];
 974                                 }
 975                         }
 976
 977                         # Replace &amp; from obsolete syntax with &.
 978                         # All HTML entities will be escaped by makeExternalLink()
 979                         # or maybeMakeImageLink()
 980                         $url = str_replace( '&amp;', '&', $url );
 981
 982                         # Process the trail (i.e. everything after this link up until start of the next link),
 983                         # replacing any non-bracketed links
 984                         $trail = $this->replaceFreeExternalLinks( $trail );
 985
 986                         # Use the encoded URL
 987                         # This means that users can paste URLs directly into the text
 988                         # Funny characters like &ouml; aren't valid in URLs anyway
 989                         # This was changed in August 2004
 990                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
 991                 }
 992
 993                 wfProfileOut( $fname );
 994                 return $s;
 995         }
 996
 997         /**
 998          * Replace anything that looks like a URL with a link
 999          * @access private
1000          */
1001         function replaceFreeExternalLinks( $text ) {
1002                 $fname = 'Parser::replaceFreeExternalLinks';
1003                 wfProfileIn( $fname );
1004
1005                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1006                 $s = array_shift( $bits );
1007                 $i = 0;
1008
1009                 $sk =& $this->mOptions->getSkin();
1010
1011                 while ( $i < count( $bits ) ){
1012                         $protocol = $bits[$i++];
1013                         $remainder = $bits[$i++];
1014
1015                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1016                                 # Found some characters after the protocol that look promising
1017                                 $url = $protocol . $m[1];
1018                                 $trail = $m[2];
1019
1020                                 # The characters '<' and '>' (which were escaped by
1021                                 # removeHTMLtags()) should not be included in
1022                                 # URLs, per RFC 2396.
1023                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1024                                         $trail = substr($url, $m2[0][1]) . $trail;
1025                                         $url = substr($url, 0, $m2[0][1]);
1026                                 }
1027
1028                                 # Move trailing punctuation to $trail
1029                                 $sep = ',;\.:!?';
1030                                 # If there is no left bracket, then consider right brackets fair game too
1031                                 if ( strpos( $url, '(' ) === false ) {
1032                                         $sep .= ')';
1033                                 }
1034
1035                                 $numSepChars = strspn( strrev( $url ), $sep );
1036                                 if ( $numSepChars ) {
1037                                         $trail = substr( $url, -$numSepChars ) . $trail;
1038                                         $url = substr( $url, 0, -$numSepChars );
1039                                 }
1040
1041                                 # Replace &amp; from obsolete syntax with &.
1042                                 # All HTML entities will be escaped by makeExternalLink()
1043                                 # or maybeMakeImageLink()
1044                                 $url = str_replace( '&amp;', '&', $url );
1045
1046                                 # Is this an external image?
1047                                 $text = $this->maybeMakeImageLink( $url );
1048                                 if ( $text === false ) {
1049                                         # Not an image, make a link
1050                                         $text = $sk->makeExternalLink( $url, $url, true, 'free' );
1051                                 }
1052                                 $s .= $text . $trail;
1053                         } else {
1054                                 $s .= $protocol . $remainder;
1055                         }
1056                 }
1057                 wfProfileOut();
1058                 return $s;
1059         }
1060
1061         /**
1062          * make an image if it's allowed
1063          * @access private
1064          */
1065         function maybeMakeImageLink( $url ) {
1066                 $sk =& $this->mOptions->getSkin();
1067                 $text = false;
1068                 if ( $this->mOptions->getAllowExternalImages() ) {
1069                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1070                                 # Image found
1071                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1072                         }
1073                 }
1074                 return $text;
1075         }
1076
1077         /**
1078          * Process [[ ]] wikilinks
1079          *
1080          * @access private
1081          */
1082         function replaceInternalLinks( $s ) {
1083                 global $wgLang, $wgContLang, $wgLinkCache;
1084                 global $wgDisableLangConversion;
1085                 static $fname = 'Parser::replaceInternalLinks' ;
1086
1087                 wfProfileIn( $fname );
1088
1089                 wfProfileIn( $fname.'-setup' );
1090                 static $tc = FALSE;
1091                 # the % is needed to support urlencoded titles as well
1092                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1093
1094                 $sk =& $this->mOptions->getSkin();
1095                 global $wgUseOldExistenceCheck;
1096                 # "Post-parse link colour check" works only on wiki text since it's now
1097                 # in Parser. Enable it, then disable it when we're done.
1098                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1099
1100                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1101
1102                 #split the entire text string on occurences of [[
1103                 $a = explode( '[[', ' ' . $s );
1104                 #get the first element (all text up to first [[), and remove the space we added
1105                 $s = array_shift( $a );
1106                 $s = substr( $s, 1 );
1107
1108                 # Match a link having the form [[namespace:link|alternate]]trail
1109                 static $e1 = FALSE;
1110                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1111                 # Match cases where there is no "]]", which might still be images
1112                 static $e1_img = FALSE;
1113                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1114                 # Match the end of a line for a word that's not followed by whitespace,
1115                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1116                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1117
1118                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1119
1120                 if( is_null( $this->mTitle ) ) {
1121                         wfDebugDieBacktrace( 'nooo' );
1122                 }
1123                 $nottalk = !$this->mTitle->isTalkPage();
1124
1125                 if ( $useLinkPrefixExtension ) {
1126                         if ( preg_match( $e2, $s, $m ) ) {
1127                                 $first_prefix = $m[2];
1128                                 $s = $m[1];
1129                         } else {
1130                                 $first_prefix = false;
1131                         }
1132                 } else {
1133                         $prefix = '';
1134                 }
1135
1136                 $selflink = $this->mTitle->getPrefixedText();
1137                 wfProfileOut( $fname.'-setup' );
1138
1139                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1140                 $useSubpages = $this->areSubpagesAllowed();
1141
1142                 # Loop for each link
1143                 for ($k = 0; isset( $a[$k] ); $k++) {
1144                         $line = $a[$k];
1145                         if ( $useLinkPrefixExtension ) {
1146                                 wfProfileIn( $fname.'-prefixhandling' );
1147                                 if ( preg_match( $e2, $s, $m ) ) {
1148                                         $prefix = $m[2];
1149                                         $s = $m[1];
1150                                 } else {
1151                                         $prefix='';
1152                                 }
1153                                 # first link
1154                                 if($first_prefix) {
1155                                         $prefix = $first_prefix;
1156                                         $first_prefix = false;
1157                                 }
1158                                 wfProfileOut( $fname.'-prefixhandling' );
1159                         }
1160
1161                         $might_be_img = false;
1162
1163                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1164                                 $text = $m[2];
1165                                 # fix up urlencoded title texts
1166                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1167                                 $trail = $m[3];
1168                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1169                                 $might_be_img = true;
1170                                 $text = $m[2];
1171                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1172                                 $trail = "";
1173                         } else { # Invalid form; output directly
1174                                 $s .= $prefix . '[[' . $line ;
1175                                 continue;
1176                         }
1177
1178                         # Don't allow internal links to pages containing
1179                         # PROTO: where PROTO is a valid URL protocol; these
1180                         # should be external links.
1181                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1182                                 $s .= $prefix . '[[' . $line ;
1183                                 continue;
1184                         }
1185
1186                         # Make subpage if necessary
1187                         if( $useSubpages ) {
1188                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1189                         } else {
1190                                 $link = $m[1];
1191                         }
1192
1193                         $noforce = (substr($m[1], 0, 1) != ':');
1194                         if (!$noforce) {
1195                                 # Strip off leading ':'
1196                                 $link = substr($link, 1);
1197                         }
1198
1199                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1200                         if( !$nt ) {
1201                                 $s .= $prefix . '[[' . $line;
1202                                 continue;
1203                         }
1204
1205                         #check other language variants of the link
1206                         #if the article does not exist
1207                         if( $checkVariantLink
1208                             && $nt->getArticleID() == 0 ) {
1209                                 $wgContLang->findVariantLink($link, $nt);
1210                         }
1211
1212                         $ns = $nt->getNamespace();
1213                         $iw = $nt->getInterWiki();
1214
1215                         if ($might_be_img) { # if this is actually an invalid link
1216                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1217                                         $found = false;
1218                                         while (isset ($a[$k+1]) ) {
1219                                                 #look at the next 'line' to see if we can close it there
1220                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1221                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1222                                                 # the first ]] closes the inner link, the second the image
1223                                                         $found = true;
1224                                                         $text .= '[[' . $m[1];
1225                                                         $trail = $m[2];
1226                                                         break;
1227                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1228                                                         #if there's exactly one ]] that's fine, we'll keep looking
1229                                                         $text .= '[[' . $m[0];
1230                                                 } else {
1231                                                         #if $next_line is invalid too, we need look no further
1232                                                         $text .= '[[' . $next_line;
1233                                                         break;
1234                                                 }
1235                                         }
1236                                         if ( !$found ) {
1237                                                 # we couldn't find the end of this imageLink, so output it raw
1238                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1239                                                 $text = $this->replaceInternalLinks($text);
1240                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1241                                                 # note: no $trail, because without an end, there *is* no trail
1242                                                 continue;
1243                                         }
1244                                 } else { #it's not an image, so output it raw
1245                                         $s .= $prefix . '[[' . $link . '|' . $text;
1246                                         # note: no $trail, because without an end, there *is* no trail
1247                                         continue;
1248                                 }
1249                         }
1250
1251                         $wasblank = ( '' == $text );
1252                         if( $wasblank ) $text = $link;
1253
1254
1255                         # Link not escaped by : , create the various objects
1256                         if( $noforce ) {
1257
1258                                 # Interwikis
1259                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1260                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1261                                         $tmp = $prefix . $trail ;
1262                                         $s .= (trim($tmp) == '')? '': $tmp;
1263                                         continue;
1264                                 }
1265
1266                                 if ( $ns == NS_IMAGE ) {
1267                                         wfProfileIn( "$fname-image" );
1268
1269                                         # recursively parse links inside the image caption
1270                                         # actually, this will parse them in any other parameters, too,
1271                                         # but it might be hard to fix that, and it doesn't matter ATM
1272                                         $text = $this->replaceExternalLinks($text);
1273                                         $text = $this->replaceInternalLinks($text);
1274
1275                                         # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1276                                         $s .= $prefix . str_replace('http://', 'http-noparse://', $sk->makeImageLinkObj( $nt, $text ) ) . $trail;
1277                                         $wgLinkCache->addImageLinkObj( $nt );
1278
1279                                         wfProfileOut( "$fname-image" );
1280                                         continue;
1281                                 }
1282
1283                                 if ( $ns == NS_CATEGORY ) {
1284                                         wfProfileIn( "$fname-category" );
1285                                         $t = $nt->getText();
1286
1287                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1288                                         $pPLC=$sk->postParseLinkColour();
1289                                         $sk->postParseLinkColour( false );
1290                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1291                                         $sk->postParseLinkColour( $pPLC );
1292                                         $wgLinkCache->resume();
1293
1294                                         if ( $wasblank ) {
1295                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1296                                                         $sortkey = $this->mTitle->getText();
1297                                                 } else {
1298                                                         $sortkey = $this->mTitle->getPrefixedText();
1299                                                 }
1300                                         } else {
1301                                                 $sortkey = $text;
1302                                         }
1303                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1304                                         $this->mOutput->addCategoryLink( $t );
1305                                         $s .= $prefix . $trail ;
1306
1307                                         wfProfileOut( "$fname-category" );
1308                                         continue;
1309                                 }
1310                         }
1311
1312                         if( ( $nt->getPrefixedText() === $selflink ) &&
1313                             ( $nt->getFragment() === '' ) ) {
1314                                 # Self-links are handled specially; generally de-link and change to bold.
1315                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1316                                 continue;
1317                         }
1318
1319                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1320                         if( $ns == NS_MEDIA ) {
1321                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1322                                 $wgLinkCache->addImageLinkObj( $nt );
1323                                 continue;
1324                         } elseif( $ns == NS_SPECIAL ) {
1325                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1326                                 continue;
1327                         }
1328                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1329                 }
1330                 $sk->postParseLinkColour( $saveParseColour );
1331                 wfProfileOut( $fname );
1332                 return $s;
1333         }
1334
1335         /**
1336          * Return true if subpage links should be expanded on this page.
1337          * @return bool
1338          */
1339         function areSubpagesAllowed() {
1340                 # Some namespaces don't allow subpages
1341                 global $wgNamespacesWithSubpages;
1342                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1343         }
1344
1345         /**
1346          * Handle link to subpage if necessary
1347          * @param string $target the source of the link
1348          * @param string &$text the link text, modified as necessary
1349          * @return string the full name of the link
1350          * @access private
1351          */
1352         function maybeDoSubpageLink($target, &$text) {
1353                 # Valid link forms:
1354                 # Foobar -- normal
1355                 # :Foobar -- override special treatment of prefix (images, language links)
1356                 # /Foobar -- convert to CurrentPage/Foobar
1357                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1358                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1359                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1360
1361                 $fname = 'Parser::maybeDoSubpageLink';
1362                 wfProfileIn( $fname );
1363                 $ret = $target; # default return value is no change
1364
1365                 # Some namespaces don't allow subpages,
1366                 # so only perform processing if subpages are allowed
1367                 if( $this->areSubpagesAllowed() ) {
1368                         # Look at the first character
1369                         if( $target != '' && $target{0} == '/' ) {
1370                                 # / at end means we don't want the slash to be shown
1371                                 if( substr( $target, -1, 1 ) == '/' ) {
1372                                         $target = substr( $target, 1, -1 );
1373                                         $noslash = $target;
1374                                 } else {
1375                                         $noslash = substr( $target, 1 );
1376                                 }
1377
1378                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1379                                 if( '' === $text ) {
1380                                         $text = $target;
1381                                 } # this might be changed for ugliness reasons
1382                         } else {
1383                                 # check for .. subpage backlinks
1384                                 $dotdotcount = 0;
1385                                 $nodotdot = $target;
1386                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1387                                         ++$dotdotcount;
1388                                         $nodotdot = substr( $nodotdot, 3 );
1389                                 }
1390                                 if($dotdotcount > 0) {
1391                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1392                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1393                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1394                                                 # / at the end means don't show full path
1395                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1396                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1397                                                         if( '' === $text ) {
1398                                                                 $text = $nodotdot;
1399                                                         }
1400                                                 }
1401                                                 $nodotdot = trim( $nodotdot );
1402                                                 if( $nodotdot != '' ) {
1403                                                         $ret .= '/' . $nodotdot;
1404                                                 }
1405                                         }
1406                                 }
1407                         }
1408                 }
1409
1410                 wfProfileOut( $fname );
1411                 return $ret;
1412         }
1413
1414         /**#@+
1415          * Used by doBlockLevels()
1416          * @access private
1417          */
1418         /* private */ function closeParagraph() {
1419                 $result = '';
1420                 if ( '' != $this->mLastSection ) {
1421                         $result = '</' . $this->mLastSection  . ">\n";
1422                 }
1423                 $this->mInPre = false;
1424                 $this->mLastSection = '';
1425                 return $result;
1426         }
1427         # getCommon() returns the length of the longest common substring
1428         # of both arguments, starting at the beginning of both.
1429         #
1430         /* private */ function getCommon( $st1, $st2 ) {
1431                 $fl = strlen( $st1 );
1432                 $shorter = strlen( $st2 );
1433                 if ( $fl < $shorter ) { $shorter = $fl; }
1434
1435                 for ( $i = 0; $i < $shorter; ++$i ) {
1436                         if ( $st1{$i} != $st2{$i} ) { break; }
1437                 }
1438                 return $i;
1439         }
1440         # These next three functions open, continue, and close the list
1441         # element appropriate to the prefix character passed into them.
1442         #
1443         /* private */ function openList( $char ) {
1444                 $result = $this->closeParagraph();
1445
1446                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1447                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1448                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1449                 else if ( ';' == $char ) {
1450                         $result .= '<dl><dt>';
1451                         $this->mDTopen = true;
1452                 }
1453                 else { $result = '<!-- ERR 1 -->'; }
1454
1455                 return $result;
1456         }
1457
1458         /* private */ function nextItem( $char ) {
1459                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1460                 else if ( ':' == $char || ';' == $char ) {
1461                         $close = '</dd>';
1462                         if ( $this->mDTopen ) { $close = '</dt>'; }
1463                         if ( ';' == $char ) {
1464                                 $this->mDTopen = true;
1465                                 return $close . '<dt>';
1466                         } else {
1467                                 $this->mDTopen = false;
1468                                 return $close . '<dd>';
1469                         }
1470                 }
1471                 return '<!-- ERR 2 -->';
1472         }
1473
1474         /* private */ function closeList( $char ) {
1475                 if ( '*' == $char ) { $text = '</li></ul>'; }
1476                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1477                 else if ( ':' == $char ) {
1478                         if ( $this->mDTopen ) {
1479                                 $this->mDTopen = false;
1480                                 $text = '</dt></dl>';
1481                         } else {
1482                                 $text = '</dd></dl>';
1483                         }
1484                 }
1485                 else {  return '<!-- ERR 3 -->'; }
1486                 return $text."\n";
1487         }
1488         /**#@-*/
1489
1490         /**
1491          * Make lists from lines starting with ':', '*', '#', etc.
1492          *
1493          * @access private
1494          * @return string the lists rendered as HTML
1495          */
1496         function doBlockLevels( $text, $linestart ) {
1497                 $fname = 'Parser::doBlockLevels';
1498                 wfProfileIn( $fname );
1499
1500                 # Parsing through the text line by line.  The main thing
1501                 # happening here is handling of block-level elements p, pre,
1502                 # and making lists from lines starting with * # : etc.
1503                 #
1504                 $textLines = explode( "\n", $text );
1505
1506                 $lastPrefix = $output = $lastLine = '';
1507                 $this->mDTopen = $inBlockElem = false;
1508                 $prefixLength = 0;
1509                 $paragraphStack = false;
1510
1511                 if ( !$linestart ) {
1512                         $output .= array_shift( $textLines );
1513                 }
1514                 foreach ( $textLines as $oLine ) {
1515                         $lastPrefixLength = strlen( $lastPrefix );
1516                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1517                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1518                         if ( !$this->mInPre ) {
1519                                 # Multiple prefixes may abut each other for nested lists.
1520                                 $prefixLength = strspn( $oLine, '*#:;' );
1521                                 $pref = substr( $oLine, 0, $prefixLength );
1522
1523                                 # eh?
1524                                 $pref2 = str_replace( ';', ':', $pref );
1525                                 $t = substr( $oLine, $prefixLength );
1526                                 $this->mInPre = !empty($preOpenMatch);
1527                         } else {
1528                                 # Don't interpret any other prefixes in preformatted text
1529                                 $prefixLength = 0;
1530                                 $pref = $pref2 = '';
1531                                 $t = $oLine;
1532                         }
1533
1534                         # List generation
1535                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1536                                 # Same as the last item, so no need to deal with nesting or opening stuff
1537                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1538                                 $paragraphStack = false;
1539
1540                                 if ( substr( $pref, -1 ) == ';') {
1541                                         # The one nasty exception: definition lists work like this:
1542                                         # ; title : definition text
1543                                         # So we check for : in the remainder text to split up the
1544                                         # title and definition, without b0rking links.
1545                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1546                                                 $t = $t2;
1547                                                 $output .= $term . $this->nextItem( ':' );
1548                                         }
1549                                 }
1550                         } elseif( $prefixLength || $lastPrefixLength ) {
1551                                 # Either open or close a level...
1552                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1553                                 $paragraphStack = false;
1554
1555                                 while( $commonPrefixLength < $lastPrefixLength ) {
1556                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1557                                         --$lastPrefixLength;
1558                                 }
1559                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1560                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1561                                 }
1562                                 while ( $prefixLength > $commonPrefixLength ) {
1563                                         $char = substr( $pref, $commonPrefixLength, 1 );
1564                                         $output .= $this->openList( $char );
1565
1566                                         if ( ';' == $char ) {
1567                                                 # FIXME: This is dupe of code above
1568                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1569                                                         $t = $t2;
1570                                                         $output .= $term . $this->nextItem( ':' );
1571                                                 }
1572                                         }
1573                                         ++$commonPrefixLength;
1574                                 }
1575                                 $lastPrefix = $pref2;
1576                         }
1577                         if( 0 == $prefixLength ) {
1578                                 wfProfileIn( "$fname-paragraph" );
1579                                 # No prefix (not in list)--go to paragraph mode
1580                                 $uniq_prefix = UNIQ_PREFIX;
1581                                 // XXX: use a stack for nestable elements like span, table and div
1582                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1583                                 $closematch = preg_match(
1584                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1585                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1586                                 if ( $openmatch or $closematch ) {
1587                                         $paragraphStack = false;
1588                                         $output .= $this->closeParagraph();
1589                                         if($preOpenMatch and !$preCloseMatch) {
1590                                                 $this->mInPre = true;
1591                                         }
1592                                         if ( $closematch ) {
1593                                                 $inBlockElem = false;
1594                                         } else {
1595                                                 $inBlockElem = true;
1596                                         }
1597                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1598                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1599                                                 // pre
1600                                                 if ($this->mLastSection != 'pre') {
1601                                                         $paragraphStack = false;
1602                                                         $output .= $this->closeParagraph().'<pre>';
1603                                                         $this->mLastSection = 'pre';
1604                                                 }
1605                                                 $t = substr( $t, 1 );
1606                                         } else {
1607                                                 // paragraph
1608                                                 if ( '' == trim($t) ) {
1609                                                         if ( $paragraphStack ) {
1610                                                                 $output .= $paragraphStack.'<br />';
1611                                                                 $paragraphStack = false;
1612                                                                 $this->mLastSection = 'p';
1613                                                         } else {
1614                                                                 if ($this->mLastSection != 'p' ) {
1615                                                                         $output .= $this->closeParagraph();
1616                                                                         $this->mLastSection = '';
1617                                                                         $paragraphStack = '<p>';
1618                                                                 } else {
1619                                                                         $paragraphStack = '</p><p>';
1620                                                                 }
1621                                                         }
1622                                                 } else {
1623                                                         if ( $paragraphStack ) {
1624                                                                 $output .= $paragraphStack;
1625                                                                 $paragraphStack = false;
1626                                                                 $this->mLastSection = 'p';
1627                                                         } else if ($this->mLastSection != 'p') {
1628                                                                 $output .= $this->closeParagraph().'<p>';
1629                                                                 $this->mLastSection = 'p';
1630                                                         }
1631                                                 }
1632                                         }
1633                                 }
1634                                 wfProfileOut( "$fname-paragraph" );
1635                         }
1636                         if ($paragraphStack === false) {
1637                                 $output .= $t."\n";
1638                         }
1639                 }
1640                 while ( $prefixLength ) {
1641                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1642                         --$prefixLength;
1643                 }
1644                 if ( '' != $this->mLastSection ) {
1645                         $output .= '</' . $this->mLastSection . '>';
1646                         $this->mLastSection = '';
1647                 }
1648
1649                 wfProfileOut( $fname );
1650                 return $output;
1651         }
1652
1653         /**
1654          * Split up a string on ':', ignoring any occurences inside
1655          * <a>..</a> or <span>...</span>
1656          * @param string $str the string to split
1657          * @param string &$before set to everything before the ':'
1658          * @param string &$after set to everything after the ':'
1659          * return string the position of the ':', or false if none found
1660          */
1661         function findColonNoLinks($str, &$before, &$after) {
1662                 # I wonder if we should make this count all tags, not just <a>
1663                 # and <span>. That would prevent us from matching a ':' that
1664                 # comes in the middle of italics other such formatting....
1665                 # -- Wil
1666                 $fname = 'Parser::findColonNoLinks';
1667                 wfProfileIn( $fname );
1668                 $pos = 0;
1669                 do {
1670                         $colon = strpos($str, ':', $pos);
1671
1672                         if ($colon !== false) {
1673                                 $before = substr($str, 0, $colon);
1674                                 $after = substr($str, $colon + 1);
1675
1676                                 # Skip any ':' within <a> or <span> pairs
1677                                 $a = substr_count($before, '<a');
1678                                 $s = substr_count($before, '<span');
1679                                 $ca = substr_count($before, '</a>');
1680                                 $cs = substr_count($before, '</span>');
1681
1682                                 if ($a <= $ca and $s <= $cs) {
1683                                         # Tags are balanced before ':'; ok
1684                                         break;
1685                                 }
1686                                 $pos = $colon + 1;
1687                         }
1688                 } while ($colon !== false);
1689                 wfProfileOut( $fname );
1690                 return $colon;
1691         }
1692
1693         /**
1694          * Return value of a magic variable (like PAGENAME)
1695          *
1696          * @access private
1697          */
1698         function getVariableValue( $index ) {
1699                 global $wgContLang, $wgSitename, $wgServer;
1700
1701                 /**
1702                  * Some of these require message or data lookups and can be
1703                  * expensive to check many times.
1704                  */
1705                 static $varCache = array();
1706                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1707
1708                 switch ( $index ) {
1709                         case MAG_CURRENTMONTH:
1710                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1711                         case MAG_CURRENTMONTHNAME:
1712                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1713                         case MAG_CURRENTMONTHNAMEGEN:
1714                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1715                         case MAG_CURRENTMONTHABBREV:
1716                                 return $varCache[$index] = $wgContLang->getMonthAbbreviation( date('n') );
1717                         case MAG_CURRENTDAY:
1718                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1719                         case MAG_PAGENAME:
1720                                 return $this->mTitle->getText();
1721                         case MAG_PAGENAMEE:
1722                                 return $this->mTitle->getPartialURL();
1723                         case MAG_NAMESPACE:
1724                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1725                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1726                         case MAG_CURRENTDAYNAME:
1727                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1728                         case MAG_CURRENTYEAR:
1729                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1730                         case MAG_CURRENTTIME:
1731                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1732                         case MAG_CURRENTWEEK:
1733                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1734                         case MAG_CURRENTDOW:
1735                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1736                         case MAG_NUMBEROFARTICLES:
1737                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1738                         case MAG_SITENAME:
1739                                 return $wgSitename;
1740                         case MAG_SERVER:
1741                                 return $wgServer;
1742                         default:
1743                                 return NULL;
1744                 }
1745         }
1746
1747         /**
1748          * initialise the magic variables (like CURRENTMONTHNAME)
1749          *
1750          * @access private
1751          */
1752         function initialiseVariables() {
1753                 $fname = 'Parser::initialiseVariables';
1754                 wfProfileIn( $fname );
1755                 global $wgVariableIDs;
1756                 $this->mVariables = array();
1757                 foreach ( $wgVariableIDs as $id ) {
1758                         $mw =& MagicWord::get( $id );
1759                         $mw->addToArray( $this->mVariables, $id );
1760                 }
1761                 wfProfileOut( $fname );
1762         }
1763
1764         /**
1765          * Replace magic variables, templates, and template arguments
1766          * with the appropriate text. Templates are substituted recursively,
1767          * taking care to avoid infinite loops.
1768          *
1769          * Note that the substitution depends on value of $mOutputType:
1770          *  OT_WIKI: only {{subst:}} templates
1771          *  OT_MSG: only magic variables
1772          *  OT_HTML: all templates and magic variables
1773          *
1774          * @param string $tex The text to transform
1775          * @param array $args Key-value pairs representing template parameters to substitute
1776          * @access private
1777          */
1778         function replaceVariables( $text, $args = array() ) {
1779                 global $wgLang, $wgScript, $wgArticlePath;
1780
1781                 # Prevent too big inclusions
1782                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1783                         return $text;
1784                 }
1785
1786                 $fname = 'Parser::replaceVariables';
1787                 wfProfileIn( $fname );
1788
1789                 $titleChars = Title::legalChars();
1790
1791                 # This function is called recursively. To keep track of arguments we need a stack:
1792                 array_push( $this->mArgStack, $args );
1793
1794                 # Variable substitution
1795                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1796
1797                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1798                         # Argument substitution
1799                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1800                 }
1801                 # Template substitution
1802                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1803                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1804
1805                 array_pop( $this->mArgStack );
1806
1807                 wfProfileOut( $fname );
1808                 return $text;
1809         }
1810
1811         /**
1812          * Replace magic variables
1813          * @access private
1814          */
1815         function variableSubstitution( $matches ) {
1816                 $fname = 'parser::variableSubstitution';
1817                 $varname = $matches[1];
1818                 wfProfileIn( $fname );
1819                 if ( !$this->mVariables ) {
1820                         $this->initialiseVariables();
1821                 }
1822                 $skip = false;
1823                 if ( $this->mOutputType == OT_WIKI ) {
1824                         # Do only magic variables prefixed by SUBST
1825                         $mwSubst =& MagicWord::get( MAG_SUBST );
1826                         if (!$mwSubst->matchStartAndRemove( $varname ))
1827                                 $skip = true;
1828                         # Note that if we don't substitute the variable below,
1829                         # we don't remove the {{subst:}} magic word, in case
1830                         # it is a template rather than a magic variable.
1831                 }
1832                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
1833                         $id = $this->mVariables[$varname];
1834                         $text = $this->getVariableValue( $id );
1835                         $this->mOutput->mContainsOldMagic = true;
1836                 } else {
1837                         $text = $matches[0];
1838                 }
1839                 wfProfileOut( $fname );
1840                 return $text;
1841         }
1842
1843         # Split template arguments
1844         function getTemplateArgs( $argsString ) {
1845                 if ( $argsString === '' ) {
1846                         return array();
1847                 }
1848
1849                 $args = explode( '|', substr( $argsString, 1 ) );
1850
1851                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1852                 # merged with the next arg because the '|' character between belongs
1853                 # to the link syntax and not the template parameter syntax.
1854                 $argc = count($args);
1855                 $i = 0;
1856                 for ( $i = 0; $i < $argc-1; $i++ ) {
1857                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1858                                 $args[$i] .= '|'.$args[$i+1];
1859                                 array_splice($args, $i+1, 1);
1860                                 $i--;
1861                                 $argc--;
1862                         }
1863                 }
1864
1865                 return $args;
1866         }
1867
1868         /**
1869          * Return the text of a template, after recursively
1870          * replacing any variables or templates within the template.
1871          *
1872          * @param array $matches The parts of the template
1873          *  $matches[1]: the title, i.e. the part before the |
1874          *  $matches[2]: the parameters (including a leading |), if  any
1875          * @return string the text of the template
1876          * @access private
1877          */
1878         function braceSubstitution( $matches ) {
1879                 global $wgLinkCache, $wgContLang;
1880                 $fname = 'Parser::braceSubstitution';
1881                 wfProfileIn( $fname );
1882
1883                 $found = false;
1884                 $nowiki = false;
1885                 $noparse = false;
1886
1887                 $title = NULL;
1888
1889                 # Need to know if the template comes at the start of a line,
1890                 # to treat the beginning of the template like the beginning
1891                 # of a line for tables and block-level elements.
1892                 $linestart = $matches[1];
1893
1894                 # $part1 is the bit before the first |, and must contain only title characters
1895                 # $args is a list of arguments, starting from index 0, not including $part1
1896
1897                 $part1 = $matches[2];
1898                 # If the third subpattern matched anything, it will start with |
1899
1900                 $args = $this->getTemplateArgs($matches[3]);
1901                 $argc = count( $args );
1902
1903                 # Don't parse {{{}}} because that's only for template arguments
1904                 if ( $linestart === '{' ) {
1905                         $text = $matches[0];
1906                         $found = true;
1907                         $noparse = true;
1908                 }
1909
1910                 # SUBST
1911                 if ( !$found ) {
1912                         $mwSubst =& MagicWord::get( MAG_SUBST );
1913                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1914                                 # One of two possibilities is true:
1915                                 # 1) Found SUBST but not in the PST phase
1916                                 # 2) Didn't find SUBST and in the PST phase
1917                                 # In either case, return without further processing
1918                                 $text = $matches[0];
1919                                 $found = true;
1920                                 $noparse = true;
1921                         }
1922                 }
1923
1924                 # MSG, MSGNW and INT
1925                 if ( !$found ) {
1926                         # Check for MSGNW:
1927                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1928                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1929                                 $nowiki = true;
1930                         } else {
1931                                 # Remove obsolete MSG:
1932                                 $mwMsg =& MagicWord::get( MAG_MSG );
1933                                 $mwMsg->matchStartAndRemove( $part1 );
1934                         }
1935
1936                         # Check if it is an internal message
1937                         $mwInt =& MagicWord::get( MAG_INT );
1938                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1939                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1940                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1941                                         $found = true;
1942                                 }
1943                         }
1944                 }
1945
1946                 # NS
1947                 if ( !$found ) {
1948                         # Check for NS: (namespace expansion)
1949                         $mwNs = MagicWord::get( MAG_NS );
1950                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1951                                 if ( intval( $part1 ) ) {
1952                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1953                                         $found = true;
1954                                 } else {
1955                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1956                                         if ( !is_null( $index ) ) {
1957                                                 $text = $linestart . $wgContLang->getNsText( $index );
1958                                                 $found = true;
1959                                         }
1960                                 }
1961                         }
1962                 }
1963
1964                 # LOCALURL and LOCALURLE
1965                 if ( !$found ) {
1966                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1967                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1968
1969                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1970                                 $func = 'getLocalURL';
1971                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1972                                 $func = 'escapeLocalURL';
1973                         } else {
1974                                 $func = '';
1975                         }
1976
1977                         if ( $func !== '' ) {
1978                                 $title = Title::newFromText( $part1 );
1979                                 if ( !is_null( $title ) ) {
1980                                         if ( $argc > 0 ) {
1981                                                 $text = $linestart . $title->$func( $args[0] );
1982                                         } else {
1983                                                 $text = $linestart . $title->$func();
1984                                         }
1985                                         $found = true;
1986                                 }
1987                         }
1988                 }
1989
1990                 # GRAMMAR
1991                 if ( !$found && $argc == 1 ) {
1992                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1993                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1994                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
1995                                 $found = true;
1996                         }
1997                 }
1998
1999                 # Template table test
2000
2001                 # Did we encounter this template already? If yes, it is in the cache
2002                 # and we need to check for loops.
2003                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2004                         $found = true;
2005
2006                         # Infinite loop test
2007                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2008                                 $noparse = true;
2009                                 $found = true;
2010                                 $text = $linestart .
2011                                         "\{\{$part1}}" .
2012                                         '<!-- WARNING: template loop detected -->';
2013                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2014                         } else {
2015                                 # set $text to cached message.
2016                                 $text = $linestart . $this->mTemplates[$part1];
2017                         }
2018                 }
2019
2020                 # Load from database
2021                 $itcamefromthedatabase = false;
2022                 $lastPathLevel = $this->mTemplatePath;
2023                 if ( !$found ) {
2024                         $ns = NS_TEMPLATE;
2025                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2026                         if ($subpage !== '') {
2027                                 $ns = $this->mTitle->getNamespace();
2028                         }
2029                         $title = Title::newFromText( $part1, $ns );
2030                         if ( !is_null( $title ) && !$title->isExternal() ) {
2031                                 # Check for excessive inclusion
2032                                 $dbk = $title->getPrefixedDBkey();
2033                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2034                                         # This should never be reached.
2035                                         $article = new Article( $title );
2036                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2037                                         if ( $articleContent !== false ) {
2038                                                 $found = true;
2039                                                 $text = $linestart . $articleContent;
2040                                                 $itcamefromthedatabase = true;
2041                                         }
2042                                 }
2043
2044                                 # If the title is valid but undisplayable, make a link to it
2045                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2046                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2047                                         $found = true;
2048                                 }
2049
2050                                 # Template cache array insertion
2051                                 if( $found ) {
2052                                         $this->mTemplates[$part1] = $text;
2053                                 }
2054                         }
2055                 }
2056
2057                 # Recursive parsing, escaping and link table handling
2058                 # Only for HTML output
2059                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2060                         $text = wfEscapeWikiText( $text );
2061                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2062                         # Clean up argument array
2063                         $assocArgs = array();
2064                         $index = 1;
2065                         foreach( $args as $arg ) {
2066                                 $eqpos = strpos( $arg, '=' );
2067                                 if ( $eqpos === false ) {
2068                                         $assocArgs[$index++] = $arg;
2069                                 } else {
2070                                         $name = trim( substr( $arg, 0, $eqpos ) );
2071                                         $value = trim( substr( $arg, $eqpos+1 ) );
2072                                         if ( $value === false ) {
2073                                                 $value = '';
2074                                         }
2075                                         if ( $name !== false ) {
2076                                                 $assocArgs[$name] = $value;
2077                                         }
2078                                 }
2079                         }
2080
2081                         # Add a new element to the templace recursion path
2082                         $this->mTemplatePath[$part1] = 1;
2083
2084                         $text = $this->strip( $text, $this->mStripState );
2085                         $text = Sanitizer::removeHTMLtags( $text );
2086                         $text = $this->replaceVariables( $text, $assocArgs );
2087
2088                         # Resume the link cache and register the inclusion as a link
2089                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2090                                 $wgLinkCache->addLinkObj( $title );
2091                         }
2092
2093                         # If the template begins with a table or block-level
2094                         # element, it should be treated as beginning a new line.
2095                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2096                                 $text = "\n" . $text;
2097                         }
2098                 }
2099                 # Prune lower levels off the recursion check path
2100                 $this->mTemplatePath = $lastPathLevel;
2101
2102                 if ( !$found ) {
2103                         wfProfileOut( $fname );
2104                         return $matches[0];
2105                 } else {
2106                         # replace ==section headers==
2107                         # XXX this needs to go away once we have a better parser.
2108                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2109                                 if( !is_null( $title ) )
2110                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2111                                 else
2112                                         $encodedname = base64_encode("");
2113                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2114                                         PREG_SPLIT_DELIM_CAPTURE);
2115                                 $text = '';
2116                                 $nsec = 0;
2117                                 for( $i = 0; $i < count($m); $i += 2 ) {
2118                                         $text .= $m[$i];
2119                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2120                                         $hl = $m[$i + 1];
2121                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2122                                                 $text .= $hl;
2123                                                 continue;
2124                                         }
2125                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2126                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2127                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2128
2129                                         $nsec++;
2130                                 }
2131                         }
2132                 }
2133                 # Prune lower levels off the recursion check path
2134                 $this->mTemplatePath = $lastPathLevel;
2135
2136                 if ( !$found ) {
2137                         wfProfileOut( $fname );
2138                         return $matches[0];
2139                 } else {
2140                         wfProfileOut( $fname );
2141                         return $text;
2142                 }
2143         }
2144
2145         /**
2146          * Triple brace replacement -- used for template arguments
2147          * @access private
2148          */
2149         function argSubstitution( $matches ) {
2150                 $arg = trim( $matches[1] );
2151                 $text = $matches[0];
2152                 $inputArgs = end( $this->mArgStack );
2153
2154                 if ( array_key_exists( $arg, $inputArgs ) ) {
2155                         $text = $inputArgs[$arg];
2156                 }
2157
2158                 return $text;
2159         }
2160
2161         /**
2162          * Returns true if the function is allowed to include this entity
2163          * @access private
2164          */
2165         function incrementIncludeCount( $dbk ) {
2166                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2167                         $this->mIncludeCount[$dbk] = 0;
2168                 }
2169                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2170                         return true;
2171                 } else {
2172                         return false;
2173                 }
2174         }
2175
2176         /**
2177          * This function accomplishes several tasks:
2178          * 1) Auto-number headings if that option is enabled
2179          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2180          * 3) Add a Table of contents on the top for users who have enabled the option
2181          * 4) Auto-anchor headings
2182          *
2183          * It loops through all headlines, collects the necessary data, then splits up the
2184          * string and re-inserts the newly formatted headlines.
2185          *
2186          * @param string $text
2187          * @param boolean $isMain
2188          * @access private
2189          */
2190         function formatHeadings( $text, $isMain=true ) {
2191                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2192
2193                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2194                 $doShowToc = $this->mOptions->getShowToc();
2195                 $forceTocHere = false;
2196                 if( !$this->mTitle->userCanEdit() ) {
2197                         $showEditLink = 0;
2198                         $rightClickHack = 0;
2199                 } else {
2200                         $showEditLink = $this->mOptions->getEditSection();
2201                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2202                 }
2203
2204                 # Inhibit editsection links if requested in the page
2205                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2206                 if( $esw->matchAndRemove( $text ) ) {
2207                         $showEditLink = 0;
2208                 }
2209                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2210                 # do not add TOC
2211                 $mw =& MagicWord::get( MAG_NOTOC );
2212                 if( $mw->matchAndRemove( $text ) ) {
2213                         $doShowToc = 0;
2214                 }
2215
2216                 # Get all headlines for numbering them and adding funky stuff like [edit]
2217                 # links - this is for later, but we need the number of headlines right now
2218                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2219
2220                 # if there are fewer than 4 headlines in the article, do not show TOC
2221                 if( $numMatches < 4 ) {
2222                         $doShowToc = 0;
2223                 }
2224
2225                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2226                 # override above conditions and always show TOC at that place
2227
2228                 $mw =& MagicWord::get( MAG_TOC );
2229                 if($mw->match( $text ) ) {
2230                         $doShowToc = 1;
2231                         $forceTocHere = true;
2232                 } else {
2233                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2234                         # override above conditions and always show TOC above first header
2235                         $mw =& MagicWord::get( MAG_FORCETOC );
2236                         if ($mw->matchAndRemove( $text ) ) {
2237                                 $doShowToc = 1;
2238                         }
2239                 }
2240
2241                 # Never ever show TOC if no headers
2242                 if( $numMatches < 1 ) {
2243                         $doShowToc = 0;
2244                 }
2245
2246                 # We need this to perform operations on the HTML
2247                 $sk =& $this->mOptions->getSkin();
2248
2249                 # headline counter
2250                 $headlineCount = 0;
2251                 $sectionCount = 0; # headlineCount excluding template sections
2252
2253                 # Ugh .. the TOC should have neat indentation levels which can be
2254                 # passed to the skin functions. These are determined here
2255                 $toc = '';
2256                 $full = '';
2257                 $head = array();
2258                 $sublevelCount = array();
2259                 $levelCount = array();
2260                 $toclevel = 0;
2261                 $level = 0;
2262                 $prevlevel = 0;
2263                 $toclevel = 0;
2264                 $prevtoclevel = 0;
2265
2266                 foreach( $matches[3] as $headline ) {
2267                         $istemplate = 0;
2268                         $templatetitle = '';
2269                         $templatesection = 0;
2270                         $numbering = '';
2271
2272                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2273                                 $istemplate = 1;
2274                                 $templatetitle = base64_decode($mat[1]);
2275                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2276                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2277                         }
2278
2279                         if( $toclevel ) {
2280                                 $prevlevel = $level;
2281                                 $prevtoclevel = $toclevel;
2282                         }
2283                         $level = $matches[1][$headlineCount];
2284
2285                         if( $doNumberHeadings || $doShowToc ) {
2286
2287                                 if ( $level > $prevlevel ) {
2288                                         # Increase TOC level
2289                                         $toclevel++;
2290                                         $sublevelCount[$toclevel] = 0;
2291                                         $toc .= $sk->tocIndent();
2292                                 }
2293                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2294                                         # Decrease TOC level, find level to jump to
2295
2296                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2297                                                 # Can only go down to level 1
2298                                                 $toclevel = 1;
2299                                         } else {
2300                                                 for ($i = $toclevel; $i > 0; $i--) {
2301                                                         if ( $levelCount[$i] == $level ) {
2302                                                                 # Found last matching level
2303                                                                 $toclevel = $i;
2304                                                                 break;
2305                                                         }
2306                                                         elseif ( $levelCount[$i] < $level ) {
2307                                                                 # Found first matching level below current level
2308                                                                 $toclevel = $i + 1;
2309                                                                 break;
2310                                                         }
2311                                                 }
2312                                         }
2313
2314                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2315                                 }
2316                                 else {
2317                                         # No change in level, end TOC line
2318                                         $toc .= $sk->tocLineEnd();
2319                                 }
2320
2321                                 $levelCount[$toclevel] = $level;
2322
2323                                 # count number of headlines for each level
2324                                 @$sublevelCount[$toclevel]++;
2325                                 $dot = 0;
2326                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2327                                         if( !empty( $sublevelCount[$i] ) ) {
2328                                                 if( $dot ) {
2329                                                         $numbering .= '.';
2330                                                 }
2331                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2332                                                 $dot = 1;
2333                                         }
2334                                 }
2335                         }
2336
2337                         # The canonized header is a version of the header text safe to use for links
2338                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2339                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2340                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2341
2342                         # Remove link placeholders by the link text.
2343                         #     <!--LINK number-->
2344                         # turns into
2345                         #     link text with suffix
2346                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2347                                                             "\$wgLinkHolders['texts'][\$1]",
2348                                                             $canonized_headline );
2349                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2350                                                             "\$wgInterwikiLinkHolders[\$1]",
2351                                                             $canonized_headline );
2352
2353                         # strip out HTML
2354                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2355                         $tocline = trim( $canonized_headline );
2356                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2357                         $replacearray = array(
2358                                 '%3A' => ':',
2359                                 '%' => '.'
2360                         );
2361                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2362                         $refer[$headlineCount] = $canonized_headline;
2363
2364                         # count how many in assoc. array so we can track dupes in anchors
2365                         @$refers[$canonized_headline]++;
2366                         $refcount[$headlineCount]=$refers[$canonized_headline];
2367
2368                         # Don't number the heading if it is the only one (looks silly)
2369                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2370                                 # the two are different if the line contains a link
2371                                 $headline=$numbering . ' ' . $headline;
2372                         }
2373
2374                         # Create the anchor for linking from the TOC to the section
2375                         $anchor = $canonized_headline;
2376                         if($refcount[$headlineCount] > 1 ) {
2377                                 $anchor .= '_' . $refcount[$headlineCount];
2378                         }
2379                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2380                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2381                         }
2382                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2383                                 if ( empty( $head[$headlineCount] ) ) {
2384                                         $head[$headlineCount] = '';
2385                                 }
2386                                 if( $istemplate )
2387                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2388                                 else
2389                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2390                         }
2391
2392                         # Add the edit section span
2393                         if( $rightClickHack ) {
2394                                 if( $istemplate )
2395                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2396                                 else
2397                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2398                         }
2399
2400                         # give headline the correct <h#> tag
2401                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2402
2403                         $headlineCount++;
2404                         if( !$istemplate )
2405                                 $sectionCount++;
2406                 }
2407
2408                 if( $doShowToc ) {
2409                         $toclines = $headlineCount;
2410                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2411                         $toc = $sk->tocList( $toc );
2412                 }
2413
2414                 # split up and insert constructed headlines
2415
2416                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2417                 $i = 0;
2418
2419                 foreach( $blocks as $block ) {
2420                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2421                                 # This is the [edit] link that appears for the top block of text when
2422                                 # section editing is enabled
2423
2424                                 # Disabled because it broke block formatting
2425                                 # For example, a bullet point in the top line
2426                                 # $full .= $sk->editSectionLink(0);
2427                         }
2428                         $full .= $block;
2429                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2430                         # Top anchor now in skin
2431                                 $full = $full.$toc;
2432                         }
2433
2434                         if( !empty( $head[$i] ) ) {
2435                                 $full .= $head[$i];
2436                         }
2437                         $i++;
2438                 }
2439                 if($forceTocHere) {
2440                         $mw =& MagicWord::get( MAG_TOC );
2441                         return $mw->replace( $toc, $full );
2442                 } else {
2443                         return $full;
2444                 }
2445         }
2446
2447         /**
2448          * Return an HTML link for the "ISBN 123456" text
2449          * @access private
2450          */
2451         function magicISBN( $text ) {
2452                 global $wgLang;
2453                 $fname = 'Parser::magicISBN';
2454                 wfProfileIn( $fname );
2455
2456                 $a = split( 'ISBN ', ' '.$text );
2457                 if ( count ( $a ) < 2 ) {
2458                         wfProfileOut( $fname );
2459                         return $text;
2460                 }
2461                 $text = substr( array_shift( $a ), 1);
2462                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2463
2464                 foreach ( $a as $x ) {
2465                         $isbn = $blank = '' ;
2466                         while ( ' ' == $x{0} ) {
2467                                 $blank .= ' ';
2468                                 $x = substr( $x, 1 );
2469                         }
2470                         if ( $x == '' ) { # blank isbn
2471                                 $text .= "ISBN $blank";
2472                                 continue;
2473                         }
2474                         while ( strstr( $valid, $x{0} ) != false ) {
2475                                 $isbn .= $x{0};
2476                                 $x = substr( $x, 1 );
2477                         }
2478                         $num = str_replace( '-', '', $isbn );
2479                         $num = str_replace( ' ', '', $num );
2480
2481                         if ( '' == $num ) {
2482                                 $text .= "ISBN $blank$x";
2483                         } else {
2484                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2485                                 $text .= '<a href="' .
2486                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2487                                         "\" class=\"internal\">ISBN $isbn</a>";
2488                                 $text .= $x;
2489                         }
2490                 }
2491                 wfProfileOut( $fname );
2492                 return $text;
2493         }
2494
2495         /**
2496          * Return an HTML link for the "RFC 1234" text
2497          * @access private
2498          * @param string $text text to be processed
2499          */
2500         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2501                 global $wgLang;
2502
2503                 $valid = '0123456789';
2504                 $internal = false;
2505
2506                 $a = split( $keyword, ' '.$text );
2507                 if ( count ( $a ) < 2 ) {
2508                         return $text;
2509                 }
2510                 $text = substr( array_shift( $a ), 1);
2511
2512                 /* Check if keyword is preceed by [[.
2513                  * This test is made here cause of the array_shift above
2514                  * that prevent the test to be done in the foreach.
2515                  */
2516                 if ( substr( $text, -2 ) == '[[' ) {
2517                         $internal = true;
2518                 }
2519
2520                 foreach ( $a as $x ) {
2521                         /* token might be empty if we have RFC RFC 1234 */
2522                         if ( $x=='' ) {
2523                                 $text.=$keyword;
2524                                 continue;
2525                                 }
2526
2527                         $id = $blank = '' ;
2528
2529                         /** remove and save whitespaces in $blank */
2530                         while ( $x{0} == ' ' ) {
2531                                 $blank .= ' ';
2532                                 $x = substr( $x, 1 );
2533                         }
2534
2535                         /** remove and save the rfc number in $id */
2536                         while ( strstr( $valid, $x{0} ) != false ) {
2537                                 $id .= $x{0};
2538                                 $x = substr( $x, 1 );
2539                         }
2540
2541                         if ( $id == '' ) {
2542                                 /* call back stripped spaces*/
2543                                 $text .= $keyword.$blank.$x;
2544                         } elseif( $internal ) {
2545                                 /* normal link */
2546                                 $text .= $keyword.$id.$x;
2547                         } else {
2548                                 /* build the external link*/
2549                                 $url = wfmsg( $urlmsg );
2550                                 $url = str_replace( '$1', $id, $url);
2551                                 $sk =& $this->mOptions->getSkin();
2552                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2553                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2554                         }
2555
2556                         /* Check if the next RFC keyword is preceed by [[ */
2557                         $internal = ( substr($x,-2) == '[[' );
2558                 }
2559                 return $text;
2560         }
2561
2562         /**
2563          * Transform wiki markup when saving a page by doing \r\n -> \n
2564          * conversion, substitting signatures, {{subst:}} templates, etc.
2565          *
2566          * @param string $text the text to transform
2567          * @param Title &$title the Title object for the current article
2568          * @param User &$user the User object describing the current user
2569          * @param ParserOptions $options parsing options
2570          * @param bool $clearState whether to clear the parser state first
2571          * @return string the altered wiki markup
2572          * @access public
2573          */
2574         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2575                 $this->mOptions = $options;
2576                 $this->mTitle =& $title;
2577                 $this->mOutputType = OT_WIKI;
2578
2579                 if ( $clearState ) {
2580                         $this->clearState();
2581                 }
2582
2583                 $stripState = false;
2584                 $pairs = array(
2585                         "\r\n" => "\n",
2586                         );
2587                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2588                 $text = $this->strip( $text, $stripState, false );
2589                 $text = $this->pstPass2( $text, $user );
2590                 $text = $this->unstrip( $text, $stripState );
2591                 $text = $this->unstripNoWiki( $text, $stripState );
2592                 return $text;
2593         }
2594
2595         /**
2596          * Pre-save transform helper function
2597          * @access private
2598          */
2599         function pstPass2( $text, &$user ) {
2600                 global $wgLang, $wgContLang, $wgLocaltimezone;
2601
2602                 # Variable replacement
2603                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2604                 $text = $this->replaceVariables( $text );
2605
2606                 # Signatures
2607                 #
2608                 $n = $user->getName();
2609                 $k = $user->getOption( 'nickname' );
2610                 if ( '' == $k ) { $k = $n; }
2611                 if ( isset( $wgLocaltimezone ) ) {
2612                         $oldtz = getenv( 'TZ' );
2613                         putenv( 'TZ='.$wgLocaltimezone );
2614                 }
2615                 /* Note: this is an ugly timezone hack for the European wikis */
2616                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2617                   ' (' . date( 'T' ) . ')';
2618                 if ( isset( $wgLocaltimezone ) ) {
2619                         putenv( 'TZ='.$oldtzs );
2620                 }
2621
2622                 if( $user->getOption( 'fancysig' ) ) {
2623                         $sigText = $k;
2624                 } else {
2625                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2626                 }
2627                 $text = preg_replace( '/~~~~~/', $d, $text );
2628                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2629                 $text = preg_replace( '/~~~/', $sigText, $text );
2630
2631                 # Context links: [[|name]] and [[name (context)|]]
2632                 #
2633                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2634                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2635                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2636                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2637
2638                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2639                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2640                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2641                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2642                 $context = '';
2643                 $t = $this->mTitle->getText();
2644                 if ( preg_match( $conpat, $t, $m ) ) {
2645                         $context = $m[2];
2646                 }
2647                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2648                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2649                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2650
2651                 if ( '' == $context ) {
2652                         $text = preg_replace( $p2, '[[\\1]]', $text );
2653                 } else {
2654                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2655                 }
2656
2657                 # Trim trailing whitespace
2658                 # MAG_END (__END__) tag allows for trailing
2659                 # whitespace to be deliberately included
2660                 $text = rtrim( $text );
2661                 $mw =& MagicWord::get( MAG_END );
2662                 $mw->matchAndRemove( $text );
2663
2664                 return $text;
2665         }
2666
2667         /**
2668          * Set up some variables which are usually set up in parse()
2669          * so that an external function can call some class members with confidence
2670          * @access public
2671          */
2672         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2673                 $this->mTitle =& $title;
2674                 $this->mOptions = $options;
2675                 $this->mOutputType = $outputType;
2676                 if ( $clearState ) {
2677                         $this->clearState();
2678                 }
2679         }
2680
2681         /**
2682          * Transform a MediaWiki message by replacing magic variables.
2683          *
2684          * @param string $text the text to transform
2685          * @param ParserOptions $options  options
2686          * @return string the text with variables substituted
2687          * @access public
2688          */
2689         function transformMsg( $text, $options ) {
2690                 global $wgTitle;
2691                 static $executing = false;
2692
2693                 # Guard against infinite recursion
2694                 if ( $executing ) {
2695                         return $text;
2696                 }
2697                 $executing = true;
2698
2699                 $this->mTitle = $wgTitle;
2700                 $this->mOptions = $options;
2701                 $this->mOutputType = OT_MSG;
2702                 $this->clearState();
2703                 $text = $this->replaceVariables( $text );
2704
2705                 $executing = false;
2706                 return $text;
2707         }
2708
2709         /**
2710          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2711          * Callback will be called with the text within
2712          * Transform and return the text within
2713          * @access public
2714          */
2715         function setHook( $tag, $callback ) {
2716                 $oldVal = @$this->mTagHooks[$tag];
2717                 $this->mTagHooks[$tag] = $callback;
2718                 return $oldVal;
2719         }
2720
2721         /**
2722          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2723          * Placeholders created in Skin::makeLinkObj()
2724          * Returns an array of links found, indexed by PDBK:
2725          *  0 - broken
2726          *  1 - normal link
2727          *  2 - stub
2728          * $options is a bit field, RLH_FOR_UPDATE to select for update
2729          */
2730         function replaceLinkHolders( &$text, $options = 0 ) {
2731                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2732                 global $wgInterwikiLinkHolders;
2733                 global $outputReplace;
2734
2735                 if ( $wgUseOldExistenceCheck ) {
2736                         return array();
2737                 }
2738
2739                 $fname = 'Parser::replaceLinkHolders';
2740                 wfProfileIn( $fname );
2741
2742                 $pdbks = array();
2743                 $colours = array();
2744
2745                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2746                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2747                         wfProfileIn( $fname.'-check' );
2748                         $dbr =& wfGetDB( DB_SLAVE );
2749                         $page = $dbr->tableName( 'page' );
2750                         $sk = $wgUser->getSkin();
2751                         $threshold = $wgUser->getOption('stubthreshold');
2752
2753                         # Sort by namespace
2754                         asort( $wgLinkHolders['namespaces'] );
2755
2756                         # Generate query
2757                         $query = false;
2758                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2759                                 # Make title object
2760                                 $title = $wgLinkHolders['titles'][$key];
2761
2762                                 # Skip invalid entries.
2763                                 # Result will be ugly, but prevents crash.
2764                                 if ( is_null( $title ) ) {
2765                                         continue;
2766                                 }
2767                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2768
2769                                 # Check if it's in the link cache already
2770                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2771                                         $colours[$pdbk] = 1;
2772                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2773                                         $colours[$pdbk] = 0;
2774                                 } else {
2775                                         # Not in the link cache, add it to the query
2776                                         if ( !isset( $current ) ) {
2777                                                 $current = $val;
2778                                                 $query =  "SELECT page_id, page_namespace, page_title";
2779                                                 if ( $threshold > 0 ) {
2780                                                         $query .= ', page_len, page_is_redirect';
2781                                                 }
2782                                                 $query .= " FROM $page WHERE (page_namespace=$val AND page_title IN(";
2783                                         } elseif ( $current != $val ) {
2784                                                 $current = $val;
2785                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
2786                                         } else {
2787                                                 $query .= ', ';
2788                                         }
2789
2790                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2791                                 }
2792                         }
2793                         if ( $query ) {
2794                                 $query .= '))';
2795                                 if ( $options & RLH_FOR_UPDATE ) {
2796                                         $query .= ' FOR UPDATE';
2797                                 }
2798
2799                                 $res = $dbr->query( $query, $fname );
2800
2801                                 # Fetch data and form into an associative array
2802                                 # non-existent = broken
2803                                 # 1 = known
2804                                 # 2 = stub
2805                                 while ( $s = $dbr->fetchObject($res) ) {
2806                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
2807                                         $pdbk = $title->getPrefixedDBkey();
2808                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
2809
2810                                         if ( $threshold >  0 ) {
2811                                                 $size = $s->page_len;
2812                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
2813                                                         $colours[$pdbk] = 1;
2814                                                 } else {
2815                                                         $colours[$pdbk] = 2;
2816                                                 }
2817                                         } else {
2818                                                 $colours[$pdbk] = 1;
2819                                         }
2820                                 }
2821                         }
2822                         wfProfileOut( $fname.'-check' );
2823
2824                         # Construct search and replace arrays
2825                         wfProfileIn( $fname.'-construct' );
2826                         $outputReplace = array();
2827                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
2828                                 $pdbk = $pdbks[$key];
2829                                 $searchkey = '<!--LINK '.$key.'-->';
2830                                 $title = $wgLinkHolders['titles'][$key];
2831                                 if ( empty( $colours[$pdbk] ) ) {
2832                                         $wgLinkCache->addBadLink( $pdbk );
2833                                         $colours[$pdbk] = 0;
2834                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
2835                                                                         $wgLinkHolders['texts'][$key],
2836                                                                         $wgLinkHolders['queries'][$key] );
2837                                 } elseif ( $colours[$pdbk] == 1 ) {
2838                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
2839                                                                         $wgLinkHolders['texts'][$key],
2840                                                                         $wgLinkHolders['queries'][$key] );
2841                                 } elseif ( $colours[$pdbk] == 2 ) {
2842                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
2843                                                                         $wgLinkHolders['texts'][$key],
2844                                                                         $wgLinkHolders['queries'][$key] );
2845                                 }
2846                         }
2847                         wfProfileOut( $fname.'-construct' );
2848
2849                         # Do the thing
2850                         wfProfileIn( $fname.'-replace' );
2851
2852                         $text = preg_replace_callback(
2853                                 '/(<!--LINK .*?-->)/',
2854                                 "outputReplaceMatches",
2855                                 $text);
2856                         wfProfileOut( $fname.'-replace' );
2857                 }
2858
2859                 if ( !empty( $wgInterwikiLinkHolders ) ) {
2860                         wfProfileIn( $fname.'-interwiki' );
2861                         $outputReplace = $wgInterwikiLinkHolders;
2862                         $text = preg_replace_callback(
2863                                 '/<!--IWLINK (.*?)-->/',
2864                                 "outputReplaceMatches",
2865                                 $text );
2866                         wfProfileOut( $fname.'-interwiki' );
2867                 }
2868
2869                 wfProfileOut( $fname );
2870                 return $colours;
2871         }
2872
2873         /**
2874          * Renders an image gallery from a text with one line per image.
2875          * text labels may be given by using |-style alternative text. E.g.
2876          *   Image:one.jpg|The number "1"
2877          *   Image:tree.jpg|A tree
2878          * given as text will return the HTML of a gallery with two images,
2879          * labeled 'The number "1"' and
2880          * 'A tree'.
2881          */
2882         function renderImageGallery( $text ) {
2883                 global $wgLinkCache;
2884                 $ig = new ImageGallery();
2885                 $ig->setShowBytes( false );
2886                 $ig->setShowFilename( false );
2887                 $lines = explode( "\n", $text );
2888
2889                 foreach ( $lines as $line ) {
2890                         # match lines like these:
2891                         # Image:someimage.jpg|This is some image
2892                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
2893                         # Skip empty lines
2894                         if ( count( $matches ) == 0 ) {
2895                                 continue;
2896                         }
2897                         $nt = Title::newFromURL( $matches[1] );
2898                         if( is_null( $nt ) ) {
2899                                 # Bogus title. Ignore these so we don't bomb out later.
2900                                 continue;
2901                         }
2902                         if ( isset( $matches[3] ) ) {
2903                                 $label = $matches[3];
2904                         } else {
2905                                 $label = '';
2906                         }
2907
2908                         # FIXME: Use the full wiki parser and add its links
2909                         # to the page's links.
2910                         $html = $this->mOptions->mSkin->formatComment( $label );
2911
2912                         $ig->add( Image::newFromTitle( $nt ), $html );
2913                         $wgLinkCache->addImageLinkObj( $nt );
2914                 }
2915                 return $ig->toHTML();
2916         }
2917 }
2918
2919 /**
2920  * @todo document
2921  * @package MediaWiki
2922  */
2923 class ParserOutput
2924 {
2925         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2926         var $mCacheTime; # Used in ParserCache
2927         var $mVersion;   # Compatibility check
2928
2929         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2930                 $containsOldMagic = false )
2931         {
2932                 $this->mText = $text;
2933                 $this->mLanguageLinks = $languageLinks;
2934                 $this->mCategoryLinks = $categoryLinks;
2935                 $this->mContainsOldMagic = $containsOldMagic;
2936                 $this->mCacheTime = '';
2937                 $this->mVersion = MW_PARSER_VERSION;
2938         }
2939
2940         function getText()                   { return $this->mText; }
2941         function getLanguageLinks()          { return $this->mLanguageLinks; }
2942         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
2943         function getCacheTime()              { return $this->mCacheTime; }
2944         function containsOldMagic()          { return $this->mContainsOldMagic; }
2945         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
2946         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
2947         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
2948         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2949         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
2950         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
2951
2952         function merge( $other ) {
2953                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2954                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2955                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2956         }
2957
2958         /**
2959          * Return true if this cached output object predates the global or
2960          * per-article cache invalidation timestamps, or if it comes from
2961          * an incompatible older version.
2962          *
2963          * @param string $touched the affected article's last touched timestamp
2964          * @return bool
2965          * @access public
2966          */
2967         function expired( $touched ) {
2968                 global $wgCacheEpoch;
2969                 return $this->getCacheTime() <= $touched ||
2970                        $this->getCacheTime() <= $wgCacheEpoch ||
2971                        !isset( $this->mVersion ) ||
2972                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
2973         }
2974 }
2975
2976 /**
2977  * Set options of the Parser
2978  * @todo document
2979  * @package MediaWiki
2980  */
2981 class ParserOptions
2982 {
2983         # All variables are private
2984         var $mUseTeX;                    # Use texvc to expand <math> tags
2985         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2986         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2987         var $mAllowExternalImages;       # Allow external images inline
2988         var $mSkin;                      # Reference to the preferred skin
2989         var $mDateFormat;                # Date format index
2990         var $mEditSection;               # Create "edit section" links
2991         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2992         var $mNumberHeadings;            # Automatically number headings
2993         var $mShowToc;                   # Show table of contents
2994
2995         function getUseTeX()                        { return $this->mUseTeX; }
2996         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2997         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2998         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2999         function getSkin()                          { return $this->mSkin; }
3000         function getDateFormat()                    { return $this->mDateFormat; }
3001         function getEditSection()                   { return $this->mEditSection; }
3002         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3003         function getNumberHeadings()                { return $this->mNumberHeadings; }
3004         function getShowToc()                       { return $this->mShowToc; }
3005
3006         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3007         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3008         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3009         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3010         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3011         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3012         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3013         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3014         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3015
3016         function setSkin( &$x ) { $this->mSkin =& $x; }
3017
3018         /**
3019          * Get parser options
3020          * @static
3021          */
3022         function newFromUser( &$user ) {
3023                 $popts = new ParserOptions;
3024                 $popts->initialiseFromUser( $user );
3025                 return $popts;
3026         }
3027
3028         /** Get user options */
3029         function initialiseFromUser( &$userInput ) {
3030                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3031                 $fname = 'ParserOptions::initialiseFromUser';
3032                 wfProfileIn( $fname );
3033                 if ( !$userInput ) {
3034                         $user = new User;
3035                         $user->setLoaded( true );
3036                 } else {
3037                         $user =& $userInput;
3038                 }
3039
3040                 $this->mUseTeX = $wgUseTeX;
3041                 $this->mUseDynamicDates = $wgUseDynamicDates;
3042                 $this->mInterwikiMagic = $wgInterwikiMagic;
3043                 $this->mAllowExternalImages = $wgAllowExternalImages;
3044                 wfProfileIn( $fname.'-skin' );
3045                 $this->mSkin =& $user->getSkin();
3046                 wfProfileOut( $fname.'-skin' );
3047                 $this->mDateFormat = $user->getOption( 'date' );
3048                 $this->mEditSection = $user->getOption( 'editsection' );
3049                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3050                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3051                 $this->mShowToc = $user->getOption( 'showtoc' );
3052                 wfProfileOut( $fname );
3053         }
3054 }
3055
3056 /**
3057  * Callback function used by Parser::replaceLinkHolders()
3058  * to substitute link placeholders.
3059  */
3060 function &outputReplaceMatches( $matches ) {
3061         global $outputReplace;
3062         return $outputReplace[$matches[1]];
3063 }
3064
3065 /**
3066  * Return the total number of articles
3067  */
3068 function wfNumberOfArticles() {
3069         global $wgNumberOfArticles;
3070
3071         wfLoadSiteStats();
3072         return $wgNumberOfArticles;
3073 }
3074
3075 /**
3076  * Get various statistics from the database
3077  * @private
3078  */
3079 function wfLoadSiteStats() {
3080         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3081         $fname = 'wfLoadSiteStats';
3082
3083         if ( -1 != $wgNumberOfArticles ) return;
3084         $dbr =& wfGetDB( DB_SLAVE );
3085         $s = $dbr->selectRow( 'site_stats',
3086                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3087                 array( 'ss_row_id' => 1 ), $fname
3088         );
3089
3090         if ( $s === false ) {
3091                 return;
3092         } else {
3093                 $wgTotalViews = $s->ss_total_views;
3094                 $wgTotalEdits = $s->ss_total_edits;
3095                 $wgNumberOfArticles = $s->ss_good_articles;
3096         }
3097 }
3098
3099 /**
3100  * Escape html tags
3101  * Basicly replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3102  *
3103  * @param string $in Text that might contain HTML tags
3104  * @return string Escaped string
3105  */
3106 function wfEscapeHTMLTagsOnly( $in ) {
3107         return str_replace(
3108                 array( '"', '>', '<' ),
3109                 array( '&quot;', '&gt;', '&lt;' ),
3110                 $in );
3111 }
3112
3113 ?>