includes/Parser.php

   1 <?php
   2
   3 /**
   4  * File for Parser and related classes
   5  *
   6  * @package MediaWiki
   7  */
   8
   9 /**
  10  * Update this version number when the ParserOutput format
  11  * changes in an incompatible way, so the parser cache
  12  * can automatically discard old data.
  13  */
  14 define( 'MW_PARSER_VERSION', '1.4.0' );
  15
  16 /**
  17  * Variable substitution O(N^2) attack
  18  *
  19  * Without countermeasures, it would be possible to attack the parser by saving
  20  * a page filled with a large number of inclusions of large pages. The size of
  21  * the generated page would be proportional to the square of the input size.
  22  * Hence, we limit the number of inclusions of any given page, thus bringing any
  23  * attack back to O(N).
  24  */
  25
  26 define( 'MAX_INCLUDE_REPEAT', 100 );
  27 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  28
  29 define( 'RLH_FOR_UPDATE', 1 );
  30
  31 # Allowed values for $mOutputType
  32 define( 'OT_HTML', 1 );
  33 define( 'OT_WIKI', 2 );
  34 define( 'OT_MSG' , 3 );
  35
  36 # string parameter for extractTags which will cause it
  37 # to strip HTML comments in addition to regular
  38 # <XML>-style tags. This should not be anything we
  39 # may want to use in wikisyntax
  40 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  41
  42 # prefix for escaping, used in two functions at least
  43 define( 'UNIQ_PREFIX', 'NaodW29');
  44
  45 # Constants needed for external link processing
  46 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  47 define( 'HTTP_PROTOCOLS', 'http|https' );
  48 # Everything except bracket, space, or control characters
  49 define( 'EXT_LINK_URL_CLASS', '[^]<>"\\x00-\\x20\\x7F]' );
  50 # Including space
  51 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  52 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  53 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  54 define( 'EXT_LINK_BRACKETED',  '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  55 define( 'EXT_IMAGE_REGEX',
  56         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  57         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  58         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  59 );
  60
  61 /**
  62  * PHP Parser
  63  *
  64  * Processes wiki markup
  65  *
  66  * <pre>
  67  * There are three main entry points into the Parser class:
  68  * parse()
  69  *   produces HTML output
  70  * preSaveTransform().
  71  *   produces altered wiki markup.
  72  * transformMsg()
  73  *   performs brace substitution on MediaWiki messages
  74  *
  75  * Globals used:
  76  *    objects:   $wgLang, $wgDateFormatter, $wgLinkCache
  77  *
  78  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  79  *
  80  * settings:
  81  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  82  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  83  *  $wgLocaltimezone
  84  *
  85  *  * only within ParserOptions
  86  * </pre>
  87  *
  88  * @package MediaWiki
  89  */
  90 class Parser
  91 {
  92         /**#@+
  93          * @access private
  94          */
  95         # Persistent:
  96         var $mTagHooks;
  97
  98         # Cleared with clearState():
  99         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
 100         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
 101
 102         # Temporary:
 103         var $mOptions, $mTitle, $mOutputType,
 104             $mTemplates,        // cache of already loaded templates, avoids
 105                                 // multiple SQL queries for the same string
 106             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 107                                 // in this path. Used for loop detection.
 108
 109         /**#@-*/
 110
 111         /**
 112          * Constructor
 113          *
 114          * @access public
 115          */
 116         function Parser() {
 117                 $this->mTemplates = array();
 118                 $this->mTemplatePath = array();
 119                 $this->mTagHooks = array();
 120                 $this->clearState();
 121         }
 122
 123         /**
 124          * Clear Parser state
 125          *
 126          * @access private
 127          */
 128         function clearState() {
 129                 $this->mOutput = new ParserOutput;
 130                 $this->mAutonumber = 0;
 131                 $this->mLastSection = "";
 132                 $this->mDTopen = false;
 133                 $this->mVariables = false;
 134                 $this->mIncludeCount = array();
 135                 $this->mStripState = array();
 136                 $this->mArgStack = array();
 137                 $this->mInPre = false;
 138         }
 139
 140         /**
 141          * First pass--just handle <nowiki> sections, pass the rest off
 142          * to internalParse() which does all the real work.
 143          *
 144          * @access private
 145          * @return ParserOutput a ParserOutput
 146          */
 147         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 148                 global $wgUseTidy, $wgContLang;
 149                 $fname = 'Parser::parse';
 150                 wfProfileIn( $fname );
 151
 152                 if ( $clearState ) {
 153                         $this->clearState();
 154                 }
 155
 156                 $this->mOptions = $options;
 157                 $this->mTitle =& $title;
 158                 $this->mOutputType = OT_HTML;
 159
 160                 $stripState = NULL;
 161                 $text = $this->strip( $text, $this->mStripState );
 162
 163                 $text = $this->internalParse( $text, $linestart );
 164                 $text = $this->unstrip( $text, $this->mStripState );
 165                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 166                 if(!$wgUseTidy) {
 167                         $fixtags = array(
 168                                 # french spaces, last one Guillemet-left
 169                                 # only if there is something before the space
 170                                 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 171                                 # french spaces, Guillemet-right
 172                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 173                                 '/<hr *>/i' => '<hr />',
 174                                 '/<br *>/i' => '<br />',
 175                                 '/<center *>/i' => '<div class="center">',
 176                                 '/<\\/center *>/i' => '</div>',
 177                                 # Clean up spare ampersands; note that we probably ought to be
 178                                 # more careful about named entities.
 179                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 180                         );
 181                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 182                 } else {
 183                         $fixtags = array(
 184                                 # french spaces, last one Guillemet-left
 185                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 186                                 # french spaces, Guillemet-right
 187                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 188                                 '/<center *>/i' => '<div class="center">',
 189                                 '/<\\/center *>/i' => '</div>'
 190                         );
 191                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 192                 }
 193                 # only once and last
 194                 $text = $this->doBlockLevels( $text, $linestart );
 195
 196                 $this->replaceLinkHolders( $text );
 197                 $text = $wgContLang->convert($text);
 198
 199                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 200                 global $wgUseTidy;
 201                 if ($wgUseTidy) {
 202                         $text = Parser::tidy($text);
 203                 }
 204
 205                 $this->mOutput->setText( $text );
 206                 wfProfileOut( $fname );
 207                 return $this->mOutput;
 208         }
 209
 210         /**
 211          * Get a random string
 212          *
 213          * @access private
 214          * @static
 215          */
 216         function getRandomString() {
 217                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 218         }
 219
 220         /**
 221          * Replaces all occurrences of <$tag>content</$tag> in the text
 222          * with a random marker and returns the new text. the output parameter
 223          * $content will be an associative array filled with data on the form
 224          * $unique_marker => content.
 225          *
 226          * If $content is already set, the additional entries will be appended
 227          * If $tag is set to STRIP_COMMENTS, the function will extract
 228          * <!-- HTML comments -->
 229          *
 230          * @access private
 231          * @static
 232          */
 233         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 234                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 235                 if ( !$content ) {
 236                         $content = array( );
 237                 }
 238                 $n = 1;
 239                 $stripped = '';
 240
 241                 while ( '' != $text ) {
 242                         if($tag==STRIP_COMMENTS) {
 243                                 $p = preg_split( '/<!--/i', $text, 2 );
 244                         } else {
 245                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 246                         }
 247                         $stripped .= $p[0];
 248                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 249                                 $text = '';
 250                         } else {
 251                                 if($tag==STRIP_COMMENTS) {
 252                                         $q = preg_split( '/-->/i', $p[1], 2 );
 253                                 } else {
 254                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 255                                 }
 256                                 $marker = $rnd . sprintf('%08X', $n++);
 257                                 $content[$marker] = $q[0];
 258                                 $stripped .= $marker;
 259                                 $text = $q[1];
 260                         }
 261                 }
 262                 return $stripped;
 263         }
 264
 265         /**
 266          * Strips and renders nowiki, pre, math, hiero
 267          * If $render is set, performs necessary rendering operations on plugins
 268          * Returns the text, and fills an array with data needed in unstrip()
 269          * If the $state is already a valid strip state, it adds to the state
 270          *
 271          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 272          *  will be stripped in addition to other tags. This is important
 273          *  for section editing, where these comments cause confusion when
 274          *  counting the sections in the wikisource
 275          *
 276          * @access private
 277          */
 278         function strip( $text, &$state, $stripcomments = false ) {
 279                 $render = ($this->mOutputType == OT_HTML);
 280                 $html_content = array();
 281                 $nowiki_content = array();
 282                 $math_content = array();
 283                 $pre_content = array();
 284                 $comment_content = array();
 285                 $ext_content = array();
 286                 $gallery_content = array();
 287
 288                 # Replace any instances of the placeholders
 289                 $uniq_prefix = UNIQ_PREFIX;
 290                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 291
 292                 # html
 293                 global $wgRawHtml, $wgWhitelistEdit;
 294                 if( $wgRawHtml && $wgWhitelistEdit ) {
 295                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 296                         foreach( $html_content as $marker => $content ) {
 297                                 if ($render ) {
 298                                         # Raw and unchecked for validity.
 299                                         $html_content[$marker] = $content;
 300                                 } else {
 301                                         $html_content[$marker] = '<html>'.$content.'</html>';
 302                                 }
 303                         }
 304                 }
 305
 306                 # nowiki
 307                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 308                 foreach( $nowiki_content as $marker => $content ) {
 309                         if( $render ){
 310                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 311                         } else {
 312                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 313                         }
 314                 }
 315
 316                 # math
 317                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 318                 foreach( $math_content as $marker => $content ){
 319                         if( $render ) {
 320                                 if( $this->mOptions->getUseTeX() ) {
 321                                         $math_content[$marker] = renderMath( $content );
 322                                 } else {
 323                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 324                                 }
 325                         } else {
 326                                 $math_content[$marker] = '<math>'.$content.'</math>';
 327                         }
 328                 }
 329
 330                 # pre
 331                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 332                 foreach( $pre_content as $marker => $content ){
 333                         if( $render ){
 334                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 335                         } else {
 336                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 337                         }
 338                 }
 339
 340                 # gallery
 341                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 342                 foreach( $gallery_content as $marker => $content ) {
 343                         require_once( 'ImageGallery.php' );
 344                         if ( $render ) {
 345                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 346                         } else {
 347                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 348                         }
 349                 }
 350
 351                 # Comments
 352                 if($stripcomments) {
 353                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 354                         foreach( $comment_content as $marker => $content ){
 355                                 $comment_content[$marker] = '<!--'.$content.'-->';
 356                         }
 357                 }
 358
 359                 # Extensions
 360                 foreach ( $this->mTagHooks as $tag => $callback ) {
 361                         $ext_contents[$tag] = array();
 362                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 363                         foreach( $ext_content[$tag] as $marker => $content ) {
 364                                 if ( $render ) {
 365                                         $ext_content[$tag][$marker] = $callback( $content );
 366                                 } else {
 367                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 368                                 }
 369                         }
 370                 }
 371
 372                 # Merge state with the pre-existing state, if there is one
 373                 if ( $state ) {
 374                         $state['html'] = $state['html'] + $html_content;
 375                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 376                         $state['math'] = $state['math'] + $math_content;
 377                         $state['pre'] = $state['pre'] + $pre_content;
 378                         $state['comment'] = $state['comment'] + $comment_content;
 379                         $state['gallery'] = $state['gallery'] + $gallery_content;
 380
 381                         foreach( $ext_content as $tag => $array ) {
 382                                 if ( array_key_exists( $tag, $state ) ) {
 383                                         $state[$tag] = $state[$tag] + $array;
 384                                 }
 385                         }
 386                 } else {
 387                         $state = array(
 388                           'html' => $html_content,
 389                           'nowiki' => $nowiki_content,
 390                           'math' => $math_content,
 391                           'pre' => $pre_content,
 392                           'comment' => $comment_content,
 393                           'gallery' => $gallery_content,
 394                         ) + $ext_content;
 395                 }
 396                 return $text;
 397         }
 398
 399         /**
 400          * restores pre, math, and hiero removed by strip()
 401          *
 402          * always call unstripNoWiki() after this one
 403          * @access private
 404          */
 405         function unstrip( $text, &$state ) {
 406                 # Must expand in reverse order, otherwise nested tags will be corrupted
 407                 $contentDict = end( $state );
 408                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 409                         if( key($state) != 'nowiki' && key($state) != 'html') {
 410                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 411                                         $text = str_replace( key( $contentDict ), $content, $text );
 412                                 }
 413                         }
 414                 }
 415
 416                 return $text;
 417         }
 418
 419         /**
 420          * always call this after unstrip() to preserve the order
 421          *
 422          * @access private
 423          */
 424         function unstripNoWiki( $text, &$state ) {
 425                 # Must expand in reverse order, otherwise nested tags will be corrupted
 426                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 427                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 428                 }
 429
 430                 global $wgRawHtml;
 431                 if ($wgRawHtml) {
 432                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 433                                 $text = str_replace( key( $state['html'] ), $content, $text );
 434                         }
 435                 }
 436
 437                 return $text;
 438         }
 439
 440         /**
 441          * Add an item to the strip state
 442          * Returns the unique tag which must be inserted into the stripped text
 443          * The tag will be replaced with the original text in unstrip()
 444          *
 445          * @access private
 446          */
 447         function insertStripItem( $text, &$state ) {
 448                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 449                 if ( !$state ) {
 450                         $state = array(
 451                           'html' => array(),
 452                           'nowiki' => array(),
 453                           'math' => array(),
 454                           'pre' => array()
 455                         );
 456                 }
 457                 $state['item'][$rnd] = $text;
 458                 return $rnd;
 459         }
 460
 461         /**
 462          * Return allowed HTML attributes
 463          *
 464          * @access private
 465          */
 466         function getHTMLattrs () {
 467                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 468                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 469                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 470                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 471                                 /* FONT */ 'type', 'start', 'value', 'compact',
 472                                 /* For various lists, mostly deprecated but safe */
 473                                 'summary', 'width', 'border', 'frame', 'rules',
 474                                 'cellspacing', 'cellpadding', 'valign', 'char',
 475                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 476                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 477                                 'id', 'class', 'name', 'style' /* For CSS */
 478                                 );
 479                 return $htmlattrs ;
 480         }
 481
 482         /**
 483          * Remove non approved attributes and javascript in css
 484          *
 485          * @access private
 486          */
 487         function fixTagAttributes ( $t ) {
 488                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 489                 $htmlattrs = $this->getHTMLattrs() ;
 490
 491                 # Strip non-approved attributes from the tag
 492                 $t = preg_replace(
 493                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 494                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 495                         $t);
 496
 497                 $t = str_replace ( '<></>' , '' , $t ) ; # This should fix bug 980557
 498
 499                 # Strip javascript "expression" from stylesheets. Brute force approach:
 500                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 501
 502                 if( preg_match(
 503                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 504                         wfMungeToUtf8( $t ) ) )
 505                 {
 506                         $t='';
 507                 }
 508
 509                 return trim ( $t ) ;
 510         }
 511
 512         /**
 513          * interface with html tidy, used if $wgUseTidy = true
 514          *
 515          * @access public
 516          * @static
 517          */
 518         function tidy ( $text ) {
 519                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 520                 global $wgInputEncoding, $wgOutputEncoding;
 521                 $fname = 'Parser::tidy';
 522                 wfProfileIn( $fname );
 523
 524                 $cleansource = '';
 525                 $opts = '';
 526                 switch(strtoupper($wgOutputEncoding)) {
 527                         case 'ISO-8859-1':
 528                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 529                                 break;
 530                         case 'UTF-8':
 531                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 532                                 break;
 533                         default:
 534                                 $opts .= ' -raw';
 535                         }
 536
 537                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 538 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 539 '<head><title>test</title></head><body>'.$text.'</body></html>';
 540                 $descriptorspec = array(
 541                         0 => array('pipe', 'r'),
 542                         1 => array('pipe', 'w'),
 543                         2 => array('file', '/dev/null', 'a')
 544                 );
 545                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 546                 if (is_resource($process)) {
 547                         fwrite($pipes[0], $wrappedtext);
 548                         fclose($pipes[0]);
 549                         while (!feof($pipes[1])) {
 550                                 $cleansource .= fgets($pipes[1], 1024);
 551                         }
 552                         fclose($pipes[1]);
 553                         $return_value = proc_close($process);
 554                 }
 555
 556                 wfProfileOut( $fname );
 557
 558                 if( $cleansource == '' && $text != '') {
 559                         wfDebug( "Tidy error detected!\n" );
 560                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 561                 } else {
 562                         return $cleansource;
 563                 }
 564         }
 565
 566         /**
 567          * parse the wiki syntax used to render tables
 568          *
 569          * @access private
 570          */
 571         function doTableStuff ( $t ) {
 572                 $fname = 'Parser::doTableStuff';
 573                 wfProfileIn( $fname );
 574
 575                 $t = explode ( "\n" , $t ) ;
 576                 $td = array () ; # Is currently a td tag open?
 577                 $ltd = array () ; # Was it TD or TH?
 578                 $tr = array () ; # Is currently a tr tag open?
 579                 $ltr = array () ; # tr attributes
 580                 $indent_level = 0; # indent level of the table
 581                 foreach ( $t AS $k => $x )
 582                 {
 583                         $x = trim ( $x ) ;
 584                         $fc = substr ( $x , 0 , 1 ) ;
 585                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 586                                 $indent_level = strlen( $matches[1] );
 587                                 $t[$k] = "\n" .
 588                                         str_repeat( '<dl><dd>', $indent_level ) .
 589                                         '<table ' . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 590                                 array_push ( $td , false ) ;
 591                                 array_push ( $ltd , '' ) ;
 592                                 array_push ( $tr , false ) ;
 593                                 array_push ( $ltr , '' ) ;
 594                         }
 595                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 596                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 597                                 $z = "</table>\n" ;
 598                                 $l = array_pop ( $ltd ) ;
 599                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 600                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 601                                 array_pop ( $ltr ) ;
 602                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 603                         }
 604                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 605                                 $x = substr ( $x , 1 ) ;
 606                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 607                                 $z = '' ;
 608                                 $l = array_pop ( $ltd ) ;
 609                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 610                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 611                                 array_pop ( $ltr ) ;
 612                                 $t[$k] = $z ;
 613                                 array_push ( $tr , false ) ;
 614                                 array_push ( $td , false ) ;
 615                                 array_push ( $ltd , '' ) ;
 616                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 617                         }
 618                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 619                                 # $x is a table row
 620                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 621                                         $fc = '+' ;
 622                                         $x = substr ( $x , 1 ) ;
 623                                 }
 624                                 $after = substr ( $x , 1 ) ;
 625                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 626                                 $after = explode ( '||' , $after ) ;
 627                                 $t[$k] = '' ;
 628
 629                                 # Loop through each table cell
 630                                 foreach ( $after AS $theline )
 631                                 {
 632                                         $z = '' ;
 633                                         if ( $fc != '+' )
 634                                         {
 635                                                 $tra = array_pop ( $ltr ) ;
 636                                                 if ( !array_pop ( $tr ) ) $z = '<tr '.$tra.">\n" ;
 637                                                 array_push ( $tr , true ) ;
 638                                                 array_push ( $ltr , '' ) ;
 639                                         }
 640
 641                                         $l = array_pop ( $ltd ) ;
 642                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 643                                         if ( $fc == '|' ) $l = 'td' ;
 644                                         else if ( $fc == '!' ) $l = 'th' ;
 645                                         else if ( $fc == '+' ) $l = 'caption' ;
 646                                         else $l = '' ;
 647                                         array_push ( $ltd , $l ) ;
 648
 649                                         # Cell parameters
 650                                         $y = explode ( '|' , $theline , 2 ) ;
 651                                         # Note that a '|' inside an invalid link should not
 652                                         # be mistaken as delimiting cell parameters
 653                                         if ( strpos( $y[0], '[[' ) !== false ) {
 654                                                 $y = array ($theline);
 655                                         }
 656                                         if ( count ( $y ) == 1 )
 657                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 658                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 659                                         $t[$k] .= $y ;
 660                                         array_push ( $td , true ) ;
 661                                 }
 662                         }
 663                 }
 664
 665                 # Closing open td, tr && table
 666                 while ( count ( $td ) > 0 )
 667                 {
 668                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 669                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 670                         $t[] = '</table>' ;
 671                 }
 672
 673                 $t = implode ( "\n" , $t ) ;
 674                 #               $t = $this->removeHTMLtags( $t );
 675                 wfProfileOut( $fname );
 676                 return $t ;
 677         }
 678
 679         /**
 680          * Helper function for parse() that transforms wiki markup into
 681          * HTML. Only called for $mOutputType == OT_HTML.
 682          *
 683          * @access private
 684          */
 685         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 686                 global $wgContLang;
 687
 688                 $fname = 'Parser::internalParse';
 689                 wfProfileIn( $fname );
 690
 691                 $text = $this->removeHTMLtags( $text );
 692                 $text = $this->replaceVariables( $text, $args );
 693
 694                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 695
 696                 $text = $this->doHeadings( $text );
 697                 if($this->mOptions->getUseDynamicDates()) {
 698                         global $wgDateFormatter;
 699                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 700                 }
 701                 $text = $this->doAllQuotes( $text );
 702                 $text = $this->replaceInternalLinks( $text );
 703                 $text = $this->replaceExternalLinks( $text );
 704
 705                 # replaceInternalLinks may sometimes leave behind
 706                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 707                 $text = str_replace("http-noparse://","http://",$text);
 708
 709                 $text = $this->doMagicLinks( $text );
 710                 $text = $this->doTableStuff( $text );
 711                 $text = $this->formatHeadings( $text, $isMain );
 712                 $sk =& $this->mOptions->getSkin();
 713                 $text = $sk->transformContent( $text );
 714
 715                 wfProfileOut( $fname );
 716                 return $text;
 717         }
 718
 719         /**
 720          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 721          * magic external links.
 722          *
 723          * @access private
 724          */
 725         function &doMagicLinks( &$text ) {
 726                 global $wgUseGeoMode;
 727                 $text = $this->magicISBN( $text );
 728                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 729                         $text = $this->magicGEO( $text );
 730                 }
 731                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 732                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 733                 return $text;
 734         }
 735
 736         /**
 737          * Parse ^^ tokens and return html
 738          *
 739          * @access private
 740          */
 741         function doExponent( $text ) {
 742                 $fname = 'Parser::doExponent';
 743                 wfProfileIn( $fname );
 744                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 745                 wfProfileOut( $fname );
 746                 return $text;
 747         }
 748
 749         /**
 750          * Parse headers and return html
 751          *
 752          * @access private
 753          */
 754         function doHeadings( $text ) {
 755                 $fname = 'Parser::doHeadings';
 756                 wfProfileIn( $fname );
 757                 for ( $i = 6; $i >= 1; --$i ) {
 758                         $h = substr( '======', 0, $i );
 759                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 760                           "<h{$i}>\\1</h{$i}>\\2", $text );
 761                 }
 762                 wfProfileOut( $fname );
 763                 return $text;
 764         }
 765
 766         /**
 767          * Replace single quotes with HTML markup
 768          * @access private
 769          * @return string the altered text
 770          */
 771         function doAllQuotes( $text ) {
 772                 $fname = 'Parser::doAllQuotes';
 773                 wfProfileIn( $fname );
 774                 $outtext = '';
 775                 $lines = explode( "\n", $text );
 776                 foreach ( $lines as $line ) {
 777                         $outtext .= $this->doQuotes ( $line ) . "\n";
 778                 }
 779                 $outtext = substr($outtext, 0,-1);
 780                 wfProfileOut( $fname );
 781                 return $outtext;
 782         }
 783
 784         /**
 785          * Helper function for doAllQuotes()
 786          * @access private
 787          */
 788         function doQuotes( $text ) {
 789                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 790                 if ( count( $arr ) == 1 )
 791                         return $text;
 792                 else
 793                 {
 794                         # First, do some preliminary work. This may shift some apostrophes from
 795                         # being mark-up to being text. It also counts the number of occurrences
 796                         # of bold and italics mark-ups.
 797                         $i = 0;
 798                         $numbold = 0;
 799                         $numitalics = 0;
 800                         foreach ( $arr as $r )
 801                         {
 802                                 if ( ( $i % 2 ) == 1 )
 803                                 {
 804                                         # If there are ever four apostrophes, assume the first is supposed to
 805                                         # be text, and the remaining three constitute mark-up for bold text.
 806                                         if ( strlen( $arr[$i] ) == 4 )
 807                                         {
 808                                                 $arr[$i-1] .= "'";
 809                                                 $arr[$i] = "'''";
 810                                         }
 811                                         # If there are more than 5 apostrophes in a row, assume they're all
 812                                         # text except for the last 5.
 813                                         else if ( strlen( $arr[$i] ) > 5 )
 814                                         {
 815                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 816                                                 $arr[$i] = "'''''";
 817                                         }
 818                                         # Count the number of occurrences of bold and italics mark-ups.
 819                                         # We are not counting sequences of five apostrophes.
 820                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 821                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 822                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 823                                 }
 824                                 $i++;
 825                         }
 826
 827                         # If there is an odd number of both bold and italics, it is likely
 828                         # that one of the bold ones was meant to be an apostrophe followed
 829                         # by italics. Which one we cannot know for certain, but it is more
 830                         # likely to be one that has a single-letter word before it.
 831                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 832                         {
 833                                 $i = 0;
 834                                 $firstsingleletterword = -1;
 835                                 $firstmultiletterword = -1;
 836                                 $firstspace = -1;
 837                                 foreach ( $arr as $r )
 838                                 {
 839                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 840                                         {
 841                                                 $x1 = substr ($arr[$i-1], -1);
 842                                                 $x2 = substr ($arr[$i-1], -2, 1);
 843                                                 if ($x1 == ' ') {
 844                                                         if ($firstspace == -1) $firstspace = $i;
 845                                                 } else if ($x2 == ' ') {
 846                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 847                                                 } else {
 848                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 849                                                 }
 850                                         }
 851                                         $i++;
 852                                 }
 853
 854                                 # If there is a single-letter word, use it!
 855                                 if ($firstsingleletterword > -1)
 856                                 {
 857                                         $arr [ $firstsingleletterword ] = "''";
 858                                         $arr [ $firstsingleletterword-1 ] .= "'";
 859                                 }
 860                                 # If not, but there's a multi-letter word, use that one.
 861                                 else if ($firstmultiletterword > -1)
 862                                 {
 863                                         $arr [ $firstmultiletterword ] = "''";
 864                                         $arr [ $firstmultiletterword-1 ] .= "'";
 865                                 }
 866                                 # ... otherwise use the first one that has neither.
 867                                 # (notice that it is possible for all three to be -1 if, for example,
 868                                 # there is only one pentuple-apostrophe in the line)
 869                                 else if ($firstspace > -1)
 870                                 {
 871                                         $arr [ $firstspace ] = "''";
 872                                         $arr [ $firstspace-1 ] .= "'";
 873                                 }
 874                         }
 875
 876                         # Now let's actually convert our apostrophic mush to HTML!
 877                         $output = '';
 878                         $buffer = '';
 879                         $state = '';
 880                         $i = 0;
 881                         foreach ($arr as $r)
 882                         {
 883                                 if (($i % 2) == 0)
 884                                 {
 885                                         if ($state == 'both')
 886                                                 $buffer .= $r;
 887                                         else
 888                                                 $output .= $r;
 889                                 }
 890                                 else
 891                                 {
 892                                         if (strlen ($r) == 2)
 893                                         {
 894                                                 if ($state == 'i')
 895                                                 { $output .= '</i>'; $state = ''; }
 896                                                 else if ($state == 'bi')
 897                                                 { $output .= '</i>'; $state = 'b'; }
 898                                                 else if ($state == 'ib')
 899                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 900                                                 else if ($state == 'both')
 901                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 902                                                 else # $state can be 'b' or ''
 903                                                 { $output .= '<i>'; $state .= 'i'; }
 904                                         }
 905                                         else if (strlen ($r) == 3)
 906                                         {
 907                                                 if ($state == 'b')
 908                                                 { $output .= '</b>'; $state = ''; }
 909                                                 else if ($state == 'bi')
 910                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 911                                                 else if ($state == 'ib')
 912                                                 { $output .= '</b>'; $state = 'i'; }
 913                                                 else if ($state == 'both')
 914                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 915                                                 else # $state can be 'i' or ''
 916                                                 { $output .= '<b>'; $state .= 'b'; }
 917                                         }
 918                                         else if (strlen ($r) == 5)
 919                                         {
 920                                                 if ($state == 'b')
 921                                                 { $output .= '</b><i>'; $state = 'i'; }
 922                                                 else if ($state == 'i')
 923                                                 { $output .= '</i><b>'; $state = 'b'; }
 924                                                 else if ($state == 'bi')
 925                                                 { $output .= '</i></b>'; $state = ''; }
 926                                                 else if ($state == 'ib')
 927                                                 { $output .= '</b></i>'; $state = ''; }
 928                                                 else if ($state == 'both')
 929                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 930                                                 else # ($state == '')
 931                                                 { $buffer = ''; $state = 'both'; }
 932                                         }
 933                                 }
 934                                 $i++;
 935                         }
 936                         # Now close all remaining tags.  Notice that the order is important.
 937                         if ($state == 'b' || $state == 'ib')
 938                                 $output .= '</b>';
 939                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 940                                 $output .= '</i>';
 941                         if ($state == 'bi')
 942                                 $output .= '</b>';
 943                         if ($state == 'both')
 944                                 $output .= '<b><i>'.$buffer.'</i></b>';
 945                         return $output;
 946                 }
 947         }
 948
 949         /**
 950          * Replace external links
 951          *
 952          * Note: this is all very hackish and the order of execution matters a lot.
 953          * Make sure to run maintenance/parserTests.php if you change this code.
 954          *
 955          * @access private
 956          */
 957         function replaceExternalLinks( $text ) {
 958                 $fname = 'Parser::replaceExternalLinks';
 959                 wfProfileIn( $fname );
 960
 961                 $sk =& $this->mOptions->getSkin();
 962                 $linktrail = wfMsgForContent('linktrail');
 963                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 964
 965                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 966
 967                 $i = 0;
 968                 while ( $i<count( $bits ) ) {
 969                         $url = $bits[$i++];
 970                         $protocol = $bits[$i++];
 971                         $text = $bits[$i++];
 972                         $trail = $bits[$i++];
 973
 974                         # The characters '<' and '>' (which were escaped by
 975                         # removeHTMLtags()) should not be included in
 976                         # URLs, per RFC 2396.
 977                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 978                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 979                                 $url = substr($url, 0, $m2[0][1]);
 980                         }
 981
 982                         # If the link text is an image URL, replace it with an <img> tag
 983                         # This happened by accident in the original parser, but some people used it extensively
 984                         $img = $this->maybeMakeImageLink( $text );
 985                         if ( $img !== false ) {
 986                                 $text = $img;
 987                         }
 988
 989                         $dtrail = '';
 990
 991                         # No link text, e.g. [http://domain.tld/some.link]
 992                         if ( $text == '' ) {
 993                                 # Autonumber if allowed
 994                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 995                                         $text = '[' . ++$this->mAutonumber . ']';
 996                                 } else {
 997                                         # Otherwise just use the URL
 998                                         $text = htmlspecialchars( $url );
 999                                 }
1000                         } else {
1001                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1002                                 # Check for trail
1003                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
1004                                         $dtrail = $m2[1];
1005                                         $trail = $m2[2];
1006                                 }
1007                         }
1008
1009                         $encUrl = htmlspecialchars( $url );
1010                         # Bit in parentheses showing the URL for the printable version
1011                         if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
1012                                 $paren = '';
1013                         } else {
1014                                 # Expand the URL for printable version
1015                                 if ( ! $sk->suppressUrlExpansion() ) {
1016                                         $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
1017                                 } else {
1018                                         $paren = '';
1019                                 }
1020                         }
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 ) . $dtrail. $paren . $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( '/((?:'.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 );
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
1123         function replaceInternalLinks( $s ) {
1124                 global $wgLang, $wgContLang, $wgLinkCache;
1125                 global $wgDisableLangConversion;
1126                 static $fname = 'Parser::replaceInternalLinks' ;
1127
1128                 wfProfileIn( $fname );
1129
1130                 wfProfileIn( $fname.'-setup' );
1131                 static $tc = FALSE;
1132                 # the % is needed to support urlencoded titles as well
1133                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1134
1135                 $sk =& $this->mOptions->getSkin();
1136                 global $wgUseOldExistenceCheck;
1137                 # "Post-parse link colour check" works only on wiki text since it's now
1138                 # in Parser. Enable it, then disable it when we're done.
1139                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1140
1141                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1142
1143                 #split the entire text string on occurences of [[
1144                 $a = explode( '[[', ' ' . $s );
1145                 #get the first element (all text up to first [[), and remove the space we added
1146                 $s = array_shift( $a );
1147                 $s = substr( $s, 1 );
1148
1149                 # Match a link having the form [[namespace:link|alternate]]trail
1150                 static $e1 = FALSE;
1151                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1152                 # Match cases where there is no "]]", which might still be images
1153                 static $e1_img = FALSE;
1154                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1155                 # Match the end of a line for a word that's not followed by whitespace,
1156                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1157                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1158
1159                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1160
1161                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1162
1163                 if ( $useLinkPrefixExtension ) {
1164                         if ( preg_match( $e2, $s, $m ) ) {
1165                                 $first_prefix = $m[2];
1166                                 $s = $m[1];
1167                         } else {
1168                                 $first_prefix = false;
1169                         }
1170                 } else {
1171                         $prefix = '';
1172                 }
1173
1174                 $selflink = $this->mTitle->getPrefixedText();
1175                 wfProfileOut( $fname.'-setup' );
1176
1177                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1178                 $useSubpages = $this->areSubpagesAllowed();
1179
1180                 # Loop for each link
1181                 for ($k = 0; isset( $a[$k] ); $k++) {
1182                         $line = $a[$k];
1183                         if ( $useLinkPrefixExtension ) {
1184                                 wfProfileIn( $fname.'-prefixhandling' );
1185                                 if ( preg_match( $e2, $s, $m ) ) {
1186                                         $prefix = $m[2];
1187                                         $s = $m[1];
1188                                 } else {
1189                                         $prefix='';
1190                                 }
1191                                 # first link
1192                                 if($first_prefix) {
1193                                         $prefix = $first_prefix;
1194                                         $first_prefix = false;
1195                                 }
1196                                 wfProfileOut( $fname.'-prefixhandling' );
1197                         }
1198
1199                         $might_be_img = false;
1200
1201                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1202                                 $text = $m[2];
1203                                 # fix up urlencoded title texts
1204                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1205                                 $trail = $m[3];
1206                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1207                                 $might_be_img = true;
1208                                 $text = $m[2];
1209                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1210                                 $trail = "";
1211                         } else { # Invalid form; output directly
1212                                 $s .= $prefix . '[[' . $line ;
1213                                 continue;
1214                         }
1215
1216                         # Don't allow internal links to pages containing
1217                         # PROTO: where PROTO is a valid URL protocol; these
1218                         # should be external links.
1219                         if (preg_match('/^((?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1220                                 $s .= $prefix . '[[' . $line ;
1221                                 continue;
1222                         }
1223
1224                         # Make subpage if necessary
1225                         if( $useSubpages ) {
1226                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1227                         } else {
1228                                 $link = $m[1];
1229                         }
1230
1231                         $noforce = (substr($m[1], 0, 1) != ':');
1232                         if (!$noforce) {
1233                                 # Strip off leading ':'
1234                                 $link = substr($link, 1);
1235                         }
1236
1237                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1238                         if( !$nt ) {
1239                                 $s .= $prefix . '[[' . $line;
1240                                 continue;
1241                         }
1242
1243                         #check other language variants of the link
1244                         #if the article does not exist
1245                         if( $checkVariantLink
1246                             && $nt->getArticleID() == 0 ) {
1247                                 $wgContLang->findVariantLink($link, $nt);
1248                         }
1249
1250                         $ns = $nt->getNamespace();
1251                         $iw = $nt->getInterWiki();
1252
1253                         if ($might_be_img) { # if this is actually an invalid link
1254                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1255                                         $found = false;
1256                                         while (isset ($a[$k+1]) ) {
1257                                                 #look at the next 'line' to see if we can close it there
1258                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1259                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1260                                                 # the first ]] closes the inner link, the second the image
1261                                                         $found = true;
1262                                                         $text .= '[[' . $m[1];
1263                                                         $trail = $m[2];
1264                                                         break;
1265                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1266                                                         #if there's exactly one ]] that's fine, we'll keep looking
1267                                                         $text .= '[[' . $m[0];
1268                                                 } else {
1269                                                         #if $next_line is invalid too, we need look no further
1270                                                         $text .= '[[' . $next_line;
1271                                                         break;
1272                                                 }
1273                                         }
1274                                         if ( !$found ) {
1275                                                 # we couldn't find the end of this imageLink, so output it raw
1276                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1277                                                 $text = $this->replaceInternalLinks($text);
1278                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1279                                                 # note: no $trail, because without an end, there *is* no trail
1280                                                 continue;
1281                                         }
1282                                 } else { #it's not an image, so output it raw
1283                                         $s .= $prefix . '[[' . $link . '|' . $text;
1284                                         # note: no $trail, because without an end, there *is* no trail
1285                                         continue;
1286                                 }
1287                         }
1288
1289                         $wasblank = ( '' == $text );
1290                         if( $wasblank ) $text = $link;
1291
1292
1293                         # Link not escaped by : , create the various objects
1294                         if( $noforce ) {
1295
1296                                 # Interwikis
1297                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1298                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1299                                         $tmp = $prefix . $trail ;
1300                                         $s .= (trim($tmp) == '')? '': $tmp;
1301                                         continue;
1302                                 }
1303
1304                                 if ( $ns == NS_IMAGE ) {
1305                                         wfProfileIn( "$fname-image" );
1306
1307                                         # recursively parse links inside the image caption
1308                                         # actually, this will parse them in any other parameters, too,
1309                                         # but it might be hard to fix that, and it doesn't matter ATM
1310                                         $text = $this->replaceExternalLinks($text);
1311                                         $text = $this->replaceInternalLinks($text);
1312
1313                                         # replace the image with a link-holder so that replaceExternalLinks() can't mess with it
1314                                         $s .= $prefix . $this->insertStripItem( $sk->makeImageLinkObj( $nt, $text ), $this->mStripState ) . $trail;
1315                                         $wgLinkCache->addImageLinkObj( $nt );
1316
1317                                         wfProfileOut( "$fname-image" );
1318                                         continue;
1319                                 }
1320
1321                                 if ( $ns == NS_CATEGORY ) {
1322                                         wfProfileIn( "$fname-category" );
1323                                         $t = $nt->getText();
1324
1325                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1326                                         $pPLC=$sk->postParseLinkColour();
1327                                         $sk->postParseLinkColour( false );
1328                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1329                                         $sk->postParseLinkColour( $pPLC );
1330                                         $wgLinkCache->resume();
1331
1332                                         if ( $wasblank ) {
1333                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1334                                                         $sortkey = $this->mTitle->getText();
1335                                                 } else {
1336                                                         $sortkey = $this->mTitle->getPrefixedText();
1337                                                 }
1338                                         } else {
1339                                                 $sortkey = $text;
1340                                         }
1341                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1342                                         $this->mOutput->addCategoryLink( $t );
1343                                         $s .= $prefix . $trail ;
1344
1345                                         wfProfileOut( "$fname-category" );
1346                                         continue;
1347                                 }
1348                         }
1349
1350                         if( ( $nt->getPrefixedText() === $selflink ) &&
1351                             ( $nt->getFragment() === '' ) ) {
1352                                 # Self-links are handled specially; generally de-link and change to bold.
1353                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1354                                 continue;
1355                         }
1356
1357                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1358                         if( $ns == NS_MEDIA ) {
1359                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1360                                 $wgLinkCache->addImageLinkObj( $nt );
1361                                 continue;
1362                         } elseif( $ns == NS_SPECIAL ) {
1363                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1364                                 continue;
1365                         }
1366                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1367                 }
1368                 $sk->postParseLinkColour( $saveParseColour );
1369                 wfProfileOut( $fname );
1370                 return $s;
1371         }
1372
1373         /**
1374          * Return true if subpage links should be expanded on this page.
1375          * @return bool
1376          */
1377         function areSubpagesAllowed() {
1378                 # Some namespaces don't allow subpages
1379                 global $wgNamespacesWithSubpages;
1380                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1381         }
1382
1383         /**
1384          * Handle link to subpage if necessary
1385          * @param string $target the source of the link
1386          * @param string &$text the link text, modified as necessary
1387          * @return string the full name of the link
1388          * @access private
1389          */
1390         function maybeDoSubpageLink($target, &$text) {
1391                 # Valid link forms:
1392                 # Foobar -- normal
1393                 # :Foobar -- override special treatment of prefix (images, language links)
1394                 # /Foobar -- convert to CurrentPage/Foobar
1395                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1396                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1397                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1398
1399                 $fname = 'Parser::maybeDoSubpageLink';
1400                 wfProfileIn( $fname );
1401                 $ret = $target; # default return value is no change
1402
1403                 # Some namespaces don't allow subpages,
1404                 # so only perform processing if subpages are allowed
1405                 if( $this->areSubpagesAllowed() ) {
1406                         # Look at the first character
1407                         if( $target != '' && $target{0} == '/' ) {
1408                                 # / at end means we don't want the slash to be shown
1409                                 if( substr( $target, -1, 1 ) == '/' ) {
1410                                         $target = substr( $target, 1, -1 );
1411                                         $noslash = $target;
1412                                 } else {
1413                                         $noslash = substr( $target, 1 );
1414                                 }
1415
1416                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1417                                 if( '' === $text ) {
1418                                         $text = $target;
1419                                 } # this might be changed for ugliness reasons
1420                         } else {
1421                                 # check for .. subpage backlinks
1422                                 $dotdotcount = 0;
1423                                 $nodotdot = $target;
1424                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1425                                         ++$dotdotcount;
1426                                         $nodotdot = substr( $nodotdot, 3 );
1427                                 }
1428                                 if($dotdotcount > 0) {
1429                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1430                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1431                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1432                                                 # / at the end means don't show full path
1433                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1434                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1435                                                         if( '' === $text ) {
1436                                                                 $text = $nodotdot;
1437                                                         }
1438                                                 }
1439                                                 $nodotdot = trim( $nodotdot );
1440                                                 if( $nodotdot != '' ) {
1441                                                         $ret .= '/' . $nodotdot;
1442                                                 }
1443                                         }
1444                                 }
1445                         }
1446                 }
1447
1448                 wfProfileOut( $fname );
1449                 return $ret;
1450         }
1451
1452         /**#@+
1453          * Used by doBlockLevels()
1454          * @access private
1455          */
1456         /* private */ function closeParagraph() {
1457                 $result = '';
1458                 if ( '' != $this->mLastSection ) {
1459                         $result = '</' . $this->mLastSection  . ">\n";
1460                 }
1461                 $this->mInPre = false;
1462                 $this->mLastSection = '';
1463                 return $result;
1464         }
1465         # getCommon() returns the length of the longest common substring
1466         # of both arguments, starting at the beginning of both.
1467         #
1468         /* private */ function getCommon( $st1, $st2 ) {
1469                 $fl = strlen( $st1 );
1470                 $shorter = strlen( $st2 );
1471                 if ( $fl < $shorter ) { $shorter = $fl; }
1472
1473                 for ( $i = 0; $i < $shorter; ++$i ) {
1474                         if ( $st1{$i} != $st2{$i} ) { break; }
1475                 }
1476                 return $i;
1477         }
1478         # These next three functions open, continue, and close the list
1479         # element appropriate to the prefix character passed into them.
1480         #
1481         /* private */ function openList( $char ) {
1482                 $result = $this->closeParagraph();
1483
1484                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1485                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1486                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1487                 else if ( ';' == $char ) {
1488                         $result .= '<dl><dt>';
1489                         $this->mDTopen = true;
1490                 }
1491                 else { $result = '<!-- ERR 1 -->'; }
1492
1493                 return $result;
1494         }
1495
1496         /* private */ function nextItem( $char ) {
1497                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1498                 else if ( ':' == $char || ';' == $char ) {
1499                         $close = '</dd>';
1500                         if ( $this->mDTopen ) { $close = '</dt>'; }
1501                         if ( ';' == $char ) {
1502                                 $this->mDTopen = true;
1503                                 return $close . '<dt>';
1504                         } else {
1505                                 $this->mDTopen = false;
1506                                 return $close . '<dd>';
1507                         }
1508                 }
1509                 return '<!-- ERR 2 -->';
1510         }
1511
1512         /* private */ function closeList( $char ) {
1513                 if ( '*' == $char ) { $text = '</li></ul>'; }
1514                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1515                 else if ( ':' == $char ) {
1516                         if ( $this->mDTopen ) {
1517                                 $this->mDTopen = false;
1518                                 $text = '</dt></dl>';
1519                         } else {
1520                                 $text = '</dd></dl>';
1521                         }
1522                 }
1523                 else {  return '<!-- ERR 3 -->'; }
1524                 return $text."\n";
1525         }
1526         /**#@-*/
1527
1528         /**
1529          * Make lists from lines starting with ':', '*', '#', etc.
1530          *
1531          * @access private
1532          * @return string the lists rendered as HTML
1533          */
1534         function doBlockLevels( $text, $linestart ) {
1535                 $fname = 'Parser::doBlockLevels';
1536                 wfProfileIn( $fname );
1537
1538                 # Parsing through the text line by line.  The main thing
1539                 # happening here is handling of block-level elements p, pre,
1540                 # and making lists from lines starting with * # : etc.
1541                 #
1542                 $textLines = explode( "\n", $text );
1543
1544                 $lastPrefix = $output = $lastLine = '';
1545                 $this->mDTopen = $inBlockElem = false;
1546                 $prefixLength = 0;
1547                 $paragraphStack = false;
1548
1549                 if ( !$linestart ) {
1550                         $output .= array_shift( $textLines );
1551                 }
1552                 foreach ( $textLines as $oLine ) {
1553                         $lastPrefixLength = strlen( $lastPrefix );
1554                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1555                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1556                         if ( !$this->mInPre ) {
1557                                 # Multiple prefixes may abut each other for nested lists.
1558                                 $prefixLength = strspn( $oLine, '*#:;' );
1559                                 $pref = substr( $oLine, 0, $prefixLength );
1560
1561                                 # eh?
1562                                 $pref2 = str_replace( ';', ':', $pref );
1563                                 $t = substr( $oLine, $prefixLength );
1564                                 $this->mInPre = !empty($preOpenMatch);
1565                         } else {
1566                                 # Don't interpret any other prefixes in preformatted text
1567                                 $prefixLength = 0;
1568                                 $pref = $pref2 = '';
1569                                 $t = $oLine;
1570                         }
1571
1572                         # List generation
1573                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1574                                 # Same as the last item, so no need to deal with nesting or opening stuff
1575                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1576                                 $paragraphStack = false;
1577
1578                                 if ( substr( $pref, -1 ) == ';') {
1579                                         # The one nasty exception: definition lists work like this:
1580                                         # ; title : definition text
1581                                         # So we check for : in the remainder text to split up the
1582                                         # title and definition, without b0rking links.
1583                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1584                                                 $t = $t2;
1585                                                 $output .= $term . $this->nextItem( ':' );
1586                                         }
1587                                 }
1588                         } elseif( $prefixLength || $lastPrefixLength ) {
1589                                 # Either open or close a level...
1590                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1591                                 $paragraphStack = false;
1592
1593                                 while( $commonPrefixLength < $lastPrefixLength ) {
1594                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1595                                         --$lastPrefixLength;
1596                                 }
1597                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1598                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1599                                 }
1600                                 while ( $prefixLength > $commonPrefixLength ) {
1601                                         $char = substr( $pref, $commonPrefixLength, 1 );
1602                                         $output .= $this->openList( $char );
1603
1604                                         if ( ';' == $char ) {
1605                                                 # FIXME: This is dupe of code above
1606                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1607                                                         $t = $t2;
1608                                                         $output .= $term . $this->nextItem( ':' );
1609                                                 }
1610                                         }
1611                                         ++$commonPrefixLength;
1612                                 }
1613                                 $lastPrefix = $pref2;
1614                         }
1615                         if( 0 == $prefixLength ) {
1616                                 wfProfileIn( "$fname-paragraph" );
1617                                 # No prefix (not in list)--go to paragraph mode
1618                                 $uniq_prefix = UNIQ_PREFIX;
1619                                 // XXX: use a stack for nestable elements like span, table and div
1620                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1621                                 $closematch = preg_match(
1622                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1623                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1624                                 if ( $openmatch or $closematch ) {
1625                                         $paragraphStack = false;
1626                                         $output .= $this->closeParagraph();
1627                                         if($preOpenMatch and !$preCloseMatch) {
1628                                                 $this->mInPre = true;
1629                                         }
1630                                         if ( $closematch ) {
1631                                                 $inBlockElem = false;
1632                                         } else {
1633                                                 $inBlockElem = true;
1634                                         }
1635                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1636                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1637                                                 // pre
1638                                                 if ($this->mLastSection != 'pre') {
1639                                                         $paragraphStack = false;
1640                                                         $output .= $this->closeParagraph().'<pre>';
1641                                                         $this->mLastSection = 'pre';
1642                                                 }
1643                                                 $t = substr( $t, 1 );
1644                                         } else {
1645                                                 // paragraph
1646                                                 if ( '' == trim($t) ) {
1647                                                         if ( $paragraphStack ) {
1648                                                                 $output .= $paragraphStack.'<br />';
1649                                                                 $paragraphStack = false;
1650                                                                 $this->mLastSection = 'p';
1651                                                         } else {
1652                                                                 if ($this->mLastSection != 'p' ) {
1653                                                                         $output .= $this->closeParagraph();
1654                                                                         $this->mLastSection = '';
1655                                                                         $paragraphStack = '<p>';
1656                                                                 } else {
1657                                                                         $paragraphStack = '</p><p>';
1658                                                                 }
1659                                                         }
1660                                                 } else {
1661                                                         if ( $paragraphStack ) {
1662                                                                 $output .= $paragraphStack;
1663                                                                 $paragraphStack = false;
1664                                                                 $this->mLastSection = 'p';
1665                                                         } else if ($this->mLastSection != 'p') {
1666                                                                 $output .= $this->closeParagraph().'<p>';
1667                                                                 $this->mLastSection = 'p';
1668                                                         }
1669                                                 }
1670                                         }
1671                                 }
1672                                 wfProfileOut( "$fname-paragraph" );
1673                         }
1674                         if ($paragraphStack === false) {
1675                                 $output .= $t."\n";
1676                         }
1677                 }
1678                 while ( $prefixLength ) {
1679                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1680                         --$prefixLength;
1681                 }
1682                 if ( '' != $this->mLastSection ) {
1683                         $output .= '</' . $this->mLastSection . '>';
1684                         $this->mLastSection = '';
1685                 }
1686
1687                 wfProfileOut( $fname );
1688                 return $output;
1689         }
1690
1691         /**
1692          * Split up a string on ':', ignoring any occurences inside
1693          * <a>..</a> or <span>...</span>
1694          * @param string $str the string to split
1695          * @param string &$before set to everything before the ':'
1696          * @param string &$after set to everything after the ':'
1697          * return string the position of the ':', or false if none found
1698          */
1699         function findColonNoLinks($str, &$before, &$after) {
1700                 # I wonder if we should make this count all tags, not just <a>
1701                 # and <span>. That would prevent us from matching a ':' that
1702                 # comes in the middle of italics other such formatting....
1703                 # -- Wil
1704                 $fname = 'Parser::findColonNoLinks';
1705                 wfProfileIn( $fname );
1706                 $pos = 0;
1707                 do {
1708                         $colon = strpos($str, ':', $pos);
1709
1710                         if ($colon !== false) {
1711                                 $before = substr($str, 0, $colon);
1712                                 $after = substr($str, $colon + 1);
1713
1714                                 # Skip any ':' within <a> or <span> pairs
1715                                 $a = substr_count($before, '<a');
1716                                 $s = substr_count($before, '<span');
1717                                 $ca = substr_count($before, '</a>');
1718                                 $cs = substr_count($before, '</span>');
1719
1720                                 if ($a <= $ca and $s <= $cs) {
1721                                         # Tags are balanced before ':'; ok
1722                                         break;
1723                                 }
1724                                 $pos = $colon + 1;
1725                         }
1726                 } while ($colon !== false);
1727                 wfProfileOut( $fname );
1728                 return $colon;
1729         }
1730
1731         /**
1732          * Return value of a magic variable (like PAGENAME)
1733          *
1734          * @access private
1735          */
1736         function getVariableValue( $index ) {
1737                 global $wgContLang, $wgSitename, $wgServer;
1738
1739                 /**
1740                  * Some of these require message or data lookups and can be
1741                  * expensive to check many times.
1742                  */
1743                 static $varCache = array();
1744                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1745
1746                 switch ( $index ) {
1747                         case MAG_CURRENTMONTH:
1748                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1749                         case MAG_CURRENTMONTHNAME:
1750                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1751                         case MAG_CURRENTMONTHNAMEGEN:
1752                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1753                         case MAG_CURRENTDAY:
1754                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1755                         case MAG_PAGENAME:
1756                                 return $this->mTitle->getText();
1757                         case MAG_PAGENAMEE:
1758                                 return $this->mTitle->getPartialURL();
1759                         case MAG_NAMESPACE:
1760                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1761                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1762                         case MAG_CURRENTDAYNAME:
1763                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1764                         case MAG_CURRENTYEAR:
1765                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1766                         case MAG_CURRENTTIME:
1767                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1768                         case MAG_NUMBEROFARTICLES:
1769                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1770                         case MAG_SITENAME:
1771                                 return $wgSitename;
1772                         case MAG_SERVER:
1773                                 return $wgServer;
1774                         default:
1775                                 return NULL;
1776                 }
1777         }
1778
1779         /**
1780          * initialise the magic variables (like CURRENTMONTHNAME)
1781          *
1782          * @access private
1783          */
1784         function initialiseVariables() {
1785                 $fname = 'Parser::initialiseVariables';
1786                 wfProfileIn( $fname );
1787                 global $wgVariableIDs;
1788                 $this->mVariables = array();
1789                 foreach ( $wgVariableIDs as $id ) {
1790                         $mw =& MagicWord::get( $id );
1791                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1792                 }
1793                 wfProfileOut( $fname );
1794         }
1795
1796         /**
1797          * Replace magic variables, templates, and template arguments
1798          * with the appropriate text. Templates are substituted recursively,
1799          * taking care to avoid infinite loops.
1800          *
1801          * Note that the substitution depends on value of $mOutputType:
1802          *  OT_WIKI: only {{subst:}} templates
1803          *  OT_MSG: only magic variables
1804          *  OT_HTML: all templates and magic variables
1805          *
1806          * @param string $tex The text to transform
1807          * @param array $args Key-value pairs representing template parameters to substitute
1808          * @access private
1809          */
1810         function replaceVariables( $text, $args = array() ) {
1811                 global $wgLang, $wgScript, $wgArticlePath;
1812
1813                 # Prevent too big inclusions
1814                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1815                         return $text;
1816                 }
1817
1818                 $fname = 'Parser::replaceVariables';
1819                 wfProfileIn( $fname );
1820
1821                 $titleChars = Title::legalChars();
1822
1823                 # This function is called recursively. To keep track of arguments we need a stack:
1824                 array_push( $this->mArgStack, $args );
1825
1826                 # Variable substitution
1827                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1828
1829                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1830                         # Argument substitution
1831                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1832                 }
1833                 # Template substitution
1834                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1835                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1836
1837                 array_pop( $this->mArgStack );
1838
1839                 wfProfileOut( $fname );
1840                 return $text;
1841         }
1842
1843         /**
1844          * Replace magic variables
1845          * @access private
1846          */
1847         function variableSubstitution( $matches ) {
1848                 $fname = 'parser::variableSubstitution';
1849                 wfProfileIn( $fname );
1850                 if ( !$this->mVariables ) {
1851                         $this->initialiseVariables();
1852                 }
1853                 $skip = false;
1854                 if ( $this->mOutputType == OT_WIKI ) {
1855                         # Do only magic variables prefixed by SUBST
1856                         $mwSubst =& MagicWord::get( MAG_SUBST );
1857                         if (!$mwSubst->matchStartAndRemove( $matches[1] ))
1858                                 $skip = true;
1859                         # Note that if we don't substitute the variable below,
1860                         # we don't remove the {{subst:}} magic word, in case
1861                         # it is a template rather than a magic variable.
1862                 }
1863                 if ( !$skip && array_key_exists( $matches[1], $this->mVariables ) ) {
1864                         $text = $this->mVariables[$matches[1]];
1865                         $this->mOutput->mContainsOldMagic = true;
1866                 } else {
1867                         $text = $matches[0];
1868                 }
1869                 wfProfileOut( $fname );
1870                 return $text;
1871         }
1872
1873         # Split template arguments
1874         function getTemplateArgs( $argsString ) {
1875                 if ( $argsString === '' ) {
1876                         return array();
1877                 }
1878
1879                 $args = explode( '|', substr( $argsString, 1 ) );
1880
1881                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1882                 # merged with the next arg because the '|' character between belongs
1883                 # to the link syntax and not the template parameter syntax.
1884                 $argc = count($args);
1885                 $i = 0;
1886                 for ( $i = 0; $i < $argc-1; $i++ ) {
1887                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1888                                 $args[$i] .= '|'.$args[$i+1];
1889                                 array_splice($args, $i+1, 1);
1890                                 $i--;
1891                                 $argc--;
1892                         }
1893                 }
1894
1895                 return $args;
1896         }
1897
1898         /**
1899          * Return the text of a template, after recursively
1900          * replacing any variables or templates within the template.
1901          *
1902          * @param array $matches The parts of the template
1903          *  $matches[1]: the title, i.e. the part before the |
1904          *  $matches[2]: the parameters (including a leading |), if  any
1905          * @return string the text of the template
1906          * @access private
1907          */
1908         function braceSubstitution( $matches ) {
1909                 global $wgLinkCache, $wgContLang;
1910                 $fname = 'Parser::braceSubstitution';
1911                 wfProfileIn( $fname );
1912
1913                 $found = false;
1914                 $nowiki = false;
1915                 $noparse = false;
1916
1917                 $title = NULL;
1918
1919                 # Need to know if the template comes at the start of a line,
1920                 # to treat the beginning of the template like the beginning
1921                 # of a line for tables and block-level elements.
1922                 $linestart = $matches[1];
1923
1924                 # $part1 is the bit before the first |, and must contain only title characters
1925                 # $args is a list of arguments, starting from index 0, not including $part1
1926
1927                 $part1 = $matches[2];
1928                 # If the third subpattern matched anything, it will start with |
1929
1930                 $args = $this->getTemplateArgs($matches[3]);
1931                 $argc = count( $args );
1932
1933                 # Don't parse {{{}}} because that's only for template arguments
1934                 if ( $linestart === '{' ) {
1935                         $text = $matches[0];
1936                         $found = true;
1937                         $noparse = true;
1938                 }
1939
1940                 # SUBST
1941                 if ( !$found ) {
1942                         $mwSubst =& MagicWord::get( MAG_SUBST );
1943                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1944                                 # One of two possibilities is true:
1945                                 # 1) Found SUBST but not in the PST phase
1946                                 # 2) Didn't find SUBST and in the PST phase
1947                                 # In either case, return without further processing
1948                                 $text = $matches[0];
1949                                 $found = true;
1950                                 $noparse = true;
1951                         }
1952                 }
1953
1954                 # MSG, MSGNW and INT
1955                 if ( !$found ) {
1956                         # Check for MSGNW:
1957                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1958                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1959                                 $nowiki = true;
1960                         } else {
1961                                 # Remove obsolete MSG:
1962                                 $mwMsg =& MagicWord::get( MAG_MSG );
1963                                 $mwMsg->matchStartAndRemove( $part1 );
1964                         }
1965
1966                         # Check if it is an internal message
1967                         $mwInt =& MagicWord::get( MAG_INT );
1968                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1969                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1970                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1971                                         $found = true;
1972                                 }
1973                         }
1974                 }
1975
1976                 # NS
1977                 if ( !$found ) {
1978                         # Check for NS: (namespace expansion)
1979                         $mwNs = MagicWord::get( MAG_NS );
1980                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1981                                 if ( intval( $part1 ) ) {
1982                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1983                                         $found = true;
1984                                 } else {
1985                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1986                                         if ( !is_null( $index ) ) {
1987                                                 $text = $linestart . $wgContLang->getNsText( $index );
1988                                                 $found = true;
1989                                         }
1990                                 }
1991                         }
1992                 }
1993
1994                 # LOCALURL and LOCALURLE
1995                 if ( !$found ) {
1996                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1997                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1998
1999                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2000                                 $func = 'getLocalURL';
2001                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2002                                 $func = 'escapeLocalURL';
2003                         } else {
2004                                 $func = '';
2005                         }
2006
2007                         if ( $func !== '' ) {
2008                                 $title = Title::newFromText( $part1 );
2009                                 if ( !is_null( $title ) ) {
2010                                         if ( $argc > 0 ) {
2011                                                 $text = $linestart . $title->$func( $args[0] );
2012                                         } else {
2013                                                 $text = $linestart . $title->$func();
2014                                         }
2015                                         $found = true;
2016                                 }
2017                         }
2018                 }
2019
2020                 # GRAMMAR
2021                 if ( !$found && $argc == 1 ) {
2022                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2023                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2024                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2025                                 $found = true;
2026                         }
2027                 }
2028
2029                 # Template table test
2030
2031                 # Did we encounter this template already? If yes, it is in the cache
2032                 # and we need to check for loops.
2033                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2034                         # set $text to cached message.
2035                         $text = $linestart . $this->mTemplates[$part1];
2036                         $found = true;
2037
2038                         # Infinite loop test
2039                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2040                                 $noparse = true;
2041                                 $found = true;
2042                                 $text .= '<!-- WARNING: template loop detected -->';
2043                         }
2044                 }
2045
2046                 # Load from database
2047                 $itcamefromthedatabase = false;
2048                 if ( !$found ) {
2049                         $ns = NS_TEMPLATE;
2050                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2051                         if ($subpage !== '') {
2052                                 $ns = $this->mTitle->getNamespace();
2053                         }
2054                         $title = Title::newFromText( $part1, $ns );
2055                         if ( !is_null( $title ) && !$title->isExternal() ) {
2056                                 # Check for excessive inclusion
2057                                 $dbk = $title->getPrefixedDBkey();
2058                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2059                                         # This should never be reached.
2060                                         $article = new Article( $title );
2061                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2062                                         if ( $articleContent !== false ) {
2063                                                 $found = true;
2064                                                 $text = $linestart . $articleContent;
2065                                                 $itcamefromthedatabase = true;
2066                                         }
2067                                 }
2068
2069                                 # If the title is valid but undisplayable, make a link to it
2070                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2071                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2072                                         $found = true;
2073                                 }
2074
2075                                 # Template cache array insertion
2076                                 $this->mTemplates[$part1] = $text;
2077                         }
2078                 }
2079
2080                 # Recursive parsing, escaping and link table handling
2081                 # Only for HTML output
2082                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2083                         $text = wfEscapeWikiText( $text );
2084                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2085                         # Clean up argument array
2086                         $assocArgs = array();
2087                         $index = 1;
2088                         foreach( $args as $arg ) {
2089                                 $eqpos = strpos( $arg, '=' );
2090                                 if ( $eqpos === false ) {
2091                                         $assocArgs[$index++] = $arg;
2092                                 } else {
2093                                         $name = trim( substr( $arg, 0, $eqpos ) );
2094                                         $value = trim( substr( $arg, $eqpos+1 ) );
2095                                         if ( $value === false ) {
2096                                                 $value = '';
2097                                         }
2098                                         if ( $name !== false ) {
2099                                                 $assocArgs[$name] = $value;
2100                                         }
2101                                 }
2102                         }
2103
2104                         # Add a new element to the templace recursion path
2105                         $this->mTemplatePath[$part1] = 1;
2106
2107                         $text = $this->strip( $text, $this->mStripState );
2108                         $text = $this->removeHTMLtags( $text );
2109                         $text = $this->replaceVariables( $text, $assocArgs );
2110
2111                         # Resume the link cache and register the inclusion as a link
2112                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2113                                 $wgLinkCache->addLinkObj( $title );
2114                         }
2115
2116                         # If the template begins with a table or block-level
2117                         # element, it should be treated as beginning a new line.
2118                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2119                                 $text = "\n" . $text;
2120                         }
2121                 }
2122
2123                 # Empties the template path
2124                 $this->mTemplatePath = array();
2125                 if ( !$found ) {
2126                         wfProfileOut( $fname );
2127                         return $matches[0];
2128                 } else {
2129                         # replace ==section headers==
2130                         # XXX this needs to go away once we have a better parser.
2131                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2132                                 if( !is_null( $title ) )
2133                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2134                                 else
2135                                         $encodedname = base64_encode("");
2136                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2137                                         PREG_SPLIT_DELIM_CAPTURE);
2138                                 $text = '';
2139                                 $nsec = 0;
2140                                 for( $i = 0; $i < count($m); $i += 2 ) {
2141                                         $text .= $m[$i];
2142                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2143                                         $hl = $m[$i + 1];
2144                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2145                                                 $text .= $hl;
2146                                                 continue;
2147                                         }
2148                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2149                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2150                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2151
2152                                         $nsec++;
2153                                 }
2154                         }
2155                 }
2156
2157                 # Empties the template path
2158                 $this->mTemplatePath = array();
2159
2160                 if ( !$found ) {
2161                         wfProfileOut( $fname );
2162                         return $matches[0];
2163                 } else {
2164                         wfProfileOut( $fname );
2165                         return $text;
2166                 }
2167         }
2168
2169         /**
2170          * Triple brace replacement -- used for template arguments
2171          * @access private
2172          */
2173         function argSubstitution( $matches ) {
2174                 $arg = trim( $matches[1] );
2175                 $text = $matches[0];
2176                 $inputArgs = end( $this->mArgStack );
2177
2178                 if ( array_key_exists( $arg, $inputArgs ) ) {
2179                         $text = $inputArgs[$arg];
2180                 }
2181
2182                 return $text;
2183         }
2184
2185         /**
2186          * Returns true if the function is allowed to include this entity
2187          * @access private
2188          */
2189         function incrementIncludeCount( $dbk ) {
2190                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2191                         $this->mIncludeCount[$dbk] = 0;
2192                 }
2193                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2194                         return true;
2195                 } else {
2196                         return false;
2197                 }
2198         }
2199
2200
2201         /**
2202          * Cleans up HTML, removes dangerous tags and attributes, and
2203          * removes HTML comments
2204          * @access private
2205          */
2206         function removeHTMLtags( $text ) {
2207                 global $wgUseTidy, $wgUserHtml;
2208                 $fname = 'Parser::removeHTMLtags';
2209                 wfProfileIn( $fname );
2210
2211                 if( $wgUserHtml ) {
2212                         $htmlpairs = array( # Tags that must be closed
2213                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
2214                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
2215                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
2216                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
2217                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
2218                         );
2219                         $htmlsingle = array(
2220                                 'br', 'hr', 'li', 'dt', 'dd'
2221                         );
2222                         $htmlnest = array( # Tags that can be nested--??
2223                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
2224                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
2225                         );
2226                         $tabletags = array( # Can only appear inside table
2227                                 'td', 'th', 'tr'
2228                         );
2229                 } else {
2230                         $htmlpairs = array();
2231                         $htmlsingle = array();
2232                         $htmlnest = array();
2233                         $tabletags = array();
2234                 }
2235
2236                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
2237                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
2238
2239                 $htmlattrs = $this->getHTMLattrs () ;
2240
2241                 # Remove HTML comments
2242                 $text = $this->removeHTMLcomments( $text );
2243
2244                 $bits = explode( '<', $text );
2245                 $text = array_shift( $bits );
2246                 if(!$wgUseTidy) {
2247                         $tagstack = array(); $tablestack = array();
2248                         foreach ( $bits as $x ) {
2249                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2250                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2251                                 $x, $regs );
2252                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2253                                 error_reporting( $prev );
2254
2255                                 $badtag = 0 ;
2256                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2257                                         # Check our stack
2258                                         if ( $slash ) {
2259                                                 # Closing a tag...
2260                                                 if ( ! in_array( $t, $htmlsingle ) &&
2261                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2262                                                         @array_push( $tagstack, $ot );
2263                                                         $badtag = 1;
2264                                                 } else {
2265                                                         if ( $t == 'table' ) {
2266                                                                 $tagstack = array_pop( $tablestack );
2267                                                         }
2268                                                         $newparams = '';
2269                                                 }
2270                                         } else {
2271                                                 # Keep track for later
2272                                                 if ( in_array( $t, $tabletags ) &&
2273                                                 ! in_array( 'table', $tagstack ) ) {
2274                                                         $badtag = 1;
2275                                                 } else if ( in_array( $t, $tagstack ) &&
2276                                                 ! in_array ( $t , $htmlnest ) ) {
2277                                                         $badtag = 1 ;
2278                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
2279                                                         if ( $t == 'table' ) {
2280                                                                 array_push( $tablestack, $tagstack );
2281                                                                 $tagstack = array();
2282                                                         }
2283                                                         array_push( $tagstack, $t );
2284                                                 }
2285                                                 # Strip non-approved attributes from the tag
2286                                                 $newparams = $this->fixTagAttributes($params);
2287
2288                                         }
2289                                         if ( ! $badtag ) {
2290                                                 $rest = str_replace( '>', '&gt;', $rest );
2291                                                 $text .= "<$slash$t $newparams$brace$rest";
2292                                                 continue;
2293                                         }
2294                                 }
2295                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2296                         }
2297                         # Close off any remaining tags
2298                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2299                                 $text .= "</$t>\n";
2300                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2301                         }
2302                 } else {
2303                         # this might be possible using tidy itself
2304                         foreach ( $bits as $x ) {
2305                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2306                                 $x, $regs );
2307                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2308                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2309                                         $newparams = $this->fixTagAttributes($params);
2310                                         $rest = str_replace( '>', '&gt;', $rest );
2311                                         $text .= "<$slash$t $newparams$brace$rest";
2312                                 } else {
2313                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2314                                 }
2315                         }
2316                 }
2317                 wfProfileOut( $fname );
2318                 return $text;
2319         }
2320
2321         /**
2322          * Remove '<!--', '-->', and everything between.
2323          * To avoid leaving blank lines, when a comment is both preceded
2324          * and followed by a newline (ignoring spaces), trim leading and
2325          * trailing spaces and one of the newlines.
2326          *
2327          * @access private
2328          */
2329         function removeHTMLcomments( $text ) {
2330                 $fname='Parser::removeHTMLcomments';
2331                 wfProfileIn( $fname );
2332                 while (($start = strpos($text, '<!--')) !== false) {
2333                         $end = strpos($text, '-->', $start + 4);
2334                         if ($end === false) {
2335                                 # Unterminated comment; bail out
2336                                 break;
2337                         }
2338
2339                         $end += 3;
2340
2341                         # Trim space and newline if the comment is both
2342                         # preceded and followed by a newline
2343                         $spaceStart = max($start - 1, 0);
2344                         $spaceLen = $end - $spaceStart;
2345                         while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2346                                 $spaceStart--;
2347                                 $spaceLen++;
2348                         }
2349                         while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2350                                 $spaceLen++;
2351                         if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2352                                 # Remove the comment, leading and trailing
2353                                 # spaces, and leave only one newline.
2354                                 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2355                         }
2356                         else {
2357                                 # Remove just the comment.
2358                                 $text = substr_replace($text, '', $start, $end - $start);
2359                         }
2360                 }
2361                 wfProfileOut( $fname );
2362                 return $text;
2363         }
2364
2365         /**
2366          * This function accomplishes several tasks:
2367          * 1) Auto-number headings if that option is enabled
2368          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2369          * 3) Add a Table of contents on the top for users who have enabled the option
2370          * 4) Auto-anchor headings
2371          *
2372          * It loops through all headlines, collects the necessary data, then splits up the
2373          * string and re-inserts the newly formatted headlines.
2374          * @access private
2375          */
2376         /* private */ function formatHeadings( $text, $isMain=true ) {
2377                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders;
2378
2379                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2380                 $doShowToc = $this->mOptions->getShowToc();
2381                 $forceTocHere = false;
2382                 if( !$this->mTitle->userCanEdit() ) {
2383                         $showEditLink = 0;
2384                         $rightClickHack = 0;
2385                 } else {
2386                         $showEditLink = $this->mOptions->getEditSection();
2387                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2388                 }
2389
2390                 # Inhibit editsection links if requested in the page
2391                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2392                 if( $esw->matchAndRemove( $text ) ) {
2393                         $showEditLink = 0;
2394                 }
2395                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2396                 # do not add TOC
2397                 $mw =& MagicWord::get( MAG_NOTOC );
2398                 if( $mw->matchAndRemove( $text ) ) {
2399                         $doShowToc = 0;
2400                 }
2401
2402                 # never add the TOC to the Main Page. This is an entry page that should not
2403                 # be more than 1-2 screens large anyway
2404                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2405                         $doShowToc = 0;
2406                 }
2407
2408                 # Get all headlines for numbering them and adding funky stuff like [edit]
2409                 # links - this is for later, but we need the number of headlines right now
2410                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
2411
2412                 # if there are fewer than 4 headlines in the article, do not show TOC
2413                 if( $numMatches < 4 ) {
2414                         $doShowToc = 0;
2415                 }
2416
2417                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2418                 # override above conditions and always show TOC at that place
2419                 $mw =& MagicWord::get( MAG_TOC );
2420                 if ($mw->match( $text ) ) {
2421                         $doShowToc = 1;
2422                         $forceTocHere = true;
2423                 } else {
2424                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2425                         # override above conditions and always show TOC above first header
2426                         $mw =& MagicWord::get( MAG_FORCETOC );
2427                         if ($mw->matchAndRemove( $text ) ) {
2428                                 $doShowToc = 1;
2429                         }
2430                 }
2431
2432
2433
2434                 # We need this to perform operations on the HTML
2435                 $sk =& $this->mOptions->getSkin();
2436
2437                 # headline counter
2438                 $headlineCount = 0;
2439                 $sectionCount = 0; # headlineCount excluding template sections
2440
2441                 # Ugh .. the TOC should have neat indentation levels which can be
2442                 # passed to the skin functions. These are determined here
2443                 $toclevel = 0;
2444                 $toc = '';
2445                 $full = '';
2446                 $head = array();
2447                 $sublevelCount = array();
2448                 $level = 0;
2449                 $prevlevel = 0;
2450                 foreach( $matches[3] as $headline ) {
2451                         $istemplate = 0;
2452                         $templatetitle = "";
2453                         $templatesection = 0;
2454
2455                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2456                                 $istemplate = 1;
2457                                 $templatetitle = base64_decode($mat[1]);
2458                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2459                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2460                         }
2461
2462                         $numbering = '';
2463                         if( $level ) {
2464                                 $prevlevel = $level;
2465                         }
2466                         $level = $matches[1][$headlineCount];
2467                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2468                                 # reset when we enter a new level
2469                                 $sublevelCount[$level] = 0;
2470                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2471                                 $toclevel += $level - $prevlevel;
2472                         }
2473                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2474                                 # reset when we step back a level
2475                                 $sublevelCount[$level+1]=0;
2476                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2477                                 $toclevel -= $prevlevel - $level;
2478                         }
2479                         # count number of headlines for each level
2480                         @$sublevelCount[$level]++;
2481                         if( $doNumberHeadings || $doShowToc ) {
2482                                 $dot = 0;
2483                                 for( $i = 1; $i <= $level; $i++ ) {
2484                                         if( !empty( $sublevelCount[$i] ) ) {
2485                                                 if( $dot ) {
2486                                                         $numbering .= '.';
2487                                                 }
2488                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2489                                                 $dot = 1;
2490                                         }
2491                                 }
2492                         }
2493
2494                         # The canonized header is a version of the header text safe to use for links
2495                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2496                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2497                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2498
2499                         # Remove link placeholders by the link text.
2500                         #     <!--LINK number-->
2501                         # turns into
2502                         #     link text with suffix
2503                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2504                                                             "\$wgLinkHolders['texts'][\$1]",
2505                                                             $canonized_headline );
2506
2507                         # strip out HTML
2508                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2509                         $tocline = trim( $canonized_headline );
2510                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2511                         $replacearray = array(
2512                                 '%3A' => ':',
2513                                 '%' => '.'
2514                         );
2515                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2516                         $refer[$headlineCount] = $canonized_headline;
2517
2518                         # count how many in assoc. array so we can track dupes in anchors
2519                         @$refers[$canonized_headline]++;
2520                         $refcount[$headlineCount]=$refers[$canonized_headline];
2521
2522                         # Prepend the number to the heading text
2523
2524                         if( $doNumberHeadings || $doShowToc ) {
2525                                 $tocline = $numbering . ' ' . $tocline;
2526
2527                                 # Don't number the heading if it is the only one (looks silly)
2528                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2529                                         # the two are different if the line contains a link
2530                                         $headline=$numbering . ' ' . $headline;
2531                                 }
2532                         }
2533
2534                         # Create the anchor for linking from the TOC to the section
2535                         $anchor = $canonized_headline;
2536                         if($refcount[$headlineCount] > 1 ) {
2537                                 $anchor .= '_' . $refcount[$headlineCount];
2538                         }
2539                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2540                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2541                         }
2542                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2543                                 if ( empty( $head[$headlineCount] ) ) {
2544                                         $head[$headlineCount] = '';
2545                                 }
2546                                 if( $istemplate )
2547                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2548                                 else
2549                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2550                         }
2551
2552                         # Add the edit section span
2553                         if( $rightClickHack ) {
2554                                 if( $istemplate )
2555                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2556                                 else
2557                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2558                         }
2559
2560                         # give headline the correct <h#> tag
2561                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2562
2563                         $headlineCount++;
2564                         if( !$istemplate )
2565                                 $sectionCount++;
2566                 }
2567
2568                 if( $doShowToc ) {
2569                         $toclines = $headlineCount;
2570                         $toc .= $sk->tocUnindent( $toclevel );
2571                         $toc = $sk->tocTable( $toc );
2572                 }
2573
2574                 # split up and insert constructed headlines
2575
2576                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2577                 $i = 0;
2578
2579                 foreach( $blocks as $block ) {
2580                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2581                                 # This is the [edit] link that appears for the top block of text when
2582                                 # section editing is enabled
2583
2584                                 # Disabled because it broke block formatting
2585                                 # For example, a bullet point in the top line
2586                                 # $full .= $sk->editSectionLink(0);
2587                         }
2588                         $full .= $block;
2589                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2590                         # Top anchor now in skin
2591                                 $full = $full.$toc;
2592                         }
2593
2594                         if( !empty( $head[$i] ) ) {
2595                                 $full .= $head[$i];
2596                         }
2597                         $i++;
2598                 }
2599                 if($forceTocHere) {
2600                         $mw =& MagicWord::get( MAG_TOC );
2601                         return $mw->replace( $toc, $full );
2602                 } else {
2603                         return $full;
2604                 }
2605         }
2606
2607         /**
2608          * Return an HTML link for the "ISBN 123456" text
2609          * @access private
2610          */
2611         function magicISBN( $text ) {
2612                 global $wgLang;
2613                 $fname = 'Parser::magicISBN';
2614                 wfProfileIn( $fname );
2615
2616                 $a = split( 'ISBN ', ' '.$text );
2617                 if ( count ( $a ) < 2 ) {
2618                         wfProfileOut( $fname );
2619                         return $text;
2620                 }
2621                 $text = substr( array_shift( $a ), 1);
2622                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2623
2624                 foreach ( $a as $x ) {
2625                         $isbn = $blank = '' ;
2626                         while ( ' ' == $x{0} ) {
2627                                 $blank .= ' ';
2628                                 $x = substr( $x, 1 );
2629                         }
2630                         if ( $x == '' ) { # blank isbn
2631                                 $text .= "ISBN $blank";
2632                                 continue;
2633                         }
2634                         while ( strstr( $valid, $x{0} ) != false ) {
2635                                 $isbn .= $x{0};
2636                                 $x = substr( $x, 1 );
2637                         }
2638                         $num = str_replace( '-', '', $isbn );
2639                         $num = str_replace( ' ', '', $num );
2640
2641                         if ( '' == $num ) {
2642                                 $text .= "ISBN $blank$x";
2643                         } else {
2644                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2645                                 $text .= '<a href="' .
2646                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2647                                         "\" class=\"internal\">ISBN $isbn</a>";
2648                                 $text .= $x;
2649                         }
2650                 }
2651                 wfProfileOut( $fname );
2652                 return $text;
2653         }
2654
2655         /**
2656          * Return an HTML link for the "GEO ..." text
2657          * @access private
2658          */
2659         function magicGEO( $text ) {
2660                 global $wgLang, $wgUseGeoMode;
2661                 $fname = 'Parser::magicGEO';
2662                 wfProfileIn( $fname );
2663
2664                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2665                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2666                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2667                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2668                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2669                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2670
2671                 $a = split( 'GEO ', ' '.$text );
2672                 if ( count ( $a ) < 2 ) {
2673                         wfProfileOut( $fname );
2674                         return $text;
2675                 }
2676                 $text = substr( array_shift( $a ), 1);
2677                 $valid = '0123456789.+-:';
2678
2679                 foreach ( $a as $x ) {
2680                         $geo = $blank = '' ;
2681                         while ( ' ' == $x{0} ) {
2682                                 $blank .= ' ';
2683                                 $x = substr( $x, 1 );
2684                         }
2685                         while ( strstr( $valid, $x{0} ) != false ) {
2686                                 $geo .= $x{0};
2687                                 $x = substr( $x, 1 );
2688                         }
2689                         $num = str_replace( '+', '', $geo );
2690                         $num = str_replace( ' ', '', $num );
2691
2692                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2693                                 $text .= "GEO $blank$x";
2694                         } else {
2695                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2696                                 $text .= '<a href="' .
2697                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2698                                         "\" class=\"internal\">GEO $geo</a>";
2699                                 $text .= $x;
2700                         }
2701                 }
2702                 wfProfileOut( $fname );
2703                 return $text;
2704         }
2705
2706         /**
2707          * Return an HTML link for the "RFC 1234" text
2708          * @access private
2709          * @param string $text text to be processed
2710          */
2711         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2712                 global $wgLang;
2713
2714                 $valid = '0123456789';
2715                 $internal = false;
2716
2717                 $a = split( $keyword, ' '.$text );
2718                 if ( count ( $a ) < 2 ) {
2719                         return $text;
2720                 }
2721                 $text = substr( array_shift( $a ), 1);
2722
2723                 /* Check if keyword is preceed by [[.
2724                  * This test is made here cause of the array_shift above
2725                  * that prevent the test to be done in the foreach.
2726                  */
2727                 if ( substr( $text, -2 ) == '[[' ) {
2728                         $internal = true;
2729                 }
2730
2731                 foreach ( $a as $x ) {
2732                         /* token might be empty if we have RFC RFC 1234 */
2733                         if ( $x=='' ) {
2734                                 $text.=$keyword;
2735                                 continue;
2736                                 }
2737
2738                         $id = $blank = '' ;
2739
2740                         /** remove and save whitespaces in $blank */
2741                         while ( $x{0} == ' ' ) {
2742                                 $blank .= ' ';
2743                                 $x = substr( $x, 1 );
2744                         }
2745
2746                         /** remove and save the rfc number in $id */
2747                         while ( strstr( $valid, $x{0} ) != false ) {
2748                                 $id .= $x{0};
2749                                 $x = substr( $x, 1 );
2750                         }
2751
2752                         if ( $id == '' ) {
2753                                 /* call back stripped spaces*/
2754                                 $text .= $keyword.$blank.$x;
2755                         } elseif( $internal ) {
2756                                 /* normal link */
2757                                 $text .= $keyword.$id.$x;
2758                         } else {
2759                                 /* build the external link*/
2760                                 $url = wfmsg( $urlmsg );
2761                                 $url = str_replace( '$1', $id, $url);
2762                                 $sk =& $this->mOptions->getSkin();
2763                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2764                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2765                         }
2766
2767                         /* Check if the next RFC keyword is preceed by [[ */
2768                         $internal = ( substr($x,-2) == '[[' );
2769                 }
2770                 return $text;
2771         }
2772
2773         /**
2774          * Transform wiki markup when saving a page by doing \r\n -> \n
2775          * conversion, substitting signatures, {{subst:}} templates, etc.
2776          *
2777          * @param string $text the text to transform
2778          * @param Title &$title the Title object for the current article
2779          * @param User &$user the User object describing the current user
2780          * @param ParserOptions $options parsing options
2781          * @param bool $clearState whether to clear the parser state first
2782          * @return string the altered wiki markup
2783          * @access public
2784          */
2785         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2786                 $this->mOptions = $options;
2787                 $this->mTitle =& $title;
2788                 $this->mOutputType = OT_WIKI;
2789
2790                 if ( $clearState ) {
2791                         $this->clearState();
2792                 }
2793
2794                 $stripState = false;
2795                 $pairs = array(
2796                         "\r\n" => "\n",
2797                         );
2798                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2799                 $text = $this->strip( $text, $stripState, false );
2800                 $text = $this->pstPass2( $text, $user );
2801                 $text = $this->unstrip( $text, $stripState );
2802                 $text = $this->unstripNoWiki( $text, $stripState );
2803                 return $text;
2804         }
2805
2806         /**
2807          * Pre-save transform helper function
2808          * @access private
2809          */
2810         function pstPass2( $text, &$user ) {
2811                 global $wgLang, $wgContLang, $wgLocaltimezone;
2812
2813                 # Variable replacement
2814                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2815                 $text = $this->replaceVariables( $text );
2816
2817                 # Signatures
2818                 #
2819                 $n = $user->getName();
2820                 $k = $user->getOption( 'nickname' );
2821                 if ( '' == $k ) { $k = $n; }
2822                 if ( isset( $wgLocaltimezone ) ) {
2823                         $oldtz = getenv( 'TZ' );
2824                         putenv( 'TZ='.$wgLocaltimezone );
2825                 }
2826                 /* Note: this is an ugly timezone hack for the European wikis */
2827                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2828                   ' (' . date( 'T' ) . ')';
2829                 if ( isset( $wgLocaltimezone ) ) {
2830                         putenv( 'TZ='.$oldtzs );
2831                 }
2832
2833                 $text = preg_replace( '/~~~~~~/', $d, $text );
2834                 $text = preg_replace( '/~~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2835                 $text = preg_replace( '/~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2836
2837                 # Context links: [[|name]] and [[name (context)|]]
2838                 #
2839                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2840                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2841                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2842                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2843
2844                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2845                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2846                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2847                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2848                 $context = '';
2849                 $t = $this->mTitle->getText();
2850                 if ( preg_match( $conpat, $t, $m ) ) {
2851                         $context = $m[2];
2852                 }
2853                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2854                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2855                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2856
2857                 if ( '' == $context ) {
2858                         $text = preg_replace( $p2, '[[\\1]]', $text );
2859                 } else {
2860                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2861                 }
2862
2863                 # Trim trailing whitespace
2864                 # MAG_END (__END__) tag allows for trailing
2865                 # whitespace to be deliberately included
2866                 $text = rtrim( $text );
2867                 $mw =& MagicWord::get( MAG_END );
2868                 $mw->matchAndRemove( $text );
2869
2870                 return $text;
2871         }
2872
2873         /**
2874          * Set up some variables which are usually set up in parse()
2875          * so that an external function can call some class members with confidence
2876          * @access public
2877          */
2878         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2879                 $this->mTitle =& $title;
2880                 $this->mOptions = $options;
2881                 $this->mOutputType = $outputType;
2882                 if ( $clearState ) {
2883                         $this->clearState();
2884                 }
2885         }
2886
2887         /**
2888          * Transform a MediaWiki message by replacing magic variables.
2889          *
2890          * @param string $text the text to transform
2891          * @param ParserOptions $options  options
2892          * @return string the text with variables substituted
2893          * @access public
2894          */
2895         function transformMsg( $text, $options ) {
2896                 global $wgTitle;
2897                 static $executing = false;
2898
2899                 # Guard against infinite recursion
2900                 if ( $executing ) {
2901                         return $text;
2902                 }
2903                 $executing = true;
2904
2905                 $this->mTitle = $wgTitle;
2906                 $this->mOptions = $options;
2907                 $this->mOutputType = OT_MSG;
2908                 $this->clearState();
2909                 $text = $this->replaceVariables( $text );
2910
2911                 $executing = false;
2912                 return $text;
2913         }
2914
2915         /**
2916          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2917          * Callback will be called with the text within
2918          * Transform and return the text within
2919          * @access public
2920          */
2921         function setHook( $tag, $callback ) {
2922                 $oldVal = @$this->mTagHooks[$tag];
2923                 $this->mTagHooks[$tag] = $callback;
2924                 return $oldVal;
2925         }
2926
2927         /**
2928          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2929          * Placeholders created in Skin::makeLinkObj()
2930          * Returns an array of links found, indexed by PDBK:
2931          *  0 - broken
2932          *  1 - normal link
2933          *  2 - stub
2934          * $options is a bit field, RLH_FOR_UPDATE to select for update
2935          */
2936         function replaceLinkHolders( &$text, $options = 0 ) {
2937                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2938                 global $wgInterwikiLinkHolders;
2939                 global $outputReplace;
2940
2941                 if ( $wgUseOldExistenceCheck ) {
2942                         return array();
2943                 }
2944
2945                 $fname = 'Parser::replaceLinkHolders';
2946                 wfProfileIn( $fname );
2947
2948                 $pdbks = array();
2949                 $colours = array();
2950
2951                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2952                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2953                         wfProfileIn( $fname.'-check' );
2954                         $dbr =& wfGetDB( DB_SLAVE );
2955                         $cur = $dbr->tableName( 'cur' );
2956                         $sk = $wgUser->getSkin();
2957                         $threshold = $wgUser->getOption('stubthreshold');
2958
2959                         # Sort by namespace
2960                         asort( $wgLinkHolders['namespaces'] );
2961
2962                         # Generate query
2963                         $query = false;
2964                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2965                                 # Make title object
2966                                 $title = $wgLinkHolders['titles'][$key];
2967
2968                                 # Skip invalid entries.
2969                                 # Result will be ugly, but prevents crash.
2970                                 if ( is_null( $title ) ) {
2971                                         continue;
2972                                 }
2973                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2974
2975                                 # Check if it's in the link cache already
2976                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2977                                         $colours[$pdbk] = 1;
2978                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2979                                         $colours[$pdbk] = 0;
2980                                 } else {
2981                                         # Not in the link cache, add it to the query
2982                                         if ( !isset( $current ) ) {
2983                                                 $current = $val;
2984                                                 $query =  "SELECT cur_id, cur_namespace, cur_title";
2985                                                 if ( $threshold > 0 ) {
2986                                                         $query .= ", LENGTH(cur_text) AS cur_len, cur_is_redirect";
2987                                                 }
2988                                                 $query .= " FROM $cur WHERE (cur_namespace=$val AND cur_title IN(";
2989                                         } elseif ( $current != $val ) {
2990                                                 $current = $val;
2991                                                 $query .= ")) OR (cur_namespace=$val AND cur_title IN(";
2992                                         } else {
2993                                                 $query .= ', ';
2994                                         }
2995
2996                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2997                                 }
2998                         }
2999                         if ( $query ) {
3000                                 $query .= '))';
3001                                 if ( $options & RLH_FOR_UPDATE ) {
3002                                         $query .= ' FOR UPDATE';
3003                                 }
3004
3005                                 $res = $dbr->query( $query, $fname );
3006
3007                                 # Fetch data and form into an associative array
3008                                 # non-existent = broken
3009                                 # 1 = known
3010                                 # 2 = stub
3011                                 while ( $s = $dbr->fetchObject($res) ) {
3012                                         $title = Title::makeTitle( $s->cur_namespace, $s->cur_title );
3013                                         $pdbk = $title->getPrefixedDBkey();
3014                                         $wgLinkCache->addGoodLink( $s->cur_id, $pdbk );
3015
3016                                         if ( $threshold >  0 ) {
3017                                                 $size = $s->cur_len;
3018                                                 if ( $s->cur_is_redirect || $s->cur_namespace != 0 || $length < $threshold ) {
3019                                                         $colours[$pdbk] = 1;
3020                                                 } else {
3021                                                         $colours[$pdbk] = 2;
3022                                                 }
3023                                         } else {
3024                                                 $colours[$pdbk] = 1;
3025                                         }
3026                                 }
3027                         }
3028                         wfProfileOut( $fname.'-check' );
3029
3030                         # Construct search and replace arrays
3031                         wfProfileIn( $fname.'-construct' );
3032                         $outputReplace = array();
3033                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
3034                                 $pdbk = $pdbks[$key];
3035                                 $searchkey = '<!--LINK '.$key.'-->';
3036                                 $title = $wgLinkHolders['titles'][$key];
3037                                 if ( empty( $colours[$pdbk] ) ) {
3038                                         $wgLinkCache->addBadLink( $pdbk );
3039                                         $colours[$pdbk] = 0;
3040                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3041                                                                         $wgLinkHolders['texts'][$key],
3042                                                                         $wgLinkHolders['queries'][$key] );
3043                                 } elseif ( $colours[$pdbk] == 1 ) {
3044                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3045                                                                         $wgLinkHolders['texts'][$key],
3046                                                                         $wgLinkHolders['queries'][$key] );
3047                                 } elseif ( $colours[$pdbk] == 2 ) {
3048                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3049                                                                         $wgLinkHolders['texts'][$key],
3050                                                                         $wgLinkHolders['queries'][$key] );
3051                                 }
3052                         }
3053                         wfProfileOut( $fname.'-construct' );
3054
3055                         # Do the thing
3056                         wfProfileIn( $fname.'-replace' );
3057
3058                         $text = preg_replace_callback(
3059                                 '/(<!--LINK .*?-->)/',
3060                                 "outputReplaceMatches",
3061                                 $text);
3062                         wfProfileOut( $fname.'-replace' );
3063                 }
3064
3065                 if ( !empty( $wgInterwikiLinkHolders ) ) {
3066                         wfProfileIn( $fname.'-interwiki' );
3067                         $outputReplace = $wgInterwikiLinkHolders;
3068                         $text = preg_replace_callback(
3069                                 '/<!--IWLINK (.*?)-->/',
3070                                 "outputReplaceMatches",
3071                                 $text );
3072                         wfProfileOut( $fname.'-interwiki' );
3073                 }
3074
3075                 wfProfileOut( $fname );
3076                 return $colours;
3077         }
3078
3079         /**
3080          * Renders an image gallery from a text with one line per image.
3081          * text labels may be given by using |-style alternative text. E.g.
3082          *   Image:one.jpg|The number "1"
3083          *   Image:tree.jpg|A tree
3084          * given as text will return the HTML of a gallery with two images,
3085          * labeled 'The number "1"' and
3086          * 'A tree'.
3087          */
3088         function renderImageGallery( $text ) {
3089                 global $wgLinkCache;
3090                 $ig = new ImageGallery();
3091                 $ig->setShowBytes( false );
3092                 $ig->setShowFilename( false );
3093                 $lines = explode( "\n", $text );
3094
3095                 foreach ( $lines as $line ) {
3096                         # match lines like these:
3097                         # Image:someimage.jpg|This is some image
3098                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3099                         # Skip empty lines
3100                         if ( count( $matches ) == 0 ) {
3101                                 continue;
3102                         }
3103                         $nt = Title::newFromURL( $matches[1] );
3104                         if ( isset( $matches[3] ) ) {
3105                                 $label = $matches[3];
3106                         } else {
3107                                 $label = '';
3108                         }
3109                         $ig->add( Image::newFromTitle( $nt ), $label );
3110                         $wgLinkCache->addImageLinkObj( $nt );
3111                 }
3112                 return $ig->toHTML();
3113         }
3114 }
3115
3116 /**
3117  * @todo document
3118  * @package MediaWiki
3119  */
3120 class ParserOutput
3121 {
3122         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3123         var $mCacheTime; # Used in ParserCache
3124         var $mVersion;   # Compatibility check
3125
3126         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3127                 $containsOldMagic = false )
3128         {
3129                 $this->mText = $text;
3130                 $this->mLanguageLinks = $languageLinks;
3131                 $this->mCategoryLinks = $categoryLinks;
3132                 $this->mContainsOldMagic = $containsOldMagic;
3133                 $this->mCacheTime = '';
3134                 $this->mVersion = MW_PARSER_VERSION;
3135         }
3136
3137         function getText()                   { return $this->mText; }
3138         function getLanguageLinks()          { return $this->mLanguageLinks; }
3139         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
3140         function getCacheTime()              { return $this->mCacheTime; }
3141         function containsOldMagic()          { return $this->mContainsOldMagic; }
3142         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3143         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3144         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3145         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3146         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3147         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3148
3149         function merge( $other ) {
3150                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3151                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3152                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3153         }
3154
3155         /**
3156          * Return true if this cached output object predates the global or
3157          * per-article cache invalidation timestamps, or if it comes from
3158          * an incompatible older version.
3159          *
3160          * @param string $touched the affected article's last touched timestamp
3161          * @return bool
3162          * @access public
3163          */
3164         function expired( $touched ) {
3165                 global $wgCacheEpoch;
3166                 return $this->getCacheTime() <= $touched ||
3167                        $this->getCacheTime() <= $wgCacheEpoch ||
3168                        !isset( $this->mVersion ) ||
3169                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3170         }
3171 }
3172
3173 /**
3174  * Set options of the Parser
3175  * @todo document
3176  * @package MediaWiki
3177  */
3178 class ParserOptions
3179 {
3180         # All variables are private
3181         var $mUseTeX;                    # Use texvc to expand <math> tags
3182         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3183         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3184         var $mAllowExternalImages;       # Allow external images inline
3185         var $mSkin;                      # Reference to the preferred skin
3186         var $mDateFormat;                # Date format index
3187         var $mEditSection;               # Create "edit section" links
3188         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3189         var $mNumberHeadings;            # Automatically number headings
3190         var $mShowToc;                   # Show table of contents
3191
3192         function getUseTeX()                        { return $this->mUseTeX; }
3193         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3194         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3195         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3196         function getSkin()                          { return $this->mSkin; }
3197         function getDateFormat()                    { return $this->mDateFormat; }
3198         function getEditSection()                   { return $this->mEditSection; }
3199         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3200         function getNumberHeadings()                { return $this->mNumberHeadings; }
3201         function getShowToc()                       { return $this->mShowToc; }
3202
3203         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3204         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3205         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3206         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3207         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3208         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3209         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3210         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3211         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3212
3213         function setSkin( &$x ) { $this->mSkin =& $x; }
3214
3215         # Get parser options
3216         /* static */ function newFromUser( &$user ) {
3217                 $popts = new ParserOptions;
3218                 $popts->initialiseFromUser( $user );
3219                 return $popts;
3220         }
3221
3222         # Get user options
3223         function initialiseFromUser( &$userInput ) {
3224                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3225                 $fname = 'ParserOptions::initialiseFromUser';
3226                 wfProfileIn( $fname );
3227                 if ( !$userInput ) {
3228                         $user = new User;
3229                         $user->setLoaded( true );
3230                 } else {
3231                         $user =& $userInput;
3232                 }
3233
3234                 $this->mUseTeX = $wgUseTeX;
3235                 $this->mUseDynamicDates = $wgUseDynamicDates;
3236                 $this->mInterwikiMagic = $wgInterwikiMagic;
3237                 $this->mAllowExternalImages = $wgAllowExternalImages;
3238                 wfProfileIn( $fname.'-skin' );
3239                 $this->mSkin =& $user->getSkin();
3240                 wfProfileOut( $fname.'-skin' );
3241                 $this->mDateFormat = $user->getOption( 'date' );
3242                 $this->mEditSection = $user->getOption( 'editsection' );
3243                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3244                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3245                 $this->mShowToc = $user->getOption( 'showtoc' );
3246                 wfProfileOut( $fname );
3247         }
3248
3249
3250 }
3251
3252 /**
3253  * Callback function used by Parser::replaceLinkHolders()
3254  * to substitute link placeholders.
3255  */
3256 function &outputReplaceMatches( $matches ) {
3257         global $outputReplace;
3258         return $outputReplace[$matches[1]];
3259 }
3260
3261 /**
3262  * Return the total number of articles
3263  */
3264 function wfNumberOfArticles() {
3265         global $wgNumberOfArticles;
3266
3267         wfLoadSiteStats();
3268         return $wgNumberOfArticles;
3269 }
3270
3271 /**
3272  * Get various statistics from the database
3273  * @private
3274  */
3275 function wfLoadSiteStats() {
3276         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3277         $fname = 'wfLoadSiteStats';
3278
3279         if ( -1 != $wgNumberOfArticles ) return;
3280         $dbr =& wfGetDB( DB_SLAVE );
3281         $s = $dbr->selectRow( 'site_stats',
3282                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3283                 array( 'ss_row_id' => 1 ), $fname
3284         );
3285
3286         if ( $s === false ) {
3287                 return;
3288         } else {
3289                 $wgTotalViews = $s->ss_total_views;
3290                 $wgTotalEdits = $s->ss_total_edits;
3291                 $wgNumberOfArticles = $s->ss_good_articles;
3292         }
3293 }
3294
3295 function wfEscapeHTMLTagsOnly( $in ) {
3296         return str_replace(
3297                 array( '"', '>', '<' ),
3298                 array( '&quot;', '&gt;', '&lt;' ),
3299                 $in );
3300 }
3301
3302 ?>