includes/Parser.php

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