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