includes/Parser.php

   1 <?php
   2
   3 /**
   4  * File for Parser and related classes
   5  *
   6  * @package MediaWiki
   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', '1.4.0' );
  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                 $text = $this->unstrip( $text, $this->mStripState );
 176                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 177                 if(!$wgUseTidy) {
 178                         $fixtags = array(
 179                                 # french spaces, last one Guillemet-left
 180                                 # only if there is something before the space
 181                                 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 182                                 # french spaces, Guillemet-right
 183                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 184                                 '/<hr *>/i' => '<hr />',
 185                                 '/<br *>/i' => '<br />',
 186                                 '/<center *>/i' => '<div class="center">',
 187                                 '/<\\/center *>/i' => '</div>',
 188                                 # Clean up spare ampersands; note that we probably ought to be
 189                                 # more careful about named entities.
 190                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 191                         );
 192                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 193                 } else {
 194                         $fixtags = array(
 195                                 # french spaces, last one Guillemet-left
 196                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 197                                 # french spaces, Guillemet-right
 198                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 199                                 '/<center *>/i' => '<div class="center">',
 200                                 '/<\\/center *>/i' => '</div>'
 201                         );
 202                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 203                 }
 204                 # only once and last
 205                 $text = $this->doBlockLevels( $text, $linestart );
 206
 207                 $this->replaceLinkHolders( $text );
 208                 $text = $wgContLang->convert($text);
 209
 210                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 211                 global $wgUseTidy;
 212                 if ($wgUseTidy) {
 213                         $text = Parser::tidy($text);
 214                 }
 215
 216                 $this->mOutput->setText( $text );
 217                 wfProfileOut( $fname );
 218                 return $this->mOutput;
 219         }
 220
 221         /**
 222          * Get a random string
 223          *
 224          * @access private
 225          * @static
 226          */
 227         function getRandomString() {
 228                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 229         }
 230
 231         /**
 232          * Replaces all occurrences of <$tag>content</$tag> in the text
 233          * with a random marker and returns the new text. the output parameter
 234          * $content will be an associative array filled with data on the form
 235          * $unique_marker => content.
 236          *
 237          * If $content is already set, the additional entries will be appended
 238          * If $tag is set to STRIP_COMMENTS, the function will extract
 239          * <!-- HTML comments -->
 240          *
 241          * @access private
 242          * @static
 243          */
 244         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 245                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 246                 if ( !$content ) {
 247                         $content = array( );
 248                 }
 249                 $n = 1;
 250                 $stripped = '';
 251
 252                 while ( '' != $text ) {
 253                         if($tag==STRIP_COMMENTS) {
 254                                 $p = preg_split( '/<!--/i', $text, 2 );
 255                         } else {
 256                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 257                         }
 258                         $stripped .= $p[0];
 259                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 260                                 $text = '';
 261                         } else {
 262                                 if($tag==STRIP_COMMENTS) {
 263                                         $q = preg_split( '/-->/i', $p[1], 2 );
 264                                 } else {
 265                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 266                                 }
 267                                 $marker = $rnd . sprintf('%08X', $n++);
 268                                 $content[$marker] = $q[0];
 269                                 $stripped .= $marker;
 270                                 $text = $q[1];
 271                         }
 272                 }
 273                 return $stripped;
 274         }
 275
 276         /**
 277          * Strips and renders nowiki, pre, math, hiero
 278          * If $render is set, performs necessary rendering operations on plugins
 279          * Returns the text, and fills an array with data needed in unstrip()
 280          * If the $state is already a valid strip state, it adds to the state
 281          *
 282          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 283          *  will be stripped in addition to other tags. This is important
 284          *  for section editing, where these comments cause confusion when
 285          *  counting the sections in the wikisource
 286          *
 287          * @access private
 288          */
 289         function strip( $text, &$state, $stripcomments = false ) {
 290                 $render = ($this->mOutputType == OT_HTML);
 291                 $html_content = array();
 292                 $nowiki_content = array();
 293                 $math_content = array();
 294                 $pre_content = array();
 295                 $comment_content = array();
 296                 $ext_content = array();
 297                 $gallery_content = array();
 298
 299                 # Replace any instances of the placeholders
 300                 $uniq_prefix = UNIQ_PREFIX;
 301                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 302
 303                 # html
 304                 global $wgRawHtml, $wgWhitelistEdit;
 305                 if( $wgRawHtml && $wgWhitelistEdit ) {
 306                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 307                         foreach( $html_content as $marker => $content ) {
 308                                 if ($render ) {
 309                                         # Raw and unchecked for validity.
 310                                         $html_content[$marker] = $content;
 311                                 } else {
 312                                         $html_content[$marker] = '<html>'.$content.'</html>';
 313                                 }
 314                         }
 315                 }
 316
 317                 # nowiki
 318                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 319                 foreach( $nowiki_content as $marker => $content ) {
 320                         if( $render ){
 321                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 322                         } else {
 323                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 324                         }
 325                 }
 326
 327                 # math
 328                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 329                 foreach( $math_content as $marker => $content ){
 330                         if( $render ) {
 331                                 if( $this->mOptions->getUseTeX() ) {
 332                                         $math_content[$marker] = renderMath( $content );
 333                                 } else {
 334                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 335                                 }
 336                         } else {
 337                                 $math_content[$marker] = '<math>'.$content.'</math>';
 338                         }
 339                 }
 340
 341                 # pre
 342                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 343                 foreach( $pre_content as $marker => $content ){
 344                         if( $render ){
 345                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 346                         } else {
 347                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 348                         }
 349                 }
 350
 351                 # gallery
 352                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 353                 foreach( $gallery_content as $marker => $content ) {
 354                         require_once( 'ImageGallery.php' );
 355                         if ( $render ) {
 356                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 357                         } else {
 358                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 359                         }
 360                 }
 361
 362                 # Comments
 363                 if($stripcomments) {
 364                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 365                         foreach( $comment_content as $marker => $content ){
 366                                 $comment_content[$marker] = '<!--'.$content.'-->';
 367                         }
 368                 }
 369
 370                 # Extensions
 371                 foreach ( $this->mTagHooks as $tag => $callback ) {
 372                         $ext_contents[$tag] = array();
 373                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 374                         foreach( $ext_content[$tag] as $marker => $content ) {
 375                                 if ( $render ) {
 376                                         $ext_content[$tag][$marker] = $callback( $content );
 377                                 } else {
 378                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 379                                 }
 380                         }
 381                 }
 382
 383                 # Merge state with the pre-existing state, if there is one
 384                 if ( $state ) {
 385                         $state['html'] = $state['html'] + $html_content;
 386                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 387                         $state['math'] = $state['math'] + $math_content;
 388                         $state['pre'] = $state['pre'] + $pre_content;
 389                         $state['comment'] = $state['comment'] + $comment_content;
 390                         $state['gallery'] = $state['gallery'] + $gallery_content;
 391
 392                         foreach( $ext_content as $tag => $array ) {
 393                                 if ( array_key_exists( $tag, $state ) ) {
 394                                         $state[$tag] = $state[$tag] + $array;
 395                                 }
 396                         }
 397                 } else {
 398                         $state = array(
 399                           'html' => $html_content,
 400                           'nowiki' => $nowiki_content,
 401                           'math' => $math_content,
 402                           'pre' => $pre_content,
 403                           'comment' => $comment_content,
 404                           'gallery' => $gallery_content,
 405                         ) + $ext_content;
 406                 }
 407                 return $text;
 408         }
 409
 410         /**
 411          * restores pre, math, and hiero removed by strip()
 412          *
 413          * always call unstripNoWiki() after this one
 414          * @access private
 415          */
 416         function unstrip( $text, &$state ) {
 417                 # Must expand in reverse order, otherwise nested tags will be corrupted
 418                 $contentDict = end( $state );
 419                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 420                         if( key($state) != 'nowiki' && key($state) != 'html') {
 421                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 422                                         $text = str_replace( key( $contentDict ), $content, $text );
 423                                 }
 424                         }
 425                 }
 426
 427                 return $text;
 428         }
 429
 430         /**
 431          * always call this after unstrip() to preserve the order
 432          *
 433          * @access private
 434          */
 435         function unstripNoWiki( $text, &$state ) {
 436                 # Must expand in reverse order, otherwise nested tags will be corrupted
 437                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 438                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 439                 }
 440
 441                 global $wgRawHtml;
 442                 if ($wgRawHtml) {
 443                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 444                                 $text = str_replace( key( $state['html'] ), $content, $text );
 445                         }
 446                 }
 447
 448                 return $text;
 449         }
 450
 451         /**
 452          * Add an item to the strip state
 453          * Returns the unique tag which must be inserted into the stripped text
 454          * The tag will be replaced with the original text in unstrip()
 455          *
 456          * @access private
 457          */
 458         function insertStripItem( $text, &$state ) {
 459                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 460                 if ( !$state ) {
 461                         $state = array(
 462                           'html' => array(),
 463                           'nowiki' => array(),
 464                           'math' => array(),
 465                           'pre' => array()
 466                         );
 467                 }
 468                 $state['item'][$rnd] = $text;
 469                 return $rnd;
 470         }
 471
 472         /**
 473          * interface with html tidy, used if $wgUseTidy = true
 474          *
 475          * @access public
 476          * @static
 477          */
 478         function tidy ( $text ) {
 479                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 480                 global $wgInputEncoding, $wgOutputEncoding;
 481                 $fname = 'Parser::tidy';
 482                 wfProfileIn( $fname );
 483
 484                 $cleansource = '';
 485                 $opts = '';
 486                 switch(strtoupper($wgOutputEncoding)) {
 487                         case 'ISO-8859-1':
 488                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 489                                 break;
 490                         case 'UTF-8':
 491                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 492                                 break;
 493                         default:
 494                                 $opts .= ' -raw';
 495                         }
 496
 497                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 498 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 499 '<head><title>test</title></head><body>'.$text.'</body></html>';
 500                 $descriptorspec = array(
 501                         0 => array('pipe', 'r'),
 502                         1 => array('pipe', 'w'),
 503                         2 => array('file', '/dev/null', 'a')
 504                 );
 505                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 506                 if (is_resource($process)) {
 507                         fwrite($pipes[0], $wrappedtext);
 508                         fclose($pipes[0]);
 509                         while (!feof($pipes[1])) {
 510                                 $cleansource .= fgets($pipes[1], 1024);
 511                         }
 512                         fclose($pipes[1]);
 513                         $return_value = proc_close($process);
 514                 }
 515
 516                 wfProfileOut( $fname );
 517
 518                 if( $cleansource == '' && $text != '') {
 519                         wfDebug( "Tidy error detected!\n" );
 520                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 521                 } else {
 522                         return $cleansource;
 523                 }
 524         }
 525
 526         /**
 527          * parse the wiki syntax used to render tables
 528          *
 529          * @access private
 530          */
 531         function doTableStuff ( $t ) {
 532                 $fname = 'Parser::doTableStuff';
 533                 wfProfileIn( $fname );
 534
 535                 $t = explode ( "\n" , $t ) ;
 536                 $td = array () ; # Is currently a td tag open?
 537                 $ltd = array () ; # Was it TD or TH?
 538                 $tr = array () ; # Is currently a tr tag open?
 539                 $ltr = array () ; # tr attributes
 540                 $indent_level = 0; # indent level of the table
 541                 foreach ( $t AS $k => $x )
 542                 {
 543                         $x = trim ( $x ) ;
 544                         $fc = substr ( $x , 0 , 1 ) ;
 545                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 546                                 $indent_level = strlen( $matches[1] );
 547                                 $t[$k] = "\n" .
 548                                         str_repeat( '<dl><dd>', $indent_level ) .
 549                                         '<table' . Sanitizer::fixTagAttributes ( $matches[2], 'table' ) . '>' ;
 550                                 array_push ( $td , false ) ;
 551                                 array_push ( $ltd , '' ) ;
 552                                 array_push ( $tr , false ) ;
 553                                 array_push ( $ltr , '' ) ;
 554                         }
 555                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 556                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 557                                 $z = "</table>\n" ;
 558                                 $l = array_pop ( $ltd ) ;
 559                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 560                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 561                                 array_pop ( $ltr ) ;
 562                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 563                         }
 564                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 565                                 $x = substr ( $x , 1 ) ;
 566                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 567                                 $z = '' ;
 568                                 $l = array_pop ( $ltd ) ;
 569                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 570                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 571                                 array_pop ( $ltr ) ;
 572                                 $t[$k] = $z ;
 573                                 array_push ( $tr , false ) ;
 574                                 array_push ( $td , false ) ;
 575                                 array_push ( $ltd , '' ) ;
 576                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $x, 'tr' ) ) ;
 577                         }
 578                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 579                                 # $x is a table row
 580                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 581                                         $fc = '+' ;
 582                                         $x = substr ( $x , 1 ) ;
 583                                 }
 584                                 $after = substr ( $x , 1 ) ;
 585                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 586                                 $after = explode ( '||' , $after ) ;
 587                                 $t[$k] = '' ;
 588
 589                                 # Loop through each table cell
 590                                 foreach ( $after AS $theline )
 591                                 {
 592                                         $z = '' ;
 593                                         if ( $fc != '+' )
 594                                         {
 595                                                 $tra = array_pop ( $ltr ) ;
 596                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 597                                                 array_push ( $tr , true ) ;
 598                                                 array_push ( $ltr , '' ) ;
 599                                         }
 600
 601                                         $l = array_pop ( $ltd ) ;
 602                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 603                                         if ( $fc == '|' ) $l = 'td' ;
 604                                         else if ( $fc == '!' ) $l = 'th' ;
 605                                         else if ( $fc == '+' ) $l = 'caption' ;
 606                                         else $l = '' ;
 607                                         array_push ( $ltd , $l ) ;
 608
 609                                         # Cell parameters
 610                                         $y = explode ( '|' , $theline , 2 ) ;
 611                                         # Note that a '|' inside an invalid link should not
 612                                         # be mistaken as delimiting cell parameters
 613                                         if ( strpos( $y[0], '[[' ) !== false ) {
 614                                                 $y = array ($theline);
 615                                         }
 616                                         if ( count ( $y ) == 1 )
 617                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 618                                         else $y = $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($y[0], $l).">{$y[1]}" ;
 619                                         $t[$k] .= $y ;
 620                                         array_push ( $td , true ) ;
 621                                 }
 622                         }
 623                 }
 624
 625                 # Closing open td, tr && table
 626                 while ( count ( $td ) > 0 )
 627                 {
 628                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 629                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 630                         $t[] = '</table>' ;
 631                 }
 632
 633                 $t = implode ( "\n" , $t ) ;
 634                 #               $t = Sanitizer::removeHTMLtags( $t );
 635                 wfProfileOut( $fname );
 636                 return $t ;
 637         }
 638
 639         /**
 640          * Helper function for parse() that transforms wiki markup into
 641          * HTML. Only called for $mOutputType == OT_HTML.
 642          *
 643          * @access private
 644          */
 645         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 646                 global $wgContLang;
 647
 648                 $fname = 'Parser::internalParse';
 649                 wfProfileIn( $fname );
 650
 651                 $text = Sanitizer::removeHTMLtags( $text );
 652                 $text = $this->replaceVariables( $text, $args );
 653
 654                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 655
 656                 $text = $this->doHeadings( $text );
 657                 if($this->mOptions->getUseDynamicDates()) {
 658                         global $wgDateFormatter;
 659                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 660                 }
 661                 $text = $this->doAllQuotes( $text );
 662                 $text = $this->replaceInternalLinks( $text );
 663                 $text = $this->replaceExternalLinks( $text );
 664
 665                 # replaceInternalLinks may sometimes leave behind
 666                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 667                 $text = str_replace("http-noparse://","http://",$text);
 668
 669                 $text = $this->doMagicLinks( $text );
 670                 $text = $this->doTableStuff( $text );
 671                 $text = $this->formatHeadings( $text, $isMain );
 672
 673                 wfProfileOut( $fname );
 674                 return $text;
 675         }
 676
 677         /**
 678          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 679          * magic external links.
 680          *
 681          * @access private
 682          */
 683         function &doMagicLinks( &$text ) {
 684                 global $wgUseGeoMode;
 685                 $text = $this->magicISBN( $text );
 686                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 687                         $text = $this->magicGEO( $text );
 688                 }
 689                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 690                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 691                 return $text;
 692         }
 693
 694         /**
 695          * Parse ^^ tokens and return html
 696          *
 697          * @access private
 698          */
 699         function doExponent( $text ) {
 700                 $fname = 'Parser::doExponent';
 701                 wfProfileIn( $fname );
 702                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 703                 wfProfileOut( $fname );
 704                 return $text;
 705         }
 706
 707         /**
 708          * Parse headers and return html
 709          *
 710          * @access private
 711          */
 712         function doHeadings( $text ) {
 713                 $fname = 'Parser::doHeadings';
 714                 wfProfileIn( $fname );
 715                 for ( $i = 6; $i >= 1; --$i ) {
 716                         $h = substr( '======', 0, $i );
 717                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 718                           "<h{$i}>\\1</h{$i}>\\2", $text );
 719                 }
 720                 wfProfileOut( $fname );
 721                 return $text;
 722         }
 723
 724         /**
 725          * Replace single quotes with HTML markup
 726          * @access private
 727          * @return string the altered text
 728          */
 729         function doAllQuotes( $text ) {
 730                 $fname = 'Parser::doAllQuotes';
 731                 wfProfileIn( $fname );
 732                 $outtext = '';
 733                 $lines = explode( "\n", $text );
 734                 foreach ( $lines as $line ) {
 735                         $outtext .= $this->doQuotes ( $line ) . "\n";
 736                 }
 737                 $outtext = substr($outtext, 0,-1);
 738                 wfProfileOut( $fname );
 739                 return $outtext;
 740         }
 741
 742         /**
 743          * Helper function for doAllQuotes()
 744          * @access private
 745          */
 746         function doQuotes( $text ) {
 747                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 748                 if ( count( $arr ) == 1 )
 749                         return $text;
 750                 else
 751                 {
 752                         # First, do some preliminary work. This may shift some apostrophes from
 753                         # being mark-up to being text. It also counts the number of occurrences
 754                         # of bold and italics mark-ups.
 755                         $i = 0;
 756                         $numbold = 0;
 757                         $numitalics = 0;
 758                         foreach ( $arr as $r )
 759                         {
 760                                 if ( ( $i % 2 ) == 1 )
 761                                 {
 762                                         # If there are ever four apostrophes, assume the first is supposed to
 763                                         # be text, and the remaining three constitute mark-up for bold text.
 764                                         if ( strlen( $arr[$i] ) == 4 )
 765                                         {
 766                                                 $arr[$i-1] .= "'";
 767                                                 $arr[$i] = "'''";
 768                                         }
 769                                         # If there are more than 5 apostrophes in a row, assume they're all
 770                                         # text except for the last 5.
 771                                         else if ( strlen( $arr[$i] ) > 5 )
 772                                         {
 773                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 774                                                 $arr[$i] = "'''''";
 775                                         }
 776                                         # Count the number of occurrences of bold and italics mark-ups.
 777                                         # We are not counting sequences of five apostrophes.
 778                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 779                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 780                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 781                                 }
 782                                 $i++;
 783                         }
 784
 785                         # If there is an odd number of both bold and italics, it is likely
 786                         # that one of the bold ones was meant to be an apostrophe followed
 787                         # by italics. Which one we cannot know for certain, but it is more
 788                         # likely to be one that has a single-letter word before it.
 789                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 790                         {
 791                                 $i = 0;
 792                                 $firstsingleletterword = -1;
 793                                 $firstmultiletterword = -1;
 794                                 $firstspace = -1;
 795                                 foreach ( $arr as $r )
 796                                 {
 797                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 798                                         {
 799                                                 $x1 = substr ($arr[$i-1], -1);
 800                                                 $x2 = substr ($arr[$i-1], -2, 1);
 801                                                 if ($x1 == ' ') {
 802                                                         if ($firstspace == -1) $firstspace = $i;
 803                                                 } else if ($x2 == ' ') {
 804                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 805                                                 } else {
 806                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 807                                                 }
 808                                         }
 809                                         $i++;
 810                                 }
 811
 812                                 # If there is a single-letter word, use it!
 813                                 if ($firstsingleletterword > -1)
 814                                 {
 815                                         $arr [ $firstsingleletterword ] = "''";
 816                                         $arr [ $firstsingleletterword-1 ] .= "'";
 817                                 }
 818                                 # If not, but there's a multi-letter word, use that one.
 819                                 else if ($firstmultiletterword > -1)
 820                                 {
 821                                         $arr [ $firstmultiletterword ] = "''";
 822                                         $arr [ $firstmultiletterword-1 ] .= "'";
 823                                 }
 824                                 # ... otherwise use the first one that has neither.
 825                                 # (notice that it is possible for all three to be -1 if, for example,
 826                                 # there is only one pentuple-apostrophe in the line)
 827                                 else if ($firstspace > -1)
 828                                 {
 829                                         $arr [ $firstspace ] = "''";
 830                                         $arr [ $firstspace-1 ] .= "'";
 831                                 }
 832                         }
 833
 834                         # Now let's actually convert our apostrophic mush to HTML!
 835                         $output = '';
 836                         $buffer = '';
 837                         $state = '';
 838                         $i = 0;
 839                         foreach ($arr as $r)
 840                         {
 841                                 if (($i % 2) == 0)
 842                                 {
 843                                         if ($state == 'both')
 844                                                 $buffer .= $r;
 845                                         else
 846                                                 $output .= $r;
 847                                 }
 848                                 else
 849                                 {
 850                                         if (strlen ($r) == 2)
 851                                         {
 852                                                 if ($state == 'i')
 853                                                 { $output .= '</i>'; $state = ''; }
 854                                                 else if ($state == 'bi')
 855                                                 { $output .= '</i>'; $state = 'b'; }
 856                                                 else if ($state == 'ib')
 857                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 858                                                 else if ($state == 'both')
 859                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 860                                                 else # $state can be 'b' or ''
 861                                                 { $output .= '<i>'; $state .= 'i'; }
 862                                         }
 863                                         else if (strlen ($r) == 3)
 864                                         {
 865                                                 if ($state == 'b')
 866                                                 { $output .= '</b>'; $state = ''; }
 867                                                 else if ($state == 'bi')
 868                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 869                                                 else if ($state == 'ib')
 870                                                 { $output .= '</b>'; $state = 'i'; }
 871                                                 else if ($state == 'both')
 872                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 873                                                 else # $state can be 'i' or ''
 874                                                 { $output .= '<b>'; $state .= 'b'; }
 875                                         }
 876                                         else if (strlen ($r) == 5)
 877                                         {
 878                                                 if ($state == 'b')
 879                                                 { $output .= '</b><i>'; $state = 'i'; }
 880                                                 else if ($state == 'i')
 881                                                 { $output .= '</i><b>'; $state = 'b'; }
 882                                                 else if ($state == 'bi')
 883                                                 { $output .= '</i></b>'; $state = ''; }
 884                                                 else if ($state == 'ib')
 885                                                 { $output .= '</b></i>'; $state = ''; }
 886                                                 else if ($state == 'both')
 887                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 888                                                 else # ($state == '')
 889                                                 { $buffer = ''; $state = 'both'; }
 890                                         }
 891                                 }
 892                                 $i++;
 893                         }
 894                         # Now close all remaining tags.  Notice that the order is important.
 895                         if ($state == 'b' || $state == 'ib')
 896                                 $output .= '</b>';
 897                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 898                                 $output .= '</i>';
 899                         if ($state == 'bi')
 900                                 $output .= '</b>';
 901                         if ($state == 'both')
 902                                 $output .= '<b><i>'.$buffer.'</i></b>';
 903                         return $output;
 904                 }
 905         }
 906
 907         /**
 908          * Replace external links
 909          *
 910          * Note: this is all very hackish and the order of execution matters a lot.
 911          * Make sure to run maintenance/parserTests.php if you change this code.
 912          *
 913          * @access private
 914          */
 915         function replaceExternalLinks( $text ) {
 916                 $fname = 'Parser::replaceExternalLinks';
 917                 wfProfileIn( $fname );
 918
 919                 $sk =& $this->mOptions->getSkin();
 920                 global $wgContLang;
 921                 $linktrail = $wgContLang->linkTrail();
 922
 923                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 924
 925                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 926
 927                 $i = 0;
 928                 while ( $i<count( $bits ) ) {
 929                         $url = $bits[$i++];
 930                         $protocol = $bits[$i++];
 931                         $text = $bits[$i++];
 932                         $trail = $bits[$i++];
 933
 934                         # The characters '<' and '>' (which were escaped by
 935                         # removeHTMLtags()) should not be included in
 936                         # URLs, per RFC 2396.
 937                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 938                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 939                                 $url = substr($url, 0, $m2[0][1]);
 940                         }
 941
 942                         # If the link text is an image URL, replace it with an <img> tag
 943                         # This happened by accident in the original parser, but some people used it extensively
 944                         $img = $this->maybeMakeImageLink( $text );
 945                         if ( $img !== false ) {
 946                                 $text = $img;
 947                         }
 948
 949                         $dtrail = '';
 950
 951                         # Set linktype for CSS - if URL==text, link is essentially free
 952                         $linktype = ($text == $url) ? 'free' : 'text';
 953
 954                         # No link text, e.g. [http://domain.tld/some.link]
 955                         if ( $text == '' ) {
 956                                 # Autonumber if allowed
 957                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 958                                         $text = '[' . ++$this->mAutonumber . ']';
 959                                         $linktype = 'autonumber';
 960                                 } else {
 961                                         # Otherwise just use the URL
 962                                         $text = htmlspecialchars( $url );
 963                                         $linktype = 'free';
 964                                 }
 965                         } else {
 966                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 967                                 # Check for trail
 968                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 969                                         $dtrail = $m2[1];
 970                                         $trail = $m2[2];
 971                                 }
 972                         }
 973
 974                         # Replace &amp; from obsolete syntax with &.
 975                         # All HTML entities will be escaped by makeExternalLink()
 976                         # or maybeMakeImageLink()
 977                         $url = str_replace( '&amp;', '&', $url );
 978
 979                         # Process the trail (i.e. everything after this link up until start of the next link),
 980                         # replacing any non-bracketed links
 981                         $trail = $this->replaceFreeExternalLinks( $trail );
 982
 983                         # Use the encoded URL
 984                         # This means that users can paste URLs directly into the text
 985                         # Funny characters like &ouml; aren't valid in URLs anyway
 986                         # This was changed in August 2004
 987                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
 988                 }
 989
 990                 wfProfileOut( $fname );
 991                 return $s;
 992         }
 993
 994         /**
 995          * Replace anything that looks like a URL with a link
 996          * @access private
 997          */
 998         function replaceFreeExternalLinks( $text ) {
 999                 $fname = 'Parser::replaceFreeExternalLinks';
1000                 wfProfileIn( $fname );
1001
1002                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1003                 $s = array_shift( $bits );
1004                 $i = 0;
1005
1006                 $sk =& $this->mOptions->getSkin();
1007
1008                 while ( $i < count( $bits ) ){
1009                         $protocol = $bits[$i++];
1010                         $remainder = $bits[$i++];
1011
1012                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1013                                 # Found some characters after the protocol that look promising
1014                                 $url = $protocol . $m[1];
1015                                 $trail = $m[2];
1016
1017                                 # The characters '<' and '>' (which were escaped by
1018                                 # removeHTMLtags()) should not be included in
1019                                 # URLs, per RFC 2396.
1020                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1021                                         $trail = substr($url, $m2[0][1]) . $trail;
1022                                         $url = substr($url, 0, $m2[0][1]);
1023                                 }
1024
1025                                 # Move trailing punctuation to $trail
1026                                 $sep = ',;\.:!?';
1027                                 # If there is no left bracket, then consider right brackets fair game too
1028                                 if ( strpos( $url, '(' ) === false ) {
1029                                         $sep .= ')';
1030                                 }
1031
1032                                 $numSepChars = strspn( strrev( $url ), $sep );
1033                                 if ( $numSepChars ) {
1034                                         $trail = substr( $url, -$numSepChars ) . $trail;
1035                                         $url = substr( $url, 0, -$numSepChars );
1036                                 }
1037
1038                                 # Replace &amp; from obsolete syntax with &.
1039                                 # All HTML entities will be escaped by makeExternalLink()
1040                                 # or maybeMakeImageLink()
1041                                 $url = str_replace( '&amp;', '&', $url );
1042
1043                                 # Is this an external image?
1044                                 $text = $this->maybeMakeImageLink( $url );
1045                                 if ( $text === false ) {
1046                                         # Not an image, make a link
1047                                         $text = $sk->makeExternalLink( $url, $url, true, 'free' );
1048                                 }
1049                                 $s .= $text . $trail;
1050                         } else {
1051                                 $s .= $protocol . $remainder;
1052                         }
1053                 }
1054                 wfProfileOut();
1055                 return $s;
1056         }
1057
1058         /**
1059          * make an image if it's allowed
1060          * @access private
1061          */
1062         function maybeMakeImageLink( $url ) {
1063                 $sk =& $this->mOptions->getSkin();
1064                 $text = false;
1065                 if ( $this->mOptions->getAllowExternalImages() ) {
1066                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1067                                 # Image found
1068                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1069                         }
1070                 }
1071                 return $text;
1072         }
1073
1074         /**
1075          * Process [[ ]] wikilinks
1076          *
1077          * @access private
1078          */
1079         function replaceInternalLinks( $s ) {
1080                 global $wgLang, $wgContLang, $wgLinkCache;
1081                 global $wgDisableLangConversion;
1082                 static $fname = 'Parser::replaceInternalLinks' ;
1083
1084                 wfProfileIn( $fname );
1085
1086                 wfProfileIn( $fname.'-setup' );
1087                 static $tc = FALSE;
1088                 # the % is needed to support urlencoded titles as well
1089                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1090
1091                 $sk =& $this->mOptions->getSkin();
1092                 global $wgUseOldExistenceCheck;
1093                 # "Post-parse link colour check" works only on wiki text since it's now
1094                 # in Parser. Enable it, then disable it when we're done.
1095                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1096
1097                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1098
1099                 #split the entire text string on occurences of [[
1100                 $a = explode( '[[', ' ' . $s );
1101                 #get the first element (all text up to first [[), and remove the space we added
1102                 $s = array_shift( $a );
1103                 $s = substr( $s, 1 );
1104
1105                 # Match a link having the form [[namespace:link|alternate]]trail
1106                 static $e1 = FALSE;
1107                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1108                 # Match cases where there is no "]]", which might still be images
1109                 static $e1_img = FALSE;
1110                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1111                 # Match the end of a line for a word that's not followed by whitespace,
1112                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1113                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1114
1115                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1116
1117                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1118
1119                 if ( $useLinkPrefixExtension ) {
1120                         if ( preg_match( $e2, $s, $m ) ) {
1121                                 $first_prefix = $m[2];
1122                                 $s = $m[1];
1123                         } else {
1124                                 $first_prefix = false;
1125                         }
1126                 } else {
1127                         $prefix = '';
1128                 }
1129
1130                 $selflink = $this->mTitle->getPrefixedText();
1131                 wfProfileOut( $fname.'-setup' );
1132
1133                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1134                 $useSubpages = $this->areSubpagesAllowed();
1135
1136                 # Loop for each link
1137                 for ($k = 0; isset( $a[$k] ); $k++) {
1138                         $line = $a[$k];
1139                         if ( $useLinkPrefixExtension ) {
1140                                 wfProfileIn( $fname.'-prefixhandling' );
1141                                 if ( preg_match( $e2, $s, $m ) ) {
1142                                         $prefix = $m[2];
1143                                         $s = $m[1];
1144                                 } else {
1145                                         $prefix='';
1146                                 }
1147                                 # first link
1148                                 if($first_prefix) {
1149                                         $prefix = $first_prefix;
1150                                         $first_prefix = false;
1151                                 }
1152                                 wfProfileOut( $fname.'-prefixhandling' );
1153                         }
1154
1155                         $might_be_img = false;
1156
1157                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1158                                 $text = $m[2];
1159                                 # fix up urlencoded title texts
1160                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1161                                 $trail = $m[3];
1162                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1163                                 $might_be_img = true;
1164                                 $text = $m[2];
1165                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1166                                 $trail = "";
1167                         } else { # Invalid form; output directly
1168                                 $s .= $prefix . '[[' . $line ;
1169                                 continue;
1170                         }
1171
1172                         # Don't allow internal links to pages containing
1173                         # PROTO: where PROTO is a valid URL protocol; these
1174                         # should be external links.
1175                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1176                                 $s .= $prefix . '[[' . $line ;
1177                                 continue;
1178                         }
1179
1180                         # Make subpage if necessary
1181                         if( $useSubpages ) {
1182                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1183                         } else {
1184                                 $link = $m[1];
1185                         }
1186
1187                         $noforce = (substr($m[1], 0, 1) != ':');
1188                         if (!$noforce) {
1189                                 # Strip off leading ':'
1190                                 $link = substr($link, 1);
1191                         }
1192
1193                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1194                         if( !$nt ) {
1195                                 $s .= $prefix . '[[' . $line;
1196                                 continue;
1197                         }
1198
1199                         #check other language variants of the link
1200                         #if the article does not exist
1201                         if( $checkVariantLink
1202                             && $nt->getArticleID() == 0 ) {
1203                                 $wgContLang->findVariantLink($link, $nt);
1204                         }
1205
1206                         $ns = $nt->getNamespace();
1207                         $iw = $nt->getInterWiki();
1208
1209                         if ($might_be_img) { # if this is actually an invalid link
1210                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1211                                         $found = false;
1212                                         while (isset ($a[$k+1]) ) {
1213                                                 #look at the next 'line' to see if we can close it there
1214                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1215                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1216                                                 # the first ]] closes the inner link, the second the image
1217                                                         $found = true;
1218                                                         $text .= '[[' . $m[1];
1219                                                         $trail = $m[2];
1220                                                         break;
1221                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1222                                                         #if there's exactly one ]] that's fine, we'll keep looking
1223                                                         $text .= '[[' . $m[0];
1224                                                 } else {
1225                                                         #if $next_line is invalid too, we need look no further
1226                                                         $text .= '[[' . $next_line;
1227                                                         break;
1228                                                 }
1229                                         }
1230                                         if ( !$found ) {
1231                                                 # we couldn't find the end of this imageLink, so output it raw
1232                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1233                                                 $text = $this->replaceInternalLinks($text);
1234                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1235                                                 # note: no $trail, because without an end, there *is* no trail
1236                                                 continue;
1237                                         }
1238                                 } else { #it's not an image, so output it raw
1239                                         $s .= $prefix . '[[' . $link . '|' . $text;
1240                                         # note: no $trail, because without an end, there *is* no trail
1241                                         continue;
1242                                 }
1243                         }
1244
1245                         $wasblank = ( '' == $text );
1246                         if( $wasblank ) $text = $link;
1247
1248
1249                         # Link not escaped by : , create the various objects
1250                         if( $noforce ) {
1251
1252                                 # Interwikis
1253                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1254                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1255                                         $tmp = $prefix . $trail ;
1256                                         $s .= (trim($tmp) == '')? '': $tmp;
1257                                         continue;
1258                                 }
1259
1260                                 if ( $ns == NS_IMAGE ) {
1261                                         wfProfileIn( "$fname-image" );
1262
1263                                         # recursively parse links inside the image caption
1264                                         # actually, this will parse them in any other parameters, too,
1265                                         # but it might be hard to fix that, and it doesn't matter ATM
1266                                         $text = $this->replaceExternalLinks($text);
1267                                         $text = $this->replaceInternalLinks($text);
1268
1269                                         # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1270                                         $s .= $prefix . str_replace('http://', 'http-noparse://', $sk->makeImageLinkObj( $nt, $text ) ) . $trail;
1271                                         $wgLinkCache->addImageLinkObj( $nt );
1272
1273                                         wfProfileOut( "$fname-image" );
1274                                         continue;
1275                                 }
1276
1277                                 if ( $ns == NS_CATEGORY ) {
1278                                         wfProfileIn( "$fname-category" );
1279                                         $t = $nt->getText();
1280
1281                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1282                                         $pPLC=$sk->postParseLinkColour();
1283                                         $sk->postParseLinkColour( false );
1284                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1285                                         $sk->postParseLinkColour( $pPLC );
1286                                         $wgLinkCache->resume();
1287
1288                                         if ( $wasblank ) {
1289                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1290                                                         $sortkey = $this->mTitle->getText();
1291                                                 } else {
1292                                                         $sortkey = $this->mTitle->getPrefixedText();
1293                                                 }
1294                                         } else {
1295                                                 $sortkey = $text;
1296                                         }
1297                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1298                                         $this->mOutput->addCategoryLink( $t );
1299                                         $s .= $prefix . $trail ;
1300
1301                                         wfProfileOut( "$fname-category" );
1302                                         continue;
1303                                 }
1304                         }
1305
1306                         if( ( $nt->getPrefixedText() === $selflink ) &&
1307                             ( $nt->getFragment() === '' ) ) {
1308                                 # Self-links are handled specially; generally de-link and change to bold.
1309                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1310                                 continue;
1311                         }
1312
1313                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1314                         if( $ns == NS_MEDIA ) {
1315                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1316                                 $wgLinkCache->addImageLinkObj( $nt );
1317                                 continue;
1318                         } elseif( $ns == NS_SPECIAL ) {
1319                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1320                                 continue;
1321                         }
1322                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1323                 }
1324                 $sk->postParseLinkColour( $saveParseColour );
1325                 wfProfileOut( $fname );
1326                 return $s;
1327         }
1328
1329         /**
1330          * Return true if subpage links should be expanded on this page.
1331          * @return bool
1332          */
1333         function areSubpagesAllowed() {
1334                 # Some namespaces don't allow subpages
1335                 global $wgNamespacesWithSubpages;
1336                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1337         }
1338
1339         /**
1340          * Handle link to subpage if necessary
1341          * @param string $target the source of the link
1342          * @param string &$text the link text, modified as necessary
1343          * @return string the full name of the link
1344          * @access private
1345          */
1346         function maybeDoSubpageLink($target, &$text) {
1347                 # Valid link forms:
1348                 # Foobar -- normal
1349                 # :Foobar -- override special treatment of prefix (images, language links)
1350                 # /Foobar -- convert to CurrentPage/Foobar
1351                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1352                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1353                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1354
1355                 $fname = 'Parser::maybeDoSubpageLink';
1356                 wfProfileIn( $fname );
1357                 $ret = $target; # default return value is no change
1358
1359                 # Some namespaces don't allow subpages,
1360                 # so only perform processing if subpages are allowed
1361                 if( $this->areSubpagesAllowed() ) {
1362                         # Look at the first character
1363                         if( $target != '' && $target{0} == '/' ) {
1364                                 # / at end means we don't want the slash to be shown
1365                                 if( substr( $target, -1, 1 ) == '/' ) {
1366                                         $target = substr( $target, 1, -1 );
1367                                         $noslash = $target;
1368                                 } else {
1369                                         $noslash = substr( $target, 1 );
1370                                 }
1371
1372                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1373                                 if( '' === $text ) {
1374                                         $text = $target;
1375                                 } # this might be changed for ugliness reasons
1376                         } else {
1377                                 # check for .. subpage backlinks
1378                                 $dotdotcount = 0;
1379                                 $nodotdot = $target;
1380                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1381                                         ++$dotdotcount;
1382                                         $nodotdot = substr( $nodotdot, 3 );
1383                                 }
1384                                 if($dotdotcount > 0) {
1385                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1386                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1387                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1388                                                 # / at the end means don't show full path
1389                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1390                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1391                                                         if( '' === $text ) {
1392                                                                 $text = $nodotdot;
1393                                                         }
1394                                                 }
1395                                                 $nodotdot = trim( $nodotdot );
1396                                                 if( $nodotdot != '' ) {
1397                                                         $ret .= '/' . $nodotdot;
1398                                                 }
1399                                         }
1400                                 }
1401                         }
1402                 }
1403
1404                 wfProfileOut( $fname );
1405                 return $ret;
1406         }
1407
1408         /**#@+
1409          * Used by doBlockLevels()
1410          * @access private
1411          */
1412         /* private */ function closeParagraph() {
1413                 $result = '';
1414                 if ( '' != $this->mLastSection ) {
1415                         $result = '</' . $this->mLastSection  . ">\n";
1416                 }
1417                 $this->mInPre = false;
1418                 $this->mLastSection = '';
1419                 return $result;
1420         }
1421         # getCommon() returns the length of the longest common substring
1422         # of both arguments, starting at the beginning of both.
1423         #
1424         /* private */ function getCommon( $st1, $st2 ) {
1425                 $fl = strlen( $st1 );
1426                 $shorter = strlen( $st2 );
1427                 if ( $fl < $shorter ) { $shorter = $fl; }
1428
1429                 for ( $i = 0; $i < $shorter; ++$i ) {
1430                         if ( $st1{$i} != $st2{$i} ) { break; }
1431                 }
1432                 return $i;
1433         }
1434         # These next three functions open, continue, and close the list
1435         # element appropriate to the prefix character passed into them.
1436         #
1437         /* private */ function openList( $char ) {
1438                 $result = $this->closeParagraph();
1439
1440                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1441                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1442                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1443                 else if ( ';' == $char ) {
1444                         $result .= '<dl><dt>';
1445                         $this->mDTopen = true;
1446                 }
1447                 else { $result = '<!-- ERR 1 -->'; }
1448
1449                 return $result;
1450         }
1451
1452         /* private */ function nextItem( $char ) {
1453                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1454                 else if ( ':' == $char || ';' == $char ) {
1455                         $close = '</dd>';
1456                         if ( $this->mDTopen ) { $close = '</dt>'; }
1457                         if ( ';' == $char ) {
1458                                 $this->mDTopen = true;
1459                                 return $close . '<dt>';
1460                         } else {
1461                                 $this->mDTopen = false;
1462                                 return $close . '<dd>';
1463                         }
1464                 }
1465                 return '<!-- ERR 2 -->';
1466         }
1467
1468         /* private */ function closeList( $char ) {
1469                 if ( '*' == $char ) { $text = '</li></ul>'; }
1470                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1471                 else if ( ':' == $char ) {
1472                         if ( $this->mDTopen ) {
1473                                 $this->mDTopen = false;
1474                                 $text = '</dt></dl>';
1475                         } else {
1476                                 $text = '</dd></dl>';
1477                         }
1478                 }
1479                 else {  return '<!-- ERR 3 -->'; }
1480                 return $text."\n";
1481         }
1482         /**#@-*/
1483
1484         /**
1485          * Make lists from lines starting with ':', '*', '#', etc.
1486          *
1487          * @access private
1488          * @return string the lists rendered as HTML
1489          */
1490         function doBlockLevels( $text, $linestart ) {
1491                 $fname = 'Parser::doBlockLevels';
1492                 wfProfileIn( $fname );
1493
1494                 # Parsing through the text line by line.  The main thing
1495                 # happening here is handling of block-level elements p, pre,
1496                 # and making lists from lines starting with * # : etc.
1497                 #
1498                 $textLines = explode( "\n", $text );
1499
1500                 $lastPrefix = $output = $lastLine = '';
1501                 $this->mDTopen = $inBlockElem = false;
1502                 $prefixLength = 0;
1503                 $paragraphStack = false;
1504
1505                 if ( !$linestart ) {
1506                         $output .= array_shift( $textLines );
1507                 }
1508                 foreach ( $textLines as $oLine ) {
1509                         $lastPrefixLength = strlen( $lastPrefix );
1510                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1511                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1512                         if ( !$this->mInPre ) {
1513                                 # Multiple prefixes may abut each other for nested lists.
1514                                 $prefixLength = strspn( $oLine, '*#:;' );
1515                                 $pref = substr( $oLine, 0, $prefixLength );
1516
1517                                 # eh?
1518                                 $pref2 = str_replace( ';', ':', $pref );
1519                                 $t = substr( $oLine, $prefixLength );
1520                                 $this->mInPre = !empty($preOpenMatch);
1521                         } else {
1522                                 # Don't interpret any other prefixes in preformatted text
1523                                 $prefixLength = 0;
1524                                 $pref = $pref2 = '';
1525                                 $t = $oLine;
1526                         }
1527
1528                         # List generation
1529                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1530                                 # Same as the last item, so no need to deal with nesting or opening stuff
1531                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1532                                 $paragraphStack = false;
1533
1534                                 if ( substr( $pref, -1 ) == ';') {
1535                                         # The one nasty exception: definition lists work like this:
1536                                         # ; title : definition text
1537                                         # So we check for : in the remainder text to split up the
1538                                         # title and definition, without b0rking links.
1539                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1540                                                 $t = $t2;
1541                                                 $output .= $term . $this->nextItem( ':' );
1542                                         }
1543                                 }
1544                         } elseif( $prefixLength || $lastPrefixLength ) {
1545                                 # Either open or close a level...
1546                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1547                                 $paragraphStack = false;
1548
1549                                 while( $commonPrefixLength < $lastPrefixLength ) {
1550                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1551                                         --$lastPrefixLength;
1552                                 }
1553                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1554                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1555                                 }
1556                                 while ( $prefixLength > $commonPrefixLength ) {
1557                                         $char = substr( $pref, $commonPrefixLength, 1 );
1558                                         $output .= $this->openList( $char );
1559
1560                                         if ( ';' == $char ) {
1561                                                 # FIXME: This is dupe of code above
1562                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1563                                                         $t = $t2;
1564                                                         $output .= $term . $this->nextItem( ':' );
1565                                                 }
1566                                         }
1567                                         ++$commonPrefixLength;
1568                                 }
1569                                 $lastPrefix = $pref2;
1570                         }
1571                         if( 0 == $prefixLength ) {
1572                                 wfProfileIn( "$fname-paragraph" );
1573                                 # No prefix (not in list)--go to paragraph mode
1574                                 $uniq_prefix = UNIQ_PREFIX;
1575                                 // XXX: use a stack for nestable elements like span, table and div
1576                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1577                                 $closematch = preg_match(
1578                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1579                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1580                                 if ( $openmatch or $closematch ) {
1581                                         $paragraphStack = false;
1582                                         $output .= $this->closeParagraph();
1583                                         if($preOpenMatch and !$preCloseMatch) {
1584                                                 $this->mInPre = true;
1585                                         }
1586                                         if ( $closematch ) {
1587                                                 $inBlockElem = false;
1588                                         } else {
1589                                                 $inBlockElem = true;
1590                                         }
1591                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1592                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1593                                                 // pre
1594                                                 if ($this->mLastSection != 'pre') {
1595                                                         $paragraphStack = false;
1596                                                         $output .= $this->closeParagraph().'<pre>';
1597                                                         $this->mLastSection = 'pre';
1598                                                 }
1599                                                 $t = substr( $t, 1 );
1600                                         } else {
1601                                                 // paragraph
1602                                                 if ( '' == trim($t) ) {
1603                                                         if ( $paragraphStack ) {
1604                                                                 $output .= $paragraphStack.'<br />';
1605                                                                 $paragraphStack = false;
1606                                                                 $this->mLastSection = 'p';
1607                                                         } else {
1608                                                                 if ($this->mLastSection != 'p' ) {
1609                                                                         $output .= $this->closeParagraph();
1610                                                                         $this->mLastSection = '';
1611                                                                         $paragraphStack = '<p>';
1612                                                                 } else {
1613                                                                         $paragraphStack = '</p><p>';
1614                                                                 }
1615                                                         }
1616                                                 } else {
1617                                                         if ( $paragraphStack ) {
1618                                                                 $output .= $paragraphStack;
1619                                                                 $paragraphStack = false;
1620                                                                 $this->mLastSection = 'p';
1621                                                         } else if ($this->mLastSection != 'p') {
1622                                                                 $output .= $this->closeParagraph().'<p>';
1623                                                                 $this->mLastSection = 'p';
1624                                                         }
1625                                                 }
1626                                         }
1627                                 }
1628                                 wfProfileOut( "$fname-paragraph" );
1629                         }
1630                         if ($paragraphStack === false) {
1631                                 $output .= $t."\n";
1632                         }
1633                 }
1634                 while ( $prefixLength ) {
1635                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1636                         --$prefixLength;
1637                 }
1638                 if ( '' != $this->mLastSection ) {
1639                         $output .= '</' . $this->mLastSection . '>';
1640                         $this->mLastSection = '';
1641                 }
1642
1643                 wfProfileOut( $fname );
1644                 return $output;
1645         }
1646
1647         /**
1648          * Split up a string on ':', ignoring any occurences inside
1649          * <a>..</a> or <span>...</span>
1650          * @param string $str the string to split
1651          * @param string &$before set to everything before the ':'
1652          * @param string &$after set to everything after the ':'
1653          * return string the position of the ':', or false if none found
1654          */
1655         function findColonNoLinks($str, &$before, &$after) {
1656                 # I wonder if we should make this count all tags, not just <a>
1657                 # and <span>. That would prevent us from matching a ':' that
1658                 # comes in the middle of italics other such formatting....
1659                 # -- Wil
1660                 $fname = 'Parser::findColonNoLinks';
1661                 wfProfileIn( $fname );
1662                 $pos = 0;
1663                 do {
1664                         $colon = strpos($str, ':', $pos);
1665
1666                         if ($colon !== false) {
1667                                 $before = substr($str, 0, $colon);
1668                                 $after = substr($str, $colon + 1);
1669
1670                                 # Skip any ':' within <a> or <span> pairs
1671                                 $a = substr_count($before, '<a');
1672                                 $s = substr_count($before, '<span');
1673                                 $ca = substr_count($before, '</a>');
1674                                 $cs = substr_count($before, '</span>');
1675
1676                                 if ($a <= $ca and $s <= $cs) {
1677                                         # Tags are balanced before ':'; ok
1678                                         break;
1679                                 }
1680                                 $pos = $colon + 1;
1681                         }
1682                 } while ($colon !== false);
1683                 wfProfileOut( $fname );
1684                 return $colon;
1685         }
1686
1687         /**
1688          * Return value of a magic variable (like PAGENAME)
1689          *
1690          * @access private
1691          */
1692         function getVariableValue( $index ) {
1693                 global $wgContLang, $wgSitename, $wgServer;
1694
1695                 /**
1696                  * Some of these require message or data lookups and can be
1697                  * expensive to check many times.
1698                  */
1699                 static $varCache = array();
1700                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1701
1702                 switch ( $index ) {
1703                         case MAG_CURRENTMONTH:
1704                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1705                         case MAG_CURRENTMONTHNAME:
1706                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1707                         case MAG_CURRENTMONTHNAMEGEN:
1708                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1709                         case MAG_CURRENTDAY:
1710                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1711                         case MAG_PAGENAME:
1712                                 return $this->mTitle->getText();
1713                         case MAG_PAGENAMEE:
1714                                 return $this->mTitle->getPartialURL();
1715                         case MAG_NAMESPACE:
1716                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1717                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1718                         case MAG_CURRENTDAYNAME:
1719                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1720                         case MAG_CURRENTYEAR:
1721                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1722                         case MAG_CURRENTTIME:
1723                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1724                         case MAG_CURRENTWEEK:
1725                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1726                         case MAG_CURRENTDOW:
1727                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1728                         case MAG_NUMBEROFARTICLES:
1729                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1730                         case MAG_SITENAME:
1731                                 return $wgSitename;
1732                         case MAG_SERVER:
1733                                 return $wgServer;
1734                         default:
1735                                 return NULL;
1736                 }
1737         }
1738
1739         /**
1740          * initialise the magic variables (like CURRENTMONTHNAME)
1741          *
1742          * @access private
1743          */
1744         function initialiseVariables() {
1745                 $fname = 'Parser::initialiseVariables';
1746                 wfProfileIn( $fname );
1747                 global $wgVariableIDs;
1748                 $this->mVariables = array();
1749                 foreach ( $wgVariableIDs as $id ) {
1750                         $mw =& MagicWord::get( $id );
1751                         $mw->addToArray( $this->mVariables, $id );
1752                 }
1753                 wfProfileOut( $fname );
1754         }
1755
1756         /**
1757          * Replace magic variables, templates, and template arguments
1758          * with the appropriate text. Templates are substituted recursively,
1759          * taking care to avoid infinite loops.
1760          *
1761          * Note that the substitution depends on value of $mOutputType:
1762          *  OT_WIKI: only {{subst:}} templates
1763          *  OT_MSG: only magic variables
1764          *  OT_HTML: all templates and magic variables
1765          *
1766          * @param string $tex The text to transform
1767          * @param array $args Key-value pairs representing template parameters to substitute
1768          * @access private
1769          */
1770         function replaceVariables( $text, $args = array() ) {
1771                 global $wgLang, $wgScript, $wgArticlePath;
1772
1773                 # Prevent too big inclusions
1774                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1775                         return $text;
1776                 }
1777
1778                 $fname = 'Parser::replaceVariables';
1779                 wfProfileIn( $fname );
1780
1781                 $titleChars = Title::legalChars();
1782
1783                 # This function is called recursively. To keep track of arguments we need a stack:
1784                 array_push( $this->mArgStack, $args );
1785
1786                 # Variable substitution
1787                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1788
1789                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1790                         # Argument substitution
1791                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1792                 }
1793                 # Template substitution
1794                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1795                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1796
1797                 array_pop( $this->mArgStack );
1798
1799                 wfProfileOut( $fname );
1800                 return $text;
1801         }
1802
1803         /**
1804          * Replace magic variables
1805          * @access private
1806          */
1807         function variableSubstitution( $matches ) {
1808                 $fname = 'parser::variableSubstitution';
1809                 $varname = $matches[1];
1810                 wfProfileIn( $fname );
1811                 if ( !$this->mVariables ) {
1812                         $this->initialiseVariables();
1813                 }
1814                 $skip = false;
1815                 if ( $this->mOutputType == OT_WIKI ) {
1816                         # Do only magic variables prefixed by SUBST
1817                         $mwSubst =& MagicWord::get( MAG_SUBST );
1818                         if (!$mwSubst->matchStartAndRemove( $varname ))
1819                                 $skip = true;
1820                         # Note that if we don't substitute the variable below,
1821                         # we don't remove the {{subst:}} magic word, in case
1822                         # it is a template rather than a magic variable.
1823                 }
1824                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
1825                         $id = $this->mVariables[$varname];
1826                         $text = $this->getVariableValue( $id );
1827                         $this->mOutput->mContainsOldMagic = true;
1828                 } else {
1829                         $text = $matches[0];
1830                 }
1831                 wfProfileOut( $fname );
1832                 return $text;
1833         }
1834
1835         # Split template arguments
1836         function getTemplateArgs( $argsString ) {
1837                 if ( $argsString === '' ) {
1838                         return array();
1839                 }
1840
1841                 $args = explode( '|', substr( $argsString, 1 ) );
1842
1843                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1844                 # merged with the next arg because the '|' character between belongs
1845                 # to the link syntax and not the template parameter syntax.
1846                 $argc = count($args);
1847                 $i = 0;
1848                 for ( $i = 0; $i < $argc-1; $i++ ) {
1849                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1850                                 $args[$i] .= '|'.$args[$i+1];
1851                                 array_splice($args, $i+1, 1);
1852                                 $i--;
1853                                 $argc--;
1854                         }
1855                 }
1856
1857                 return $args;
1858         }
1859
1860         /**
1861          * Return the text of a template, after recursively
1862          * replacing any variables or templates within the template.
1863          *
1864          * @param array $matches The parts of the template
1865          *  $matches[1]: the title, i.e. the part before the |
1866          *  $matches[2]: the parameters (including a leading |), if  any
1867          * @return string the text of the template
1868          * @access private
1869          */
1870         function braceSubstitution( $matches ) {
1871                 global $wgLinkCache, $wgContLang;
1872                 $fname = 'Parser::braceSubstitution';
1873                 wfProfileIn( $fname );
1874
1875                 $found = false;
1876                 $nowiki = false;
1877                 $noparse = false;
1878
1879                 $title = NULL;
1880
1881                 # Need to know if the template comes at the start of a line,
1882                 # to treat the beginning of the template like the beginning
1883                 # of a line for tables and block-level elements.
1884                 $linestart = $matches[1];
1885
1886                 # $part1 is the bit before the first |, and must contain only title characters
1887                 # $args is a list of arguments, starting from index 0, not including $part1
1888
1889                 $part1 = $matches[2];
1890                 # If the third subpattern matched anything, it will start with |
1891
1892                 $args = $this->getTemplateArgs($matches[3]);
1893                 $argc = count( $args );
1894
1895                 # Don't parse {{{}}} because that's only for template arguments
1896                 if ( $linestart === '{' ) {
1897                         $text = $matches[0];
1898                         $found = true;
1899                         $noparse = true;
1900                 }
1901
1902                 # SUBST
1903                 if ( !$found ) {
1904                         $mwSubst =& MagicWord::get( MAG_SUBST );
1905                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1906                                 # One of two possibilities is true:
1907                                 # 1) Found SUBST but not in the PST phase
1908                                 # 2) Didn't find SUBST and in the PST phase
1909                                 # In either case, return without further processing
1910                                 $text = $matches[0];
1911                                 $found = true;
1912                                 $noparse = true;
1913                         }
1914                 }
1915
1916                 # MSG, MSGNW and INT
1917                 if ( !$found ) {
1918                         # Check for MSGNW:
1919                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1920                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1921                                 $nowiki = true;
1922                         } else {
1923                                 # Remove obsolete MSG:
1924                                 $mwMsg =& MagicWord::get( MAG_MSG );
1925                                 $mwMsg->matchStartAndRemove( $part1 );
1926                         }
1927
1928                         # Check if it is an internal message
1929                         $mwInt =& MagicWord::get( MAG_INT );
1930                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1931                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1932                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1933                                         $found = true;
1934                                 }
1935                         }
1936                 }
1937
1938                 # NS
1939                 if ( !$found ) {
1940                         # Check for NS: (namespace expansion)
1941                         $mwNs = MagicWord::get( MAG_NS );
1942                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1943                                 if ( intval( $part1 ) ) {
1944                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1945                                         $found = true;
1946                                 } else {
1947                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1948                                         if ( !is_null( $index ) ) {
1949                                                 $text = $linestart . $wgContLang->getNsText( $index );
1950                                                 $found = true;
1951                                         }
1952                                 }
1953                         }
1954                 }
1955
1956                 # LOCALURL and LOCALURLE
1957                 if ( !$found ) {
1958                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1959                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1960
1961                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1962                                 $func = 'getLocalURL';
1963                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1964                                 $func = 'escapeLocalURL';
1965                         } else {
1966                                 $func = '';
1967                         }
1968
1969                         if ( $func !== '' ) {
1970                                 $title = Title::newFromText( $part1 );
1971                                 if ( !is_null( $title ) ) {
1972                                         if ( $argc > 0 ) {
1973                                                 $text = $linestart . $title->$func( $args[0] );
1974                                         } else {
1975                                                 $text = $linestart . $title->$func();
1976                                         }
1977                                         $found = true;
1978                                 }
1979                         }
1980                 }
1981
1982                 # GRAMMAR
1983                 if ( !$found && $argc == 1 ) {
1984                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1985                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1986                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
1987                                 $found = true;
1988                         }
1989                 }
1990
1991                 # Template table test
1992
1993                 # Did we encounter this template already? If yes, it is in the cache
1994                 # and we need to check for loops.
1995                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
1996                         # set $text to cached message.
1997                         $text = $linestart . $this->mTemplates[$part1];
1998                         $found = true;
1999
2000                         # Infinite loop test
2001                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2002                                 $noparse = true;
2003                                 $found = true;
2004                                 $text .= '<!-- WARNING: template loop detected -->';
2005                         }
2006                 }
2007
2008                 # Load from database
2009                 $itcamefromthedatabase = false;
2010                 if ( !$found ) {
2011                         $ns = NS_TEMPLATE;
2012                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2013                         if ($subpage !== '') {
2014                                 $ns = $this->mTitle->getNamespace();
2015                         }
2016                         $title = Title::newFromText( $part1, $ns );
2017                         if ( !is_null( $title ) && !$title->isExternal() ) {
2018                                 # Check for excessive inclusion
2019                                 $dbk = $title->getPrefixedDBkey();
2020                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2021                                         # This should never be reached.
2022                                         $article = new Article( $title );
2023                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2024                                         if ( $articleContent !== false ) {
2025                                                 $found = true;
2026                                                 $text = $linestart . $articleContent;
2027                                                 $itcamefromthedatabase = true;
2028                                         }
2029                                 }
2030
2031                                 # If the title is valid but undisplayable, make a link to it
2032                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2033                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2034                                         $found = true;
2035                                 }
2036
2037                                 # Template cache array insertion
2038                                 if( $found ) {
2039                                         $this->mTemplates[$part1] = $text;
2040                                 }
2041                         }
2042                 }
2043
2044                 # Recursive parsing, escaping and link table handling
2045                 # Only for HTML output
2046                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2047                         $text = wfEscapeWikiText( $text );
2048                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2049                         # Clean up argument array
2050                         $assocArgs = array();
2051                         $index = 1;
2052                         foreach( $args as $arg ) {
2053                                 $eqpos = strpos( $arg, '=' );
2054                                 if ( $eqpos === false ) {
2055                                         $assocArgs[$index++] = $arg;
2056                                 } else {
2057                                         $name = trim( substr( $arg, 0, $eqpos ) );
2058                                         $value = trim( substr( $arg, $eqpos+1 ) );
2059                                         if ( $value === false ) {
2060                                                 $value = '';
2061                                         }
2062                                         if ( $name !== false ) {
2063                                                 $assocArgs[$name] = $value;
2064                                         }
2065                                 }
2066                         }
2067
2068                         # Add a new element to the templace recursion path
2069                         $this->mTemplatePath[$part1] = 1;
2070
2071                         $text = $this->strip( $text, $this->mStripState );
2072                         $text = Sanitizer::removeHTMLtags( $text );
2073                         $text = $this->replaceVariables( $text, $assocArgs );
2074
2075                         # Resume the link cache and register the inclusion as a link
2076                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2077                                 $wgLinkCache->addLinkObj( $title );
2078                         }
2079
2080                         # If the template begins with a table or block-level
2081                         # element, it should be treated as beginning a new line.
2082                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2083                                 $text = "\n" . $text;
2084                         }
2085                 }
2086
2087                 # Empties the template path
2088                 $this->mTemplatePath = array();
2089                 if ( !$found ) {
2090                         wfProfileOut( $fname );
2091                         return $matches[0];
2092                 } else {
2093                         # replace ==section headers==
2094                         # XXX this needs to go away once we have a better parser.
2095                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2096                                 if( !is_null( $title ) )
2097                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2098                                 else
2099                                         $encodedname = base64_encode("");
2100                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2101                                         PREG_SPLIT_DELIM_CAPTURE);
2102                                 $text = '';
2103                                 $nsec = 0;
2104                                 for( $i = 0; $i < count($m); $i += 2 ) {
2105                                         $text .= $m[$i];
2106                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2107                                         $hl = $m[$i + 1];
2108                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2109                                                 $text .= $hl;
2110                                                 continue;
2111                                         }
2112                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2113                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2114                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2115
2116                                         $nsec++;
2117                                 }
2118                         }
2119                 }
2120
2121                 # Empties the template path
2122                 $this->mTemplatePath = array();
2123
2124                 if ( !$found ) {
2125                         wfProfileOut( $fname );
2126                         return $matches[0];
2127                 } else {
2128                         wfProfileOut( $fname );
2129                         return $text;
2130                 }
2131         }
2132
2133         /**
2134          * Triple brace replacement -- used for template arguments
2135          * @access private
2136          */
2137         function argSubstitution( $matches ) {
2138                 $arg = trim( $matches[1] );
2139                 $text = $matches[0];
2140                 $inputArgs = end( $this->mArgStack );
2141
2142                 if ( array_key_exists( $arg, $inputArgs ) ) {
2143                         $text = $inputArgs[$arg];
2144                 }
2145
2146                 return $text;
2147         }
2148
2149         /**
2150          * Returns true if the function is allowed to include this entity
2151          * @access private
2152          */
2153         function incrementIncludeCount( $dbk ) {
2154                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2155                         $this->mIncludeCount[$dbk] = 0;
2156                 }
2157                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2158                         return true;
2159                 } else {
2160                         return false;
2161                 }
2162         }
2163
2164         /**
2165          * This function accomplishes several tasks:
2166          * 1) Auto-number headings if that option is enabled
2167          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2168          * 3) Add a Table of contents on the top for users who have enabled the option
2169          * 4) Auto-anchor headings
2170          *
2171          * It loops through all headlines, collects the necessary data, then splits up the
2172          * string and re-inserts the newly formatted headlines.
2173          *
2174          * @param string $text
2175          * @param boolean $isMain
2176          * @access private
2177          */
2178         function formatHeadings( $text, $isMain=true ) {
2179                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2180
2181                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2182                 $doShowToc = $this->mOptions->getShowToc();
2183                 $forceTocHere = false;
2184                 if( !$this->mTitle->userCanEdit() ) {
2185                         $showEditLink = 0;
2186                         $rightClickHack = 0;
2187                 } else {
2188                         $showEditLink = $this->mOptions->getEditSection();
2189                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2190                 }
2191
2192                 # Inhibit editsection links if requested in the page
2193                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2194                 if( $esw->matchAndRemove( $text ) ) {
2195                         $showEditLink = 0;
2196                 }
2197                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2198                 # do not add TOC
2199                 $mw =& MagicWord::get( MAG_NOTOC );
2200                 if( $mw->matchAndRemove( $text ) ) {
2201                         $doShowToc = 0;
2202                 }
2203
2204                 # never add the TOC to the Main Page. This is an entry page that should not
2205                 # be more than 1-2 screens large anyway
2206                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2207                         $doShowToc = 0;
2208                 }
2209
2210                 # Get all headlines for numbering them and adding funky stuff like [edit]
2211                 # links - this is for later, but we need the number of headlines right now
2212                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2213
2214                 # if there are fewer than 4 headlines in the article, do not show TOC
2215                 if( $numMatches < 4 ) {
2216                         $doShowToc = 0;
2217                 }
2218
2219                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2220                 # override above conditions and always show TOC at that place
2221
2222                 $mw =& MagicWord::get( MAG_TOC );
2223                 if($mw->match( $text ) ) {
2224                         $doShowToc = 1;
2225                         $forceTocHere = true;
2226                 } else {
2227                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2228                         # override above conditions and always show TOC above first header
2229                         $mw =& MagicWord::get( MAG_FORCETOC );
2230                         if ($mw->matchAndRemove( $text ) ) {
2231                                 $doShowToc = 1;
2232                         }
2233                 }
2234
2235                 # Never ever show TOC if no headers
2236                 if( $numMatches < 1 ) {
2237                         $doShowToc = 0;
2238                 }
2239
2240                 # We need this to perform operations on the HTML
2241                 $sk =& $this->mOptions->getSkin();
2242
2243                 # headline counter
2244                 $headlineCount = 0;
2245                 $sectionCount = 0; # headlineCount excluding template sections
2246
2247                 # Ugh .. the TOC should have neat indentation levels which can be
2248                 # passed to the skin functions. These are determined here
2249                 $toc = '';
2250                 $full = '';
2251                 $head = array();
2252                 $sublevelCount = array();
2253                 $levelCount = array();
2254                 $toclevel = 0;
2255                 $level = 0;
2256                 $prevlevel = 0;
2257                 $toclevel = 0;
2258                 $prevtoclevel = 0;
2259
2260                 foreach( $matches[3] as $headline ) {
2261                         $istemplate = 0;
2262                         $templatetitle = '';
2263                         $templatesection = 0;
2264                         $numbering = '';
2265
2266                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2267                                 $istemplate = 1;
2268                                 $templatetitle = base64_decode($mat[1]);
2269                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2270                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2271                         }
2272
2273                         if( $toclevel ) {
2274                                 $prevlevel = $level;
2275                                 $prevtoclevel = $toclevel;
2276                         }
2277                         $level = $matches[1][$headlineCount];
2278
2279                         if( $doNumberHeadings || $doShowToc ) {
2280
2281                                 if ( $level > $prevlevel ) {
2282                                         # Increase TOC level
2283                                         $toclevel++;
2284                                         $sublevelCount[$toclevel] = 0;
2285                                         $toc .= $sk->tocIndent();
2286                                 }
2287                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2288                                         # Decrease TOC level, find level to jump to
2289
2290                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2291                                                 # Can only go down to level 1
2292                                                 $toclevel = 1;
2293                                         } else {
2294                                                 for ($i = $toclevel; $i > 0; $i--) {
2295                                                         if ( $levelCount[$i] == $level ) {
2296                                                                 # Found last matching level
2297                                                                 $toclevel = $i;
2298                                                                 break;
2299                                                         }
2300                                                         elseif ( $levelCount[$i] < $level ) {
2301                                                                 # Found first matching level below current level
2302                                                                 $toclevel = $i + 1;
2303                                                                 break;
2304                                                         }
2305                                                 }
2306                                         }
2307
2308                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2309                                 }
2310                                 else {
2311                                         # No change in level, end TOC line
2312                                         $toc .= $sk->tocLineEnd();
2313                                 }
2314
2315                                 $levelCount[$toclevel] = $level;
2316
2317                                 # count number of headlines for each level
2318                                 @$sublevelCount[$toclevel]++;
2319                                 $dot = 0;
2320                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2321                                         if( !empty( $sublevelCount[$i] ) ) {
2322                                                 if( $dot ) {
2323                                                         $numbering .= '.';
2324                                                 }
2325                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2326                                                 $dot = 1;
2327                                         }
2328                                 }
2329                         }
2330
2331                         # The canonized header is a version of the header text safe to use for links
2332                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2333                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2334                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2335
2336                         # Remove link placeholders by the link text.
2337                         #     <!--LINK number-->
2338                         # turns into
2339                         #     link text with suffix
2340                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2341                                                             "\$wgLinkHolders['texts'][\$1]",
2342                                                             $canonized_headline );
2343                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2344                                                             "\$wgInterwikiLinkHolders[\$1]",
2345                                                             $canonized_headline );
2346
2347                         # strip out HTML
2348                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2349                         $tocline = trim( $canonized_headline );
2350                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2351                         $replacearray = array(
2352                                 '%3A' => ':',
2353                                 '%' => '.'
2354                         );
2355                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2356                         $refer[$headlineCount] = $canonized_headline;
2357
2358                         # count how many in assoc. array so we can track dupes in anchors
2359                         @$refers[$canonized_headline]++;
2360                         $refcount[$headlineCount]=$refers[$canonized_headline];
2361
2362                         # Don't number the heading if it is the only one (looks silly)
2363                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2364                                 # the two are different if the line contains a link
2365                                 $headline=$numbering . ' ' . $headline;
2366                         }
2367
2368                         # Create the anchor for linking from the TOC to the section
2369                         $anchor = $canonized_headline;
2370                         if($refcount[$headlineCount] > 1 ) {
2371                                 $anchor .= '_' . $refcount[$headlineCount];
2372                         }
2373                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2374                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2375                         }
2376                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2377                                 if ( empty( $head[$headlineCount] ) ) {
2378                                         $head[$headlineCount] = '';
2379                                 }
2380                                 if( $istemplate )
2381                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2382                                 else
2383                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2384                         }
2385
2386                         # Add the edit section span
2387                         if( $rightClickHack ) {
2388                                 if( $istemplate )
2389                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2390                                 else
2391                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2392                         }
2393
2394                         # give headline the correct <h#> tag
2395                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2396
2397                         $headlineCount++;
2398                         if( !$istemplate )
2399                                 $sectionCount++;
2400                 }
2401
2402                 if( $doShowToc ) {
2403                         $toclines = $headlineCount;
2404                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2405                         $toc = $sk->tocList( $toc );
2406                 }
2407
2408                 # split up and insert constructed headlines
2409
2410                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2411                 $i = 0;
2412
2413                 foreach( $blocks as $block ) {
2414                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2415                                 # This is the [edit] link that appears for the top block of text when
2416                                 # section editing is enabled
2417
2418                                 # Disabled because it broke block formatting
2419                                 # For example, a bullet point in the top line
2420                                 # $full .= $sk->editSectionLink(0);
2421                         }
2422                         $full .= $block;
2423                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2424                         # Top anchor now in skin
2425                                 $full = $full.$toc;
2426                         }
2427
2428                         if( !empty( $head[$i] ) ) {
2429                                 $full .= $head[$i];
2430                         }
2431                         $i++;
2432                 }
2433                 if($forceTocHere) {
2434                         $mw =& MagicWord::get( MAG_TOC );
2435                         return $mw->replace( $toc, $full );
2436                 } else {
2437                         return $full;
2438                 }
2439         }
2440
2441         /**
2442          * Return an HTML link for the "ISBN 123456" text
2443          * @access private
2444          */
2445         function magicISBN( $text ) {
2446                 global $wgLang;
2447                 $fname = 'Parser::magicISBN';
2448                 wfProfileIn( $fname );
2449
2450                 $a = split( 'ISBN ', ' '.$text );
2451                 if ( count ( $a ) < 2 ) {
2452                         wfProfileOut( $fname );
2453                         return $text;
2454                 }
2455                 $text = substr( array_shift( $a ), 1);
2456                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2457
2458                 foreach ( $a as $x ) {
2459                         $isbn = $blank = '' ;
2460                         while ( ' ' == $x{0} ) {
2461                                 $blank .= ' ';
2462                                 $x = substr( $x, 1 );
2463                         }
2464                         if ( $x == '' ) { # blank isbn
2465                                 $text .= "ISBN $blank";
2466                                 continue;
2467                         }
2468                         while ( strstr( $valid, $x{0} ) != false ) {
2469                                 $isbn .= $x{0};
2470                                 $x = substr( $x, 1 );
2471                         }
2472                         $num = str_replace( '-', '', $isbn );
2473                         $num = str_replace( ' ', '', $num );
2474
2475                         if ( '' == $num ) {
2476                                 $text .= "ISBN $blank$x";
2477                         } else {
2478                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2479                                 $text .= '<a href="' .
2480                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2481                                         "\" class=\"internal\">ISBN $isbn</a>";
2482                                 $text .= $x;
2483                         }
2484                 }
2485                 wfProfileOut( $fname );
2486                 return $text;
2487         }
2488
2489         /**
2490          * Return an HTML link for the "GEO ..." text
2491          * @access private
2492          */
2493         function magicGEO( $text ) {
2494                 global $wgLang, $wgUseGeoMode;
2495                 $fname = 'Parser::magicGEO';
2496                 wfProfileIn( $fname );
2497
2498                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2499                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2500                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2501                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2502                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2503                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2504
2505                 $a = split( 'GEO ', ' '.$text );
2506                 if ( count ( $a ) < 2 ) {
2507                         wfProfileOut( $fname );
2508                         return $text;
2509                 }
2510                 $text = substr( array_shift( $a ), 1);
2511                 $valid = '0123456789.+-:';
2512
2513                 foreach ( $a as $x ) {
2514                         $geo = $blank = '' ;
2515                         while ( ' ' == $x{0} ) {
2516                                 $blank .= ' ';
2517                                 $x = substr( $x, 1 );
2518                         }
2519                         while ( strstr( $valid, $x{0} ) != false ) {
2520                                 $geo .= $x{0};
2521                                 $x = substr( $x, 1 );
2522                         }
2523                         $num = str_replace( '+', '', $geo );
2524                         $num = str_replace( ' ', '', $num );
2525
2526                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2527                                 $text .= "GEO $blank$x";
2528                         } else {
2529                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2530                                 $text .= '<a href="' .
2531                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2532                                         "\" class=\"internal\">GEO $geo</a>";
2533                                 $text .= $x;
2534                         }
2535                 }
2536                 wfProfileOut( $fname );
2537                 return $text;
2538         }
2539
2540         /**
2541          * Return an HTML link for the "RFC 1234" text
2542          * @access private
2543          * @param string $text text to be processed
2544          */
2545         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2546                 global $wgLang;
2547
2548                 $valid = '0123456789';
2549                 $internal = false;
2550
2551                 $a = split( $keyword, ' '.$text );
2552                 if ( count ( $a ) < 2 ) {
2553                         return $text;
2554                 }
2555                 $text = substr( array_shift( $a ), 1);
2556
2557                 /* Check if keyword is preceed by [[.
2558                  * This test is made here cause of the array_shift above
2559                  * that prevent the test to be done in the foreach.
2560                  */
2561                 if ( substr( $text, -2 ) == '[[' ) {
2562                         $internal = true;
2563                 }
2564
2565                 foreach ( $a as $x ) {
2566                         /* token might be empty if we have RFC RFC 1234 */
2567                         if ( $x=='' ) {
2568                                 $text.=$keyword;
2569                                 continue;
2570                                 }
2571
2572                         $id = $blank = '' ;
2573
2574                         /** remove and save whitespaces in $blank */
2575                         while ( $x{0} == ' ' ) {
2576                                 $blank .= ' ';
2577                                 $x = substr( $x, 1 );
2578                         }
2579
2580                         /** remove and save the rfc number in $id */
2581                         while ( strstr( $valid, $x{0} ) != false ) {
2582                                 $id .= $x{0};
2583                                 $x = substr( $x, 1 );
2584                         }
2585
2586                         if ( $id == '' ) {
2587                                 /* call back stripped spaces*/
2588                                 $text .= $keyword.$blank.$x;
2589                         } elseif( $internal ) {
2590                                 /* normal link */
2591                                 $text .= $keyword.$id.$x;
2592                         } else {
2593                                 /* build the external link*/
2594                                 $url = wfmsg( $urlmsg );
2595                                 $url = str_replace( '$1', $id, $url);
2596                                 $sk =& $this->mOptions->getSkin();
2597                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2598                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2599                         }
2600
2601                         /* Check if the next RFC keyword is preceed by [[ */
2602                         $internal = ( substr($x,-2) == '[[' );
2603                 }
2604                 return $text;
2605         }
2606
2607         /**
2608          * Transform wiki markup when saving a page by doing \r\n -> \n
2609          * conversion, substitting signatures, {{subst:}} templates, etc.
2610          *
2611          * @param string $text the text to transform
2612          * @param Title &$title the Title object for the current article
2613          * @param User &$user the User object describing the current user
2614          * @param ParserOptions $options parsing options
2615          * @param bool $clearState whether to clear the parser state first
2616          * @return string the altered wiki markup
2617          * @access public
2618          */
2619         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2620                 $this->mOptions = $options;
2621                 $this->mTitle =& $title;
2622                 $this->mOutputType = OT_WIKI;
2623
2624                 if ( $clearState ) {
2625                         $this->clearState();
2626                 }
2627
2628                 $stripState = false;
2629                 $pairs = array(
2630                         "\r\n" => "\n",
2631                         );
2632                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2633                 $text = $this->strip( $text, $stripState, false );
2634                 $text = $this->pstPass2( $text, $user );
2635                 $text = $this->unstrip( $text, $stripState );
2636                 $text = $this->unstripNoWiki( $text, $stripState );
2637                 return $text;
2638         }
2639
2640         /**
2641          * Pre-save transform helper function
2642          * @access private
2643          */
2644         function pstPass2( $text, &$user ) {
2645                 global $wgLang, $wgContLang, $wgLocaltimezone;
2646
2647                 # Variable replacement
2648                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2649                 $text = $this->replaceVariables( $text );
2650
2651                 # Signatures
2652                 #
2653                 $n = $user->getName();
2654                 $k = $user->getOption( 'nickname' );
2655                 if ( '' == $k ) { $k = $n; }
2656                 if ( isset( $wgLocaltimezone ) ) {
2657                         $oldtz = getenv( 'TZ' );
2658                         putenv( 'TZ='.$wgLocaltimezone );
2659                 }
2660                 /* Note: this is an ugly timezone hack for the European wikis */
2661                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2662                   ' (' . date( 'T' ) . ')';
2663                 if ( isset( $wgLocaltimezone ) ) {
2664                         putenv( 'TZ='.$oldtzs );
2665                 }
2666
2667                 if( $user->getOption( 'fancysig' ) ) {
2668                         $sigText = $k;
2669                 } else {
2670                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2671                 }
2672                 $text = preg_replace( '/~~~~~/', $d, $text );
2673                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2674                 $text = preg_replace( '/~~~/', $sigText, $text );
2675
2676                 # Context links: [[|name]] and [[name (context)|]]
2677                 #
2678                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2679                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2680                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2681                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2682
2683                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2684                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2685                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2686                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2687                 $context = '';
2688                 $t = $this->mTitle->getText();
2689                 if ( preg_match( $conpat, $t, $m ) ) {
2690                         $context = $m[2];
2691                 }
2692                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2693                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2694                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2695
2696                 if ( '' == $context ) {
2697                         $text = preg_replace( $p2, '[[\\1]]', $text );
2698                 } else {
2699                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2700                 }
2701
2702                 # Trim trailing whitespace
2703                 # MAG_END (__END__) tag allows for trailing
2704                 # whitespace to be deliberately included
2705                 $text = rtrim( $text );
2706                 $mw =& MagicWord::get( MAG_END );
2707                 $mw->matchAndRemove( $text );
2708
2709                 return $text;
2710         }
2711
2712         /**
2713          * Set up some variables which are usually set up in parse()
2714          * so that an external function can call some class members with confidence
2715          * @access public
2716          */
2717         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2718                 $this->mTitle =& $title;
2719                 $this->mOptions = $options;
2720                 $this->mOutputType = $outputType;
2721                 if ( $clearState ) {
2722                         $this->clearState();
2723                 }
2724         }
2725
2726         /**
2727          * Transform a MediaWiki message by replacing magic variables.
2728          *
2729          * @param string $text the text to transform
2730          * @param ParserOptions $options  options
2731          * @return string the text with variables substituted
2732          * @access public
2733          */
2734         function transformMsg( $text, $options ) {
2735                 global $wgTitle;
2736                 static $executing = false;
2737
2738                 # Guard against infinite recursion
2739                 if ( $executing ) {
2740                         return $text;
2741                 }
2742                 $executing = true;
2743
2744                 $this->mTitle = $wgTitle;
2745                 $this->mOptions = $options;
2746                 $this->mOutputType = OT_MSG;
2747                 $this->clearState();
2748                 $text = $this->replaceVariables( $text );
2749
2750                 $executing = false;
2751                 return $text;
2752         }
2753
2754         /**
2755          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2756          * Callback will be called with the text within
2757          * Transform and return the text within
2758          * @access public
2759          */
2760         function setHook( $tag, $callback ) {
2761                 $oldVal = @$this->mTagHooks[$tag];
2762                 $this->mTagHooks[$tag] = $callback;
2763                 return $oldVal;
2764         }
2765
2766         /**
2767          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2768          * Placeholders created in Skin::makeLinkObj()
2769          * Returns an array of links found, indexed by PDBK:
2770          *  0 - broken
2771          *  1 - normal link
2772          *  2 - stub
2773          * $options is a bit field, RLH_FOR_UPDATE to select for update
2774          */
2775         function replaceLinkHolders( &$text, $options = 0 ) {
2776                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2777                 global $wgInterwikiLinkHolders;
2778                 global $outputReplace;
2779
2780                 if ( $wgUseOldExistenceCheck ) {
2781                         return array();
2782                 }
2783
2784                 $fname = 'Parser::replaceLinkHolders';
2785                 wfProfileIn( $fname );
2786
2787                 $pdbks = array();
2788                 $colours = array();
2789
2790                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2791                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2792                         wfProfileIn( $fname.'-check' );
2793                         $dbr =& wfGetDB( DB_SLAVE );
2794                         $page = $dbr->tableName( 'page' );
2795                         $sk = $wgUser->getSkin();
2796                         $threshold = $wgUser->getOption('stubthreshold');
2797
2798                         # Sort by namespace
2799                         asort( $wgLinkHolders['namespaces'] );
2800
2801                         # Generate query
2802                         $query = false;
2803                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2804                                 # Make title object
2805                                 $title = $wgLinkHolders['titles'][$key];
2806
2807                                 # Skip invalid entries.
2808                                 # Result will be ugly, but prevents crash.
2809                                 if ( is_null( $title ) ) {
2810                                         continue;
2811                                 }
2812                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2813
2814                                 # Check if it's in the link cache already
2815                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2816                                         $colours[$pdbk] = 1;
2817                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2818                                         $colours[$pdbk] = 0;
2819                                 } else {
2820                                         # Not in the link cache, add it to the query
2821                                         if ( !isset( $current ) ) {
2822                                                 $current = $val;
2823                                                 $tables = $page;
2824                                                 $join = '';
2825                                                 $query =  "SELECT page_id, page_namespace, page_title";
2826                                                 if ( $threshold > 0 ) {
2827                                                         $textTable = $dbr->tableName( 'text' );
2828                                                         $query .= ', LENGTH(old_text) AS page_len, page_is_redirect';
2829                                                         $tables .= ", $textTable";
2830                                                         $join = 'page_latest=old_id AND';
2831                                                 }
2832                                                 $query .= " FROM $tables WHERE $join (page_namespace=$val AND page_title IN(";
2833                                         } elseif ( $current != $val ) {
2834                                                 $current = $val;
2835                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
2836                                         } else {
2837                                                 $query .= ', ';
2838                                         }
2839
2840                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2841                                 }
2842                         }
2843                         if ( $query ) {
2844                                 $query .= '))';
2845                                 if ( $options & RLH_FOR_UPDATE ) {
2846                                         $query .= ' FOR UPDATE';
2847                                 }
2848
2849                                 $res = $dbr->query( $query, $fname );
2850
2851                                 # Fetch data and form into an associative array
2852                                 # non-existent = broken
2853                                 # 1 = known
2854                                 # 2 = stub
2855                                 while ( $s = $dbr->fetchObject($res) ) {
2856                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
2857                                         $pdbk = $title->getPrefixedDBkey();
2858                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
2859
2860                                         if ( $threshold >  0 ) {
2861                                                 $size = $s->page_len;
2862                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $length < $threshold ) {
2863                                                         $colours[$pdbk] = 1;
2864                                                 } else {
2865                                                         $colours[$pdbk] = 2;
2866                                                 }
2867                                         } else {
2868                                                 $colours[$pdbk] = 1;
2869                                         }
2870                                 }
2871                         }
2872                         wfProfileOut( $fname.'-check' );
2873
2874                         # Construct search and replace arrays
2875                         wfProfileIn( $fname.'-construct' );
2876                         $outputReplace = array();
2877                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
2878                                 $pdbk = $pdbks[$key];
2879                                 $searchkey = '<!--LINK '.$key.'-->';
2880                                 $title = $wgLinkHolders['titles'][$key];
2881                                 if ( empty( $colours[$pdbk] ) ) {
2882                                         $wgLinkCache->addBadLink( $pdbk );
2883                                         $colours[$pdbk] = 0;
2884                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
2885                                                                         $wgLinkHolders['texts'][$key],
2886                                                                         $wgLinkHolders['queries'][$key] );
2887                                 } elseif ( $colours[$pdbk] == 1 ) {
2888                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
2889                                                                         $wgLinkHolders['texts'][$key],
2890                                                                         $wgLinkHolders['queries'][$key] );
2891                                 } elseif ( $colours[$pdbk] == 2 ) {
2892                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
2893                                                                         $wgLinkHolders['texts'][$key],
2894                                                                         $wgLinkHolders['queries'][$key] );
2895                                 }
2896                         }
2897                         wfProfileOut( $fname.'-construct' );
2898
2899                         # Do the thing
2900                         wfProfileIn( $fname.'-replace' );
2901
2902                         $text = preg_replace_callback(
2903                                 '/(<!--LINK .*?-->)/',
2904                                 "outputReplaceMatches",
2905                                 $text);
2906                         wfProfileOut( $fname.'-replace' );
2907                 }
2908
2909                 if ( !empty( $wgInterwikiLinkHolders ) ) {
2910                         wfProfileIn( $fname.'-interwiki' );
2911                         $outputReplace = $wgInterwikiLinkHolders;
2912                         $text = preg_replace_callback(
2913                                 '/<!--IWLINK (.*?)-->/',
2914                                 "outputReplaceMatches",
2915                                 $text );
2916                         wfProfileOut( $fname.'-interwiki' );
2917                 }
2918
2919                 wfProfileOut( $fname );
2920                 return $colours;
2921         }
2922
2923         /**
2924          * Renders an image gallery from a text with one line per image.
2925          * text labels may be given by using |-style alternative text. E.g.
2926          *   Image:one.jpg|The number "1"
2927          *   Image:tree.jpg|A tree
2928          * given as text will return the HTML of a gallery with two images,
2929          * labeled 'The number "1"' and
2930          * 'A tree'.
2931          */
2932         function renderImageGallery( $text ) {
2933                 global $wgLinkCache;
2934                 $ig = new ImageGallery();
2935                 $ig->setShowBytes( false );
2936                 $ig->setShowFilename( false );
2937                 $lines = explode( "\n", $text );
2938
2939                 foreach ( $lines as $line ) {
2940                         # match lines like these:
2941                         # Image:someimage.jpg|This is some image
2942                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
2943                         # Skip empty lines
2944                         if ( count( $matches ) == 0 ) {
2945                                 continue;
2946                         }
2947                         $nt = Title::newFromURL( $matches[1] );
2948                         if( is_null( $nt ) ) {
2949                                 # Bogus title. Ignore these so we don't bomb out later.
2950                                 continue;
2951                         }
2952                         if ( isset( $matches[3] ) ) {
2953                                 $label = $matches[3];
2954                         } else {
2955                                 $label = '';
2956                         }
2957
2958                         # FIXME: Use the full wiki parser and add its links
2959                         # to the page's links.
2960                         $html = $this->mOptions->mSkin->formatComment( $label );
2961
2962                         $ig->add( Image::newFromTitle( $nt ), $html );
2963                         $wgLinkCache->addImageLinkObj( $nt );
2964                 }
2965                 return $ig->toHTML();
2966         }
2967 }
2968
2969 /**
2970  * @todo document
2971  * @package MediaWiki
2972  */
2973 class ParserOutput
2974 {
2975         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2976         var $mCacheTime; # Used in ParserCache
2977         var $mVersion;   # Compatibility check
2978
2979         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2980                 $containsOldMagic = false )
2981         {
2982                 $this->mText = $text;
2983                 $this->mLanguageLinks = $languageLinks;
2984                 $this->mCategoryLinks = $categoryLinks;
2985                 $this->mContainsOldMagic = $containsOldMagic;
2986                 $this->mCacheTime = '';
2987                 $this->mVersion = MW_PARSER_VERSION;
2988         }
2989
2990         function getText()                   { return $this->mText; }
2991         function getLanguageLinks()          { return $this->mLanguageLinks; }
2992         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
2993         function getCacheTime()              { return $this->mCacheTime; }
2994         function containsOldMagic()          { return $this->mContainsOldMagic; }
2995         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
2996         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
2997         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
2998         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2999         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3000         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3001
3002         function merge( $other ) {
3003                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3004                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3005                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3006         }
3007
3008         /**
3009          * Return true if this cached output object predates the global or
3010          * per-article cache invalidation timestamps, or if it comes from
3011          * an incompatible older version.
3012          *
3013          * @param string $touched the affected article's last touched timestamp
3014          * @return bool
3015          * @access public
3016          */
3017         function expired( $touched ) {
3018                 global $wgCacheEpoch;
3019                 return $this->getCacheTime() <= $touched ||
3020                        $this->getCacheTime() <= $wgCacheEpoch ||
3021                        !isset( $this->mVersion ) ||
3022                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3023         }
3024 }
3025
3026 /**
3027  * Set options of the Parser
3028  * @todo document
3029  * @package MediaWiki
3030  */
3031 class ParserOptions
3032 {
3033         # All variables are private
3034         var $mUseTeX;                    # Use texvc to expand <math> tags
3035         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3036         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3037         var $mAllowExternalImages;       # Allow external images inline
3038         var $mSkin;                      # Reference to the preferred skin
3039         var $mDateFormat;                # Date format index
3040         var $mEditSection;               # Create "edit section" links
3041         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3042         var $mNumberHeadings;            # Automatically number headings
3043         var $mShowToc;                   # Show table of contents
3044
3045         function getUseTeX()                        { return $this->mUseTeX; }
3046         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3047         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3048         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3049         function getSkin()                          { return $this->mSkin; }
3050         function getDateFormat()                    { return $this->mDateFormat; }
3051         function getEditSection()                   { return $this->mEditSection; }
3052         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3053         function getNumberHeadings()                { return $this->mNumberHeadings; }
3054         function getShowToc()                       { return $this->mShowToc; }
3055
3056         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3057         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3058         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3059         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3060         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3061         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3062         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3063         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3064         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3065
3066         function setSkin( &$x ) { $this->mSkin =& $x; }
3067
3068         /**
3069          * Get parser options
3070          * @static
3071          */
3072         function newFromUser( &$user ) {
3073                 $popts = new ParserOptions;
3074                 $popts->initialiseFromUser( $user );
3075                 return $popts;
3076         }
3077
3078         /** Get user options */
3079         function initialiseFromUser( &$userInput ) {
3080                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3081                 $fname = 'ParserOptions::initialiseFromUser';
3082                 wfProfileIn( $fname );
3083                 if ( !$userInput ) {
3084                         $user = new User;
3085                         $user->setLoaded( true );
3086                 } else {
3087                         $user =& $userInput;
3088                 }
3089
3090                 $this->mUseTeX = $wgUseTeX;
3091                 $this->mUseDynamicDates = $wgUseDynamicDates;
3092                 $this->mInterwikiMagic = $wgInterwikiMagic;
3093                 $this->mAllowExternalImages = $wgAllowExternalImages;
3094                 wfProfileIn( $fname.'-skin' );
3095                 $this->mSkin =& $user->getSkin();
3096                 wfProfileOut( $fname.'-skin' );
3097                 $this->mDateFormat = $user->getOption( 'date' );
3098                 $this->mEditSection = $user->getOption( 'editsection' );
3099                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3100                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3101                 $this->mShowToc = $user->getOption( 'showtoc' );
3102                 wfProfileOut( $fname );
3103         }
3104 }
3105
3106 /**
3107  * Callback function used by Parser::replaceLinkHolders()
3108  * to substitute link placeholders.
3109  */
3110 function &outputReplaceMatches( $matches ) {
3111         global $outputReplace;
3112         return $outputReplace[$matches[1]];
3113 }
3114
3115 /**
3116  * Return the total number of articles
3117  */
3118 function wfNumberOfArticles() {
3119         global $wgNumberOfArticles;
3120
3121         wfLoadSiteStats();
3122         return $wgNumberOfArticles;
3123 }
3124
3125 /**
3126  * Get various statistics from the database
3127  * @private
3128  */
3129 function wfLoadSiteStats() {
3130         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3131         $fname = 'wfLoadSiteStats';
3132
3133         if ( -1 != $wgNumberOfArticles ) return;
3134         $dbr =& wfGetDB( DB_SLAVE );
3135         $s = $dbr->selectRow( 'site_stats',
3136                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3137                 array( 'ss_row_id' => 1 ), $fname
3138         );
3139
3140         if ( $s === false ) {
3141                 return;
3142         } else {
3143                 $wgTotalViews = $s->ss_total_views;
3144                 $wgTotalEdits = $s->ss_total_edits;
3145                 $wgNumberOfArticles = $s->ss_good_articles;
3146         }
3147 }
3148
3149 /**
3150  * Escape html tags
3151  * Basicly replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3152  *
3153  * @param string $in Text that might contain HTML tags
3154  * @return string Escaped string
3155  */
3156 function wfEscapeHTMLTagsOnly( $in ) {
3157         return str_replace(
3158                 array( '"', '>', '<' ),
3159                 array( '&quot;', '&gt;', '&lt;' ),
3160                 $in );
3161 }
3162
3163 ?>