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, $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                 $this->mStripState = 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 );
 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
 187                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 188                 $fixtags = array(
 189                         # french spaces, last one Guillemet-left
 190                         # only if there is something before the space
 191                         '/(.) (?=\\?|:|;|!|\\302\\273)/' => '\\1&nbsp;\\2',
 192                         # french spaces, Guillemet-right
 193                         '/(\\302\\253) /' => '\\1&nbsp;',
 194                         '/<hr *>/i' => '<hr />',
 195                         '/<br *>/i' => '<br />',
 196                         '/<center *>/i' => '<div class="center">',
 197                         '/<\\/center *>/i' => '</div>',
 198                 );
 199                 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 200
 201                 # only once and last
 202                 $text = $this->doBlockLevels( $text, $linestart );
 203
 204                 $this->replaceLinkHolders( $text );
 205
 206                 # the position of the convert() call should not be changed. it
 207                 # assumes that the links are all replaces and the only thing left
 208                 # is the <nowiki> mark.
 209                 $text = $wgContLang->convert($text);
 210                 $this->mOutput->setTitleText($wgContLang->getParsedTitle());
 211
 212                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 213
 214                 $text = Sanitizer::normalizeCharReferences( $text );
 215                 global $wgUseTidy;
 216                 if ($wgUseTidy) {
 217                         $text = Parser::tidy($text);
 218                 }
 219
 220                 $this->mOutput->setText( $text );
 221                 wfProfileOut( $fname );
 222                 return $this->mOutput;
 223         }
 224
 225         /**
 226          * Get a random string
 227          *
 228          * @access private
 229          * @static
 230          */
 231         function getRandomString() {
 232                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 233         }
 234
 235         /**
 236          * Replaces all occurrences of <$tag>content</$tag> in the text
 237          * with a random marker and returns the new text. the output parameter
 238          * $content will be an associative array filled with data on the form
 239          * $unique_marker => content.
 240          *
 241          * If $content is already set, the additional entries will be appended
 242          * If $tag is set to STRIP_COMMENTS, the function will extract
 243          * <!-- HTML comments -->
 244          *
 245          * @access private
 246          * @static
 247          */
 248         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 249                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 250                 if ( !$content ) {
 251                         $content = array( );
 252                 }
 253                 $n = 1;
 254                 $stripped = '';
 255
 256                 while ( '' != $text ) {
 257                         if($tag==STRIP_COMMENTS) {
 258                                 $p = preg_split( '/<!--/', $text, 2 );
 259                         } else {
 260                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 261                         }
 262                         $stripped .= $p[0];
 263                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 264                                 $text = '';
 265                         } else {
 266                                 if($tag==STRIP_COMMENTS) {
 267                                         $q = preg_split( '/-->/i', $p[1], 2 );
 268                                 } else {
 269                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 270                                 }
 271                                 $marker = $rnd . sprintf('%08X', $n++);
 272                                 $content[$marker] = $q[0];
 273                                 $stripped .= $marker;
 274                                 $text = $q[1];
 275                         }
 276                 }
 277                 return $stripped;
 278         }
 279
 280         /**
 281          * Strips and renders nowiki, pre, math, hiero
 282          * If $render is set, performs necessary rendering operations on plugins
 283          * Returns the text, and fills an array with data needed in unstrip()
 284          * If the $state is already a valid strip state, it adds to the state
 285          *
 286          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 287          *  will be stripped in addition to other tags. This is important
 288          *  for section editing, where these comments cause confusion when
 289          *  counting the sections in the wikisource
 290          *
 291          * @access private
 292          */
 293         function strip( $text, &$state, $stripcomments = false ) {
 294                 $render = ($this->mOutputType == OT_HTML);
 295                 $html_content = array();
 296                 $nowiki_content = array();
 297                 $math_content = array();
 298                 $pre_content = array();
 299                 $comment_content = array();
 300                 $ext_content = array();
 301                 $gallery_content = array();
 302
 303                 # Replace any instances of the placeholders
 304                 $uniq_prefix = UNIQ_PREFIX;
 305                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 306
 307                 # html
 308                 global $wgRawHtml, $wgWhitelistEdit;
 309                 if( $wgRawHtml && $wgWhitelistEdit ) {
 310                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 311                         foreach( $html_content as $marker => $content ) {
 312                                 if ($render ) {
 313                                         # Raw and unchecked for validity.
 314                                         $html_content[$marker] = $content;
 315                                 } else {
 316                                         $html_content[$marker] = '<html>'.$content.'</html>';
 317                                 }
 318                         }
 319                 }
 320
 321                 # nowiki
 322                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 323                 foreach( $nowiki_content as $marker => $content ) {
 324                         if( $render ){
 325                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 326                         } else {
 327                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 328                         }
 329                 }
 330
 331                 # math
 332                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 333                 foreach( $math_content as $marker => $content ){
 334                         if( $render ) {
 335                                 if( $this->mOptions->getUseTeX() ) {
 336                                         $math_content[$marker] = renderMath( $content );
 337                                 } else {
 338                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 339                                 }
 340                         } else {
 341                                 $math_content[$marker] = '<math>'.$content.'</math>';
 342                         }
 343                 }
 344
 345                 # pre
 346                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 347                 foreach( $pre_content as $marker => $content ){
 348                         if( $render ){
 349                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 350                         } else {
 351                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 352                         }
 353                 }
 354
 355                 # gallery
 356                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 357                 foreach( $gallery_content as $marker => $content ) {
 358                         require_once( 'ImageGallery.php' );
 359                         if ( $render ) {
 360                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 361                         } else {
 362                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 363                         }
 364                 }
 365
 366                 # Comments
 367                 if($stripcomments) {
 368                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 369                         foreach( $comment_content as $marker => $content ){
 370                                 $comment_content[$marker] = '<!--'.$content.'-->';
 371                         }
 372                 }
 373
 374                 # Extensions
 375                 foreach ( $this->mTagHooks as $tag => $callback ) {
 376                         $ext_content[$tag] = array();
 377                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 378                         foreach( $ext_content[$tag] as $marker => $content ) {
 379                                 if ( $render ) {
 380                                         $ext_content[$tag][$marker] = $callback( $content );
 381                                 } else {
 382                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 383                                 }
 384                         }
 385                 }
 386
 387                 # Merge state with the pre-existing state, if there is one
 388                 if ( $state ) {
 389                         $state['html'] = $state['html'] + $html_content;
 390                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 391                         $state['math'] = $state['math'] + $math_content;
 392                         $state['pre'] = $state['pre'] + $pre_content;
 393                         $state['comment'] = $state['comment'] + $comment_content;
 394                         $state['gallery'] = $state['gallery'] + $gallery_content;
 395
 396                         foreach( $ext_content as $tag => $array ) {
 397                                 if ( array_key_exists( $tag, $state ) ) {
 398                                         $state[$tag] = $state[$tag] + $array;
 399                                 }
 400                         }
 401                 } else {
 402                         $state = array(
 403                           'html' => $html_content,
 404                           'nowiki' => $nowiki_content,
 405                           'math' => $math_content,
 406                           'pre' => $pre_content,
 407                           'comment' => $comment_content,
 408                           'gallery' => $gallery_content,
 409                         ) + $ext_content;
 410                 }
 411                 return $text;
 412         }
 413
 414         /**
 415          * restores pre, math, and hiero removed by strip()
 416          *
 417          * always call unstripNoWiki() after this one
 418          * @access private
 419          */
 420         function unstrip( $text, &$state ) {
 421                 # Must expand in reverse order, otherwise nested tags will be corrupted
 422                 $contentDict = end( $state );
 423                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 424                         if( key($state) != 'nowiki' && key($state) != 'html') {
 425                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 426                                         $text = str_replace( key( $contentDict ), $content, $text );
 427                                 }
 428                         }
 429                 }
 430
 431                 return $text;
 432         }
 433
 434         /**
 435          * always call this after unstrip() to preserve the order
 436          *
 437          * @access private
 438          */
 439         function unstripNoWiki( $text, &$state ) {
 440                 # Must expand in reverse order, otherwise nested tags will be corrupted
 441                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 442                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 443                 }
 444
 445                 global $wgRawHtml;
 446                 if ($wgRawHtml) {
 447                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 448                                 $text = str_replace( key( $state['html'] ), $content, $text );
 449                         }
 450                 }
 451
 452                 return $text;
 453         }
 454
 455         /**
 456          * Add an item to the strip state
 457          * Returns the unique tag which must be inserted into the stripped text
 458          * The tag will be replaced with the original text in unstrip()
 459          *
 460          * @access private
 461          */
 462         function insertStripItem( $text, &$state ) {
 463                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 464                 if ( !$state ) {
 465                         $state = array(
 466                           'html' => array(),
 467                           'nowiki' => array(),
 468                           'math' => array(),
 469                           'pre' => array()
 470                         );
 471                 }
 472                 $state['item'][$rnd] = $text;
 473                 return $rnd;
 474         }
 475
 476         /**
 477          * Interface with html tidy, used if $wgUseTidy = true.
 478          * If tidy isn't able to correct the markup, the original will be
 479          * returned in all its glory with a warning comment appended.
 480          *
 481          * Either the external tidy program or the in-process tidy extension
 482          * will be used depending on availability. Override the default
 483          * $wgTidyInternal setting to disable the internal if it's not working.
 484          *
 485          * @param string $text Hideous HTML input
 486          * @return string Corrected HTML output
 487          * @access public
 488          * @static
 489          */
 490         function tidy( $text ) {
 491                 global $wgTidyInternal;
 492                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 493 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 494 '<head><title>test</title></head><body>'.$text.'</body></html>';
 495                 if( $wgTidyInternal ) {
 496                         $correctedtext = Parser::internalTidy( $wrappedtext );
 497                 } else {
 498                         $correctedtext = Parser::externalTidy( $wrappedtext );
 499                 }
 500                 if( is_null( $correctedtext ) ) {
 501                         wfDebug( "Tidy error detected!\n" );
 502                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 503                 }
 504                 return $correctedtext;
 505         }
 506
 507         /**
 508          * Spawn an external HTML tidy process and get corrected markup back from it.
 509          *
 510          * @access private
 511          * @static
 512          */
 513         function externalTidy( $text ) {
 514                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 515                 $fname = 'Parser::externalTidy';
 516                 wfProfileIn( $fname );
 517
 518                 $cleansource = '';
 519                 $opts = ' -utf8';
 520
 521                 $descriptorspec = array(
 522                         0 => array('pipe', 'r'),
 523                         1 => array('pipe', 'w'),
 524                         2 => array('file', '/dev/null', 'a')
 525                 );
 526                 $pipes = array();
 527                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 528                 if (is_resource($process)) {
 529                         fwrite($pipes[0], $text);
 530                         fclose($pipes[0]);
 531                         while (!feof($pipes[1])) {
 532                                 $cleansource .= fgets($pipes[1], 1024);
 533                         }
 534                         fclose($pipes[1]);
 535                         proc_close($process);
 536                 }
 537
 538                 wfProfileOut( $fname );
 539
 540                 if( $cleansource == '' && $text != '') {
 541                         // Some kind of error happened, so we couldn't get the corrected text.
 542                         // Just give up; we'll use the source text and append a warning.
 543                         return null;
 544                 } else {
 545                         return $cleansource;
 546                 }
 547         }
 548
 549         /**
 550          * Use the HTML tidy PECL extension to use the tidy library in-process,
 551          * saving the overhead of spawning a new process. Currently written to
 552          * the PHP 4.3.x version of the extension, may not work on PHP 5.
 553          *
 554          * 'pear install tidy' should be able to compile the extension module.
 555          *
 556          * @access private
 557          * @static
 558          */
 559         function internalTidy( $text ) {
 560                 global $wgTidyConf;
 561                 $fname = 'Parser::internalTidy';
 562                 wfProfileIn( $fname );
 563
 564                 tidy_load_config( $wgTidyConf );
 565                 tidy_set_encoding( 'utf8' );
 566                 tidy_parse_string( $text );
 567                 tidy_clean_repair();
 568                 if( tidy_get_status() == 2 ) {
 569                         // 2 is magic number for fatal error
 570                         // http://www.php.net/manual/en/function.tidy-get-status.php
 571                         $cleansource = null;
 572                 } else {
 573                         $cleansource = tidy_get_output();
 574                 }
 575                 wfProfileOut( $fname );
 576                 return $cleansource;
 577         }
 578
 579         /**
 580          * parse the wiki syntax used to render tables
 581          *
 582          * @access private
 583          */
 584         function doTableStuff ( $t ) {
 585                 $fname = 'Parser::doTableStuff';
 586                 wfProfileIn( $fname );
 587
 588                 $t = explode ( "\n" , $t ) ;
 589                 $td = array () ; # Is currently a td tag open?
 590                 $ltd = array () ; # Was it TD or TH?
 591                 $tr = array () ; # Is currently a tr tag open?
 592                 $ltr = array () ; # tr attributes
 593                 $indent_level = 0; # indent level of the table
 594                 foreach ( $t AS $k => $x )
 595                 {
 596                         $x = trim ( $x ) ;
 597                         $fc = substr ( $x , 0 , 1 ) ;
 598                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 599                                 $indent_level = strlen( $matches[1] );
 600                                 $t[$k] = str_repeat( '<dl><dd>', $indent_level ) .
 601                                         '<table' . Sanitizer::fixTagAttributes ( $matches[2], 'table' ) . '>' ;
 602                                 array_push ( $td , false ) ;
 603                                 array_push ( $ltd , '' ) ;
 604                                 array_push ( $tr , false ) ;
 605                                 array_push ( $ltr , '' ) ;
 606                         }
 607                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 608                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 609                                 $z = "</table>" . substr ( $x , 2);
 610                                 $l = array_pop ( $ltd ) ;
 611                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 612                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 613                                 array_pop ( $ltr ) ;
 614                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 615                         }
 616                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 617                                 $x = substr ( $x , 1 ) ;
 618                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 619                                 $z = '' ;
 620                                 $l = array_pop ( $ltd ) ;
 621                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 622                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 623                                 array_pop ( $ltr ) ;
 624                                 $t[$k] = $z ;
 625                                 array_push ( $tr , false ) ;
 626                                 array_push ( $td , false ) ;
 627                                 array_push ( $ltd , '' ) ;
 628                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $x, 'tr' ) ) ;
 629                         }
 630                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 631                                 # $x is a table row
 632                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 633                                         $fc = '+' ;
 634                                         $x = substr ( $x , 1 ) ;
 635                                 }
 636                                 $after = substr ( $x , 1 ) ;
 637                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 638                                 $after = explode ( '||' , $after ) ;
 639                                 $t[$k] = '' ;
 640
 641                                 # Loop through each table cell
 642                                 foreach ( $after AS $theline )
 643                                 {
 644                                         $z = '' ;
 645                                         if ( $fc != '+' )
 646                                         {
 647                                                 $tra = array_pop ( $ltr ) ;
 648                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 649                                                 array_push ( $tr , true ) ;
 650                                                 array_push ( $ltr , '' ) ;
 651                                         }
 652
 653                                         $l = array_pop ( $ltd ) ;
 654                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 655                                         if ( $fc == '|' ) $l = 'td' ;
 656                                         else if ( $fc == '!' ) $l = 'th' ;
 657                                         else if ( $fc == '+' ) $l = 'caption' ;
 658                                         else $l = '' ;
 659                                         array_push ( $ltd , $l ) ;
 660
 661                                         # Cell parameters
 662                                         $y = explode ( '|' , $theline , 2 ) ;
 663                                         # Note that a '|' inside an invalid link should not
 664                                         # be mistaken as delimiting cell parameters
 665                                         if ( strpos( $y[0], '[[' ) !== false ) {
 666                                                 $y = array ($theline);
 667                                         }
 668                                         if ( count ( $y ) == 1 )
 669                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 670                                         else $y = $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($y[0], $l).">{$y[1]}" ;
 671                                         $t[$k] .= $y ;
 672                                         array_push ( $td , true ) ;
 673                                 }
 674                         }
 675                 }
 676
 677                 # Closing open td, tr && table
 678                 while ( count ( $td ) > 0 )
 679                 {
 680                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 681                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 682                         $t[] = '</table>' ;
 683                 }
 684
 685                 $t = implode ( "\n" , $t ) ;
 686                 wfProfileOut( $fname );
 687                 return $t ;
 688         }
 689
 690         /**
 691          * Helper function for parse() that transforms wiki markup into
 692          * HTML. Only called for $mOutputType == OT_HTML.
 693          *
 694          * @access private
 695          */
 696         function internalParse( $text ) {
 697                 global $wgContLang;
 698                 $args = array();
 699                 $isMain = true;
 700                 $fname = 'Parser::internalParse';
 701                 wfProfileIn( $fname );
 702
 703                 $text = Sanitizer::removeHTMLtags( $text );
 704                 $text = $this->replaceVariables( $text, $args );
 705
 706                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 707
 708                 $text = $this->doHeadings( $text );
 709                 if($this->mOptions->getUseDynamicDates()) {
 710                         $df =& DateFormatter::getInstance();
 711                         $text = $df->reformat( $this->mOptions->getDateFormat(), $text );
 712                 }
 713                 $text = $this->doAllQuotes( $text );
 714                 $text = $this->replaceInternalLinks( $text );
 715                 $text = $this->replaceExternalLinks( $text );
 716
 717                 # replaceInternalLinks may sometimes leave behind
 718                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 719                 $text = str_replace("http-noparse://","http://",$text);
 720
 721                 $text = $this->doMagicLinks( $text );
 722                 $text = $this->doTableStuff( $text );
 723                 $text = $this->formatHeadings( $text, $isMain );
 724
 725                 wfProfileOut( $fname );
 726                 return $text;
 727         }
 728
 729         /**
 730          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 731          * magic external links.
 732          *
 733          * @access private
 734          */
 735         function &doMagicLinks( &$text ) {
 736                 $text = $this->magicISBN( $text );
 737                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 738                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 739                 return $text;
 740         }
 741
 742         /**
 743          * Parse ^^ tokens and return html
 744          *
 745          * @access private
 746          */
 747         function doExponent( $text ) {
 748                 $fname = 'Parser::doExponent';
 749                 wfProfileIn( $fname );
 750                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 751                 wfProfileOut( $fname );
 752                 return $text;
 753         }
 754
 755         /**
 756          * Parse headers and return html
 757          *
 758          * @access private
 759          */
 760         function doHeadings( $text ) {
 761                 $fname = 'Parser::doHeadings';
 762                 wfProfileIn( $fname );
 763                 for ( $i = 6; $i >= 1; --$i ) {
 764                         $h = substr( '======', 0, $i );
 765                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 766                           "<h{$i}>\\1</h{$i}>\\2", $text );
 767                 }
 768                 wfProfileOut( $fname );
 769                 return $text;
 770         }
 771
 772         /**
 773          * Replace single quotes with HTML markup
 774          * @access private
 775          * @return string the altered text
 776          */
 777         function doAllQuotes( $text ) {
 778                 $fname = 'Parser::doAllQuotes';
 779                 wfProfileIn( $fname );
 780                 $outtext = '';
 781                 $lines = explode( "\n", $text );
 782                 foreach ( $lines as $line ) {
 783                         $outtext .= $this->doQuotes ( $line ) . "\n";
 784                 }
 785                 $outtext = substr($outtext, 0,-1);
 786                 wfProfileOut( $fname );
 787                 return $outtext;
 788         }
 789
 790         /**
 791          * Helper function for doAllQuotes()
 792          * @access private
 793          */
 794         function doQuotes( $text ) {
 795                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 796                 if ( count( $arr ) == 1 )
 797                         return $text;
 798                 else
 799                 {
 800                         # First, do some preliminary work. This may shift some apostrophes from
 801                         # being mark-up to being text. It also counts the number of occurrences
 802                         # of bold and italics mark-ups.
 803                         $i = 0;
 804                         $numbold = 0;
 805                         $numitalics = 0;
 806                         foreach ( $arr as $r )
 807                         {
 808                                 if ( ( $i % 2 ) == 1 )
 809                                 {
 810                                         # If there are ever four apostrophes, assume the first is supposed to
 811                                         # be text, and the remaining three constitute mark-up for bold text.
 812                                         if ( strlen( $arr[$i] ) == 4 )
 813                                         {
 814                                                 $arr[$i-1] .= "'";
 815                                                 $arr[$i] = "'''";
 816                                         }
 817                                         # If there are more than 5 apostrophes in a row, assume they're all
 818                                         # text except for the last 5.
 819                                         else if ( strlen( $arr[$i] ) > 5 )
 820                                         {
 821                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 822                                                 $arr[$i] = "'''''";
 823                                         }
 824                                         # Count the number of occurrences of bold and italics mark-ups.
 825                                         # We are not counting sequences of five apostrophes.
 826                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 827                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 828                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 829                                 }
 830                                 $i++;
 831                         }
 832
 833                         # If there is an odd number of both bold and italics, it is likely
 834                         # that one of the bold ones was meant to be an apostrophe followed
 835                         # by italics. Which one we cannot know for certain, but it is more
 836                         # likely to be one that has a single-letter word before it.
 837                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 838                         {
 839                                 $i = 0;
 840                                 $firstsingleletterword = -1;
 841                                 $firstmultiletterword = -1;
 842                                 $firstspace = -1;
 843                                 foreach ( $arr as $r )
 844                                 {
 845                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 846                                         {
 847                                                 $x1 = substr ($arr[$i-1], -1);
 848                                                 $x2 = substr ($arr[$i-1], -2, 1);
 849                                                 if ($x1 == ' ') {
 850                                                         if ($firstspace == -1) $firstspace = $i;
 851                                                 } else if ($x2 == ' ') {
 852                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 853                                                 } else {
 854                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 855                                                 }
 856                                         }
 857                                         $i++;
 858                                 }
 859
 860                                 # If there is a single-letter word, use it!
 861                                 if ($firstsingleletterword > -1)
 862                                 {
 863                                         $arr [ $firstsingleletterword ] = "''";
 864                                         $arr [ $firstsingleletterword-1 ] .= "'";
 865                                 }
 866                                 # If not, but there's a multi-letter word, use that one.
 867                                 else if ($firstmultiletterword > -1)
 868                                 {
 869                                         $arr [ $firstmultiletterword ] = "''";
 870                                         $arr [ $firstmultiletterword-1 ] .= "'";
 871                                 }
 872                                 # ... otherwise use the first one that has neither.
 873                                 # (notice that it is possible for all three to be -1 if, for example,
 874                                 # there is only one pentuple-apostrophe in the line)
 875                                 else if ($firstspace > -1)
 876                                 {
 877                                         $arr [ $firstspace ] = "''";
 878                                         $arr [ $firstspace-1 ] .= "'";
 879                                 }
 880                         }
 881
 882                         # Now let's actually convert our apostrophic mush to HTML!
 883                         $output = '';
 884                         $buffer = '';
 885                         $state = '';
 886                         $i = 0;
 887                         foreach ($arr as $r)
 888                         {
 889                                 if (($i % 2) == 0)
 890                                 {
 891                                         if ($state == 'both')
 892                                                 $buffer .= $r;
 893                                         else
 894                                                 $output .= $r;
 895                                 }
 896                                 else
 897                                 {
 898                                         if (strlen ($r) == 2)
 899                                         {
 900                                                 if ($state == 'i')
 901                                                 { $output .= '</i>'; $state = ''; }
 902                                                 else if ($state == 'bi')
 903                                                 { $output .= '</i>'; $state = 'b'; }
 904                                                 else if ($state == 'ib')
 905                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 906                                                 else if ($state == 'both')
 907                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 908                                                 else # $state can be 'b' or ''
 909                                                 { $output .= '<i>'; $state .= 'i'; }
 910                                         }
 911                                         else if (strlen ($r) == 3)
 912                                         {
 913                                                 if ($state == 'b')
 914                                                 { $output .= '</b>'; $state = ''; }
 915                                                 else if ($state == 'bi')
 916                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 917                                                 else if ($state == 'ib')
 918                                                 { $output .= '</b>'; $state = 'i'; }
 919                                                 else if ($state == 'both')
 920                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 921                                                 else # $state can be 'i' or ''
 922                                                 { $output .= '<b>'; $state .= 'b'; }
 923                                         }
 924                                         else if (strlen ($r) == 5)
 925                                         {
 926                                                 if ($state == 'b')
 927                                                 { $output .= '</b><i>'; $state = 'i'; }
 928                                                 else if ($state == 'i')
 929                                                 { $output .= '</i><b>'; $state = 'b'; }
 930                                                 else if ($state == 'bi')
 931                                                 { $output .= '</i></b>'; $state = ''; }
 932                                                 else if ($state == 'ib')
 933                                                 { $output .= '</b></i>'; $state = ''; }
 934                                                 else if ($state == 'both')
 935                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 936                                                 else # ($state == '')
 937                                                 { $buffer = ''; $state = 'both'; }
 938                                         }
 939                                 }
 940                                 $i++;
 941                         }
 942                         # Now close all remaining tags.  Notice that the order is important.
 943                         if ($state == 'b' || $state == 'ib')
 944                                 $output .= '</b>';
 945                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 946                                 $output .= '</i>';
 947                         if ($state == 'bi')
 948                                 $output .= '</b>';
 949                         if ($state == 'both')
 950                                 $output .= '<b><i>'.$buffer.'</i></b>';
 951                         return $output;
 952                 }
 953         }
 954
 955         /**
 956          * Replace external links
 957          *
 958          * Note: this is all very hackish and the order of execution matters a lot.
 959          * Make sure to run maintenance/parserTests.php if you change this code.
 960          *
 961          * @access private
 962          */
 963         function replaceExternalLinks( $text ) {
 964                 $fname = 'Parser::replaceExternalLinks';
 965                 wfProfileIn( $fname );
 966
 967                 $sk =& $this->mOptions->getSkin();
 968                 global $wgContLang;
 969                 $linktrail = $wgContLang->linkTrail();
 970
 971                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 972
 973                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 974
 975                 $i = 0;
 976                 while ( $i<count( $bits ) ) {
 977                         $url = $bits[$i++];
 978                         $protocol = $bits[$i++];
 979                         $text = $bits[$i++];
 980                         $trail = $bits[$i++];
 981
 982                         # The characters '<' and '>' (which were escaped by
 983                         # removeHTMLtags()) should not be included in
 984                         # URLs, per RFC 2396.
 985                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 986                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 987                                 $url = substr($url, 0, $m2[0][1]);
 988                         }
 989
 990                         # If the link text is an image URL, replace it with an <img> tag
 991                         # This happened by accident in the original parser, but some people used it extensively
 992                         $img = $this->maybeMakeImageLink( $text );
 993                         if ( $img !== false ) {
 994                                 $text = $img;
 995                         }
 996
 997                         $dtrail = '';
 998
 999                         # Set linktype for CSS - if URL==text, link is essentially free
1000                         $linktype = ($text == $url) ? 'free' : 'text';
1001
1002                         # No link text, e.g. [http://domain.tld/some.link]
1003                         if ( $text == '' ) {
1004                                 # Autonumber if allowed
1005                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
1006                                         $text = '[' . ++$this->mAutonumber . ']';
1007                                         $linktype = 'autonumber';
1008                                 } else {
1009                                         # Otherwise just use the URL
1010                                         $text = htmlspecialchars( $url );
1011                                         $linktype = 'free';
1012                                 }
1013                         } else {
1014                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1015                                 # Check for trail
1016                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
1017                                         $dtrail = $m2[1];
1018                                         $trail = $m2[2];
1019                                 }
1020                         }
1021
1022                         $text = $wgContLang->markNoConversion($text);
1023
1024                         # Replace &amp; from obsolete syntax with &.
1025                         # All HTML entities will be escaped by makeExternalLink()
1026                         # or maybeMakeImageLink()
1027                         $url = str_replace( '&amp;', '&', $url );
1028
1029                         # Process the trail (i.e. everything after this link up until start of the next link),
1030                         # replacing any non-bracketed links
1031                         $trail = $this->replaceFreeExternalLinks( $trail );
1032
1033
1034                         # Use the encoded URL
1035                         # This means that users can paste URLs directly into the text
1036                         # Funny characters like &ouml; aren't valid in URLs anyway
1037                         # This was changed in August 2004
1038                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
1039                 }
1040
1041                 wfProfileOut( $fname );
1042                 return $s;
1043         }
1044
1045         /**
1046          * Replace anything that looks like a URL with a link
1047          * @access private
1048          */
1049         function replaceFreeExternalLinks( $text ) {
1050                 global $wgContLang;
1051                 $fname = 'Parser::replaceFreeExternalLinks';
1052                 wfProfileIn( $fname );
1053
1054                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1055                 $s = array_shift( $bits );
1056                 $i = 0;
1057
1058                 $sk =& $this->mOptions->getSkin();
1059
1060                 while ( $i < count( $bits ) ){
1061                         $protocol = $bits[$i++];
1062                         $remainder = $bits[$i++];
1063
1064                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1065                                 # Found some characters after the protocol that look promising
1066                                 $url = $protocol . $m[1];
1067                                 $trail = $m[2];
1068
1069                                 # The characters '<' and '>' (which were escaped by
1070                                 # removeHTMLtags()) should not be included in
1071                                 # URLs, per RFC 2396.
1072                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1073                                         $trail = substr($url, $m2[0][1]) . $trail;
1074                                         $url = substr($url, 0, $m2[0][1]);
1075                                 }
1076
1077                                 # Move trailing punctuation to $trail
1078                                 $sep = ',;\.:!?';
1079                                 # If there is no left bracket, then consider right brackets fair game too
1080                                 if ( strpos( $url, '(' ) === false ) {
1081                                         $sep .= ')';
1082                                 }
1083
1084                                 $numSepChars = strspn( strrev( $url ), $sep );
1085                                 if ( $numSepChars ) {
1086                                         $trail = substr( $url, -$numSepChars ) . $trail;
1087                                         $url = substr( $url, 0, -$numSepChars );
1088                                 }
1089
1090                                 # Replace &amp; from obsolete syntax with &.
1091                                 # All HTML entities will be escaped by makeExternalLink()
1092                                 # or maybeMakeImageLink()
1093                                 $url = str_replace( '&amp;', '&', $url );
1094
1095                                 # Is this an external image?
1096                                 $text = $this->maybeMakeImageLink( $url );
1097                                 if ( $text === false ) {
1098                                         # Not an image, make a link
1099                                         $text = $sk->makeExternalLink( $url, $wgContLang->markNoConversion($url), true, 'free' );
1100                                 }
1101                                 $s .= $text . $trail;
1102                         } else {
1103                                 $s .= $protocol . $remainder;
1104                         }
1105                 }
1106                 wfProfileOut();
1107                 return $s;
1108         }
1109
1110         /**
1111          * make an image if it's allowed
1112          * @access private
1113          */
1114         function maybeMakeImageLink( $url ) {
1115                 $sk =& $this->mOptions->getSkin();
1116                 $text = false;
1117                 if ( $this->mOptions->getAllowExternalImages() ) {
1118                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1119                                 # Image found
1120                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1121                         }
1122                 }
1123                 return $text;
1124         }
1125
1126         /**
1127          * Process [[ ]] wikilinks
1128          *
1129          * @access private
1130          */
1131         function replaceInternalLinks( $s ) {
1132                 global $wgContLang, $wgLinkCache;
1133                 static $fname = 'Parser::replaceInternalLinks' ;
1134
1135                 wfProfileIn( $fname );
1136
1137                 wfProfileIn( $fname.'-setup' );
1138                 static $tc = FALSE;
1139                 # the % is needed to support urlencoded titles as well
1140                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1141
1142                 $sk =& $this->mOptions->getSkin();
1143                 global $wgUseOldExistenceCheck;
1144                 # "Post-parse link colour check" works only on wiki text since it's now
1145                 # in Parser. Enable it, then disable it when we're done.
1146                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1147
1148                 #split the entire text string on occurences of [[
1149                 $a = explode( '[[', ' ' . $s );
1150                 #get the first element (all text up to first [[), and remove the space we added
1151                 $s = array_shift( $a );
1152                 $s = substr( $s, 1 );
1153
1154                 # Match a link having the form [[namespace:link|alternate]]trail
1155                 static $e1 = FALSE;
1156                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1157                 # Match cases where there is no "]]", which might still be images
1158                 static $e1_img = FALSE;
1159                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1160                 # Match the end of a line for a word that's not followed by whitespace,
1161                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1162                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1163
1164                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1165
1166                 if( is_null( $this->mTitle ) ) {
1167                         wfDebugDieBacktrace( 'nooo' );
1168                 }
1169                 $nottalk = !$this->mTitle->isTalkPage();
1170
1171                 if ( $useLinkPrefixExtension ) {
1172                         if ( preg_match( $e2, $s, $m ) ) {
1173                                 $first_prefix = $m[2];
1174                                 $s = $m[1];
1175                         } else {
1176                                 $first_prefix = false;
1177                         }
1178                 } else {
1179                         $prefix = '';
1180                 }
1181
1182                 $selflink = $this->mTitle->getPrefixedText();
1183                 wfProfileOut( $fname.'-setup' );
1184
1185                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1186                 $useSubpages = $this->areSubpagesAllowed();
1187
1188                 # Loop for each link
1189                 for ($k = 0; isset( $a[$k] ); $k++) {
1190                         $line = $a[$k];
1191                         if ( $useLinkPrefixExtension ) {
1192                                 wfProfileIn( $fname.'-prefixhandling' );
1193                                 if ( preg_match( $e2, $s, $m ) ) {
1194                                         $prefix = $m[2];
1195                                         $s = $m[1];
1196                                 } else {
1197                                         $prefix='';
1198                                 }
1199                                 # first link
1200                                 if($first_prefix) {
1201                                         $prefix = $first_prefix;
1202                                         $first_prefix = false;
1203                                 }
1204                                 wfProfileOut( $fname.'-prefixhandling' );
1205                         }
1206
1207                         $might_be_img = false;
1208
1209                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1210                                 $text = $m[2];
1211                                 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1212                                 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1213                                 # the real problem is with the $e1 regex
1214                                 # See bug 1300.
1215                                 if (preg_match( "/^\](.*)/s", $m[3], $n ) ) {
1216                                         $text .= ']'; # so that replaceExternalLinks($text) works later
1217                                         $m[3] = $n[1];
1218                                 }
1219                                 # fix up urlencoded title texts
1220                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1221                                 $trail = $m[3];
1222                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1223                                 $might_be_img = true;
1224                                 $text = $m[2];
1225                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1226                                 $trail = "";
1227                         } else { # Invalid form; output directly
1228                                 $s .= $prefix . '[[' . $line ;
1229                                 continue;
1230                         }
1231
1232                         # Don't allow internal links to pages containing
1233                         # PROTO: where PROTO is a valid URL protocol; these
1234                         # should be external links.
1235                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1236                                 $s .= $prefix . '[[' . $line ;
1237                                 continue;
1238                         }
1239
1240                         # Make subpage if necessary
1241                         if( $useSubpages ) {
1242                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1243                         } else {
1244                                 $link = $m[1];
1245                         }
1246
1247                         $noforce = (substr($m[1], 0, 1) != ':');
1248                         if (!$noforce) {
1249                                 # Strip off leading ':'
1250                                 $link = substr($link, 1);
1251                         }
1252
1253                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1254                         if( !$nt ) {
1255                                 $s .= $prefix . '[[' . $line;
1256                                 continue;
1257                         }
1258
1259                         #check other language variants of the link
1260                         #if the article does not exist
1261                         if( $checkVariantLink
1262                             && $nt->getArticleID() == 0 ) {
1263                                 $wgContLang->findVariantLink($link, $nt);
1264                         }
1265
1266                         $ns = $nt->getNamespace();
1267                         $iw = $nt->getInterWiki();
1268
1269                         if ($might_be_img) { # if this is actually an invalid link
1270                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1271                                         $found = false;
1272                                         while (isset ($a[$k+1]) ) {
1273                                                 #look at the next 'line' to see if we can close it there
1274                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1275                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1276                                                 # the first ]] closes the inner link, the second the image
1277                                                         $found = true;
1278                                                         $text .= '[[' . $m[1];
1279                                                         $trail = $m[2];
1280                                                         break;
1281                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1282                                                         #if there's exactly one ]] that's fine, we'll keep looking
1283                                                         $text .= '[[' . $m[0];
1284                                                 } else {
1285                                                         #if $next_line is invalid too, we need look no further
1286                                                         $text .= '[[' . $next_line;
1287                                                         break;
1288                                                 }
1289                                         }
1290                                         if ( !$found ) {
1291                                                 # we couldn't find the end of this imageLink, so output it raw
1292                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1293                                                 $text = $this->replaceInternalLinks($text);
1294                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1295                                                 # note: no $trail, because without an end, there *is* no trail
1296                                                 continue;
1297                                         }
1298                                 } else { #it's not an image, so output it raw
1299                                         $s .= $prefix . '[[' . $link . '|' . $text;
1300                                         # note: no $trail, because without an end, there *is* no trail
1301                                         continue;
1302                                 }
1303                         }
1304
1305                         $wasblank = ( '' == $text );
1306                         if( $wasblank ) $text = $link;
1307
1308
1309                         # Link not escaped by : , create the various objects
1310                         if( $noforce ) {
1311
1312                                 # Interwikis
1313                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1314                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1315                                         $s = rtrim($s . "\n");
1316                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1317                                         continue;
1318                                 }
1319
1320                                 if ( $ns == NS_IMAGE ) {
1321                                         wfProfileIn( "$fname-image" );
1322                                         if ( !wfIsBadImage( $nt->getDBkey() ) ) {
1323                                                 # recursively parse links inside the image caption
1324                                                 # actually, this will parse them in any other parameters, too,
1325                                                 # but it might be hard to fix that, and it doesn't matter ATM
1326                                                 $text = $this->replaceExternalLinks($text);
1327                                                 $text = $this->replaceInternalLinks($text);
1328
1329                                                 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1330                                                 $s .= $prefix . str_replace('http://', 'http-noparse://', $sk->makeImageLinkObj( $nt, $text ) ) . $trail;
1331                                                 $wgLinkCache->addImageLinkObj( $nt );
1332
1333                                                 wfProfileOut( "$fname-image" );
1334                                                 continue;
1335                                         }
1336                                         wfProfileOut( "$fname-image" );
1337
1338                                 }
1339
1340                                 if ( $ns == NS_CATEGORY ) {
1341                                         wfProfileIn( "$fname-category" );
1342                                         $t = $nt->getText();
1343                                         $s = rtrim($s . "\n"); # bug 87
1344
1345                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1346                                         $pPLC=$sk->postParseLinkColour();
1347                                         $sk->postParseLinkColour( false );
1348                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1349                                         $sk->postParseLinkColour( $pPLC );
1350                                         $wgLinkCache->resume();
1351
1352                                         if ( $wasblank ) {
1353                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1354                                                         $sortkey = $this->mTitle->getText();
1355                                                 } else {
1356                                                         $sortkey = $this->mTitle->getPrefixedText();
1357                                                 }
1358                                         } else {
1359                                                 $sortkey = $text;
1360                                         }
1361                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1362                                         $this->mOutput->addCategoryLink( $t );
1363
1364                                         /**
1365                                          * Strip the whitespace Category links produce, see bug 87
1366                                          * @todo We might want to use trim($tmp, "\n") here.
1367                                          */
1368                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1369
1370                                         wfProfileOut( "$fname-category" );
1371                                         continue;
1372                                 }
1373                         }
1374
1375                         if( ( $nt->getPrefixedText() === $selflink ) &&
1376                             ( $nt->getFragment() === '' ) ) {
1377                                 # Self-links are handled specially; generally de-link and change to bold.
1378                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1379                                 continue;
1380                         }
1381
1382                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1383                         if( $ns == NS_MEDIA ) {
1384                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1385                                 $wgLinkCache->addImageLinkObj( $nt );
1386                                 continue;
1387                         } elseif( $ns == NS_SPECIAL ) {
1388                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1389                                 continue;
1390                         }
1391                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1392                 }
1393                 $sk->postParseLinkColour( $saveParseColour );
1394                 wfProfileOut( $fname );
1395                 return $s;
1396         }
1397
1398         /**
1399          * Return true if subpage links should be expanded on this page.
1400          * @return bool
1401          */
1402         function areSubpagesAllowed() {
1403                 # Some namespaces don't allow subpages
1404                 global $wgNamespacesWithSubpages;
1405                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1406         }
1407
1408         /**
1409          * Handle link to subpage if necessary
1410          * @param string $target the source of the link
1411          * @param string &$text the link text, modified as necessary
1412          * @return string the full name of the link
1413          * @access private
1414          */
1415         function maybeDoSubpageLink($target, &$text) {
1416                 # Valid link forms:
1417                 # Foobar -- normal
1418                 # :Foobar -- override special treatment of prefix (images, language links)
1419                 # /Foobar -- convert to CurrentPage/Foobar
1420                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1421                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1422                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1423
1424                 $fname = 'Parser::maybeDoSubpageLink';
1425                 wfProfileIn( $fname );
1426                 $ret = $target; # default return value is no change
1427
1428                 # Some namespaces don't allow subpages,
1429                 # so only perform processing if subpages are allowed
1430                 if( $this->areSubpagesAllowed() ) {
1431                         # Look at the first character
1432                         if( $target != '' && $target{0} == '/' ) {
1433                                 # / at end means we don't want the slash to be shown
1434                                 if( substr( $target, -1, 1 ) == '/' ) {
1435                                         $target = substr( $target, 1, -1 );
1436                                         $noslash = $target;
1437                                 } else {
1438                                         $noslash = substr( $target, 1 );
1439                                 }
1440
1441                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1442                                 if( '' === $text ) {
1443                                         $text = $target;
1444                                 } # this might be changed for ugliness reasons
1445                         } else {
1446                                 # check for .. subpage backlinks
1447                                 $dotdotcount = 0;
1448                                 $nodotdot = $target;
1449                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1450                                         ++$dotdotcount;
1451                                         $nodotdot = substr( $nodotdot, 3 );
1452                                 }
1453                                 if($dotdotcount > 0) {
1454                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1455                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1456                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1457                                                 # / at the end means don't show full path
1458                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1459                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1460                                                         if( '' === $text ) {
1461                                                                 $text = $nodotdot;
1462                                                         }
1463                                                 }
1464                                                 $nodotdot = trim( $nodotdot );
1465                                                 if( $nodotdot != '' ) {
1466                                                         $ret .= '/' . $nodotdot;
1467                                                 }
1468                                         }
1469                                 }
1470                         }
1471                 }
1472
1473                 wfProfileOut( $fname );
1474                 return $ret;
1475         }
1476
1477         /**#@+
1478          * Used by doBlockLevels()
1479          * @access private
1480          */
1481         /* private */ function closeParagraph() {
1482                 $result = '';
1483                 if ( '' != $this->mLastSection ) {
1484                         $result = '</' . $this->mLastSection  . ">\n";
1485                 }
1486                 $this->mInPre = false;
1487                 $this->mLastSection = '';
1488                 return $result;
1489         }
1490         # getCommon() returns the length of the longest common substring
1491         # of both arguments, starting at the beginning of both.
1492         #
1493         /* private */ function getCommon( $st1, $st2 ) {
1494                 $fl = strlen( $st1 );
1495                 $shorter = strlen( $st2 );
1496                 if ( $fl < $shorter ) { $shorter = $fl; }
1497
1498                 for ( $i = 0; $i < $shorter; ++$i ) {
1499                         if ( $st1{$i} != $st2{$i} ) { break; }
1500                 }
1501                 return $i;
1502         }
1503         # These next three functions open, continue, and close the list
1504         # element appropriate to the prefix character passed into them.
1505         #
1506         /* private */ function openList( $char ) {
1507                 $result = $this->closeParagraph();
1508
1509                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1510                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1511                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1512                 else if ( ';' == $char ) {
1513                         $result .= '<dl><dt>';
1514                         $this->mDTopen = true;
1515                 }
1516                 else { $result = '<!-- ERR 1 -->'; }
1517
1518                 return $result;
1519         }
1520
1521         /* private */ function nextItem( $char ) {
1522                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1523                 else if ( ':' == $char || ';' == $char ) {
1524                         $close = '</dd>';
1525                         if ( $this->mDTopen ) { $close = '</dt>'; }
1526                         if ( ';' == $char ) {
1527                                 $this->mDTopen = true;
1528                                 return $close . '<dt>';
1529                         } else {
1530                                 $this->mDTopen = false;
1531                                 return $close . '<dd>';
1532                         }
1533                 }
1534                 return '<!-- ERR 2 -->';
1535         }
1536
1537         /* private */ function closeList( $char ) {
1538                 if ( '*' == $char ) { $text = '</li></ul>'; }
1539                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1540                 else if ( ':' == $char ) {
1541                         if ( $this->mDTopen ) {
1542                                 $this->mDTopen = false;
1543                                 $text = '</dt></dl>';
1544                         } else {
1545                                 $text = '</dd></dl>';
1546                         }
1547                 }
1548                 else {  return '<!-- ERR 3 -->'; }
1549                 return $text."\n";
1550         }
1551         /**#@-*/
1552
1553         /**
1554          * Make lists from lines starting with ':', '*', '#', etc.
1555          *
1556          * @access private
1557          * @return string the lists rendered as HTML
1558          */
1559         function doBlockLevels( $text, $linestart ) {
1560                 $fname = 'Parser::doBlockLevels';
1561                 wfProfileIn( $fname );
1562
1563                 # Parsing through the text line by line.  The main thing
1564                 # happening here is handling of block-level elements p, pre,
1565                 # and making lists from lines starting with * # : etc.
1566                 #
1567                 $textLines = explode( "\n", $text );
1568
1569                 $lastPrefix = $output = '';
1570                 $this->mDTopen = $inBlockElem = false;
1571                 $prefixLength = 0;
1572                 $paragraphStack = false;
1573
1574                 if ( !$linestart ) {
1575                         $output .= array_shift( $textLines );
1576                 }
1577                 foreach ( $textLines as $oLine ) {
1578                         $lastPrefixLength = strlen( $lastPrefix );
1579                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1580                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1581                         if ( !$this->mInPre ) {
1582                                 # Multiple prefixes may abut each other for nested lists.
1583                                 $prefixLength = strspn( $oLine, '*#:;' );
1584                                 $pref = substr( $oLine, 0, $prefixLength );
1585
1586                                 # eh?
1587                                 $pref2 = str_replace( ';', ':', $pref );
1588                                 $t = substr( $oLine, $prefixLength );
1589                                 $this->mInPre = !empty($preOpenMatch);
1590                         } else {
1591                                 # Don't interpret any other prefixes in preformatted text
1592                                 $prefixLength = 0;
1593                                 $pref = $pref2 = '';
1594                                 $t = $oLine;
1595                         }
1596
1597                         # List generation
1598                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1599                                 # Same as the last item, so no need to deal with nesting or opening stuff
1600                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1601                                 $paragraphStack = false;
1602
1603                                 if ( substr( $pref, -1 ) == ';') {
1604                                         # The one nasty exception: definition lists work like this:
1605                                         # ; title : definition text
1606                                         # So we check for : in the remainder text to split up the
1607                                         # title and definition, without b0rking links.
1608                                         $term = $t2 = '';
1609                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1610                                                 $t = $t2;
1611                                                 $output .= $term . $this->nextItem( ':' );
1612                                         }
1613                                 }
1614                         } elseif( $prefixLength || $lastPrefixLength ) {
1615                                 # Either open or close a level...
1616                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1617                                 $paragraphStack = false;
1618
1619                                 while( $commonPrefixLength < $lastPrefixLength ) {
1620                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1621                                         --$lastPrefixLength;
1622                                 }
1623                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1624                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1625                                 }
1626                                 while ( $prefixLength > $commonPrefixLength ) {
1627                                         $char = substr( $pref, $commonPrefixLength, 1 );
1628                                         $output .= $this->openList( $char );
1629
1630                                         if ( ';' == $char ) {
1631                                                 # FIXME: This is dupe of code above
1632                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1633                                                         $t = $t2;
1634                                                         $output .= $term . $this->nextItem( ':' );
1635                                                 }
1636                                         }
1637                                         ++$commonPrefixLength;
1638                                 }
1639                                 $lastPrefix = $pref2;
1640                         }
1641                         if( 0 == $prefixLength ) {
1642                                 wfProfileIn( "$fname-paragraph" );
1643                                 # No prefix (not in list)--go to paragraph mode
1644                                 $uniq_prefix = UNIQ_PREFIX;
1645                                 // XXX: use a stack for nestable elements like span, table and div
1646                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1647                                 $closematch = preg_match(
1648                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1649                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1650                                 if ( $openmatch or $closematch ) {
1651                                         $paragraphStack = false;
1652                                         $output .= $this->closeParagraph();
1653                                         if($preOpenMatch and !$preCloseMatch) {
1654                                                 $this->mInPre = true;
1655                                         }
1656                                         if ( $closematch ) {
1657                                                 $inBlockElem = false;
1658                                         } else {
1659                                                 $inBlockElem = true;
1660                                         }
1661                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1662                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1663                                                 // pre
1664                                                 if ($this->mLastSection != 'pre') {
1665                                                         $paragraphStack = false;
1666                                                         $output .= $this->closeParagraph().'<pre>';
1667                                                         $this->mLastSection = 'pre';
1668                                                 }
1669                                                 $t = substr( $t, 1 );
1670                                         } else {
1671                                                 // paragraph
1672                                                 if ( '' == trim($t) ) {
1673                                                         if ( $paragraphStack ) {
1674                                                                 $output .= $paragraphStack.'<br />';
1675                                                                 $paragraphStack = false;
1676                                                                 $this->mLastSection = 'p';
1677                                                         } else {
1678                                                                 if ($this->mLastSection != 'p' ) {
1679                                                                         $output .= $this->closeParagraph();
1680                                                                         $this->mLastSection = '';
1681                                                                         $paragraphStack = '<p>';
1682                                                                 } else {
1683                                                                         $paragraphStack = '</p><p>';
1684                                                                 }
1685                                                         }
1686                                                 } else {
1687                                                         if ( $paragraphStack ) {
1688                                                                 $output .= $paragraphStack;
1689                                                                 $paragraphStack = false;
1690                                                                 $this->mLastSection = 'p';
1691                                                         } else if ($this->mLastSection != 'p') {
1692                                                                 $output .= $this->closeParagraph().'<p>';
1693                                                                 $this->mLastSection = 'p';
1694                                                         }
1695                                                 }
1696                                         }
1697                                 }
1698                                 wfProfileOut( "$fname-paragraph" );
1699                         }
1700                         if ($paragraphStack === false) {
1701                                 $output .= $t."\n";
1702                         }
1703                 }
1704                 while ( $prefixLength ) {
1705                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1706                         --$prefixLength;
1707                 }
1708                 if ( '' != $this->mLastSection ) {
1709                         $output .= '</' . $this->mLastSection . '>';
1710                         $this->mLastSection = '';
1711                 }
1712
1713                 wfProfileOut( $fname );
1714                 return $output;
1715         }
1716
1717         /**
1718          * Split up a string on ':', ignoring any occurences inside
1719          * <a>..</a> or <span>...</span>
1720          * @param string $str the string to split
1721          * @param string &$before set to everything before the ':'
1722          * @param string &$after set to everything after the ':'
1723          * return string the position of the ':', or false if none found
1724          */
1725         function findColonNoLinks($str, &$before, &$after) {
1726                 # I wonder if we should make this count all tags, not just <a>
1727                 # and <span>. That would prevent us from matching a ':' that
1728                 # comes in the middle of italics other such formatting....
1729                 # -- Wil
1730                 $fname = 'Parser::findColonNoLinks';
1731                 wfProfileIn( $fname );
1732                 $pos = 0;
1733                 do {
1734                         $colon = strpos($str, ':', $pos);
1735
1736                         if ($colon !== false) {
1737                                 $before = substr($str, 0, $colon);
1738                                 $after = substr($str, $colon + 1);
1739
1740                                 # Skip any ':' within <a> or <span> pairs
1741                                 $a = substr_count($before, '<a');
1742                                 $s = substr_count($before, '<span');
1743                                 $ca = substr_count($before, '</a>');
1744                                 $cs = substr_count($before, '</span>');
1745
1746                                 if ($a <= $ca and $s <= $cs) {
1747                                         # Tags are balanced before ':'; ok
1748                                         break;
1749                                 }
1750                                 $pos = $colon + 1;
1751                         }
1752                 } while ($colon !== false);
1753                 wfProfileOut( $fname );
1754                 return $colon;
1755         }
1756
1757         /**
1758          * Return value of a magic variable (like PAGENAME)
1759          *
1760          * @access private
1761          */
1762         function getVariableValue( $index ) {
1763                 global $wgContLang, $wgSitename, $wgServer, $wgArticle;
1764
1765                 /**
1766                  * Some of these require message or data lookups and can be
1767                  * expensive to check many times.
1768                  */
1769                 static $varCache = array();
1770                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1771
1772                 switch ( $index ) {
1773                         case MAG_CURRENTMONTH:
1774                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1775                         case MAG_CURRENTMONTHNAME:
1776                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1777                         case MAG_CURRENTMONTHNAMEGEN:
1778                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1779                         case MAG_CURRENTMONTHABBREV:
1780                                 return $varCache[$index] = $wgContLang->getMonthAbbreviation( date('n') );
1781                         case MAG_CURRENTDAY:
1782                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1783                         case MAG_PAGENAME:
1784                                 return $this->mTitle->getText();
1785                         case MAG_PAGENAMEE:
1786                                 return $this->mTitle->getPartialURL();
1787                         case MAG_REVISIONID:
1788                                 return $wgArticle->getRevIdFetched();
1789                         case MAG_NAMESPACE:
1790                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1791                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1792                         case MAG_CURRENTDAYNAME:
1793                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1794                         case MAG_CURRENTYEAR:
1795                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ), true );
1796                         case MAG_CURRENTTIME:
1797                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1798                         case MAG_CURRENTWEEK:
1799                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1800                         case MAG_CURRENTDOW:
1801                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1802                         case MAG_NUMBEROFARTICLES:
1803                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1804                         case MAG_SITENAME:
1805                                 return $wgSitename;
1806                         case MAG_SERVER:
1807                                 return $wgServer;
1808                         default:
1809                                 return NULL;
1810                 }
1811         }
1812
1813         /**
1814          * initialise the magic variables (like CURRENTMONTHNAME)
1815          *
1816          * @access private
1817          */
1818         function initialiseVariables() {
1819                 $fname = 'Parser::initialiseVariables';
1820                 wfProfileIn( $fname );
1821                 global $wgVariableIDs;
1822                 $this->mVariables = array();
1823                 foreach ( $wgVariableIDs as $id ) {
1824                         $mw =& MagicWord::get( $id );
1825                         $mw->addToArray( $this->mVariables, $id );
1826                 }
1827                 wfProfileOut( $fname );
1828         }
1829
1830         /**
1831          * Replace magic variables, templates, and template arguments
1832          * with the appropriate text. Templates are substituted recursively,
1833          * taking care to avoid infinite loops.
1834          *
1835          * Note that the substitution depends on value of $mOutputType:
1836          *  OT_WIKI: only {{subst:}} templates
1837          *  OT_MSG: only magic variables
1838          *  OT_HTML: all templates and magic variables
1839          *
1840          * @param string $tex The text to transform
1841          * @param array $args Key-value pairs representing template parameters to substitute
1842          * @access private
1843          */
1844         function replaceVariables( $text, $args = array() ) {
1845
1846                 # Prevent too big inclusions
1847                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1848                         return $text;
1849                 }
1850
1851                 $fname = 'Parser::replaceVariables';
1852                 wfProfileIn( $fname );
1853
1854                 $titleChars = Title::legalChars();
1855
1856                 # This function is called recursively. To keep track of arguments we need a stack:
1857                 array_push( $this->mArgStack, $args );
1858
1859                 # Variable substitution
1860                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1861
1862                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1863                         # Argument substitution
1864                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1865                 }
1866                 # Template substitution
1867                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1868                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1869
1870                 array_pop( $this->mArgStack );
1871
1872                 wfProfileOut( $fname );
1873                 return $text;
1874         }
1875
1876         /**
1877          * Replace magic variables
1878          * @access private
1879          */
1880         function variableSubstitution( $matches ) {
1881                 $fname = 'parser::variableSubstitution';
1882                 $varname = $matches[1];
1883                 wfProfileIn( $fname );
1884                 if ( !$this->mVariables ) {
1885                         $this->initialiseVariables();
1886                 }
1887                 $skip = false;
1888                 if ( $this->mOutputType == OT_WIKI ) {
1889                         # Do only magic variables prefixed by SUBST
1890                         $mwSubst =& MagicWord::get( MAG_SUBST );
1891                         if (!$mwSubst->matchStartAndRemove( $varname ))
1892                                 $skip = true;
1893                         # Note that if we don't substitute the variable below,
1894                         # we don't remove the {{subst:}} magic word, in case
1895                         # it is a template rather than a magic variable.
1896                 }
1897                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
1898                         $id = $this->mVariables[$varname];
1899                         $text = $this->getVariableValue( $id );
1900                         $this->mOutput->mContainsOldMagic = true;
1901                 } else {
1902                         $text = $matches[0];
1903                 }
1904                 wfProfileOut( $fname );
1905                 return $text;
1906         }
1907
1908         # Split template arguments
1909         function getTemplateArgs( $argsString ) {
1910                 if ( $argsString === '' ) {
1911                         return array();
1912                 }
1913
1914                 $args = explode( '|', substr( $argsString, 1 ) );
1915
1916                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1917                 # merged with the next arg because the '|' character between belongs
1918                 # to the link syntax and not the template parameter syntax.
1919                 $argc = count($args);
1920
1921                 for ( $i = 0; $i < $argc-1; $i++ ) {
1922                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1923                                 $args[$i] .= '|'.$args[$i+1];
1924                                 array_splice($args, $i+1, 1);
1925                                 $i--;
1926                                 $argc--;
1927                         }
1928                 }
1929
1930                 return $args;
1931         }
1932
1933         /**
1934          * Return the text of a template, after recursively
1935          * replacing any variables or templates within the template.
1936          *
1937          * @param array $matches The parts of the template
1938          *  $matches[1]: the title, i.e. the part before the |
1939          *  $matches[2]: the parameters (including a leading |), if  any
1940          * @return string the text of the template
1941          * @access private
1942          */
1943         function braceSubstitution( $matches ) {
1944                 global $wgLinkCache, $wgContLang;
1945                 $fname = 'Parser::braceSubstitution';
1946                 wfProfileIn( $fname );
1947
1948                 $found = false;
1949                 $nowiki = false;
1950                 $noparse = false;
1951
1952                 $title = NULL;
1953
1954                 # Need to know if the template comes at the start of a line,
1955                 # to treat the beginning of the template like the beginning
1956                 # of a line for tables and block-level elements.
1957                 $linestart = $matches[1];
1958
1959                 # $part1 is the bit before the first |, and must contain only title characters
1960                 # $args is a list of arguments, starting from index 0, not including $part1
1961
1962                 $part1 = $matches[2];
1963                 # If the third subpattern matched anything, it will start with |
1964
1965                 $args = $this->getTemplateArgs($matches[3]);
1966                 $argc = count( $args );
1967
1968                 # Don't parse {{{}}} because that's only for template arguments
1969                 if ( $linestart === '{' ) {
1970                         $text = $matches[0];
1971                         $found = true;
1972                         $noparse = true;
1973                 }
1974
1975                 # SUBST
1976                 if ( !$found ) {
1977                         $mwSubst =& MagicWord::get( MAG_SUBST );
1978                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1979                                 # One of two possibilities is true:
1980                                 # 1) Found SUBST but not in the PST phase
1981                                 # 2) Didn't find SUBST and in the PST phase
1982                                 # In either case, return without further processing
1983                                 $text = $matches[0];
1984                                 $found = true;
1985                                 $noparse = true;
1986                         }
1987                 }
1988
1989                 # MSG, MSGNW and INT
1990                 if ( !$found ) {
1991                         # Check for MSGNW:
1992                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1993                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1994                                 $nowiki = true;
1995                         } else {
1996                                 # Remove obsolete MSG:
1997                                 $mwMsg =& MagicWord::get( MAG_MSG );
1998                                 $mwMsg->matchStartAndRemove( $part1 );
1999                         }
2000
2001                         # Check if it is an internal message
2002                         $mwInt =& MagicWord::get( MAG_INT );
2003                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
2004                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
2005                                         $text = $linestart . wfMsgReal( $part1, $args, true );
2006                                         $found = true;
2007                                 }
2008                         }
2009                 }
2010
2011                 # NS
2012                 if ( !$found ) {
2013                         # Check for NS: (namespace expansion)
2014                         $mwNs = MagicWord::get( MAG_NS );
2015                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
2016                                 if ( intval( $part1 ) ) {
2017                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
2018                                         $found = true;
2019                                 } else {
2020                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
2021                                         if ( !is_null( $index ) ) {
2022                                                 $text = $linestart . $wgContLang->getNsText( $index );
2023                                                 $found = true;
2024                                         }
2025                                 }
2026                         }
2027                 }
2028
2029                 # LOCALURL and LOCALURLE
2030                 if ( !$found ) {
2031                         $mwLocal = MagicWord::get( MAG_LOCALURL );
2032                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
2033
2034                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2035                                 $func = 'getLocalURL';
2036                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2037                                 $func = 'escapeLocalURL';
2038                         } else {
2039                                 $func = '';
2040                         }
2041
2042                         if ( $func !== '' ) {
2043                                 $title = Title::newFromText( $part1 );
2044                                 if ( !is_null( $title ) ) {
2045                                         if ( $argc > 0 ) {
2046                                                 $text = $linestart . $title->$func( $args[0] );
2047                                         } else {
2048                                                 $text = $linestart . $title->$func();
2049                                         }
2050                                         $found = true;
2051                                 }
2052                         }
2053                 }
2054
2055                 # GRAMMAR
2056                 if ( !$found && $argc == 1 ) {
2057                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2058                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2059                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2060                                 $found = true;
2061                         }
2062                 }
2063
2064                 # Template table test
2065
2066                 # Did we encounter this template already? If yes, it is in the cache
2067                 # and we need to check for loops.
2068                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2069                         $found = true;
2070
2071                         # Infinite loop test
2072                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2073                                 $noparse = true;
2074                                 $found = true;
2075                                 $text = $linestart .
2076                                         "\{\{$part1}}" .
2077                                         '<!-- WARNING: template loop detected -->';
2078                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2079                         } else {
2080                                 # set $text to cached message.
2081                                 $text = $linestart . $this->mTemplates[$part1];
2082                         }
2083                 }
2084
2085                 # Load from database
2086                 $itcamefromthedatabase = false;
2087                 $lastPathLevel = $this->mTemplatePath;
2088                 if ( !$found ) {
2089                         $ns = NS_TEMPLATE;
2090                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2091                         if ($subpage !== '') {
2092                                 $ns = $this->mTitle->getNamespace();
2093                         }
2094                         $title = Title::newFromText( $part1, $ns );
2095                         if ( !is_null( $title ) && !$title->isExternal() ) {
2096                                 # Check for excessive inclusion
2097                                 $dbk = $title->getPrefixedDBkey();
2098                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2099                                         # This should never be reached.
2100                                         $article = new Article( $title );
2101                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2102                                         if ( $articleContent !== false ) {
2103                                                 $found = true;
2104                                                 $text = $linestart . $articleContent;
2105                                                 $itcamefromthedatabase = true;
2106                                         }
2107                                 }
2108
2109                                 # If the title is valid but undisplayable, make a link to it
2110                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2111                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2112                                         $found = true;
2113                                 }
2114
2115                                 # Template cache array insertion
2116                                 if( $found ) {
2117                                         $this->mTemplates[$part1] = $text;
2118                                 }
2119                         }
2120                 }
2121
2122                 # Recursive parsing, escaping and link table handling
2123                 # Only for HTML output
2124                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2125                         $text = wfEscapeWikiText( $text );
2126                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2127                         # Clean up argument array
2128                         $assocArgs = array();
2129                         $index = 1;
2130                         foreach( $args as $arg ) {
2131                                 $eqpos = strpos( $arg, '=' );
2132                                 if ( $eqpos === false ) {
2133                                         $assocArgs[$index++] = $arg;
2134                                 } else {
2135                                         $name = trim( substr( $arg, 0, $eqpos ) );
2136                                         $value = trim( substr( $arg, $eqpos+1 ) );
2137                                         if ( $value === false ) {
2138                                                 $value = '';
2139                                         }
2140                                         if ( $name !== false ) {
2141                                                 $assocArgs[$name] = $value;
2142                                         }
2143                                 }
2144                         }
2145
2146                         # Add a new element to the templace recursion path
2147                         $this->mTemplatePath[$part1] = 1;
2148
2149                         if( $this->mOutputType == OT_HTML ) {
2150                                 $text = $this->strip( $text, $this->mStripState );
2151                                 $text = Sanitizer::removeHTMLtags( $text );
2152                         }
2153                         $text = $this->replaceVariables( $text, $assocArgs );
2154
2155                         # Resume the link cache and register the inclusion as a link
2156                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2157                                 $wgLinkCache->addLinkObj( $title );
2158                         }
2159
2160                         # If the template begins with a table or block-level
2161                         # element, it should be treated as beginning a new line.
2162                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2163                                 $text = "\n" . $text;
2164                         }
2165                 }
2166                 # Prune lower levels off the recursion check path
2167                 $this->mTemplatePath = $lastPathLevel;
2168
2169                 if ( !$found ) {
2170                         wfProfileOut( $fname );
2171                         return $matches[0];
2172                 } else {
2173                         # replace ==section headers==
2174                         # XXX this needs to go away once we have a better parser.
2175                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2176                                 if( !is_null( $title ) )
2177                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2178                                 else
2179                                         $encodedname = base64_encode("");
2180                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2181                                         PREG_SPLIT_DELIM_CAPTURE);
2182                                 $text = '';
2183                                 $nsec = 0;
2184                                 for( $i = 0; $i < count($m); $i += 2 ) {
2185                                         $text .= $m[$i];
2186                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2187                                         $hl = $m[$i + 1];
2188                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2189                                                 $text .= $hl;
2190                                                 continue;
2191                                         }
2192                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2193                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2194                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2195
2196                                         $nsec++;
2197                                 }
2198                         }
2199                 }
2200                 # Prune lower levels off the recursion check path
2201                 $this->mTemplatePath = $lastPathLevel;
2202
2203                 if ( !$found ) {
2204                         wfProfileOut( $fname );
2205                         return $matches[0];
2206                 } else {
2207                         wfProfileOut( $fname );
2208                         return $text;
2209                 }
2210         }
2211
2212         /**
2213          * Triple brace replacement -- used for template arguments
2214          * @access private
2215          */
2216         function argSubstitution( $matches ) {
2217                 $arg = trim( $matches[1] );
2218                 $text = $matches[0];
2219                 $inputArgs = end( $this->mArgStack );
2220
2221                 if ( array_key_exists( $arg, $inputArgs ) ) {
2222                         $text = $inputArgs[$arg];
2223                 }
2224
2225                 return $text;
2226         }
2227
2228         /**
2229          * Returns true if the function is allowed to include this entity
2230          * @access private
2231          */
2232         function incrementIncludeCount( $dbk ) {
2233                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2234                         $this->mIncludeCount[$dbk] = 0;
2235                 }
2236                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2237                         return true;
2238                 } else {
2239                         return false;
2240                 }
2241         }
2242
2243         /**
2244          * This function accomplishes several tasks:
2245          * 1) Auto-number headings if that option is enabled
2246          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2247          * 3) Add a Table of contents on the top for users who have enabled the option
2248          * 4) Auto-anchor headings
2249          *
2250          * It loops through all headlines, collects the necessary data, then splits up the
2251          * string and re-inserts the newly formatted headlines.
2252          *
2253          * @param string $text
2254          * @param boolean $isMain
2255          * @access private
2256          */
2257         function formatHeadings( $text, $isMain=true ) {
2258                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2259
2260                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2261                 $doShowToc = true;
2262                 $forceTocHere = false;
2263                 if( !$this->mTitle->userCanEdit() ) {
2264                         $showEditLink = 0;
2265                 } else {
2266                         $showEditLink = $this->mOptions->getEditSection();
2267                 }
2268
2269                 # Inhibit editsection links if requested in the page
2270                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2271                 if( $esw->matchAndRemove( $text ) ) {
2272                         $showEditLink = 0;
2273                 }
2274                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2275                 # do not add TOC
2276                 $mw =& MagicWord::get( MAG_NOTOC );
2277                 if( $mw->matchAndRemove( $text ) ) {
2278                         $doShowToc = false;
2279                 }
2280
2281                 # Get all headlines for numbering them and adding funky stuff like [edit]
2282                 # links - this is for later, but we need the number of headlines right now
2283                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2284
2285                 # if there are fewer than 4 headlines in the article, do not show TOC
2286                 if( $numMatches < 4 ) {
2287                         $doShowToc = false;
2288                 }
2289
2290                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2291                 # override above conditions and always show TOC at that place
2292
2293                 $mw =& MagicWord::get( MAG_TOC );
2294                 if($mw->match( $text ) ) {
2295                         $doShowToc = true;
2296                         $forceTocHere = true;
2297                 } else {
2298                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2299                         # override above conditions and always show TOC above first header
2300                         $mw =& MagicWord::get( MAG_FORCETOC );
2301                         if ($mw->matchAndRemove( $text ) ) {
2302                                 $doShowToc = true;
2303                         }
2304                 }
2305
2306                 # Never ever show TOC if no headers
2307                 if( $numMatches < 1 ) {
2308                         $doShowToc = false;
2309                 }
2310
2311                 # We need this to perform operations on the HTML
2312                 $sk =& $this->mOptions->getSkin();
2313
2314                 # headline counter
2315                 $headlineCount = 0;
2316                 $sectionCount = 0; # headlineCount excluding template sections
2317
2318                 # Ugh .. the TOC should have neat indentation levels which can be
2319                 # passed to the skin functions. These are determined here
2320                 $toc = '';
2321                 $full = '';
2322                 $head = array();
2323                 $sublevelCount = array();
2324                 $levelCount = array();
2325                 $toclevel = 0;
2326                 $level = 0;
2327                 $prevlevel = 0;
2328                 $toclevel = 0;
2329                 $prevtoclevel = 0;
2330
2331                 foreach( $matches[3] as $headline ) {
2332                         $istemplate = 0;
2333                         $templatetitle = '';
2334                         $templatesection = 0;
2335                         $numbering = '';
2336
2337                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2338                                 $istemplate = 1;
2339                                 $templatetitle = base64_decode($mat[1]);
2340                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2341                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2342                         }
2343
2344                         if( $toclevel ) {
2345                                 $prevlevel = $level;
2346                                 $prevtoclevel = $toclevel;
2347                         }
2348                         $level = $matches[1][$headlineCount];
2349
2350                         if( $doNumberHeadings || $doShowToc ) {
2351
2352                                 if ( $level > $prevlevel ) {
2353                                         # Increase TOC level
2354                                         $toclevel++;
2355                                         $sublevelCount[$toclevel] = 0;
2356                                         $toc .= $sk->tocIndent();
2357                                 }
2358                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2359                                         # Decrease TOC level, find level to jump to
2360
2361                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2362                                                 # Can only go down to level 1
2363                                                 $toclevel = 1;
2364                                         } else {
2365                                                 for ($i = $toclevel; $i > 0; $i--) {
2366                                                         if ( $levelCount[$i] == $level ) {
2367                                                                 # Found last matching level
2368                                                                 $toclevel = $i;
2369                                                                 break;
2370                                                         }
2371                                                         elseif ( $levelCount[$i] < $level ) {
2372                                                                 # Found first matching level below current level
2373                                                                 $toclevel = $i + 1;
2374                                                                 break;
2375                                                         }
2376                                                 }
2377                                         }
2378
2379                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2380                                 }
2381                                 else {
2382                                         # No change in level, end TOC line
2383                                         $toc .= $sk->tocLineEnd();
2384                                 }
2385
2386                                 $levelCount[$toclevel] = $level;
2387
2388                                 # count number of headlines for each level
2389                                 @$sublevelCount[$toclevel]++;
2390                                 $dot = 0;
2391                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2392                                         if( !empty( $sublevelCount[$i] ) ) {
2393                                                 if( $dot ) {
2394                                                         $numbering .= '.';
2395                                                 }
2396                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2397                                                 $dot = 1;
2398                                         }
2399                                 }
2400                         }
2401
2402                         # The canonized header is a version of the header text safe to use for links
2403                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2404                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2405                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2406
2407                         # Remove link placeholders by the link text.
2408                         #     <!--LINK number-->
2409                         # turns into
2410                         #     link text with suffix
2411                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2412                                                             "\$wgLinkHolders['texts'][\$1]",
2413                                                             $canonized_headline );
2414                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2415                                                             "\$wgInterwikiLinkHolders[\$1]",
2416                                                             $canonized_headline );
2417
2418                         # strip out HTML
2419                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2420                         $tocline = trim( $canonized_headline );
2421                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2422                         $replacearray = array(
2423                                 '%3A' => ':',
2424                                 '%' => '.'
2425                         );
2426                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2427                         $refers[$headlineCount] = $canonized_headline;
2428
2429                         # count how many in assoc. array so we can track dupes in anchors
2430                         @$refers[$canonized_headline]++;
2431                         $refcount[$headlineCount]=$refers[$canonized_headline];
2432
2433                         # Don't number the heading if it is the only one (looks silly)
2434                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2435                                 # the two are different if the line contains a link
2436                                 $headline=$numbering . ' ' . $headline;
2437                         }
2438
2439                         # Create the anchor for linking from the TOC to the section
2440                         $anchor = $canonized_headline;
2441                         if($refcount[$headlineCount] > 1 ) {
2442                                 $anchor .= '_' . $refcount[$headlineCount];
2443                         }
2444                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2445                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2446                         }
2447                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2448                                 if ( empty( $head[$headlineCount] ) ) {
2449                                         $head[$headlineCount] = '';
2450                                 }
2451                                 if( $istemplate )
2452                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2453                                 else
2454                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2455                         }
2456
2457                         # give headline the correct <h#> tag
2458                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2459
2460                         $headlineCount++;
2461                         if( !$istemplate )
2462                                 $sectionCount++;
2463                 }
2464
2465                 if( $doShowToc ) {
2466                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2467                         $toc = $sk->tocList( $toc );
2468                 }
2469
2470                 # split up and insert constructed headlines
2471
2472                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2473                 $i = 0;
2474
2475                 foreach( $blocks as $block ) {
2476                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2477                                 # This is the [edit] link that appears for the top block of text when
2478                                 # section editing is enabled
2479
2480                                 # Disabled because it broke block formatting
2481                                 # For example, a bullet point in the top line
2482                                 # $full .= $sk->editSectionLink(0);
2483                         }
2484                         $full .= $block;
2485                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2486                         # Top anchor now in skin
2487                                 $full = $full.$toc;
2488                         }
2489
2490                         if( !empty( $head[$i] ) ) {
2491                                 $full .= $head[$i];
2492                         }
2493                         $i++;
2494                 }
2495                 if($forceTocHere) {
2496                         $mw =& MagicWord::get( MAG_TOC );
2497                         return $mw->replace( $toc, $full );
2498                 } else {
2499                         return $full;
2500                 }
2501         }
2502
2503         /**
2504          * Return an HTML link for the "ISBN 123456" text
2505          * @access private
2506          */
2507         function magicISBN( $text ) {
2508                 $fname = 'Parser::magicISBN';
2509                 wfProfileIn( $fname );
2510
2511                 $a = split( 'ISBN ', ' '.$text );
2512                 if ( count ( $a ) < 2 ) {
2513                         wfProfileOut( $fname );
2514                         return $text;
2515                 }
2516                 $text = substr( array_shift( $a ), 1);
2517                 $valid = '0123456789-Xx';
2518
2519                 foreach ( $a as $x ) {
2520                         $isbn = $blank = '' ;
2521                         while ( ' ' == $x{0} ) {
2522                                 $blank .= ' ';
2523                                 $x = substr( $x, 1 );
2524                         }
2525                         if ( $x == '' ) { # blank isbn
2526                                 $text .= "ISBN $blank";
2527                                 continue;
2528                         }
2529                         while ( strstr( $valid, $x{0} ) != false ) {
2530                                 $isbn .= $x{0};
2531                                 $x = substr( $x, 1 );
2532                         }
2533                         $num = str_replace( '-', '', $isbn );
2534                         $num = str_replace( ' ', '', $num );
2535                         $num = str_replace( 'x', 'X', $num );
2536
2537                         if ( '' == $num ) {
2538                                 $text .= "ISBN $blank$x";
2539                         } else {
2540                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2541                                 $text .= '<a href="' .
2542                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2543                                         "\" class=\"internal\">ISBN $isbn</a>";
2544                                 $text .= $x;
2545                         }
2546                 }
2547                 wfProfileOut( $fname );
2548                 return $text;
2549         }
2550
2551         /**
2552          * Return an HTML link for the "RFC 1234" text
2553          *
2554          * @access private
2555          * @param string $text     Text to be processed
2556          * @param string $keyword  Magic keyword to use (default RFC)
2557          * @param string $urlmsg   Interface message to use (default rfcurl)
2558          * @return string
2559          */
2560         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2561
2562                 $valid = '0123456789';
2563                 $internal = false;
2564
2565                 $a = split( $keyword, ' '.$text );
2566                 if ( count ( $a ) < 2 ) {
2567                         return $text;
2568                 }
2569                 $text = substr( array_shift( $a ), 1);
2570
2571                 /* Check if keyword is preceed by [[.
2572                  * This test is made here cause of the array_shift above
2573                  * that prevent the test to be done in the foreach.
2574                  */
2575                 if ( substr( $text, -2 ) == '[[' ) {
2576                         $internal = true;
2577                 }
2578
2579                 foreach ( $a as $x ) {
2580                         /* token might be empty if we have RFC RFC 1234 */
2581                         if ( $x=='' ) {
2582                                 $text.=$keyword;
2583                                 continue;
2584                                 }
2585
2586                         $id = $blank = '' ;
2587
2588                         /** remove and save whitespaces in $blank */
2589                         while ( $x{0} == ' ' ) {
2590                                 $blank .= ' ';
2591                                 $x = substr( $x, 1 );
2592                         }
2593
2594                         /** remove and save the rfc number in $id */
2595                         while ( strstr( $valid, $x{0} ) != false ) {
2596                                 $id .= $x{0};
2597                                 $x = substr( $x, 1 );
2598                         }
2599
2600                         if ( $id == '' ) {
2601                                 /* call back stripped spaces*/
2602                                 $text .= $keyword.$blank.$x;
2603                         } elseif( $internal ) {
2604                                 /* normal link */
2605                                 $text .= $keyword.$id.$x;
2606                         } else {
2607                                 /* build the external link*/
2608                                 $url = wfMsg( $urlmsg, $id);
2609                                 $sk =& $this->mOptions->getSkin();
2610                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2611                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2612                         }
2613
2614                         /* Check if the next RFC keyword is preceed by [[ */
2615                         $internal = ( substr($x,-2) == '[[' );
2616                 }
2617                 return $text;
2618         }
2619
2620         /**
2621          * Transform wiki markup when saving a page by doing \r\n -> \n
2622          * conversion, substitting signatures, {{subst:}} templates, etc.
2623          *
2624          * @param string $text the text to transform
2625          * @param Title &$title the Title object for the current article
2626          * @param User &$user the User object describing the current user
2627          * @param ParserOptions $options parsing options
2628          * @param bool $clearState whether to clear the parser state first
2629          * @return string the altered wiki markup
2630          * @access public
2631          */
2632         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2633                 $this->mOptions = $options;
2634                 $this->mTitle =& $title;
2635                 $this->mOutputType = OT_WIKI;
2636
2637                 if ( $clearState ) {
2638                         $this->clearState();
2639                 }
2640
2641                 $stripState = false;
2642                 $pairs = array(
2643                         "\r\n" => "\n",
2644                 );
2645                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2646                 $text = $this->strip( $text, $stripState, false );
2647                 $text = $this->pstPass2( $text, $user );
2648                 $text = $this->unstrip( $text, $stripState );
2649                 $text = $this->unstripNoWiki( $text, $stripState );
2650                 return $text;
2651         }
2652
2653         /**
2654          * Pre-save transform helper function
2655          * @access private
2656          */
2657         function pstPass2( $text, &$user ) {
2658                 global $wgContLang, $wgLocaltimezone;
2659
2660                 # Variable replacement
2661                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2662                 $text = $this->replaceVariables( $text );
2663
2664                 # Signatures
2665                 #
2666                 $n = $user->getName();
2667                 $k = $user->getOption( 'nickname' );
2668                 if ( '' == $k ) { $k = $n; }
2669                 if ( isset( $wgLocaltimezone ) ) {
2670                         $oldtz = getenv( 'TZ' );
2671                         putenv( 'TZ='.$wgLocaltimezone );
2672                 }
2673
2674                 /* Note: This is the timestamp saved as hardcoded wikitext to
2675                  * the database, we use $wgContLang here in order to give
2676                  * everyone the same signiture and use the default one rather
2677                  * than the one selected in each users preferences.
2678                  */
2679                 $d = $wgContLang->timeanddate( wfTimestampNow(), false, false) .
2680                   ' (' . date( 'T' ) . ')';
2681                 if ( isset( $wgLocaltimezone ) ) {
2682                         putenv( 'TZ='.$oldtz );
2683                 }
2684
2685                 if( $user->getOption( 'fancysig' ) ) {
2686                         $sigText = $k;
2687                 } else {
2688                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2689                 }
2690                 $text = preg_replace( '/~~~~~/', $d, $text );
2691                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2692                 $text = preg_replace( '/~~~/', $sigText, $text );
2693
2694                 # Context links: [[|name]] and [[name (context)|]]
2695                 #
2696                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2697                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2698                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2699                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2700
2701                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2702                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2703                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2704                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2705                 $context = '';
2706                 $t = $this->mTitle->getText();
2707                 if ( preg_match( $conpat, $t, $m ) ) {
2708                         $context = $m[2];
2709                 }
2710                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2711                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2712                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2713
2714                 if ( '' == $context ) {
2715                         $text = preg_replace( $p2, '[[\\1]]', $text );
2716                 } else {
2717                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2718                 }
2719
2720                 # Trim trailing whitespace
2721                 # MAG_END (__END__) tag allows for trailing
2722                 # whitespace to be deliberately included
2723                 $text = rtrim( $text );
2724                 $mw =& MagicWord::get( MAG_END );
2725                 $mw->matchAndRemove( $text );
2726
2727                 return $text;
2728         }
2729
2730         /**
2731          * Set up some variables which are usually set up in parse()
2732          * so that an external function can call some class members with confidence
2733          * @access public
2734          */
2735         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2736                 $this->mTitle =& $title;
2737                 $this->mOptions = $options;
2738                 $this->mOutputType = $outputType;
2739                 if ( $clearState ) {
2740                         $this->clearState();
2741                 }
2742         }
2743
2744         /**
2745          * Transform a MediaWiki message by replacing magic variables.
2746          *
2747          * @param string $text the text to transform
2748          * @param ParserOptions $options  options
2749          * @return string the text with variables substituted
2750          * @access public
2751          */
2752         function transformMsg( $text, $options ) {
2753                 global $wgTitle;
2754                 static $executing = false;
2755
2756                 # Guard against infinite recursion
2757                 if ( $executing ) {
2758                         return $text;
2759                 }
2760                 $executing = true;
2761
2762                 $this->mTitle = $wgTitle;
2763                 $this->mOptions = $options;
2764                 $this->mOutputType = OT_MSG;
2765                 $this->clearState();
2766                 $text = $this->replaceVariables( $text );
2767
2768                 $executing = false;
2769                 return $text;
2770         }
2771
2772         /**
2773          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2774          * Callback will be called with the text within
2775          * Transform and return the text within
2776          * @access public
2777          */
2778         function setHook( $tag, $callback ) {
2779                 $oldVal = @$this->mTagHooks[$tag];
2780                 $this->mTagHooks[$tag] = $callback;
2781                 return $oldVal;
2782         }
2783
2784         /**
2785          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2786          * Placeholders created in Skin::makeLinkObj()
2787          * Returns an array of links found, indexed by PDBK:
2788          *  0 - broken
2789          *  1 - normal link
2790          *  2 - stub
2791          * $options is a bit field, RLH_FOR_UPDATE to select for update
2792          */
2793         function replaceLinkHolders( &$text, $options = 0 ) {
2794                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2795                 global $wgInterwikiLinkHolders;
2796                 global $outputReplace;
2797
2798                 if ( $wgUseOldExistenceCheck ) {
2799                         return array();
2800                 }
2801
2802                 $fname = 'Parser::replaceLinkHolders';
2803                 wfProfileIn( $fname );
2804
2805                 $pdbks = array();
2806                 $colours = array();
2807
2808                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2809                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2810                         wfProfileIn( $fname.'-check' );
2811                         $dbr =& wfGetDB( DB_SLAVE );
2812                         $page = $dbr->tableName( 'page' );
2813                         $sk = $wgUser->getSkin();
2814                         $threshold = $wgUser->getOption('stubthreshold');
2815
2816                         # Sort by namespace
2817                         asort( $wgLinkHolders['namespaces'] );
2818
2819                         # Generate query
2820                         $query = false;
2821                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2822                                 # Make title object
2823                                 $title = $wgLinkHolders['titles'][$key];
2824
2825                                 # Skip invalid entries.
2826                                 # Result will be ugly, but prevents crash.
2827                                 if ( is_null( $title ) ) {
2828                                         continue;
2829                                 }
2830                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2831
2832                                 # Check if it's in the link cache already
2833                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2834                                         $colours[$pdbk] = 1;
2835                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2836                                         $colours[$pdbk] = 0;
2837                                 } else {
2838                                         # Not in the link cache, add it to the query
2839                                         if ( !isset( $current ) ) {
2840                                                 $current = $val;
2841                                                 $query =  "SELECT page_id, page_namespace, page_title";
2842                                                 if ( $threshold > 0 ) {
2843                                                         $query .= ', page_len, page_is_redirect';
2844                                                 }
2845                                                 $query .= " FROM $page WHERE (page_namespace=$val AND page_title IN(";
2846                                         } elseif ( $current != $val ) {
2847                                                 $current = $val;
2848                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
2849                                         } else {
2850                                                 $query .= ', ';
2851                                         }
2852
2853                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2854                                 }
2855                         }
2856                         if ( $query ) {
2857                                 $query .= '))';
2858                                 if ( $options & RLH_FOR_UPDATE ) {
2859                                         $query .= ' FOR UPDATE';
2860                                 }
2861
2862                                 $res = $dbr->query( $query, $fname );
2863
2864                                 # Fetch data and form into an associative array
2865                                 # non-existent = broken
2866                                 # 1 = known
2867                                 # 2 = stub
2868                                 while ( $s = $dbr->fetchObject($res) ) {
2869                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
2870                                         $pdbk = $title->getPrefixedDBkey();
2871                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
2872
2873                                         if ( $threshold >  0 ) {
2874                                                 $size = $s->page_len;
2875                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
2876                                                         $colours[$pdbk] = 1;
2877                                                 } else {
2878                                                         $colours[$pdbk] = 2;
2879                                                 }
2880                                         } else {
2881                                                 $colours[$pdbk] = 1;
2882                                         }
2883                                 }
2884                         }
2885                         wfProfileOut( $fname.'-check' );
2886
2887                         # Construct search and replace arrays
2888                         wfProfileIn( $fname.'-construct' );
2889                         $outputReplace = array();
2890                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
2891                                 $pdbk = $pdbks[$key];
2892                                 $searchkey = '<!--LINK '.$key.'-->';
2893                                 $title = $wgLinkHolders['titles'][$key];
2894                                 if ( empty( $colours[$pdbk] ) ) {
2895                                         $wgLinkCache->addBadLink( $pdbk );
2896                                         $colours[$pdbk] = 0;
2897                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
2898                                                                         $wgLinkHolders['texts'][$key],
2899                                                                         $wgLinkHolders['queries'][$key] );
2900                                 } elseif ( $colours[$pdbk] == 1 ) {
2901                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
2902                                                                         $wgLinkHolders['texts'][$key],
2903                                                                         $wgLinkHolders['queries'][$key] );
2904                                 } elseif ( $colours[$pdbk] == 2 ) {
2905                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
2906                                                                         $wgLinkHolders['texts'][$key],
2907                                                                         $wgLinkHolders['queries'][$key] );
2908                                 }
2909                         }
2910                         wfProfileOut( $fname.'-construct' );
2911
2912                         # Do the thing
2913                         wfProfileIn( $fname.'-replace' );
2914
2915                         $text = preg_replace_callback(
2916                                 '/(<!--LINK .*?-->)/',
2917                                 "outputReplaceMatches",
2918                                 $text);
2919                         wfProfileOut( $fname.'-replace' );
2920                 }
2921
2922                 if ( !empty( $wgInterwikiLinkHolders ) ) {
2923                         wfProfileIn( $fname.'-interwiki' );
2924                         $outputReplace = $wgInterwikiLinkHolders;
2925                         $text = preg_replace_callback(
2926                                 '/<!--IWLINK (.*?)-->/',
2927                                 "outputReplaceMatches",
2928                                 $text );
2929                         wfProfileOut( $fname.'-interwiki' );
2930                 }
2931
2932                 # Clear link holders, no need to fetch these ones again in a subsequent invocation
2933                 $wgLinkHolders = array(
2934                         'namespaces' => array(),
2935                         'dbkeys' => array(),
2936                         'queries' => array(),
2937                         'texts' => array(),
2938                         'titles' => array()
2939                 );
2940                 $wgInterwikiLinkHolders = array();
2941
2942
2943                 wfProfileOut( $fname );
2944                 return $colours;
2945         }
2946
2947         /**
2948          * Renders an image gallery from a text with one line per image.
2949          * text labels may be given by using |-style alternative text. E.g.
2950          *   Image:one.jpg|The number "1"
2951          *   Image:tree.jpg|A tree
2952          * given as text will return the HTML of a gallery with two images,
2953          * labeled 'The number "1"' and
2954          * 'A tree'.
2955          */
2956         function renderImageGallery( $text ) {
2957                 # Setup the parser
2958                 global $wgUser, $wgParser, $wgTitle;
2959                 $parserOptions = ParserOptions::newFromUser( $wgUser );
2960
2961                 global $wgLinkCache;
2962                 $ig = new ImageGallery();
2963                 $ig->setShowBytes( false );
2964                 $ig->setShowFilename( false );
2965                 $lines = explode( "\n", $text );
2966
2967                 foreach ( $lines as $line ) {
2968                         # match lines like these:
2969                         # Image:someimage.jpg|This is some image
2970                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
2971                         # Skip empty lines
2972                         if ( count( $matches ) == 0 ) {
2973                                 continue;
2974                         }
2975                         $nt = Title::newFromURL( $matches[1] );
2976                         if( is_null( $nt ) ) {
2977                                 # Bogus title. Ignore these so we don't bomb out later.
2978                                 continue;
2979                         }
2980                         if ( isset( $matches[3] ) ) {
2981                                 $label = $matches[3];
2982                         } else {
2983                                 $label = '';
2984                         }
2985
2986                         $html = $wgParser->parse( $label , $wgTitle, $parserOptions );
2987                         $html = $html->mText;
2988
2989                         $ig->add( new Image( $nt ), $html );
2990                         $wgLinkCache->addImageLinkObj( $nt );
2991                 }
2992                 return $ig->toHTML();
2993         }
2994 }
2995
2996 /**
2997  * @todo document
2998  * @package MediaWiki
2999  */
3000 class ParserOutput
3001 {
3002         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3003         var $mCacheTime; # Used in ParserCache
3004         var $mVersion;   # Compatibility check
3005         var $mTitleText; # title text of the chosen language variant
3006
3007         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3008                 $containsOldMagic = false, $titletext = '' )
3009         {
3010                 $this->mText = $text;
3011                 $this->mLanguageLinks = $languageLinks;
3012                 $this->mCategoryLinks = $categoryLinks;
3013                 $this->mContainsOldMagic = $containsOldMagic;
3014                 $this->mCacheTime = '';
3015                 $this->mVersion = MW_PARSER_VERSION;
3016                 $this->mTitleText = $titletext;
3017         }
3018
3019         function getText()                   { return $this->mText; }
3020         function getLanguageLinks()          { return $this->mLanguageLinks; }
3021         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
3022         function getCacheTime()              { return $this->mCacheTime; }
3023         function getTitleText()              { return $this->mTitleText; }
3024         function containsOldMagic()          { return $this->mContainsOldMagic; }
3025         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3026         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3027         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3028         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3029         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3030         function setTitleText( $t )          { return wfSetVar ($this->mTitleText, $t); }
3031
3032         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3033
3034         function merge( $other ) {
3035                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3036                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3037                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3038         }
3039
3040         /**
3041          * Return true if this cached output object predates the global or
3042          * per-article cache invalidation timestamps, or if it comes from
3043          * an incompatible older version.
3044          *
3045          * @param string $touched the affected article's last touched timestamp
3046          * @return bool
3047          * @access public
3048          */
3049         function expired( $touched ) {
3050                 global $wgCacheEpoch;
3051                 return $this->getCacheTime() <= $touched ||
3052                        $this->getCacheTime() <= $wgCacheEpoch ||
3053                        !isset( $this->mVersion ) ||
3054                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3055         }
3056 }
3057
3058 /**
3059  * Set options of the Parser
3060  * @todo document
3061  * @package MediaWiki
3062  */
3063 class ParserOptions
3064 {
3065         # All variables are private
3066         var $mUseTeX;                    # Use texvc to expand <math> tags
3067         var $mUseDynamicDates;           # Use DateFormatter to format dates
3068         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3069         var $mAllowExternalImages;       # Allow external images inline
3070         var $mSkin;                      # Reference to the preferred skin
3071         var $mDateFormat;                # Date format index
3072         var $mEditSection;               # Create "edit section" links
3073         var $mNumberHeadings;            # Automatically number headings
3074
3075         function getUseTeX()                        { return $this->mUseTeX; }
3076         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3077         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3078         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3079         function getSkin()                          { return $this->mSkin; }
3080         function getDateFormat()                    { return $this->mDateFormat; }
3081         function getEditSection()                   { return $this->mEditSection; }
3082         function getNumberHeadings()                { return $this->mNumberHeadings; }
3083
3084         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3085         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3086         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3087         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3088         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3089         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3090         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3091
3092         function setSkin( &$x ) { $this->mSkin =& $x; }
3093
3094         /**
3095          * Get parser options
3096          * @static
3097          */
3098         function newFromUser( &$user ) {
3099                 $popts = new ParserOptions;
3100                 $popts->initialiseFromUser( $user );
3101                 return $popts;
3102         }
3103
3104         /** Get user options */
3105         function initialiseFromUser( &$userInput ) {
3106                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3107                 $fname = 'ParserOptions::initialiseFromUser';
3108                 wfProfileIn( $fname );
3109                 if ( !$userInput ) {
3110                         $user = new User;
3111                         $user->setLoaded( true );
3112                 } else {
3113                         $user =& $userInput;
3114                 }
3115
3116                 $this->mUseTeX = $wgUseTeX;
3117                 $this->mUseDynamicDates = $wgUseDynamicDates;
3118                 $this->mInterwikiMagic = $wgInterwikiMagic;
3119                 $this->mAllowExternalImages = $wgAllowExternalImages;
3120                 wfProfileIn( $fname.'-skin' );
3121                 $this->mSkin =& $user->getSkin();
3122                 wfProfileOut( $fname.'-skin' );
3123                 $this->mDateFormat = $user->getOption( 'date' );
3124                 $this->mEditSection = $user->getOption( 'editsection' );
3125                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3126                 wfProfileOut( $fname );
3127         }
3128 }
3129
3130 /**
3131  * Callback function used by Parser::replaceLinkHolders()
3132  * to substitute link placeholders.
3133  */
3134 function &outputReplaceMatches( $matches ) {
3135         global $outputReplace;
3136         return $outputReplace[$matches[1]];
3137 }
3138
3139 /**
3140  * Return the total number of articles
3141  */
3142 function wfNumberOfArticles() {
3143         global $wgNumberOfArticles;
3144
3145         wfLoadSiteStats();
3146         return $wgNumberOfArticles;
3147 }
3148
3149 /**
3150  * Get various statistics from the database
3151  * @private
3152  */
3153 function wfLoadSiteStats() {
3154         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3155         $fname = 'wfLoadSiteStats';
3156
3157         if ( -1 != $wgNumberOfArticles ) return;
3158         $dbr =& wfGetDB( DB_SLAVE );
3159         $s = $dbr->selectRow( 'site_stats',
3160                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3161                 array( 'ss_row_id' => 1 ), $fname
3162         );
3163
3164         if ( $s === false ) {
3165                 return;
3166         } else {
3167                 $wgTotalViews = $s->ss_total_views;
3168                 $wgTotalEdits = $s->ss_total_edits;
3169                 $wgNumberOfArticles = $s->ss_good_articles;
3170         }
3171 }
3172
3173 /**
3174  * Escape html tags
3175  * Basicly replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3176  *
3177  * @param string $in Text that might contain HTML tags
3178  * @return string Escaped string
3179  */
3180 function wfEscapeHTMLTagsOnly( $in ) {
3181         return str_replace(
3182                 array( '"', '>', '<' ),
3183                 array( '&quot;', '&gt;', '&lt;' ),
3184                 $in );
3185 }
3186
3187 ?>