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                         );
 189                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 190                         $text = Sanitizer::normalizeCharReferences( $text );
 191                 } else {
 192                         $fixtags = array(
 193                                 # french spaces, last one Guillemet-left
 194                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 195                                 # french spaces, Guillemet-right
 196                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 197                                 '/<center *>/i' => '<div class="center">',
 198                                 '/<\\/center *>/i' => '</div>'
 199                         );
 200                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 201                 }
 202                 # only once and last
 203                 $text = $this->doBlockLevels( $text, $linestart );
 204
 205                 $this->replaceLinkHolders( $text );
 206                 $text = $wgContLang->convert($text);
 207
 208                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 209                 global $wgUseTidy;
 210                 if ($wgUseTidy) {
 211                         $text = Parser::tidy($text);
 212                 }
 213
 214                 $this->mOutput->setText( $text );
 215                 wfProfileOut( $fname );
 216                 return $this->mOutput;
 217         }
 218
 219         /**
 220          * Get a random string
 221          *
 222          * @access private
 223          * @static
 224          */
 225         function getRandomString() {
 226                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 227         }
 228
 229         /**
 230          * Replaces all occurrences of <$tag>content</$tag> in the text
 231          * with a random marker and returns the new text. the output parameter
 232          * $content will be an associative array filled with data on the form
 233          * $unique_marker => content.
 234          *
 235          * If $content is already set, the additional entries will be appended
 236          * If $tag is set to STRIP_COMMENTS, the function will extract
 237          * <!-- HTML comments -->
 238          *
 239          * @access private
 240          * @static
 241          */
 242         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 243                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 244                 if ( !$content ) {
 245                         $content = array( );
 246                 }
 247                 $n = 1;
 248                 $stripped = '';
 249
 250                 while ( '' != $text ) {
 251                         if($tag==STRIP_COMMENTS) {
 252                                 $p = preg_split( '/<!--/i', $text, 2 );
 253                         } else {
 254                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 255                         }
 256                         $stripped .= $p[0];
 257                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 258                                 $text = '';
 259                         } else {
 260                                 if($tag==STRIP_COMMENTS) {
 261                                         $q = preg_split( '/-->/i', $p[1], 2 );
 262                                 } else {
 263                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 264                                 }
 265                                 $marker = $rnd . sprintf('%08X', $n++);
 266                                 $content[$marker] = $q[0];
 267                                 $stripped .= $marker;
 268                                 $text = $q[1];
 269                         }
 270                 }
 271                 return $stripped;
 272         }
 273
 274         /**
 275          * Strips and renders nowiki, pre, math, hiero
 276          * If $render is set, performs necessary rendering operations on plugins
 277          * Returns the text, and fills an array with data needed in unstrip()
 278          * If the $state is already a valid strip state, it adds to the state
 279          *
 280          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 281          *  will be stripped in addition to other tags. This is important
 282          *  for section editing, where these comments cause confusion when
 283          *  counting the sections in the wikisource
 284          *
 285          * @access private
 286          */
 287         function strip( $text, &$state, $stripcomments = false ) {
 288                 $render = ($this->mOutputType == OT_HTML);
 289                 $html_content = array();
 290                 $nowiki_content = array();
 291                 $math_content = array();
 292                 $pre_content = array();
 293                 $comment_content = array();
 294                 $ext_content = array();
 295                 $gallery_content = array();
 296
 297                 # Replace any instances of the placeholders
 298                 $uniq_prefix = UNIQ_PREFIX;
 299                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 300
 301                 # html
 302                 global $wgRawHtml, $wgWhitelistEdit;
 303                 if( $wgRawHtml && $wgWhitelistEdit ) {
 304                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 305                         foreach( $html_content as $marker => $content ) {
 306                                 if ($render ) {
 307                                         # Raw and unchecked for validity.
 308                                         $html_content[$marker] = $content;
 309                                 } else {
 310                                         $html_content[$marker] = '<html>'.$content.'</html>';
 311                                 }
 312                         }
 313                 }
 314
 315                 # nowiki
 316                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 317                 foreach( $nowiki_content as $marker => $content ) {
 318                         if( $render ){
 319                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 320                         } else {
 321                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 322                         }
 323                 }
 324
 325                 # math
 326                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 327                 foreach( $math_content as $marker => $content ){
 328                         if( $render ) {
 329                                 if( $this->mOptions->getUseTeX() ) {
 330                                         $math_content[$marker] = renderMath( $content );
 331                                 } else {
 332                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 333                                 }
 334                         } else {
 335                                 $math_content[$marker] = '<math>'.$content.'</math>';
 336                         }
 337                 }
 338
 339                 # pre
 340                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 341                 foreach( $pre_content as $marker => $content ){
 342                         if( $render ){
 343                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 344                         } else {
 345                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 346                         }
 347                 }
 348
 349                 # gallery
 350                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 351                 foreach( $gallery_content as $marker => $content ) {
 352                         require_once( 'ImageGallery.php' );
 353                         if ( $render ) {
 354                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 355                         } else {
 356                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 357                         }
 358                 }
 359
 360                 # Comments
 361                 if($stripcomments) {
 362                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 363                         foreach( $comment_content as $marker => $content ){
 364                                 $comment_content[$marker] = '<!--'.$content.'-->';
 365                         }
 366                 }
 367
 368                 # Extensions
 369                 foreach ( $this->mTagHooks as $tag => $callback ) {
 370                         $ext_contents[$tag] = array();
 371                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 372                         foreach( $ext_content[$tag] as $marker => $content ) {
 373                                 if ( $render ) {
 374                                         $ext_content[$tag][$marker] = $callback( $content );
 375                                 } else {
 376                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 377                                 }
 378                         }
 379                 }
 380
 381                 # Merge state with the pre-existing state, if there is one
 382                 if ( $state ) {
 383                         $state['html'] = $state['html'] + $html_content;
 384                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 385                         $state['math'] = $state['math'] + $math_content;
 386                         $state['pre'] = $state['pre'] + $pre_content;
 387                         $state['comment'] = $state['comment'] + $comment_content;
 388                         $state['gallery'] = $state['gallery'] + $gallery_content;
 389
 390                         foreach( $ext_content as $tag => $array ) {
 391                                 if ( array_key_exists( $tag, $state ) ) {
 392                                         $state[$tag] = $state[$tag] + $array;
 393                                 }
 394                         }
 395                 } else {
 396                         $state = array(
 397                           'html' => $html_content,
 398                           'nowiki' => $nowiki_content,
 399                           'math' => $math_content,
 400                           'pre' => $pre_content,
 401                           'comment' => $comment_content,
 402                           'gallery' => $gallery_content,
 403                         ) + $ext_content;
 404                 }
 405                 return $text;
 406         }
 407
 408         /**
 409          * restores pre, math, and hiero removed by strip()
 410          *
 411          * always call unstripNoWiki() after this one
 412          * @access private
 413          */
 414         function unstrip( $text, &$state ) {
 415                 # Must expand in reverse order, otherwise nested tags will be corrupted
 416                 $contentDict = end( $state );
 417                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 418                         if( key($state) != 'nowiki' && key($state) != 'html') {
 419                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 420                                         $text = str_replace( key( $contentDict ), $content, $text );
 421                                 }
 422                         }
 423                 }
 424
 425                 return $text;
 426         }
 427
 428         /**
 429          * always call this after unstrip() to preserve the order
 430          *
 431          * @access private
 432          */
 433         function unstripNoWiki( $text, &$state ) {
 434                 # Must expand in reverse order, otherwise nested tags will be corrupted
 435                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 436                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 437                 }
 438
 439                 global $wgRawHtml;
 440                 if ($wgRawHtml) {
 441                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 442                                 $text = str_replace( key( $state['html'] ), $content, $text );
 443                         }
 444                 }
 445
 446                 return $text;
 447         }
 448
 449         /**
 450          * Add an item to the strip state
 451          * Returns the unique tag which must be inserted into the stripped text
 452          * The tag will be replaced with the original text in unstrip()
 453          *
 454          * @access private
 455          */
 456         function insertStripItem( $text, &$state ) {
 457                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 458                 if ( !$state ) {
 459                         $state = array(
 460                           'html' => array(),
 461                           'nowiki' => array(),
 462                           'math' => array(),
 463                           'pre' => array()
 464                         );
 465                 }
 466                 $state['item'][$rnd] = $text;
 467                 return $rnd;
 468         }
 469
 470         /**
 471          * interface with html tidy, used if $wgUseTidy = true
 472          *
 473          * @access public
 474          * @static
 475          */
 476         function tidy ( $text ) {
 477                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 478                 global $wgInputEncoding, $wgOutputEncoding;
 479                 $fname = 'Parser::tidy';
 480                 wfProfileIn( $fname );
 481
 482                 $cleansource = '';
 483                 $opts = '';
 484                 switch(strtoupper($wgOutputEncoding)) {
 485                         case 'ISO-8859-1':
 486                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 487                                 break;
 488                         case 'UTF-8':
 489                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 490                                 break;
 491                         default:
 492                                 $opts .= ' -raw';
 493                         }
 494
 495                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 496 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 497 '<head><title>test</title></head><body>'.$text.'</body></html>';
 498                 $descriptorspec = array(
 499                         0 => array('pipe', 'r'),
 500                         1 => array('pipe', 'w'),
 501                         2 => array('file', '/dev/null', 'a')
 502                 );
 503                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 504                 if (is_resource($process)) {
 505                         fwrite($pipes[0], $wrappedtext);
 506                         fclose($pipes[0]);
 507                         while (!feof($pipes[1])) {
 508                                 $cleansource .= fgets($pipes[1], 1024);
 509                         }
 510                         fclose($pipes[1]);
 511                         $return_value = proc_close($process);
 512                 }
 513
 514                 wfProfileOut( $fname );
 515
 516                 if( $cleansource == '' && $text != '') {
 517                         wfDebug( "Tidy error detected!\n" );
 518                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 519                 } else {
 520                         return $cleansource;
 521                 }
 522         }
 523
 524         /**
 525          * parse the wiki syntax used to render tables
 526          *
 527          * @access private
 528          */
 529         function doTableStuff ( $t ) {
 530                 $fname = 'Parser::doTableStuff';
 531                 wfProfileIn( $fname );
 532
 533                 $t = explode ( "\n" , $t ) ;
 534                 $td = array () ; # Is currently a td tag open?
 535                 $ltd = array () ; # Was it TD or TH?
 536                 $tr = array () ; # Is currently a tr tag open?
 537                 $ltr = array () ; # tr attributes
 538                 $indent_level = 0; # indent level of the table
 539                 foreach ( $t AS $k => $x )
 540                 {
 541                         $x = trim ( $x ) ;
 542                         $fc = substr ( $x , 0 , 1 ) ;
 543                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 544                                 $indent_level = strlen( $matches[1] );
 545                                 $t[$k] = "\n" .
 546                                         str_repeat( '<dl><dd>', $indent_level ) .
 547                                         '<table' . Sanitizer::fixTagAttributes ( $matches[2], 'table' ) . '>' ;
 548                                 array_push ( $td , false ) ;
 549                                 array_push ( $ltd , '' ) ;
 550                                 array_push ( $tr , false ) ;
 551                                 array_push ( $ltr , '' ) ;
 552                         }
 553                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 554                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 555                                 $z = "</table>\n" ;
 556                                 $l = array_pop ( $ltd ) ;
 557                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 558                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 559                                 array_pop ( $ltr ) ;
 560                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 561                         }
 562                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 563                                 $x = substr ( $x , 1 ) ;
 564                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 565                                 $z = '' ;
 566                                 $l = array_pop ( $ltd ) ;
 567                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 568                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 569                                 array_pop ( $ltr ) ;
 570                                 $t[$k] = $z ;
 571                                 array_push ( $tr , false ) ;
 572                                 array_push ( $td , false ) ;
 573                                 array_push ( $ltd , '' ) ;
 574                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $x, 'tr' ) ) ;
 575                         }
 576                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 577                                 # $x is a table row
 578                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 579                                         $fc = '+' ;
 580                                         $x = substr ( $x , 1 ) ;
 581                                 }
 582                                 $after = substr ( $x , 1 ) ;
 583                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 584                                 $after = explode ( '||' , $after ) ;
 585                                 $t[$k] = '' ;
 586
 587                                 # Loop through each table cell
 588                                 foreach ( $after AS $theline )
 589                                 {
 590                                         $z = '' ;
 591                                         if ( $fc != '+' )
 592                                         {
 593                                                 $tra = array_pop ( $ltr ) ;
 594                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 595                                                 array_push ( $tr , true ) ;
 596                                                 array_push ( $ltr , '' ) ;
 597                                         }
 598
 599                                         $l = array_pop ( $ltd ) ;
 600                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 601                                         if ( $fc == '|' ) $l = 'td' ;
 602                                         else if ( $fc == '!' ) $l = 'th' ;
 603                                         else if ( $fc == '+' ) $l = 'caption' ;
 604                                         else $l = '' ;
 605                                         array_push ( $ltd , $l ) ;
 606
 607                                         # Cell parameters
 608                                         $y = explode ( '|' , $theline , 2 ) ;
 609                                         # Note that a '|' inside an invalid link should not
 610                                         # be mistaken as delimiting cell parameters
 611                                         if ( strpos( $y[0], '[[' ) !== false ) {
 612                                                 $y = array ($theline);
 613                                         }
 614                                         if ( count ( $y ) == 1 )
 615                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 616                                         else $y = $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($y[0], $l).">{$y[1]}" ;
 617                                         $t[$k] .= $y ;
 618                                         array_push ( $td , true ) ;
 619                                 }
 620                         }
 621                 }
 622
 623                 # Closing open td, tr && table
 624                 while ( count ( $td ) > 0 )
 625                 {
 626                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 627                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 628                         $t[] = '</table>' ;
 629                 }
 630
 631                 $t = implode ( "\n" , $t ) ;
 632                 #               $t = Sanitizer::removeHTMLtags( $t );
 633                 wfProfileOut( $fname );
 634                 return $t ;
 635         }
 636
 637         /**
 638          * Helper function for parse() that transforms wiki markup into
 639          * HTML. Only called for $mOutputType == OT_HTML.
 640          *
 641          * @access private
 642          */
 643         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 644                 global $wgContLang;
 645
 646                 $fname = 'Parser::internalParse';
 647                 wfProfileIn( $fname );
 648
 649                 $text = Sanitizer::removeHTMLtags( $text );
 650                 $text = $this->replaceVariables( $text, $args );
 651
 652                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 653
 654                 $text = $this->doHeadings( $text );
 655                 if($this->mOptions->getUseDynamicDates()) {
 656                         global $wgDateFormatter;
 657                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 658                 }
 659                 $text = $this->doAllQuotes( $text );
 660                 $text = $this->replaceInternalLinks( $text );
 661                 $text = $this->replaceExternalLinks( $text );
 662
 663                 # replaceInternalLinks may sometimes leave behind
 664                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 665                 $text = str_replace("http-noparse://","http://",$text);
 666
 667                 $text = $this->doMagicLinks( $text );
 668                 $text = $this->doTableStuff( $text );
 669                 $text = $this->formatHeadings( $text, $isMain );
 670
 671                 wfProfileOut( $fname );
 672                 return $text;
 673         }
 674
 675         /**
 676          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 677          * magic external links.
 678          *
 679          * @access private
 680          */
 681         function &doMagicLinks( &$text ) {
 682                 global $wgUseGeoMode;
 683                 $text = $this->magicISBN( $text );
 684                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 685                         $text = $this->magicGEO( $text );
 686                 }
 687                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 688                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 689                 return $text;
 690         }
 691
 692         /**
 693          * Parse ^^ tokens and return html
 694          *
 695          * @access private
 696          */
 697         function doExponent( $text ) {
 698                 $fname = 'Parser::doExponent';
 699                 wfProfileIn( $fname );
 700                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 701                 wfProfileOut( $fname );
 702                 return $text;
 703         }
 704
 705         /**
 706          * Parse headers and return html
 707          *
 708          * @access private
 709          */
 710         function doHeadings( $text ) {
 711                 $fname = 'Parser::doHeadings';
 712                 wfProfileIn( $fname );
 713                 for ( $i = 6; $i >= 1; --$i ) {
 714                         $h = substr( '======', 0, $i );
 715                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 716                           "<h{$i}>\\1</h{$i}>\\2", $text );
 717                 }
 718                 wfProfileOut( $fname );
 719                 return $text;
 720         }
 721
 722         /**
 723          * Replace single quotes with HTML markup
 724          * @access private
 725          * @return string the altered text
 726          */
 727         function doAllQuotes( $text ) {
 728                 $fname = 'Parser::doAllQuotes';
 729                 wfProfileIn( $fname );
 730                 $outtext = '';
 731                 $lines = explode( "\n", $text );
 732                 foreach ( $lines as $line ) {
 733                         $outtext .= $this->doQuotes ( $line ) . "\n";
 734                 }
 735                 $outtext = substr($outtext, 0,-1);
 736                 wfProfileOut( $fname );
 737                 return $outtext;
 738         }
 739
 740         /**
 741          * Helper function for doAllQuotes()
 742          * @access private
 743          */
 744         function doQuotes( $text ) {
 745                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 746                 if ( count( $arr ) == 1 )
 747                         return $text;
 748                 else
 749                 {
 750                         # First, do some preliminary work. This may shift some apostrophes from
 751                         # being mark-up to being text. It also counts the number of occurrences
 752                         # of bold and italics mark-ups.
 753                         $i = 0;
 754                         $numbold = 0;
 755                         $numitalics = 0;
 756                         foreach ( $arr as $r )
 757                         {
 758                                 if ( ( $i % 2 ) == 1 )
 759                                 {
 760                                         # If there are ever four apostrophes, assume the first is supposed to
 761                                         # be text, and the remaining three constitute mark-up for bold text.
 762                                         if ( strlen( $arr[$i] ) == 4 )
 763                                         {
 764                                                 $arr[$i-1] .= "'";
 765                                                 $arr[$i] = "'''";
 766                                         }
 767                                         # If there are more than 5 apostrophes in a row, assume they're all
 768                                         # text except for the last 5.
 769                                         else if ( strlen( $arr[$i] ) > 5 )
 770                                         {
 771                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 772                                                 $arr[$i] = "'''''";
 773                                         }
 774                                         # Count the number of occurrences of bold and italics mark-ups.
 775                                         # We are not counting sequences of five apostrophes.
 776                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 777                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 778                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 779                                 }
 780                                 $i++;
 781                         }
 782
 783                         # If there is an odd number of both bold and italics, it is likely
 784                         # that one of the bold ones was meant to be an apostrophe followed
 785                         # by italics. Which one we cannot know for certain, but it is more
 786                         # likely to be one that has a single-letter word before it.
 787                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 788                         {
 789                                 $i = 0;
 790                                 $firstsingleletterword = -1;
 791                                 $firstmultiletterword = -1;
 792                                 $firstspace = -1;
 793                                 foreach ( $arr as $r )
 794                                 {
 795                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 796                                         {
 797                                                 $x1 = substr ($arr[$i-1], -1);
 798                                                 $x2 = substr ($arr[$i-1], -2, 1);
 799                                                 if ($x1 == ' ') {
 800                                                         if ($firstspace == -1) $firstspace = $i;
 801                                                 } else if ($x2 == ' ') {
 802                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 803                                                 } else {
 804                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 805                                                 }
 806                                         }
 807                                         $i++;
 808                                 }
 809
 810                                 # If there is a single-letter word, use it!
 811                                 if ($firstsingleletterword > -1)
 812                                 {
 813                                         $arr [ $firstsingleletterword ] = "''";
 814                                         $arr [ $firstsingleletterword-1 ] .= "'";
 815                                 }
 816                                 # If not, but there's a multi-letter word, use that one.
 817                                 else if ($firstmultiletterword > -1)
 818                                 {
 819                                         $arr [ $firstmultiletterword ] = "''";
 820                                         $arr [ $firstmultiletterword-1 ] .= "'";
 821                                 }
 822                                 # ... otherwise use the first one that has neither.
 823                                 # (notice that it is possible for all three to be -1 if, for example,
 824                                 # there is only one pentuple-apostrophe in the line)
 825                                 else if ($firstspace > -1)
 826                                 {
 827                                         $arr [ $firstspace ] = "''";
 828                                         $arr [ $firstspace-1 ] .= "'";
 829                                 }
 830                         }
 831
 832                         # Now let's actually convert our apostrophic mush to HTML!
 833                         $output = '';
 834                         $buffer = '';
 835                         $state = '';
 836                         $i = 0;
 837                         foreach ($arr as $r)
 838                         {
 839                                 if (($i % 2) == 0)
 840                                 {
 841                                         if ($state == 'both')
 842                                                 $buffer .= $r;
 843                                         else
 844                                                 $output .= $r;
 845                                 }
 846                                 else
 847                                 {
 848                                         if (strlen ($r) == 2)
 849                                         {
 850                                                 if ($state == 'i')
 851                                                 { $output .= '</i>'; $state = ''; }
 852                                                 else if ($state == 'bi')
 853                                                 { $output .= '</i>'; $state = 'b'; }
 854                                                 else if ($state == 'ib')
 855                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 856                                                 else if ($state == 'both')
 857                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 858                                                 else # $state can be 'b' or ''
 859                                                 { $output .= '<i>'; $state .= 'i'; }
 860                                         }
 861                                         else if (strlen ($r) == 3)
 862                                         {
 863                                                 if ($state == 'b')
 864                                                 { $output .= '</b>'; $state = ''; }
 865                                                 else if ($state == 'bi')
 866                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 867                                                 else if ($state == 'ib')
 868                                                 { $output .= '</b>'; $state = 'i'; }
 869                                                 else if ($state == 'both')
 870                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 871                                                 else # $state can be 'i' or ''
 872                                                 { $output .= '<b>'; $state .= 'b'; }
 873                                         }
 874                                         else if (strlen ($r) == 5)
 875                                         {
 876                                                 if ($state == 'b')
 877                                                 { $output .= '</b><i>'; $state = 'i'; }
 878                                                 else if ($state == 'i')
 879                                                 { $output .= '</i><b>'; $state = 'b'; }
 880                                                 else if ($state == 'bi')
 881                                                 { $output .= '</i></b>'; $state = ''; }
 882                                                 else if ($state == 'ib')
 883                                                 { $output .= '</b></i>'; $state = ''; }
 884                                                 else if ($state == 'both')
 885                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 886                                                 else # ($state == '')
 887                                                 { $buffer = ''; $state = 'both'; }
 888                                         }
 889                                 }
 890                                 $i++;
 891                         }
 892                         # Now close all remaining tags.  Notice that the order is important.
 893                         if ($state == 'b' || $state == 'ib')
 894                                 $output .= '</b>';
 895                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 896                                 $output .= '</i>';
 897                         if ($state == 'bi')
 898                                 $output .= '</b>';
 899                         if ($state == 'both')
 900                                 $output .= '<b><i>'.$buffer.'</i></b>';
 901                         return $output;
 902                 }
 903         }
 904
 905         /**
 906          * Replace external links
 907          *
 908          * Note: this is all very hackish and the order of execution matters a lot.
 909          * Make sure to run maintenance/parserTests.php if you change this code.
 910          *
 911          * @access private
 912          */
 913         function replaceExternalLinks( $text ) {
 914                 $fname = 'Parser::replaceExternalLinks';
 915                 wfProfileIn( $fname );
 916
 917                 $sk =& $this->mOptions->getSkin();
 918                 global $wgContLang;
 919                 $linktrail = $wgContLang->linkTrail();
 920
 921                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 922
 923                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 924
 925                 $i = 0;
 926                 while ( $i<count( $bits ) ) {
 927                         $url = $bits[$i++];
 928                         $protocol = $bits[$i++];
 929                         $text = $bits[$i++];
 930                         $trail = $bits[$i++];
 931
 932                         # The characters '<' and '>' (which were escaped by
 933                         # removeHTMLtags()) should not be included in
 934                         # URLs, per RFC 2396.
 935                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 936                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 937                                 $url = substr($url, 0, $m2[0][1]);
 938                         }
 939
 940                         # If the link text is an image URL, replace it with an <img> tag
 941                         # This happened by accident in the original parser, but some people used it extensively
 942                         $img = $this->maybeMakeImageLink( $text );
 943                         if ( $img !== false ) {
 944                                 $text = $img;
 945                         }
 946
 947                         $dtrail = '';
 948
 949                         # Set linktype for CSS - if URL==text, link is essentially free
 950                         $linktype = ($text == $url) ? 'free' : 'text';
 951
 952                         # No link text, e.g. [http://domain.tld/some.link]
 953                         if ( $text == '' ) {
 954                                 # Autonumber if allowed
 955                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 956                                         $text = '[' . ++$this->mAutonumber . ']';
 957                                         $linktype = 'autonumber';
 958                                 } else {
 959                                         # Otherwise just use the URL
 960                                         $text = htmlspecialchars( $url );
 961                                         $linktype = 'free';
 962                                 }
 963                         } else {
 964                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 965                                 # Check for trail
 966                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 967                                         $dtrail = $m2[1];
 968                                         $trail = $m2[2];
 969                                 }
 970                         }
 971
 972                         # Replace &amp; from obsolete syntax with &.
 973                         # All HTML entities will be escaped by makeExternalLink()
 974                         # or maybeMakeImageLink()
 975                         $url = str_replace( '&amp;', '&', $url );
 976
 977                         # Process the trail (i.e. everything after this link up until start of the next link),
 978                         # replacing any non-bracketed links
 979                         $trail = $this->replaceFreeExternalLinks( $trail );
 980
 981                         # Use the encoded URL
 982                         # This means that users can paste URLs directly into the text
 983                         # Funny characters like &ouml; aren't valid in URLs anyway
 984                         # This was changed in August 2004
 985                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
 986                 }
 987
 988                 wfProfileOut( $fname );
 989                 return $s;
 990         }
 991
 992         /**
 993          * Replace anything that looks like a URL with a link
 994          * @access private
 995          */
 996         function replaceFreeExternalLinks( $text ) {
 997                 $fname = 'Parser::replaceFreeExternalLinks';
 998                 wfProfileIn( $fname );
 999
1000                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1001                 $s = array_shift( $bits );
1002                 $i = 0;
1003
1004                 $sk =& $this->mOptions->getSkin();
1005
1006                 while ( $i < count( $bits ) ){
1007                         $protocol = $bits[$i++];
1008                         $remainder = $bits[$i++];
1009
1010                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1011                                 # Found some characters after the protocol that look promising
1012                                 $url = $protocol . $m[1];
1013                                 $trail = $m[2];
1014
1015                                 # The characters '<' and '>' (which were escaped by
1016                                 # removeHTMLtags()) should not be included in
1017                                 # URLs, per RFC 2396.
1018                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1019                                         $trail = substr($url, $m2[0][1]) . $trail;
1020                                         $url = substr($url, 0, $m2[0][1]);
1021                                 }
1022
1023                                 # Move trailing punctuation to $trail
1024                                 $sep = ',;\.:!?';
1025                                 # If there is no left bracket, then consider right brackets fair game too
1026                                 if ( strpos( $url, '(' ) === false ) {
1027                                         $sep .= ')';
1028                                 }
1029
1030                                 $numSepChars = strspn( strrev( $url ), $sep );
1031                                 if ( $numSepChars ) {
1032                                         $trail = substr( $url, -$numSepChars ) . $trail;
1033                                         $url = substr( $url, 0, -$numSepChars );
1034                                 }
1035
1036                                 # Replace &amp; from obsolete syntax with &.
1037                                 # All HTML entities will be escaped by makeExternalLink()
1038                                 # or maybeMakeImageLink()
1039                                 $url = str_replace( '&amp;', '&', $url );
1040
1041                                 # Is this an external image?
1042                                 $text = $this->maybeMakeImageLink( $url );
1043                                 if ( $text === false ) {
1044                                         # Not an image, make a link
1045                                         $text = $sk->makeExternalLink( $url, $url, true, 'free' );
1046                                 }
1047                                 $s .= $text . $trail;
1048                         } else {
1049                                 $s .= $protocol . $remainder;
1050                         }
1051                 }
1052                 wfProfileOut();
1053                 return $s;
1054         }
1055
1056         /**
1057          * make an image if it's allowed
1058          * @access private
1059          */
1060         function maybeMakeImageLink( $url ) {
1061                 $sk =& $this->mOptions->getSkin();
1062                 $text = false;
1063                 if ( $this->mOptions->getAllowExternalImages() ) {
1064                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1065                                 # Image found
1066                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1067                         }
1068                 }
1069                 return $text;
1070         }
1071
1072         /**
1073          * Process [[ ]] wikilinks
1074          *
1075          * @access private
1076          */
1077         function replaceInternalLinks( $s ) {
1078                 global $wgLang, $wgContLang, $wgLinkCache;
1079                 global $wgDisableLangConversion;
1080                 static $fname = 'Parser::replaceInternalLinks' ;
1081
1082                 wfProfileIn( $fname );
1083
1084                 wfProfileIn( $fname.'-setup' );
1085                 static $tc = FALSE;
1086                 # the % is needed to support urlencoded titles as well
1087                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1088
1089                 $sk =& $this->mOptions->getSkin();
1090                 global $wgUseOldExistenceCheck;
1091                 # "Post-parse link colour check" works only on wiki text since it's now
1092                 # in Parser. Enable it, then disable it when we're done.
1093                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1094
1095                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1096
1097                 #split the entire text string on occurences of [[
1098                 $a = explode( '[[', ' ' . $s );
1099                 #get the first element (all text up to first [[), and remove the space we added
1100                 $s = array_shift( $a );
1101                 $s = substr( $s, 1 );
1102
1103                 # Match a link having the form [[namespace:link|alternate]]trail
1104                 static $e1 = FALSE;
1105                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1106                 # Match cases where there is no "]]", which might still be images
1107                 static $e1_img = FALSE;
1108                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1109                 # Match the end of a line for a word that's not followed by whitespace,
1110                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1111                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1112
1113                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1114
1115                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1116
1117                 if ( $useLinkPrefixExtension ) {
1118                         if ( preg_match( $e2, $s, $m ) ) {
1119                                 $first_prefix = $m[2];
1120                                 $s = $m[1];
1121                         } else {
1122                                 $first_prefix = false;
1123                         }
1124                 } else {
1125                         $prefix = '';
1126                 }
1127
1128                 $selflink = $this->mTitle->getPrefixedText();
1129                 wfProfileOut( $fname.'-setup' );
1130
1131                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1132                 $useSubpages = $this->areSubpagesAllowed();
1133
1134                 # Loop for each link
1135                 for ($k = 0; isset( $a[$k] ); $k++) {
1136                         $line = $a[$k];
1137                         if ( $useLinkPrefixExtension ) {
1138                                 wfProfileIn( $fname.'-prefixhandling' );
1139                                 if ( preg_match( $e2, $s, $m ) ) {
1140                                         $prefix = $m[2];
1141                                         $s = $m[1];
1142                                 } else {
1143                                         $prefix='';
1144                                 }
1145                                 # first link
1146                                 if($first_prefix) {
1147                                         $prefix = $first_prefix;
1148                                         $first_prefix = false;
1149                                 }
1150                                 wfProfileOut( $fname.'-prefixhandling' );
1151                         }
1152
1153                         $might_be_img = false;
1154
1155                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1156                                 $text = $m[2];
1157                                 # fix up urlencoded title texts
1158                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1159                                 $trail = $m[3];
1160                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1161                                 $might_be_img = true;
1162                                 $text = $m[2];
1163                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1164                                 $trail = "";
1165                         } else { # Invalid form; output directly
1166                                 $s .= $prefix . '[[' . $line ;
1167                                 continue;
1168                         }
1169
1170                         # Don't allow internal links to pages containing
1171                         # PROTO: where PROTO is a valid URL protocol; these
1172                         # should be external links.
1173                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1174                                 $s .= $prefix . '[[' . $line ;
1175                                 continue;
1176                         }
1177
1178                         # Make subpage if necessary
1179                         if( $useSubpages ) {
1180                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1181                         } else {
1182                                 $link = $m[1];
1183                         }
1184
1185                         $noforce = (substr($m[1], 0, 1) != ':');
1186                         if (!$noforce) {
1187                                 # Strip off leading ':'
1188                                 $link = substr($link, 1);
1189                         }
1190
1191                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1192                         if( !$nt ) {
1193                                 $s .= $prefix . '[[' . $line;
1194                                 continue;
1195                         }
1196
1197                         #check other language variants of the link
1198                         #if the article does not exist
1199                         if( $checkVariantLink
1200                             && $nt->getArticleID() == 0 ) {
1201                                 $wgContLang->findVariantLink($link, $nt);
1202                         }
1203
1204                         $ns = $nt->getNamespace();
1205                         $iw = $nt->getInterWiki();
1206
1207                         if ($might_be_img) { # if this is actually an invalid link
1208                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1209                                         $found = false;
1210                                         while (isset ($a[$k+1]) ) {
1211                                                 #look at the next 'line' to see if we can close it there
1212                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1213                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1214                                                 # the first ]] closes the inner link, the second the image
1215                                                         $found = true;
1216                                                         $text .= '[[' . $m[1];
1217                                                         $trail = $m[2];
1218                                                         break;
1219                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1220                                                         #if there's exactly one ]] that's fine, we'll keep looking
1221                                                         $text .= '[[' . $m[0];
1222                                                 } else {
1223                                                         #if $next_line is invalid too, we need look no further
1224                                                         $text .= '[[' . $next_line;
1225                                                         break;
1226                                                 }
1227                                         }
1228                                         if ( !$found ) {
1229                                                 # we couldn't find the end of this imageLink, so output it raw
1230                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1231                                                 $text = $this->replaceInternalLinks($text);
1232                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1233                                                 # note: no $trail, because without an end, there *is* no trail
1234                                                 continue;
1235                                         }
1236                                 } else { #it's not an image, so output it raw
1237                                         $s .= $prefix . '[[' . $link . '|' . $text;
1238                                         # note: no $trail, because without an end, there *is* no trail
1239                                         continue;
1240                                 }
1241                         }
1242
1243                         $wasblank = ( '' == $text );
1244                         if( $wasblank ) $text = $link;
1245
1246
1247                         # Link not escaped by : , create the various objects
1248                         if( $noforce ) {
1249
1250                                 # Interwikis
1251                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1252                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1253                                         $tmp = $prefix . $trail ;
1254                                         $s .= (trim($tmp) == '')? '': $tmp;
1255                                         continue;
1256                                 }
1257
1258                                 if ( $ns == NS_IMAGE ) {
1259                                         wfProfileIn( "$fname-image" );
1260
1261                                         # recursively parse links inside the image caption
1262                                         # actually, this will parse them in any other parameters, too,
1263                                         # but it might be hard to fix that, and it doesn't matter ATM
1264                                         $text = $this->replaceExternalLinks($text);
1265                                         $text = $this->replaceInternalLinks($text);
1266
1267                                         # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1268                                         $s .= $prefix . str_replace('http://', 'http-noparse://', $sk->makeImageLinkObj( $nt, $text ) ) . $trail;
1269                                         $wgLinkCache->addImageLinkObj( $nt );
1270
1271                                         wfProfileOut( "$fname-image" );
1272                                         continue;
1273                                 }
1274
1275                                 if ( $ns == NS_CATEGORY ) {
1276                                         wfProfileIn( "$fname-category" );
1277                                         $t = $nt->getText();
1278
1279                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1280                                         $pPLC=$sk->postParseLinkColour();
1281                                         $sk->postParseLinkColour( false );
1282                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1283                                         $sk->postParseLinkColour( $pPLC );
1284                                         $wgLinkCache->resume();
1285
1286                                         if ( $wasblank ) {
1287                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1288                                                         $sortkey = $this->mTitle->getText();
1289                                                 } else {
1290                                                         $sortkey = $this->mTitle->getPrefixedText();
1291                                                 }
1292                                         } else {
1293                                                 $sortkey = $text;
1294                                         }
1295                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1296                                         $this->mOutput->addCategoryLink( $t );
1297                                         $s .= $prefix . $trail ;
1298
1299                                         wfProfileOut( "$fname-category" );
1300                                         continue;
1301                                 }
1302                         }
1303
1304                         if( ( $nt->getPrefixedText() === $selflink ) &&
1305                             ( $nt->getFragment() === '' ) ) {
1306                                 # Self-links are handled specially; generally de-link and change to bold.
1307                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1308                                 continue;
1309                         }
1310
1311                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1312                         if( $ns == NS_MEDIA ) {
1313                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1314                                 $wgLinkCache->addImageLinkObj( $nt );
1315                                 continue;
1316                         } elseif( $ns == NS_SPECIAL ) {
1317                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1318                                 continue;
1319                         }
1320                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1321                 }
1322                 $sk->postParseLinkColour( $saveParseColour );
1323                 wfProfileOut( $fname );
1324                 return $s;
1325         }
1326
1327         /**
1328          * Return true if subpage links should be expanded on this page.
1329          * @return bool
1330          */
1331         function areSubpagesAllowed() {
1332                 # Some namespaces don't allow subpages
1333                 global $wgNamespacesWithSubpages;
1334                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1335         }
1336
1337         /**
1338          * Handle link to subpage if necessary
1339          * @param string $target the source of the link
1340          * @param string &$text the link text, modified as necessary
1341          * @return string the full name of the link
1342          * @access private
1343          */
1344         function maybeDoSubpageLink($target, &$text) {
1345                 # Valid link forms:
1346                 # Foobar -- normal
1347                 # :Foobar -- override special treatment of prefix (images, language links)
1348                 # /Foobar -- convert to CurrentPage/Foobar
1349                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1350                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1351                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1352
1353                 $fname = 'Parser::maybeDoSubpageLink';
1354                 wfProfileIn( $fname );
1355                 $ret = $target; # default return value is no change
1356
1357                 # Some namespaces don't allow subpages,
1358                 # so only perform processing if subpages are allowed
1359                 if( $this->areSubpagesAllowed() ) {
1360                         # Look at the first character
1361                         if( $target != '' && $target{0} == '/' ) {
1362                                 # / at end means we don't want the slash to be shown
1363                                 if( substr( $target, -1, 1 ) == '/' ) {
1364                                         $target = substr( $target, 1, -1 );
1365                                         $noslash = $target;
1366                                 } else {
1367                                         $noslash = substr( $target, 1 );
1368                                 }
1369
1370                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1371                                 if( '' === $text ) {
1372                                         $text = $target;
1373                                 } # this might be changed for ugliness reasons
1374                         } else {
1375                                 # check for .. subpage backlinks
1376                                 $dotdotcount = 0;
1377                                 $nodotdot = $target;
1378                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1379                                         ++$dotdotcount;
1380                                         $nodotdot = substr( $nodotdot, 3 );
1381                                 }
1382                                 if($dotdotcount > 0) {
1383                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1384                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1385                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1386                                                 # / at the end means don't show full path
1387                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1388                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1389                                                         if( '' === $text ) {
1390                                                                 $text = $nodotdot;
1391                                                         }
1392                                                 }
1393                                                 $nodotdot = trim( $nodotdot );
1394                                                 if( $nodotdot != '' ) {
1395                                                         $ret .= '/' . $nodotdot;
1396                                                 }
1397                                         }
1398                                 }
1399                         }
1400                 }
1401
1402                 wfProfileOut( $fname );
1403                 return $ret;
1404         }
1405
1406         /**#@+
1407          * Used by doBlockLevels()
1408          * @access private
1409          */
1410         /* private */ function closeParagraph() {
1411                 $result = '';
1412                 if ( '' != $this->mLastSection ) {
1413                         $result = '</' . $this->mLastSection  . ">\n";
1414                 }
1415                 $this->mInPre = false;
1416                 $this->mLastSection = '';
1417                 return $result;
1418         }
1419         # getCommon() returns the length of the longest common substring
1420         # of both arguments, starting at the beginning of both.
1421         #
1422         /* private */ function getCommon( $st1, $st2 ) {
1423                 $fl = strlen( $st1 );
1424                 $shorter = strlen( $st2 );
1425                 if ( $fl < $shorter ) { $shorter = $fl; }
1426
1427                 for ( $i = 0; $i < $shorter; ++$i ) {
1428                         if ( $st1{$i} != $st2{$i} ) { break; }
1429                 }
1430                 return $i;
1431         }
1432         # These next three functions open, continue, and close the list
1433         # element appropriate to the prefix character passed into them.
1434         #
1435         /* private */ function openList( $char ) {
1436                 $result = $this->closeParagraph();
1437
1438                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1439                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1440                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1441                 else if ( ';' == $char ) {
1442                         $result .= '<dl><dt>';
1443                         $this->mDTopen = true;
1444                 }
1445                 else { $result = '<!-- ERR 1 -->'; }
1446
1447                 return $result;
1448         }
1449
1450         /* private */ function nextItem( $char ) {
1451                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1452                 else if ( ':' == $char || ';' == $char ) {
1453                         $close = '</dd>';
1454                         if ( $this->mDTopen ) { $close = '</dt>'; }
1455                         if ( ';' == $char ) {
1456                                 $this->mDTopen = true;
1457                                 return $close . '<dt>';
1458                         } else {
1459                                 $this->mDTopen = false;
1460                                 return $close . '<dd>';
1461                         }
1462                 }
1463                 return '<!-- ERR 2 -->';
1464         }
1465
1466         /* private */ function closeList( $char ) {
1467                 if ( '*' == $char ) { $text = '</li></ul>'; }
1468                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1469                 else if ( ':' == $char ) {
1470                         if ( $this->mDTopen ) {
1471                                 $this->mDTopen = false;
1472                                 $text = '</dt></dl>';
1473                         } else {
1474                                 $text = '</dd></dl>';
1475                         }
1476                 }
1477                 else {  return '<!-- ERR 3 -->'; }
1478                 return $text."\n";
1479         }
1480         /**#@-*/
1481
1482         /**
1483          * Make lists from lines starting with ':', '*', '#', etc.
1484          *
1485          * @access private
1486          * @return string the lists rendered as HTML
1487          */
1488         function doBlockLevels( $text, $linestart ) {
1489                 $fname = 'Parser::doBlockLevels';
1490                 wfProfileIn( $fname );
1491
1492                 # Parsing through the text line by line.  The main thing
1493                 # happening here is handling of block-level elements p, pre,
1494                 # and making lists from lines starting with * # : etc.
1495                 #
1496                 $textLines = explode( "\n", $text );
1497
1498                 $lastPrefix = $output = $lastLine = '';
1499                 $this->mDTopen = $inBlockElem = false;
1500                 $prefixLength = 0;
1501                 $paragraphStack = false;
1502
1503                 if ( !$linestart ) {
1504                         $output .= array_shift( $textLines );
1505                 }
1506                 foreach ( $textLines as $oLine ) {
1507                         $lastPrefixLength = strlen( $lastPrefix );
1508                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1509                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1510                         if ( !$this->mInPre ) {
1511                                 # Multiple prefixes may abut each other for nested lists.
1512                                 $prefixLength = strspn( $oLine, '*#:;' );
1513                                 $pref = substr( $oLine, 0, $prefixLength );
1514
1515                                 # eh?
1516                                 $pref2 = str_replace( ';', ':', $pref );
1517                                 $t = substr( $oLine, $prefixLength );
1518                                 $this->mInPre = !empty($preOpenMatch);
1519                         } else {
1520                                 # Don't interpret any other prefixes in preformatted text
1521                                 $prefixLength = 0;
1522                                 $pref = $pref2 = '';
1523                                 $t = $oLine;
1524                         }
1525
1526                         # List generation
1527                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1528                                 # Same as the last item, so no need to deal with nesting or opening stuff
1529                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1530                                 $paragraphStack = false;
1531
1532                                 if ( substr( $pref, -1 ) == ';') {
1533                                         # The one nasty exception: definition lists work like this:
1534                                         # ; title : definition text
1535                                         # So we check for : in the remainder text to split up the
1536                                         # title and definition, without b0rking links.
1537                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1538                                                 $t = $t2;
1539                                                 $output .= $term . $this->nextItem( ':' );
1540                                         }
1541                                 }
1542                         } elseif( $prefixLength || $lastPrefixLength ) {
1543                                 # Either open or close a level...
1544                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1545                                 $paragraphStack = false;
1546
1547                                 while( $commonPrefixLength < $lastPrefixLength ) {
1548                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1549                                         --$lastPrefixLength;
1550                                 }
1551                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1552                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1553                                 }
1554                                 while ( $prefixLength > $commonPrefixLength ) {
1555                                         $char = substr( $pref, $commonPrefixLength, 1 );
1556                                         $output .= $this->openList( $char );
1557
1558                                         if ( ';' == $char ) {
1559                                                 # FIXME: This is dupe of code above
1560                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1561                                                         $t = $t2;
1562                                                         $output .= $term . $this->nextItem( ':' );
1563                                                 }
1564                                         }
1565                                         ++$commonPrefixLength;
1566                                 }
1567                                 $lastPrefix = $pref2;
1568                         }
1569                         if( 0 == $prefixLength ) {
1570                                 wfProfileIn( "$fname-paragraph" );
1571                                 # No prefix (not in list)--go to paragraph mode
1572                                 $uniq_prefix = UNIQ_PREFIX;
1573                                 // XXX: use a stack for nestable elements like span, table and div
1574                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1575                                 $closematch = preg_match(
1576                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1577                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1578                                 if ( $openmatch or $closematch ) {
1579                                         $paragraphStack = false;
1580                                         $output .= $this->closeParagraph();
1581                                         if($preOpenMatch and !$preCloseMatch) {
1582                                                 $this->mInPre = true;
1583                                         }
1584                                         if ( $closematch ) {
1585                                                 $inBlockElem = false;
1586                                         } else {
1587                                                 $inBlockElem = true;
1588                                         }
1589                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1590                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1591                                                 // pre
1592                                                 if ($this->mLastSection != 'pre') {
1593                                                         $paragraphStack = false;
1594                                                         $output .= $this->closeParagraph().'<pre>';
1595                                                         $this->mLastSection = 'pre';
1596                                                 }
1597                                                 $t = substr( $t, 1 );
1598                                         } else {
1599                                                 // paragraph
1600                                                 if ( '' == trim($t) ) {
1601                                                         if ( $paragraphStack ) {
1602                                                                 $output .= $paragraphStack.'<br />';
1603                                                                 $paragraphStack = false;
1604                                                                 $this->mLastSection = 'p';
1605                                                         } else {
1606                                                                 if ($this->mLastSection != 'p' ) {
1607                                                                         $output .= $this->closeParagraph();
1608                                                                         $this->mLastSection = '';
1609                                                                         $paragraphStack = '<p>';
1610                                                                 } else {
1611                                                                         $paragraphStack = '</p><p>';
1612                                                                 }
1613                                                         }
1614                                                 } else {
1615                                                         if ( $paragraphStack ) {
1616                                                                 $output .= $paragraphStack;
1617                                                                 $paragraphStack = false;
1618                                                                 $this->mLastSection = 'p';
1619                                                         } else if ($this->mLastSection != 'p') {
1620                                                                 $output .= $this->closeParagraph().'<p>';
1621                                                                 $this->mLastSection = 'p';
1622                                                         }
1623                                                 }
1624                                         }
1625                                 }
1626                                 wfProfileOut( "$fname-paragraph" );
1627                         }
1628                         if ($paragraphStack === false) {
1629                                 $output .= $t."\n";
1630                         }
1631                 }
1632                 while ( $prefixLength ) {
1633                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1634                         --$prefixLength;
1635                 }
1636                 if ( '' != $this->mLastSection ) {
1637                         $output .= '</' . $this->mLastSection . '>';
1638                         $this->mLastSection = '';
1639                 }
1640
1641                 wfProfileOut( $fname );
1642                 return $output;
1643         }
1644
1645         /**
1646          * Split up a string on ':', ignoring any occurences inside
1647          * <a>..</a> or <span>...</span>
1648          * @param string $str the string to split
1649          * @param string &$before set to everything before the ':'
1650          * @param string &$after set to everything after the ':'
1651          * return string the position of the ':', or false if none found
1652          */
1653         function findColonNoLinks($str, &$before, &$after) {
1654                 # I wonder if we should make this count all tags, not just <a>
1655                 # and <span>. That would prevent us from matching a ':' that
1656                 # comes in the middle of italics other such formatting....
1657                 # -- Wil
1658                 $fname = 'Parser::findColonNoLinks';
1659                 wfProfileIn( $fname );
1660                 $pos = 0;
1661                 do {
1662                         $colon = strpos($str, ':', $pos);
1663
1664                         if ($colon !== false) {
1665                                 $before = substr($str, 0, $colon);
1666                                 $after = substr($str, $colon + 1);
1667
1668                                 # Skip any ':' within <a> or <span> pairs
1669                                 $a = substr_count($before, '<a');
1670                                 $s = substr_count($before, '<span');
1671                                 $ca = substr_count($before, '</a>');
1672                                 $cs = substr_count($before, '</span>');
1673
1674                                 if ($a <= $ca and $s <= $cs) {
1675                                         # Tags are balanced before ':'; ok
1676                                         break;
1677                                 }
1678                                 $pos = $colon + 1;
1679                         }
1680                 } while ($colon !== false);
1681                 wfProfileOut( $fname );
1682                 return $colon;
1683         }
1684
1685         /**
1686          * Return value of a magic variable (like PAGENAME)
1687          *
1688          * @access private
1689          */
1690         function getVariableValue( $index ) {
1691                 global $wgContLang, $wgSitename, $wgServer;
1692
1693                 /**
1694                  * Some of these require message or data lookups and can be
1695                  * expensive to check many times.
1696                  */
1697                 static $varCache = array();
1698                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1699
1700                 switch ( $index ) {
1701                         case MAG_CURRENTMONTH:
1702                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1703                         case MAG_CURRENTMONTHNAME:
1704                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1705                         case MAG_CURRENTMONTHNAMEGEN:
1706                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1707                         case MAG_CURRENTDAY:
1708                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1709                         case MAG_PAGENAME:
1710                                 return $this->mTitle->getText();
1711                         case MAG_PAGENAMEE:
1712                                 return $this->mTitle->getPartialURL();
1713                         case MAG_NAMESPACE:
1714                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1715                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1716                         case MAG_CURRENTDAYNAME:
1717                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1718                         case MAG_CURRENTYEAR:
1719                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1720                         case MAG_CURRENTTIME:
1721                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1722                         case MAG_CURRENTWEEK:
1723                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1724                         case MAG_CURRENTDOW:
1725                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1726                         case MAG_NUMBEROFARTICLES:
1727                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1728                         case MAG_SITENAME:
1729                                 return $wgSitename;
1730                         case MAG_SERVER:
1731                                 return $wgServer;
1732                         default:
1733                                 return NULL;
1734                 }
1735         }
1736
1737         /**
1738          * initialise the magic variables (like CURRENTMONTHNAME)
1739          *
1740          * @access private
1741          */
1742         function initialiseVariables() {
1743                 $fname = 'Parser::initialiseVariables';
1744                 wfProfileIn( $fname );
1745                 global $wgVariableIDs;
1746                 $this->mVariables = array();
1747                 foreach ( $wgVariableIDs as $id ) {
1748                         $mw =& MagicWord::get( $id );
1749                         $mw->addToArray( $this->mVariables, $id );
1750                 }
1751                 wfProfileOut( $fname );
1752         }
1753
1754         /**
1755          * Replace magic variables, templates, and template arguments
1756          * with the appropriate text. Templates are substituted recursively,
1757          * taking care to avoid infinite loops.
1758          *
1759          * Note that the substitution depends on value of $mOutputType:
1760          *  OT_WIKI: only {{subst:}} templates
1761          *  OT_MSG: only magic variables
1762          *  OT_HTML: all templates and magic variables
1763          *
1764          * @param string $tex The text to transform
1765          * @param array $args Key-value pairs representing template parameters to substitute
1766          * @access private
1767          */
1768         function replaceVariables( $text, $args = array() ) {
1769                 global $wgLang, $wgScript, $wgArticlePath;
1770
1771                 # Prevent too big inclusions
1772                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1773                         return $text;
1774                 }
1775
1776                 $fname = 'Parser::replaceVariables';
1777                 wfProfileIn( $fname );
1778
1779                 $titleChars = Title::legalChars();
1780
1781                 # This function is called recursively. To keep track of arguments we need a stack:
1782                 array_push( $this->mArgStack, $args );
1783
1784                 # Variable substitution
1785                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1786
1787                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1788                         # Argument substitution
1789                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1790                 }
1791                 # Template substitution
1792                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1793                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1794
1795                 array_pop( $this->mArgStack );
1796
1797                 wfProfileOut( $fname );
1798                 return $text;
1799         }
1800
1801         /**
1802          * Replace magic variables
1803          * @access private
1804          */
1805         function variableSubstitution( $matches ) {
1806                 $fname = 'parser::variableSubstitution';
1807                 $varname = $matches[1];
1808                 wfProfileIn( $fname );
1809                 if ( !$this->mVariables ) {
1810                         $this->initialiseVariables();
1811                 }
1812                 $skip = false;
1813                 if ( $this->mOutputType == OT_WIKI ) {
1814                         # Do only magic variables prefixed by SUBST
1815                         $mwSubst =& MagicWord::get( MAG_SUBST );
1816                         if (!$mwSubst->matchStartAndRemove( $varname ))
1817                                 $skip = true;
1818                         # Note that if we don't substitute the variable below,
1819                         # we don't remove the {{subst:}} magic word, in case
1820                         # it is a template rather than a magic variable.
1821                 }
1822                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
1823                         $id = $this->mVariables[$varname];
1824                         $text = $this->getVariableValue( $id );
1825                         $this->mOutput->mContainsOldMagic = true;
1826                 } else {
1827                         $text = $matches[0];
1828                 }
1829                 wfProfileOut( $fname );
1830                 return $text;
1831         }
1832
1833         # Split template arguments
1834         function getTemplateArgs( $argsString ) {
1835                 if ( $argsString === '' ) {
1836                         return array();
1837                 }
1838
1839                 $args = explode( '|', substr( $argsString, 1 ) );
1840
1841                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1842                 # merged with the next arg because the '|' character between belongs
1843                 # to the link syntax and not the template parameter syntax.
1844                 $argc = count($args);
1845                 $i = 0;
1846                 for ( $i = 0; $i < $argc-1; $i++ ) {
1847                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1848                                 $args[$i] .= '|'.$args[$i+1];
1849                                 array_splice($args, $i+1, 1);
1850                                 $i--;
1851                                 $argc--;
1852                         }
1853                 }
1854
1855                 return $args;
1856         }
1857
1858         /**
1859          * Return the text of a template, after recursively
1860          * replacing any variables or templates within the template.
1861          *
1862          * @param array $matches The parts of the template
1863          *  $matches[1]: the title, i.e. the part before the |
1864          *  $matches[2]: the parameters (including a leading |), if  any
1865          * @return string the text of the template
1866          * @access private
1867          */
1868         function braceSubstitution( $matches ) {
1869                 global $wgLinkCache, $wgContLang;
1870                 $fname = 'Parser::braceSubstitution';
1871                 wfProfileIn( $fname );
1872
1873                 $found = false;
1874                 $nowiki = false;
1875                 $noparse = false;
1876
1877                 $title = NULL;
1878
1879                 # Need to know if the template comes at the start of a line,
1880                 # to treat the beginning of the template like the beginning
1881                 # of a line for tables and block-level elements.
1882                 $linestart = $matches[1];
1883
1884                 # $part1 is the bit before the first |, and must contain only title characters
1885                 # $args is a list of arguments, starting from index 0, not including $part1
1886
1887                 $part1 = $matches[2];
1888                 # If the third subpattern matched anything, it will start with |
1889
1890                 $args = $this->getTemplateArgs($matches[3]);
1891                 $argc = count( $args );
1892
1893                 # Don't parse {{{}}} because that's only for template arguments
1894                 if ( $linestart === '{' ) {
1895                         $text = $matches[0];
1896                         $found = true;
1897                         $noparse = true;
1898                 }
1899
1900                 # SUBST
1901                 if ( !$found ) {
1902                         $mwSubst =& MagicWord::get( MAG_SUBST );
1903                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1904                                 # One of two possibilities is true:
1905                                 # 1) Found SUBST but not in the PST phase
1906                                 # 2) Didn't find SUBST and in the PST phase
1907                                 # In either case, return without further processing
1908                                 $text = $matches[0];
1909                                 $found = true;
1910                                 $noparse = true;
1911                         }
1912                 }
1913
1914                 # MSG, MSGNW and INT
1915                 if ( !$found ) {
1916                         # Check for MSGNW:
1917                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1918                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1919                                 $nowiki = true;
1920                         } else {
1921                                 # Remove obsolete MSG:
1922                                 $mwMsg =& MagicWord::get( MAG_MSG );
1923                                 $mwMsg->matchStartAndRemove( $part1 );
1924                         }
1925
1926                         # Check if it is an internal message
1927                         $mwInt =& MagicWord::get( MAG_INT );
1928                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1929                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1930                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1931                                         $found = true;
1932                                 }
1933                         }
1934                 }
1935
1936                 # NS
1937                 if ( !$found ) {
1938                         # Check for NS: (namespace expansion)
1939                         $mwNs = MagicWord::get( MAG_NS );
1940                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1941                                 if ( intval( $part1 ) ) {
1942                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1943                                         $found = true;
1944                                 } else {
1945                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1946                                         if ( !is_null( $index ) ) {
1947                                                 $text = $linestart . $wgContLang->getNsText( $index );
1948                                                 $found = true;
1949                                         }
1950                                 }
1951                         }
1952                 }
1953
1954                 # LOCALURL and LOCALURLE
1955                 if ( !$found ) {
1956                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1957                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1958
1959                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1960                                 $func = 'getLocalURL';
1961                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1962                                 $func = 'escapeLocalURL';
1963                         } else {
1964                                 $func = '';
1965                         }
1966
1967                         if ( $func !== '' ) {
1968                                 $title = Title::newFromText( $part1 );
1969                                 if ( !is_null( $title ) ) {
1970                                         if ( $argc > 0 ) {
1971                                                 $text = $linestart . $title->$func( $args[0] );
1972                                         } else {
1973                                                 $text = $linestart . $title->$func();
1974                                         }
1975                                         $found = true;
1976                                 }
1977                         }
1978                 }
1979
1980                 # GRAMMAR
1981                 if ( !$found && $argc == 1 ) {
1982                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1983                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1984                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
1985                                 $found = true;
1986                         }
1987                 }
1988
1989                 # Template table test
1990
1991                 # Did we encounter this template already? If yes, it is in the cache
1992                 # and we need to check for loops.
1993                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
1994                         $found = true;
1995
1996                         # Infinite loop test
1997                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1998                                 $noparse = true;
1999                                 $found = true;
2000                                 $text = $linestart .
2001                                         "\{\{$part1}}" .
2002                                         '<!-- WARNING: template loop detected -->';
2003                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2004                         } else {
2005                                 # set $text to cached message.
2006                                 $text = $linestart . $this->mTemplates[$part1];
2007                         }
2008                 }
2009
2010                 # Load from database
2011                 $itcamefromthedatabase = false;
2012                 $lastPathLevel = $this->mTemplatePath;
2013                 if ( !$found ) {
2014                         $ns = NS_TEMPLATE;
2015                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2016                         if ($subpage !== '') {
2017                                 $ns = $this->mTitle->getNamespace();
2018                         }
2019                         $title = Title::newFromText( $part1, $ns );
2020                         if ( !is_null( $title ) && !$title->isExternal() ) {
2021                                 # Check for excessive inclusion
2022                                 $dbk = $title->getPrefixedDBkey();
2023                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2024                                         # This should never be reached.
2025                                         $article = new Article( $title );
2026                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2027                                         if ( $articleContent !== false ) {
2028                                                 $found = true;
2029                                                 $text = $linestart . $articleContent;
2030                                                 $itcamefromthedatabase = true;
2031                                         }
2032                                 }
2033
2034                                 # If the title is valid but undisplayable, make a link to it
2035                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2036                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2037                                         $found = true;
2038                                 }
2039
2040                                 # Template cache array insertion
2041                                 if( $found ) {
2042                                         $this->mTemplates[$part1] = $text;
2043                                 }
2044                         }
2045                 }
2046
2047                 # Recursive parsing, escaping and link table handling
2048                 # Only for HTML output
2049                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2050                         $text = wfEscapeWikiText( $text );
2051                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2052                         # Clean up argument array
2053                         $assocArgs = array();
2054                         $index = 1;
2055                         foreach( $args as $arg ) {
2056                                 $eqpos = strpos( $arg, '=' );
2057                                 if ( $eqpos === false ) {
2058                                         $assocArgs[$index++] = $arg;
2059                                 } else {
2060                                         $name = trim( substr( $arg, 0, $eqpos ) );
2061                                         $value = trim( substr( $arg, $eqpos+1 ) );
2062                                         if ( $value === false ) {
2063                                                 $value = '';
2064                                         }
2065                                         if ( $name !== false ) {
2066                                                 $assocArgs[$name] = $value;
2067                                         }
2068                                 }
2069                         }
2070
2071                         # Add a new element to the templace recursion path
2072                         $this->mTemplatePath[$part1] = 1;
2073
2074                         $text = $this->strip( $text, $this->mStripState );
2075                         $text = Sanitizer::removeHTMLtags( $text );
2076                         $text = $this->replaceVariables( $text, $assocArgs );
2077
2078                         # Resume the link cache and register the inclusion as a link
2079                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2080                                 $wgLinkCache->addLinkObj( $title );
2081                         }
2082
2083                         # If the template begins with a table or block-level
2084                         # element, it should be treated as beginning a new line.
2085                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2086                                 $text = "\n" . $text;
2087                         }
2088                 }
2089                 # Prune lower levels off the recursion check path
2090                 $this->mTemplatePath = $lastPathLevel;
2091
2092                 if ( !$found ) {
2093                         wfProfileOut( $fname );
2094                         return $matches[0];
2095                 } else {
2096                         # replace ==section headers==
2097                         # XXX this needs to go away once we have a better parser.
2098                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2099                                 if( !is_null( $title ) )
2100                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2101                                 else
2102                                         $encodedname = base64_encode("");
2103                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2104                                         PREG_SPLIT_DELIM_CAPTURE);
2105                                 $text = '';
2106                                 $nsec = 0;
2107                                 for( $i = 0; $i < count($m); $i += 2 ) {
2108                                         $text .= $m[$i];
2109                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2110                                         $hl = $m[$i + 1];
2111                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2112                                                 $text .= $hl;
2113                                                 continue;
2114                                         }
2115                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2116                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2117                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2118
2119                                         $nsec++;
2120                                 }
2121                         }
2122                 }
2123                 # Prune lower levels off the recursion check path
2124                 $this->mTemplatePath = $lastPathLevel;
2125
2126                 if ( !$found ) {
2127                         wfProfileOut( $fname );
2128                         return $matches[0];
2129                 } else {
2130                         wfProfileOut( $fname );
2131                         return $text;
2132                 }
2133         }
2134
2135         /**
2136          * Triple brace replacement -- used for template arguments
2137          * @access private
2138          */
2139         function argSubstitution( $matches ) {
2140                 $arg = trim( $matches[1] );
2141                 $text = $matches[0];
2142                 $inputArgs = end( $this->mArgStack );
2143
2144                 if ( array_key_exists( $arg, $inputArgs ) ) {
2145                         $text = $inputArgs[$arg];
2146                 }
2147
2148                 return $text;
2149         }
2150
2151         /**
2152          * Returns true if the function is allowed to include this entity
2153          * @access private
2154          */
2155         function incrementIncludeCount( $dbk ) {
2156                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2157                         $this->mIncludeCount[$dbk] = 0;
2158                 }
2159                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2160                         return true;
2161                 } else {
2162                         return false;
2163                 }
2164         }
2165
2166         /**
2167          * This function accomplishes several tasks:
2168          * 1) Auto-number headings if that option is enabled
2169          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2170          * 3) Add a Table of contents on the top for users who have enabled the option
2171          * 4) Auto-anchor headings
2172          *
2173          * It loops through all headlines, collects the necessary data, then splits up the
2174          * string and re-inserts the newly formatted headlines.
2175          *
2176          * @param string $text
2177          * @param boolean $isMain
2178          * @access private
2179          */
2180         function formatHeadings( $text, $isMain=true ) {
2181                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2182
2183                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2184                 $doShowToc = $this->mOptions->getShowToc();
2185                 $forceTocHere = false;
2186                 if( !$this->mTitle->userCanEdit() ) {
2187                         $showEditLink = 0;
2188                         $rightClickHack = 0;
2189                 } else {
2190                         $showEditLink = $this->mOptions->getEditSection();
2191                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2192                 }
2193
2194                 # Inhibit editsection links if requested in the page
2195                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2196                 if( $esw->matchAndRemove( $text ) ) {
2197                         $showEditLink = 0;
2198                 }
2199                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2200                 # do not add TOC
2201                 $mw =& MagicWord::get( MAG_NOTOC );
2202                 if( $mw->matchAndRemove( $text ) ) {
2203                         $doShowToc = 0;
2204                 }
2205
2206                 # never add the TOC to the Main Page. This is an entry page that should not
2207                 # be more than 1-2 screens large anyway
2208                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2209                         $doShowToc = 0;
2210                 }
2211
2212                 # Get all headlines for numbering them and adding funky stuff like [edit]
2213                 # links - this is for later, but we need the number of headlines right now
2214                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2215
2216                 # if there are fewer than 4 headlines in the article, do not show TOC
2217                 if( $numMatches < 4 ) {
2218                         $doShowToc = 0;
2219                 }
2220
2221                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2222                 # override above conditions and always show TOC at that place
2223
2224                 $mw =& MagicWord::get( MAG_TOC );
2225                 if($mw->match( $text ) ) {
2226                         $doShowToc = 1;
2227                         $forceTocHere = true;
2228                 } else {
2229                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2230                         # override above conditions and always show TOC above first header
2231                         $mw =& MagicWord::get( MAG_FORCETOC );
2232                         if ($mw->matchAndRemove( $text ) ) {
2233                                 $doShowToc = 1;
2234                         }
2235                 }
2236
2237                 # Never ever show TOC if no headers
2238                 if( $numMatches < 1 ) {
2239                         $doShowToc = 0;
2240                 }
2241
2242                 # We need this to perform operations on the HTML
2243                 $sk =& $this->mOptions->getSkin();
2244
2245                 # headline counter
2246                 $headlineCount = 0;
2247                 $sectionCount = 0; # headlineCount excluding template sections
2248
2249                 # Ugh .. the TOC should have neat indentation levels which can be
2250                 # passed to the skin functions. These are determined here
2251                 $toc = '';
2252                 $full = '';
2253                 $head = array();
2254                 $sublevelCount = array();
2255                 $levelCount = array();
2256                 $toclevel = 0;
2257                 $level = 0;
2258                 $prevlevel = 0;
2259                 $toclevel = 0;
2260                 $prevtoclevel = 0;
2261
2262                 foreach( $matches[3] as $headline ) {
2263                         $istemplate = 0;
2264                         $templatetitle = '';
2265                         $templatesection = 0;
2266                         $numbering = '';
2267
2268                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2269                                 $istemplate = 1;
2270                                 $templatetitle = base64_decode($mat[1]);
2271                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2272                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2273                         }
2274
2275                         if( $toclevel ) {
2276                                 $prevlevel = $level;
2277                                 $prevtoclevel = $toclevel;
2278                         }
2279                         $level = $matches[1][$headlineCount];
2280
2281                         if( $doNumberHeadings || $doShowToc ) {
2282
2283                                 if ( $level > $prevlevel ) {
2284                                         # Increase TOC level
2285                                         $toclevel++;
2286                                         $sublevelCount[$toclevel] = 0;
2287                                         $toc .= $sk->tocIndent();
2288                                 }
2289                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2290                                         # Decrease TOC level, find level to jump to
2291
2292                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2293                                                 # Can only go down to level 1
2294                                                 $toclevel = 1;
2295                                         } else {
2296                                                 for ($i = $toclevel; $i > 0; $i--) {
2297                                                         if ( $levelCount[$i] == $level ) {
2298                                                                 # Found last matching level
2299                                                                 $toclevel = $i;
2300                                                                 break;
2301                                                         }
2302                                                         elseif ( $levelCount[$i] < $level ) {
2303                                                                 # Found first matching level below current level
2304                                                                 $toclevel = $i + 1;
2305                                                                 break;
2306                                                         }
2307                                                 }
2308                                         }
2309
2310                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2311                                 }
2312                                 else {
2313                                         # No change in level, end TOC line
2314                                         $toc .= $sk->tocLineEnd();
2315                                 }
2316
2317                                 $levelCount[$toclevel] = $level;
2318
2319                                 # count number of headlines for each level
2320                                 @$sublevelCount[$toclevel]++;
2321                                 $dot = 0;
2322                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2323                                         if( !empty( $sublevelCount[$i] ) ) {
2324                                                 if( $dot ) {
2325                                                         $numbering .= '.';
2326                                                 }
2327                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2328                                                 $dot = 1;
2329                                         }
2330                                 }
2331                         }
2332
2333                         # The canonized header is a version of the header text safe to use for links
2334                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2335                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2336                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2337
2338                         # Remove link placeholders by the link text.
2339                         #     <!--LINK number-->
2340                         # turns into
2341                         #     link text with suffix
2342                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2343                                                             "\$wgLinkHolders['texts'][\$1]",
2344                                                             $canonized_headline );
2345                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2346                                                             "\$wgInterwikiLinkHolders[\$1]",
2347                                                             $canonized_headline );
2348
2349                         # strip out HTML
2350                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2351                         $tocline = trim( $canonized_headline );
2352                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2353                         $replacearray = array(
2354                                 '%3A' => ':',
2355                                 '%' => '.'
2356                         );
2357                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2358                         $refer[$headlineCount] = $canonized_headline;
2359
2360                         # count how many in assoc. array so we can track dupes in anchors
2361                         @$refers[$canonized_headline]++;
2362                         $refcount[$headlineCount]=$refers[$canonized_headline];
2363
2364                         # Don't number the heading if it is the only one (looks silly)
2365                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2366                                 # the two are different if the line contains a link
2367                                 $headline=$numbering . ' ' . $headline;
2368                         }
2369
2370                         # Create the anchor for linking from the TOC to the section
2371                         $anchor = $canonized_headline;
2372                         if($refcount[$headlineCount] > 1 ) {
2373                                 $anchor .= '_' . $refcount[$headlineCount];
2374                         }
2375                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2376                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2377                         }
2378                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2379                                 if ( empty( $head[$headlineCount] ) ) {
2380                                         $head[$headlineCount] = '';
2381                                 }
2382                                 if( $istemplate )
2383                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2384                                 else
2385                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2386                         }
2387
2388                         # Add the edit section span
2389                         if( $rightClickHack ) {
2390                                 if( $istemplate )
2391                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2392                                 else
2393                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2394                         }
2395
2396                         # give headline the correct <h#> tag
2397                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2398
2399                         $headlineCount++;
2400                         if( !$istemplate )
2401                                 $sectionCount++;
2402                 }
2403
2404                 if( $doShowToc ) {
2405                         $toclines = $headlineCount;
2406                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2407                         $toc = $sk->tocList( $toc );
2408                 }
2409
2410                 # split up and insert constructed headlines
2411
2412                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2413                 $i = 0;
2414
2415                 foreach( $blocks as $block ) {
2416                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2417                                 # This is the [edit] link that appears for the top block of text when
2418                                 # section editing is enabled
2419
2420                                 # Disabled because it broke block formatting
2421                                 # For example, a bullet point in the top line
2422                                 # $full .= $sk->editSectionLink(0);
2423                         }
2424                         $full .= $block;
2425                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2426                         # Top anchor now in skin
2427                                 $full = $full.$toc;
2428                         }
2429
2430                         if( !empty( $head[$i] ) ) {
2431                                 $full .= $head[$i];
2432                         }
2433                         $i++;
2434                 }
2435                 if($forceTocHere) {
2436                         $mw =& MagicWord::get( MAG_TOC );
2437                         return $mw->replace( $toc, $full );
2438                 } else {
2439                         return $full;
2440                 }
2441         }
2442
2443         /**
2444          * Return an HTML link for the "ISBN 123456" text
2445          * @access private
2446          */
2447         function magicISBN( $text ) {
2448                 global $wgLang;
2449                 $fname = 'Parser::magicISBN';
2450                 wfProfileIn( $fname );
2451
2452                 $a = split( 'ISBN ', ' '.$text );
2453                 if ( count ( $a ) < 2 ) {
2454                         wfProfileOut( $fname );
2455                         return $text;
2456                 }
2457                 $text = substr( array_shift( $a ), 1);
2458                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2459
2460                 foreach ( $a as $x ) {
2461                         $isbn = $blank = '' ;
2462                         while ( ' ' == $x{0} ) {
2463                                 $blank .= ' ';
2464                                 $x = substr( $x, 1 );
2465                         }
2466                         if ( $x == '' ) { # blank isbn
2467                                 $text .= "ISBN $blank";
2468                                 continue;
2469                         }
2470                         while ( strstr( $valid, $x{0} ) != false ) {
2471                                 $isbn .= $x{0};
2472                                 $x = substr( $x, 1 );
2473                         }
2474                         $num = str_replace( '-', '', $isbn );
2475                         $num = str_replace( ' ', '', $num );
2476
2477                         if ( '' == $num ) {
2478                                 $text .= "ISBN $blank$x";
2479                         } else {
2480                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2481                                 $text .= '<a href="' .
2482                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2483                                         "\" class=\"internal\">ISBN $isbn</a>";
2484                                 $text .= $x;
2485                         }
2486                 }
2487                 wfProfileOut( $fname );
2488                 return $text;
2489         }
2490
2491         /**
2492          * Return an HTML link for the "GEO ..." text
2493          * @access private
2494          */
2495         function magicGEO( $text ) {
2496                 global $wgLang, $wgUseGeoMode;
2497                 $fname = 'Parser::magicGEO';
2498                 wfProfileIn( $fname );
2499
2500                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2501                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2502                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2503                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2504                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2505                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2506
2507                 $a = split( 'GEO ', ' '.$text );
2508                 if ( count ( $a ) < 2 ) {
2509                         wfProfileOut( $fname );
2510                         return $text;
2511                 }
2512                 $text = substr( array_shift( $a ), 1);
2513                 $valid = '0123456789.+-:';
2514
2515                 foreach ( $a as $x ) {
2516                         $geo = $blank = '' ;
2517                         while ( ' ' == $x{0} ) {
2518                                 $blank .= ' ';
2519                                 $x = substr( $x, 1 );
2520                         }
2521                         while ( strstr( $valid, $x{0} ) != false ) {
2522                                 $geo .= $x{0};
2523                                 $x = substr( $x, 1 );
2524                         }
2525                         $num = str_replace( '+', '', $geo );
2526                         $num = str_replace( ' ', '', $num );
2527
2528                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2529                                 $text .= "GEO $blank$x";
2530                         } else {
2531                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2532                                 $text .= '<a href="' .
2533                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2534                                         "\" class=\"internal\">GEO $geo</a>";
2535                                 $text .= $x;
2536                         }
2537                 }
2538                 wfProfileOut( $fname );
2539                 return $text;
2540         }
2541
2542         /**
2543          * Return an HTML link for the "RFC 1234" text
2544          * @access private
2545          * @param string $text text to be processed
2546          */
2547         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2548                 global $wgLang;
2549
2550                 $valid = '0123456789';
2551                 $internal = false;
2552
2553                 $a = split( $keyword, ' '.$text );
2554                 if ( count ( $a ) < 2 ) {
2555                         return $text;
2556                 }
2557                 $text = substr( array_shift( $a ), 1);
2558
2559                 /* Check if keyword is preceed by [[.
2560                  * This test is made here cause of the array_shift above
2561                  * that prevent the test to be done in the foreach.
2562                  */
2563                 if ( substr( $text, -2 ) == '[[' ) {
2564                         $internal = true;
2565                 }
2566
2567                 foreach ( $a as $x ) {
2568                         /* token might be empty if we have RFC RFC 1234 */
2569                         if ( $x=='' ) {
2570                                 $text.=$keyword;
2571                                 continue;
2572                                 }
2573
2574                         $id = $blank = '' ;
2575
2576                         /** remove and save whitespaces in $blank */
2577                         while ( $x{0} == ' ' ) {
2578                                 $blank .= ' ';
2579                                 $x = substr( $x, 1 );
2580                         }
2581
2582                         /** remove and save the rfc number in $id */
2583                         while ( strstr( $valid, $x{0} ) != false ) {
2584                                 $id .= $x{0};
2585                                 $x = substr( $x, 1 );
2586                         }
2587
2588                         if ( $id == '' ) {
2589                                 /* call back stripped spaces*/
2590                                 $text .= $keyword.$blank.$x;
2591                         } elseif( $internal ) {
2592                                 /* normal link */
2593                                 $text .= $keyword.$id.$x;
2594                         } else {
2595                                 /* build the external link*/
2596                                 $url = wfmsg( $urlmsg );
2597                                 $url = str_replace( '$1', $id, $url);
2598                                 $sk =& $this->mOptions->getSkin();
2599                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2600                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2601                         }
2602
2603                         /* Check if the next RFC keyword is preceed by [[ */
2604                         $internal = ( substr($x,-2) == '[[' );
2605                 }
2606                 return $text;
2607         }
2608
2609         /**
2610          * Transform wiki markup when saving a page by doing \r\n -> \n
2611          * conversion, substitting signatures, {{subst:}} templates, etc.
2612          *
2613          * @param string $text the text to transform
2614          * @param Title &$title the Title object for the current article
2615          * @param User &$user the User object describing the current user
2616          * @param ParserOptions $options parsing options
2617          * @param bool $clearState whether to clear the parser state first
2618          * @return string the altered wiki markup
2619          * @access public
2620          */
2621         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2622                 $this->mOptions = $options;
2623                 $this->mTitle =& $title;
2624                 $this->mOutputType = OT_WIKI;
2625
2626                 if ( $clearState ) {
2627                         $this->clearState();
2628                 }
2629
2630                 $stripState = false;
2631                 $pairs = array(
2632                         "\r\n" => "\n",
2633                         );
2634                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2635                 $text = $this->strip( $text, $stripState, false );
2636                 $text = $this->pstPass2( $text, $user );
2637                 $text = $this->unstrip( $text, $stripState );
2638                 $text = $this->unstripNoWiki( $text, $stripState );
2639                 return $text;
2640         }
2641
2642         /**
2643          * Pre-save transform helper function
2644          * @access private
2645          */
2646         function pstPass2( $text, &$user ) {
2647                 global $wgLang, $wgContLang, $wgLocaltimezone;
2648
2649                 # Variable replacement
2650                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2651                 $text = $this->replaceVariables( $text );
2652
2653                 # Signatures
2654                 #
2655                 $n = $user->getName();
2656                 $k = $user->getOption( 'nickname' );
2657                 if ( '' == $k ) { $k = $n; }
2658                 if ( isset( $wgLocaltimezone ) ) {
2659                         $oldtz = getenv( 'TZ' );
2660                         putenv( 'TZ='.$wgLocaltimezone );
2661                 }
2662                 /* Note: this is an ugly timezone hack for the European wikis */
2663                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2664                   ' (' . date( 'T' ) . ')';
2665                 if ( isset( $wgLocaltimezone ) ) {
2666                         putenv( 'TZ='.$oldtzs );
2667                 }
2668
2669                 if( $user->getOption( 'fancysig' ) ) {
2670                         $sigText = $k;
2671                 } else {
2672                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2673                 }
2674                 $text = preg_replace( '/~~~~~/', $d, $text );
2675                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2676                 $text = preg_replace( '/~~~/', $sigText, $text );
2677
2678                 # Context links: [[|name]] and [[name (context)|]]
2679                 #
2680                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2681                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2682                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2683                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2684
2685                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2686                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2687                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2688                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2689                 $context = '';
2690                 $t = $this->mTitle->getText();
2691                 if ( preg_match( $conpat, $t, $m ) ) {
2692                         $context = $m[2];
2693                 }
2694                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2695                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2696                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2697
2698                 if ( '' == $context ) {
2699                         $text = preg_replace( $p2, '[[\\1]]', $text );
2700                 } else {
2701                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2702                 }
2703
2704                 # Trim trailing whitespace
2705                 # MAG_END (__END__) tag allows for trailing
2706                 # whitespace to be deliberately included
2707                 $text = rtrim( $text );
2708                 $mw =& MagicWord::get( MAG_END );
2709                 $mw->matchAndRemove( $text );
2710
2711                 return $text;
2712         }
2713
2714         /**
2715          * Set up some variables which are usually set up in parse()
2716          * so that an external function can call some class members with confidence
2717          * @access public
2718          */
2719         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2720                 $this->mTitle =& $title;
2721                 $this->mOptions = $options;
2722                 $this->mOutputType = $outputType;
2723                 if ( $clearState ) {
2724                         $this->clearState();
2725                 }
2726         }
2727
2728         /**
2729          * Transform a MediaWiki message by replacing magic variables.
2730          *
2731          * @param string $text the text to transform
2732          * @param ParserOptions $options  options
2733          * @return string the text with variables substituted
2734          * @access public
2735          */
2736         function transformMsg( $text, $options ) {
2737                 global $wgTitle;
2738                 static $executing = false;
2739
2740                 # Guard against infinite recursion
2741                 if ( $executing ) {
2742                         return $text;
2743                 }
2744                 $executing = true;
2745
2746                 $this->mTitle = $wgTitle;
2747                 $this->mOptions = $options;
2748                 $this->mOutputType = OT_MSG;
2749                 $this->clearState();
2750                 $text = $this->replaceVariables( $text );
2751
2752                 $executing = false;
2753                 return $text;
2754         }
2755
2756         /**
2757          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2758          * Callback will be called with the text within
2759          * Transform and return the text within
2760          * @access public
2761          */
2762         function setHook( $tag, $callback ) {
2763                 $oldVal = @$this->mTagHooks[$tag];
2764                 $this->mTagHooks[$tag] = $callback;
2765                 return $oldVal;
2766         }
2767
2768         /**
2769          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2770          * Placeholders created in Skin::makeLinkObj()
2771          * Returns an array of links found, indexed by PDBK:
2772          *  0 - broken
2773          *  1 - normal link
2774          *  2 - stub
2775          * $options is a bit field, RLH_FOR_UPDATE to select for update
2776          */
2777         function replaceLinkHolders( &$text, $options = 0 ) {
2778                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2779                 global $wgInterwikiLinkHolders;
2780                 global $outputReplace;
2781
2782                 if ( $wgUseOldExistenceCheck ) {
2783                         return array();
2784                 }
2785
2786                 $fname = 'Parser::replaceLinkHolders';
2787                 wfProfileIn( $fname );
2788
2789                 $pdbks = array();
2790                 $colours = array();
2791
2792                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2793                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2794                         wfProfileIn( $fname.'-check' );
2795                         $dbr =& wfGetDB( DB_SLAVE );
2796                         $page = $dbr->tableName( 'page' );
2797                         $sk = $wgUser->getSkin();
2798                         $threshold = $wgUser->getOption('stubthreshold');
2799
2800                         # Sort by namespace
2801                         asort( $wgLinkHolders['namespaces'] );
2802
2803                         # Generate query
2804                         $query = false;
2805                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2806                                 # Make title object
2807                                 $title = $wgLinkHolders['titles'][$key];
2808
2809                                 # Skip invalid entries.
2810                                 # Result will be ugly, but prevents crash.
2811                                 if ( is_null( $title ) ) {
2812                                         continue;
2813                                 }
2814                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2815
2816                                 # Check if it's in the link cache already
2817                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2818                                         $colours[$pdbk] = 1;
2819                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2820                                         $colours[$pdbk] = 0;
2821                                 } else {
2822                                         # Not in the link cache, add it to the query
2823                                         if ( !isset( $current ) ) {
2824                                                 $current = $val;
2825                                                 $tables = $page;
2826                                                 $join = '';
2827                                                 $query =  "SELECT page_id, page_namespace, page_title";
2828                                                 if ( $threshold > 0 ) {
2829                                                         $textTable = $dbr->tableName( 'text' );
2830                                                         $query .= ', LENGTH(old_text) AS page_len, page_is_redirect';
2831                                                         $tables .= ", $textTable";
2832                                                         $join = 'page_latest=old_id AND';
2833                                                 }
2834                                                 $query .= " FROM $tables WHERE $join (page_namespace=$val AND page_title IN(";
2835                                         } elseif ( $current != $val ) {
2836                                                 $current = $val;
2837                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
2838                                         } else {
2839                                                 $query .= ', ';
2840                                         }
2841
2842                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2843                                 }
2844                         }
2845                         if ( $query ) {
2846                                 $query .= '))';
2847                                 if ( $options & RLH_FOR_UPDATE ) {
2848                                         $query .= ' FOR UPDATE';
2849                                 }
2850
2851                                 $res = $dbr->query( $query, $fname );
2852
2853                                 # Fetch data and form into an associative array
2854                                 # non-existent = broken
2855                                 # 1 = known
2856                                 # 2 = stub
2857                                 while ( $s = $dbr->fetchObject($res) ) {
2858                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
2859                                         $pdbk = $title->getPrefixedDBkey();
2860                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
2861
2862                                         if ( $threshold >  0 ) {
2863                                                 $size = $s->page_len;
2864                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $length < $threshold ) {
2865                                                         $colours[$pdbk] = 1;
2866                                                 } else {
2867                                                         $colours[$pdbk] = 2;
2868                                                 }
2869                                         } else {
2870                                                 $colours[$pdbk] = 1;
2871                                         }
2872                                 }
2873                         }
2874                         wfProfileOut( $fname.'-check' );
2875
2876                         # Construct search and replace arrays
2877                         wfProfileIn( $fname.'-construct' );
2878                         $outputReplace = array();
2879                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
2880                                 $pdbk = $pdbks[$key];
2881                                 $searchkey = '<!--LINK '.$key.'-->';
2882                                 $title = $wgLinkHolders['titles'][$key];
2883                                 if ( empty( $colours[$pdbk] ) ) {
2884                                         $wgLinkCache->addBadLink( $pdbk );
2885                                         $colours[$pdbk] = 0;
2886                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
2887                                                                         $wgLinkHolders['texts'][$key],
2888                                                                         $wgLinkHolders['queries'][$key] );
2889                                 } elseif ( $colours[$pdbk] == 1 ) {
2890                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
2891                                                                         $wgLinkHolders['texts'][$key],
2892                                                                         $wgLinkHolders['queries'][$key] );
2893                                 } elseif ( $colours[$pdbk] == 2 ) {
2894                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
2895                                                                         $wgLinkHolders['texts'][$key],
2896                                                                         $wgLinkHolders['queries'][$key] );
2897                                 }
2898                         }
2899                         wfProfileOut( $fname.'-construct' );
2900
2901                         # Do the thing
2902                         wfProfileIn( $fname.'-replace' );
2903
2904                         $text = preg_replace_callback(
2905                                 '/(<!--LINK .*?-->)/',
2906                                 "outputReplaceMatches",
2907                                 $text);
2908                         wfProfileOut( $fname.'-replace' );
2909                 }
2910
2911                 if ( !empty( $wgInterwikiLinkHolders ) ) {
2912                         wfProfileIn( $fname.'-interwiki' );
2913                         $outputReplace = $wgInterwikiLinkHolders;
2914                         $text = preg_replace_callback(
2915                                 '/<!--IWLINK (.*?)-->/',
2916                                 "outputReplaceMatches",
2917                                 $text );
2918                         wfProfileOut( $fname.'-interwiki' );
2919                 }
2920
2921                 wfProfileOut( $fname );
2922                 return $colours;
2923         }
2924
2925         /**
2926          * Renders an image gallery from a text with one line per image.
2927          * text labels may be given by using |-style alternative text. E.g.
2928          *   Image:one.jpg|The number "1"
2929          *   Image:tree.jpg|A tree
2930          * given as text will return the HTML of a gallery with two images,
2931          * labeled 'The number "1"' and
2932          * 'A tree'.
2933          */
2934         function renderImageGallery( $text ) {
2935                 global $wgLinkCache;
2936                 $ig = new ImageGallery();
2937                 $ig->setShowBytes( false );
2938                 $ig->setShowFilename( false );
2939                 $lines = explode( "\n", $text );
2940
2941                 foreach ( $lines as $line ) {
2942                         # match lines like these:
2943                         # Image:someimage.jpg|This is some image
2944                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
2945                         # Skip empty lines
2946                         if ( count( $matches ) == 0 ) {
2947                                 continue;
2948                         }
2949                         $nt = Title::newFromURL( $matches[1] );
2950                         if( is_null( $nt ) ) {
2951                                 # Bogus title. Ignore these so we don't bomb out later.
2952                                 continue;
2953                         }
2954                         if ( isset( $matches[3] ) ) {
2955                                 $label = $matches[3];
2956                         } else {
2957                                 $label = '';
2958                         }
2959
2960                         # FIXME: Use the full wiki parser and add its links
2961                         # to the page's links.
2962                         $html = $this->mOptions->mSkin->formatComment( $label );
2963
2964                         $ig->add( Image::newFromTitle( $nt ), $html );
2965                         $wgLinkCache->addImageLinkObj( $nt );
2966                 }
2967                 return $ig->toHTML();
2968         }
2969 }
2970
2971 /**
2972  * @todo document
2973  * @package MediaWiki
2974  */
2975 class ParserOutput
2976 {
2977         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2978         var $mCacheTime; # Used in ParserCache
2979         var $mVersion;   # Compatibility check
2980
2981         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2982                 $containsOldMagic = false )
2983         {
2984                 $this->mText = $text;
2985                 $this->mLanguageLinks = $languageLinks;
2986                 $this->mCategoryLinks = $categoryLinks;
2987                 $this->mContainsOldMagic = $containsOldMagic;
2988                 $this->mCacheTime = '';
2989                 $this->mVersion = MW_PARSER_VERSION;
2990         }
2991
2992         function getText()                   { return $this->mText; }
2993         function getLanguageLinks()          { return $this->mLanguageLinks; }
2994         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
2995         function getCacheTime()              { return $this->mCacheTime; }
2996         function containsOldMagic()          { return $this->mContainsOldMagic; }
2997         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
2998         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
2999         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3000         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3001         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3002         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3003
3004         function merge( $other ) {
3005                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3006                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3007                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3008         }
3009
3010         /**
3011          * Return true if this cached output object predates the global or
3012          * per-article cache invalidation timestamps, or if it comes from
3013          * an incompatible older version.
3014          *
3015          * @param string $touched the affected article's last touched timestamp
3016          * @return bool
3017          * @access public
3018          */
3019         function expired( $touched ) {
3020                 global $wgCacheEpoch;
3021                 return $this->getCacheTime() <= $touched ||
3022                        $this->getCacheTime() <= $wgCacheEpoch ||
3023                        !isset( $this->mVersion ) ||
3024                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3025         }
3026 }
3027
3028 /**
3029  * Set options of the Parser
3030  * @todo document
3031  * @package MediaWiki
3032  */
3033 class ParserOptions
3034 {
3035         # All variables are private
3036         var $mUseTeX;                    # Use texvc to expand <math> tags
3037         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3038         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3039         var $mAllowExternalImages;       # Allow external images inline
3040         var $mSkin;                      # Reference to the preferred skin
3041         var $mDateFormat;                # Date format index
3042         var $mEditSection;               # Create "edit section" links
3043         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3044         var $mNumberHeadings;            # Automatically number headings
3045         var $mShowToc;                   # Show table of contents
3046
3047         function getUseTeX()                        { return $this->mUseTeX; }
3048         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3049         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3050         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3051         function getSkin()                          { return $this->mSkin; }
3052         function getDateFormat()                    { return $this->mDateFormat; }
3053         function getEditSection()                   { return $this->mEditSection; }
3054         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3055         function getNumberHeadings()                { return $this->mNumberHeadings; }
3056         function getShowToc()                       { return $this->mShowToc; }
3057
3058         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3059         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3060         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3061         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3062         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3063         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3064         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3065         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3066         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3067
3068         function setSkin( &$x ) { $this->mSkin =& $x; }
3069
3070         /**
3071          * Get parser options
3072          * @static
3073          */
3074         function newFromUser( &$user ) {
3075                 $popts = new ParserOptions;
3076                 $popts->initialiseFromUser( $user );
3077                 return $popts;
3078         }
3079
3080         /** Get user options */
3081         function initialiseFromUser( &$userInput ) {
3082                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3083                 $fname = 'ParserOptions::initialiseFromUser';
3084                 wfProfileIn( $fname );
3085                 if ( !$userInput ) {
3086                         $user = new User;
3087                         $user->setLoaded( true );
3088                 } else {
3089                         $user =& $userInput;
3090                 }
3091
3092                 $this->mUseTeX = $wgUseTeX;
3093                 $this->mUseDynamicDates = $wgUseDynamicDates;
3094                 $this->mInterwikiMagic = $wgInterwikiMagic;
3095                 $this->mAllowExternalImages = $wgAllowExternalImages;
3096                 wfProfileIn( $fname.'-skin' );
3097                 $this->mSkin =& $user->getSkin();
3098                 wfProfileOut( $fname.'-skin' );
3099                 $this->mDateFormat = $user->getOption( 'date' );
3100                 $this->mEditSection = $user->getOption( 'editsection' );
3101                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3102                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3103                 $this->mShowToc = $user->getOption( 'showtoc' );
3104                 wfProfileOut( $fname );
3105         }
3106 }
3107
3108 /**
3109  * Callback function used by Parser::replaceLinkHolders()
3110  * to substitute link placeholders.
3111  */
3112 function &outputReplaceMatches( $matches ) {
3113         global $outputReplace;
3114         return $outputReplace[$matches[1]];
3115 }
3116
3117 /**
3118  * Return the total number of articles
3119  */
3120 function wfNumberOfArticles() {
3121         global $wgNumberOfArticles;
3122
3123         wfLoadSiteStats();
3124         return $wgNumberOfArticles;
3125 }
3126
3127 /**
3128  * Get various statistics from the database
3129  * @private
3130  */
3131 function wfLoadSiteStats() {
3132         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3133         $fname = 'wfLoadSiteStats';
3134
3135         if ( -1 != $wgNumberOfArticles ) return;
3136         $dbr =& wfGetDB( DB_SLAVE );
3137         $s = $dbr->selectRow( 'site_stats',
3138                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3139                 array( 'ss_row_id' => 1 ), $fname
3140         );
3141
3142         if ( $s === false ) {
3143                 return;
3144         } else {
3145                 $wgTotalViews = $s->ss_total_views;
3146                 $wgTotalEdits = $s->ss_total_edits;
3147                 $wgNumberOfArticles = $s->ss_good_articles;
3148         }
3149 }
3150
3151 /**
3152  * Escape html tags
3153  * Basicly replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3154  *
3155  * @param string $in Text that might contain HTML tags
3156  * @return string Escaped string
3157  */
3158 function wfEscapeHTMLTagsOnly( $in ) {
3159         return str_replace(
3160                 array( '"', '>', '<' ),
3161                 array( '&quot;', '&gt;', '&lt;' ),
3162                 $in );
3163 }
3164
3165 ?>