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
1397                 $fname = 'Parser::maybeDoSubpageLink';
1398                 wfProfileIn( $fname );
1399                 # Look at the first character
1400                 if( $target != '' && $target{0} == '/' ) {
1401                         # / at end means we don't want the slash to be shown
1402                         if(substr($target,-1,1)=='/') {
1403                                 $target=substr($target,1,-1);
1404                                 $noslash=$target;
1405                         } else {
1406                                 $noslash=substr($target,1);
1407                         }
1408
1409                         # Some namespaces don't allow subpages
1410                         if( $this->areSubpagesAllowed() ) {
1411                                 # subpages allowed here
1412                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1413                                 if( '' === $text ) {
1414                                         $text = $target;
1415                                 } # this might be changed for ugliness reasons
1416                         } else {
1417                                 # no subpage allowed, use standard link
1418                                 $ret = $target;
1419                         }
1420                 } else {
1421                         # no subpage
1422                         $ret = $target;
1423                 }
1424
1425                 wfProfileOut( $fname );
1426                 return $ret;
1427         }
1428
1429         /**#@+
1430          * Used by doBlockLevels()
1431          * @access private
1432          */
1433         /* private */ function closeParagraph() {
1434                 $result = '';
1435                 if ( '' != $this->mLastSection ) {
1436                         $result = '</' . $this->mLastSection  . ">\n";
1437                 }
1438                 $this->mInPre = false;
1439                 $this->mLastSection = '';
1440                 return $result;
1441         }
1442         # getCommon() returns the length of the longest common substring
1443         # of both arguments, starting at the beginning of both.
1444         #
1445         /* private */ function getCommon( $st1, $st2 ) {
1446                 $fl = strlen( $st1 );
1447                 $shorter = strlen( $st2 );
1448                 if ( $fl < $shorter ) { $shorter = $fl; }
1449
1450                 for ( $i = 0; $i < $shorter; ++$i ) {
1451                         if ( $st1{$i} != $st2{$i} ) { break; }
1452                 }
1453                 return $i;
1454         }
1455         # These next three functions open, continue, and close the list
1456         # element appropriate to the prefix character passed into them.
1457         #
1458         /* private */ function openList( $char ) {
1459                 $result = $this->closeParagraph();
1460
1461                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1462                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1463                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1464                 else if ( ';' == $char ) {
1465                         $result .= '<dl><dt>';
1466                         $this->mDTopen = true;
1467                 }
1468                 else { $result = '<!-- ERR 1 -->'; }
1469
1470                 return $result;
1471         }
1472
1473         /* private */ function nextItem( $char ) {
1474                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1475                 else if ( ':' == $char || ';' == $char ) {
1476                         $close = '</dd>';
1477                         if ( $this->mDTopen ) { $close = '</dt>'; }
1478                         if ( ';' == $char ) {
1479                                 $this->mDTopen = true;
1480                                 return $close . '<dt>';
1481                         } else {
1482                                 $this->mDTopen = false;
1483                                 return $close . '<dd>';
1484                         }
1485                 }
1486                 return '<!-- ERR 2 -->';
1487         }
1488
1489         /* private */ function closeList( $char ) {
1490                 if ( '*' == $char ) { $text = '</li></ul>'; }
1491                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1492                 else if ( ':' == $char ) {
1493                         if ( $this->mDTopen ) {
1494                                 $this->mDTopen = false;
1495                                 $text = '</dt></dl>';
1496                         } else {
1497                                 $text = '</dd></dl>';
1498                         }
1499                 }
1500                 else {  return '<!-- ERR 3 -->'; }
1501                 return $text."\n";
1502         }
1503         /**#@-*/
1504
1505         /**
1506          * Make lists from lines starting with ':', '*', '#', etc.
1507          *
1508          * @access private
1509          * @return string the lists rendered as HTML
1510          */
1511         function doBlockLevels( $text, $linestart ) {
1512                 $fname = 'Parser::doBlockLevels';
1513                 wfProfileIn( $fname );
1514
1515                 # Parsing through the text line by line.  The main thing
1516                 # happening here is handling of block-level elements p, pre,
1517                 # and making lists from lines starting with * # : etc.
1518                 #
1519                 $textLines = explode( "\n", $text );
1520
1521                 $lastPrefix = $output = $lastLine = '';
1522                 $this->mDTopen = $inBlockElem = false;
1523                 $prefixLength = 0;
1524                 $paragraphStack = false;
1525
1526                 if ( !$linestart ) {
1527                         $output .= array_shift( $textLines );
1528                 }
1529                 foreach ( $textLines as $oLine ) {
1530                         $lastPrefixLength = strlen( $lastPrefix );
1531                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1532                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1533                         if ( !$this->mInPre ) {
1534                                 # Multiple prefixes may abut each other for nested lists.
1535                                 $prefixLength = strspn( $oLine, '*#:;' );
1536                                 $pref = substr( $oLine, 0, $prefixLength );
1537
1538                                 # eh?
1539                                 $pref2 = str_replace( ';', ':', $pref );
1540                                 $t = substr( $oLine, $prefixLength );
1541                                 $this->mInPre = !empty($preOpenMatch);
1542                         } else {
1543                                 # Don't interpret any other prefixes in preformatted text
1544                                 $prefixLength = 0;
1545                                 $pref = $pref2 = '';
1546                                 $t = $oLine;
1547                         }
1548
1549                         # List generation
1550                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1551                                 # Same as the last item, so no need to deal with nesting or opening stuff
1552                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1553                                 $paragraphStack = false;
1554
1555                                 if ( substr( $pref, -1 ) == ';') {
1556                                         # The one nasty exception: definition lists work like this:
1557                                         # ; title : definition text
1558                                         # So we check for : in the remainder text to split up the
1559                                         # title and definition, without b0rking links.
1560                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1561                                                 $t = $t2;
1562                                                 $output .= $term . $this->nextItem( ':' );
1563                                         }
1564                                 }
1565                         } elseif( $prefixLength || $lastPrefixLength ) {
1566                                 # Either open or close a level...
1567                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1568                                 $paragraphStack = false;
1569
1570                                 while( $commonPrefixLength < $lastPrefixLength ) {
1571                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1572                                         --$lastPrefixLength;
1573                                 }
1574                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1575                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1576                                 }
1577                                 while ( $prefixLength > $commonPrefixLength ) {
1578                                         $char = substr( $pref, $commonPrefixLength, 1 );
1579                                         $output .= $this->openList( $char );
1580
1581                                         if ( ';' == $char ) {
1582                                                 # FIXME: This is dupe of code above
1583                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1584                                                         $t = $t2;
1585                                                         $output .= $term . $this->nextItem( ':' );
1586                                                 }
1587                                         }
1588                                         ++$commonPrefixLength;
1589                                 }
1590                                 $lastPrefix = $pref2;
1591                         }
1592                         if( 0 == $prefixLength ) {
1593                                 wfProfileIn( "$fname-paragraph" );
1594                                 # No prefix (not in list)--go to paragraph mode
1595                                 $uniq_prefix = UNIQ_PREFIX;
1596                                 // XXX: use a stack for nestable elements like span, table and div
1597                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1598                                 $closematch = preg_match(
1599                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1600                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1601                                 if ( $openmatch or $closematch ) {
1602                                         $paragraphStack = false;
1603                                         $output .= $this->closeParagraph();
1604                                         if($preOpenMatch and !$preCloseMatch) {
1605                                                 $this->mInPre = true;
1606                                         }
1607                                         if ( $closematch ) {
1608                                                 $inBlockElem = false;
1609                                         } else {
1610                                                 $inBlockElem = true;
1611                                         }
1612                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1613                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1614                                                 // pre
1615                                                 if ($this->mLastSection != 'pre') {
1616                                                         $paragraphStack = false;
1617                                                         $output .= $this->closeParagraph().'<pre>';
1618                                                         $this->mLastSection = 'pre';
1619                                                 }
1620                                                 $t = substr( $t, 1 );
1621                                         } else {
1622                                                 // paragraph
1623                                                 if ( '' == trim($t) ) {
1624                                                         if ( $paragraphStack ) {
1625                                                                 $output .= $paragraphStack.'<br />';
1626                                                                 $paragraphStack = false;
1627                                                                 $this->mLastSection = 'p';
1628                                                         } else {
1629                                                                 if ($this->mLastSection != 'p' ) {
1630                                                                         $output .= $this->closeParagraph();
1631                                                                         $this->mLastSection = '';
1632                                                                         $paragraphStack = '<p>';
1633                                                                 } else {
1634                                                                         $paragraphStack = '</p><p>';
1635                                                                 }
1636                                                         }
1637                                                 } else {
1638                                                         if ( $paragraphStack ) {
1639                                                                 $output .= $paragraphStack;
1640                                                                 $paragraphStack = false;
1641                                                                 $this->mLastSection = 'p';
1642                                                         } else if ($this->mLastSection != 'p') {
1643                                                                 $output .= $this->closeParagraph().'<p>';
1644                                                                 $this->mLastSection = 'p';
1645                                                         }
1646                                                 }
1647                                         }
1648                                 }
1649                                 wfProfileOut( "$fname-paragraph" );
1650                         }
1651                         if ($paragraphStack === false) {
1652                                 $output .= $t."\n";
1653                         }
1654                 }
1655                 while ( $prefixLength ) {
1656                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1657                         --$prefixLength;
1658                 }
1659                 if ( '' != $this->mLastSection ) {
1660                         $output .= '</' . $this->mLastSection . '>';
1661                         $this->mLastSection = '';
1662                 }
1663
1664                 wfProfileOut( $fname );
1665                 return $output;
1666         }
1667
1668         /**
1669          * Split up a string on ':', ignoring any occurences inside
1670          * <a>..</a> or <span>...</span>
1671          * @param string $str the string to split
1672          * @param string &$before set to everything before the ':'
1673          * @param string &$after set to everything after the ':'
1674          * return string the position of the ':', or false if none found
1675          */
1676         function findColonNoLinks($str, &$before, &$after) {
1677                 # I wonder if we should make this count all tags, not just <a>
1678                 # and <span>. That would prevent us from matching a ':' that
1679                 # comes in the middle of italics other such formatting....
1680                 # -- Wil
1681                 $fname = 'Parser::findColonNoLinks';
1682                 wfProfileIn( $fname );
1683                 $pos = 0;
1684                 do {
1685                         $colon = strpos($str, ':', $pos);
1686
1687                         if ($colon !== false) {
1688                                 $before = substr($str, 0, $colon);
1689                                 $after = substr($str, $colon + 1);
1690
1691                                 # Skip any ':' within <a> or <span> pairs
1692                                 $a = substr_count($before, '<a');
1693                                 $s = substr_count($before, '<span');
1694                                 $ca = substr_count($before, '</a>');
1695                                 $cs = substr_count($before, '</span>');
1696
1697                                 if ($a <= $ca and $s <= $cs) {
1698                                         # Tags are balanced before ':'; ok
1699                                         break;
1700                                 }
1701                                 $pos = $colon + 1;
1702                         }
1703                 } while ($colon !== false);
1704                 wfProfileOut( $fname );
1705                 return $colon;
1706         }
1707
1708         /**
1709          * Return value of a magic variable (like PAGENAME)
1710          *
1711          * @access private
1712          */
1713         function getVariableValue( $index ) {
1714                 global $wgContLang, $wgSitename, $wgServer;
1715
1716                 /**
1717                  * Some of these require message or data lookups and can be
1718                  * expensive to check many times.
1719                  */
1720                 static $varCache = array();
1721                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1722
1723                 switch ( $index ) {
1724                         case MAG_CURRENTMONTH:
1725                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1726                         case MAG_CURRENTMONTHNAME:
1727                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1728                         case MAG_CURRENTMONTHNAMEGEN:
1729                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1730                         case MAG_CURRENTDAY:
1731                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1732                         case MAG_PAGENAME:
1733                                 return $this->mTitle->getText();
1734                         case MAG_PAGENAMEE:
1735                                 return $this->mTitle->getPartialURL();
1736                         case MAG_NAMESPACE:
1737                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1738                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1739                         case MAG_CURRENTDAYNAME:
1740                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1741                         case MAG_CURRENTYEAR:
1742                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1743                         case MAG_CURRENTTIME:
1744                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1745                         case MAG_NUMBEROFARTICLES:
1746                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1747                         case MAG_SITENAME:
1748                                 return $wgSitename;
1749                         case MAG_SERVER:
1750                                 return $wgServer;
1751                         default:
1752                                 return NULL;
1753                 }
1754         }
1755
1756         /**
1757          * initialise the magic variables (like CURRENTMONTHNAME)
1758          *
1759          * @access private
1760          */
1761         function initialiseVariables() {
1762                 $fname = 'Parser::initialiseVariables';
1763                 wfProfileIn( $fname );
1764                 global $wgVariableIDs;
1765                 $this->mVariables = array();
1766                 foreach ( $wgVariableIDs as $id ) {
1767                         $mw =& MagicWord::get( $id );
1768                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1769                 }
1770                 wfProfileOut( $fname );
1771         }
1772
1773         /**
1774          * Replace magic variables, templates, and template arguments
1775          * with the appropriate text. Templates are substituted recursively,
1776          * taking care to avoid infinite loops.
1777          *
1778          * Note that the substitution depends on value of $mOutputType:
1779          *  OT_WIKI: only {{subst:}} templates
1780          *  OT_MSG: only magic variables
1781          *  OT_HTML: all templates and magic variables
1782          *
1783          * @param string $tex The text to transform
1784          * @param array $args Key-value pairs representing template parameters to substitute
1785          * @access private
1786          */
1787         function replaceVariables( $text, $args = array() ) {
1788                 global $wgLang, $wgScript, $wgArticlePath;
1789
1790                 # Prevent too big inclusions
1791                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1792                         return $text;
1793                 }
1794
1795                 $fname = 'Parser::replaceVariables';
1796                 wfProfileIn( $fname );
1797
1798                 $titleChars = Title::legalChars();
1799
1800                 # This function is called recursively. To keep track of arguments we need a stack:
1801                 array_push( $this->mArgStack, $args );
1802
1803                 # Variable substitution
1804                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1805
1806                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1807                         # Argument substitution
1808                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1809                 }
1810                 # Template substitution
1811                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1812                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1813
1814                 array_pop( $this->mArgStack );
1815
1816                 wfProfileOut( $fname );
1817                 return $text;
1818         }
1819
1820         /**
1821          * Replace magic variables
1822          * @access private
1823          */
1824         function variableSubstitution( $matches ) {
1825                 $fname = 'parser::variableSubstitution';
1826                 wfProfileIn( $fname );
1827                 if ( !$this->mVariables ) {
1828                         $this->initialiseVariables();
1829                 }
1830                 $skip = false;
1831                 if ( $this->mOutputType == OT_WIKI ) {
1832                         # Do only magic variables prefixed by SUBST
1833                         $mwSubst =& MagicWord::get( MAG_SUBST );
1834                         if (!$mwSubst->matchStartAndRemove( $matches[1] ))
1835                                 $skip = true;
1836                         # Note that if we don't substitute the variable below,
1837                         # we don't remove the {{subst:}} magic word, in case
1838                         # it is a template rather than a magic variable.
1839                 }
1840                 if ( !$skip && array_key_exists( $matches[1], $this->mVariables ) ) {
1841                         $text = $this->mVariables[$matches[1]];
1842                         $this->mOutput->mContainsOldMagic = true;
1843                 } else {
1844                         $text = $matches[0];
1845                 }
1846                 wfProfileOut( $fname );
1847                 return $text;
1848         }
1849
1850         # Split template arguments
1851         function getTemplateArgs( $argsString ) {
1852                 if ( $argsString === '' ) {
1853                         return array();
1854                 }
1855
1856                 $args = explode( '|', substr( $argsString, 1 ) );
1857
1858                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1859                 # merged with the next arg because the '|' character between belongs
1860                 # to the link syntax and not the template parameter syntax.
1861                 $argc = count($args);
1862                 $i = 0;
1863                 for ( $i = 0; $i < $argc-1; $i++ ) {
1864                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1865                                 $args[$i] .= '|'.$args[$i+1];
1866                                 array_splice($args, $i+1, 1);
1867                                 $i--;
1868                                 $argc--;
1869                         }
1870                 }
1871
1872                 return $args;
1873         }
1874
1875         /**
1876          * Return the text of a template, after recursively
1877          * replacing any variables or templates within the template.
1878          *
1879          * @param array $matches The parts of the template
1880          *  $matches[1]: the title, i.e. the part before the |
1881          *  $matches[2]: the parameters (including a leading |), if  any
1882          * @return string the text of the template
1883          * @access private
1884          */
1885         function braceSubstitution( $matches ) {
1886                 global $wgLinkCache, $wgContLang;
1887                 $fname = 'Parser::braceSubstitution';
1888                 wfProfileIn( $fname );
1889
1890                 $found = false;
1891                 $nowiki = false;
1892                 $noparse = false;
1893
1894                 $title = NULL;
1895
1896                 # Need to know if the template comes at the start of a line,
1897                 # to treat the beginning of the template like the beginning
1898                 # of a line for tables and block-level elements.
1899                 $linestart = $matches[1];
1900
1901                 # $part1 is the bit before the first |, and must contain only title characters
1902                 # $args is a list of arguments, starting from index 0, not including $part1
1903
1904                 $part1 = $matches[2];
1905                 # If the third subpattern matched anything, it will start with |
1906
1907                 $args = $this->getTemplateArgs($matches[3]);
1908                 $argc = count( $args );
1909
1910                 # Don't parse {{{}}} because that's only for template arguments
1911                 if ( $linestart === '{' ) {
1912                         $text = $matches[0];
1913                         $found = true;
1914                         $noparse = true;
1915                 }
1916
1917                 # SUBST
1918                 if ( !$found ) {
1919                         $mwSubst =& MagicWord::get( MAG_SUBST );
1920                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1921                                 # One of two possibilities is true:
1922                                 # 1) Found SUBST but not in the PST phase
1923                                 # 2) Didn't find SUBST and in the PST phase
1924                                 # In either case, return without further processing
1925                                 $text = $matches[0];
1926                                 $found = true;
1927                                 $noparse = true;
1928                         }
1929                 }
1930
1931                 # MSG, MSGNW and INT
1932                 if ( !$found ) {
1933                         # Check for MSGNW:
1934                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1935                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1936                                 $nowiki = true;
1937                         } else {
1938                                 # Remove obsolete MSG:
1939                                 $mwMsg =& MagicWord::get( MAG_MSG );
1940                                 $mwMsg->matchStartAndRemove( $part1 );
1941                         }
1942
1943                         # Check if it is an internal message
1944                         $mwInt =& MagicWord::get( MAG_INT );
1945                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1946                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1947                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1948                                         $found = true;
1949                                 }
1950                         }
1951                 }
1952
1953                 # NS
1954                 if ( !$found ) {
1955                         # Check for NS: (namespace expansion)
1956                         $mwNs = MagicWord::get( MAG_NS );
1957                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1958                                 if ( intval( $part1 ) ) {
1959                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1960                                         $found = true;
1961                                 } else {
1962                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1963                                         if ( !is_null( $index ) ) {
1964                                                 $text = $linestart . $wgContLang->getNsText( $index );
1965                                                 $found = true;
1966                                         }
1967                                 }
1968                         }
1969                 }
1970
1971                 # LOCALURL and LOCALURLE
1972                 if ( !$found ) {
1973                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1974                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1975
1976                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1977                                 $func = 'getLocalURL';
1978                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1979                                 $func = 'escapeLocalURL';
1980                         } else {
1981                                 $func = '';
1982                         }
1983
1984                         if ( $func !== '' ) {
1985                                 $title = Title::newFromText( $part1 );
1986                                 if ( !is_null( $title ) ) {
1987                                         if ( $argc > 0 ) {
1988                                                 $text = $linestart . $title->$func( $args[0] );
1989                                         } else {
1990                                                 $text = $linestart . $title->$func();
1991                                         }
1992                                         $found = true;
1993                                 }
1994                         }
1995                 }
1996
1997                 # GRAMMAR
1998                 if ( !$found && $argc == 1 ) {
1999                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2000                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2001                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2002                                 $found = true;
2003                         }
2004                 }
2005
2006                 # Template table test
2007
2008                 # Did we encounter this template already? If yes, it is in the cache
2009                 # and we need to check for loops.
2010                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2011                         # set $text to cached message.
2012                         $text = $linestart . $this->mTemplates[$part1];
2013                         $found = true;
2014
2015                         # Infinite loop test
2016                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2017                                 $noparse = true;
2018                                 $found = true;
2019                                 $text .= '<!-- WARNING: template loop detected -->';
2020                         }
2021                 }
2022
2023                 # Load from database
2024                 $itcamefromthedatabase = false;
2025                 if ( !$found ) {
2026                         $ns = NS_TEMPLATE;
2027                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2028                         if ($subpage !== '') {
2029                                 $ns = $this->mTitle->getNamespace();
2030                         }
2031                         $title = Title::newFromText( $part1, $ns );
2032                         if ( !is_null( $title ) && !$title->isExternal() ) {
2033                                 # Check for excessive inclusion
2034                                 $dbk = $title->getPrefixedDBkey();
2035                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2036                                         # This should never be reached.
2037                                         $article = new Article( $title );
2038                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2039                                         if ( $articleContent !== false ) {
2040                                                 $found = true;
2041                                                 $text = $linestart . $articleContent;
2042                                                 $itcamefromthedatabase = true;
2043                                         }
2044                                 }
2045
2046                                 # If the title is valid but undisplayable, make a link to it
2047                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2048                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2049                                         $found = true;
2050                                 }
2051
2052                                 # Template cache array insertion
2053                                 $this->mTemplates[$part1] = $text;
2054                         }
2055                 }
2056
2057                 # Recursive parsing, escaping and link table handling
2058                 # Only for HTML output
2059                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2060                         $text = wfEscapeWikiText( $text );
2061                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2062                         # Clean up argument array
2063                         $assocArgs = array();
2064                         $index = 1;
2065                         foreach( $args as $arg ) {
2066                                 $eqpos = strpos( $arg, '=' );
2067                                 if ( $eqpos === false ) {
2068                                         $assocArgs[$index++] = $arg;
2069                                 } else {
2070                                         $name = trim( substr( $arg, 0, $eqpos ) );
2071                                         $value = trim( substr( $arg, $eqpos+1 ) );
2072                                         if ( $value === false ) {
2073                                                 $value = '';
2074                                         }
2075                                         if ( $name !== false ) {
2076                                                 $assocArgs[$name] = $value;
2077                                         }
2078                                 }
2079                         }
2080
2081                         # Add a new element to the templace recursion path
2082                         $this->mTemplatePath[$part1] = 1;
2083
2084                         $text = $this->strip( $text, $this->mStripState );
2085                         $text = $this->removeHTMLtags( $text );
2086                         $text = $this->replaceVariables( $text, $assocArgs );
2087
2088                         # Resume the link cache and register the inclusion as a link
2089                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2090                                 $wgLinkCache->addLinkObj( $title );
2091                         }
2092
2093                         # If the template begins with a table or block-level
2094                         # element, it should be treated as beginning a new line.
2095                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2096                                 $text = "\n" . $text;
2097                         }
2098                 }
2099
2100                 # Empties the template path
2101                 $this->mTemplatePath = array();
2102                 if ( !$found ) {
2103                         wfProfileOut( $fname );
2104                         return $matches[0];
2105                 } else {
2106                         # replace ==section headers==
2107                         # XXX this needs to go away once we have a better parser.
2108                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2109                                 if( !is_null( $title ) )
2110                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2111                                 else
2112                                         $encodedname = base64_encode("");
2113                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2114                                         PREG_SPLIT_DELIM_CAPTURE);
2115                                 $text = '';
2116                                 $nsec = 0;
2117                                 for( $i = 0; $i < count($m); $i += 2 ) {
2118                                         $text .= $m[$i];
2119                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2120                                         $hl = $m[$i + 1];
2121                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2122                                                 $text .= $hl;
2123                                                 continue;
2124                                         }
2125                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2126                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2127                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2128
2129                                         $nsec++;
2130                                 }
2131                         }
2132                 }
2133
2134                 # Empties the template path
2135                 $this->mTemplatePath = array();
2136
2137                 if ( !$found ) {
2138                         wfProfileOut( $fname );
2139                         return $matches[0];
2140                 } else {
2141                         wfProfileOut( $fname );
2142                         return $text;
2143                 }
2144         }
2145
2146         /**
2147          * Triple brace replacement -- used for template arguments
2148          * @access private
2149          */
2150         function argSubstitution( $matches ) {
2151                 $arg = trim( $matches[1] );
2152                 $text = $matches[0];
2153                 $inputArgs = end( $this->mArgStack );
2154
2155                 if ( array_key_exists( $arg, $inputArgs ) ) {
2156                         $text = $inputArgs[$arg];
2157                 }
2158
2159                 return $text;
2160         }
2161
2162         /**
2163          * Returns true if the function is allowed to include this entity
2164          * @access private
2165          */
2166         function incrementIncludeCount( $dbk ) {
2167                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2168                         $this->mIncludeCount[$dbk] = 0;
2169                 }
2170                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2171                         return true;
2172                 } else {
2173                         return false;
2174                 }
2175         }
2176
2177
2178         /**
2179          * Cleans up HTML, removes dangerous tags and attributes, and
2180          * removes HTML comments
2181          * @access private
2182          */
2183         function removeHTMLtags( $text ) {
2184                 global $wgUseTidy, $wgUserHtml;
2185                 $fname = 'Parser::removeHTMLtags';
2186                 wfProfileIn( $fname );
2187
2188                 if( $wgUserHtml ) {
2189                         $htmlpairs = array( # Tags that must be closed
2190                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
2191                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
2192                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
2193                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
2194                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
2195                         );
2196                         $htmlsingle = array(
2197                                 'br', 'hr', 'li', 'dt', 'dd'
2198                         );
2199                         $htmlnest = array( # Tags that can be nested--??
2200                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
2201                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
2202                         );
2203                         $tabletags = array( # Can only appear inside table
2204                                 'td', 'th', 'tr'
2205                         );
2206                 } else {
2207                         $htmlpairs = array();
2208                         $htmlsingle = array();
2209                         $htmlnest = array();
2210                         $tabletags = array();
2211                 }
2212
2213                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
2214                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
2215
2216                 $htmlattrs = $this->getHTMLattrs () ;
2217
2218                 # Remove HTML comments
2219                 $text = $this->removeHTMLcomments( $text );
2220
2221                 $bits = explode( '<', $text );
2222                 $text = array_shift( $bits );
2223                 if(!$wgUseTidy) {
2224                         $tagstack = array(); $tablestack = array();
2225                         foreach ( $bits as $x ) {
2226                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2227                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2228                                 $x, $regs );
2229                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2230                                 error_reporting( $prev );
2231
2232                                 $badtag = 0 ;
2233                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2234                                         # Check our stack
2235                                         if ( $slash ) {
2236                                                 # Closing a tag...
2237                                                 if ( ! in_array( $t, $htmlsingle ) &&
2238                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2239                                                         @array_push( $tagstack, $ot );
2240                                                         $badtag = 1;
2241                                                 } else {
2242                                                         if ( $t == 'table' ) {
2243                                                                 $tagstack = array_pop( $tablestack );
2244                                                         }
2245                                                         $newparams = '';
2246                                                 }
2247                                         } else {
2248                                                 # Keep track for later
2249                                                 if ( in_array( $t, $tabletags ) &&
2250                                                 ! in_array( 'table', $tagstack ) ) {
2251                                                         $badtag = 1;
2252                                                 } else if ( in_array( $t, $tagstack ) &&
2253                                                 ! in_array ( $t , $htmlnest ) ) {
2254                                                         $badtag = 1 ;
2255                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
2256                                                         if ( $t == 'table' ) {
2257                                                                 array_push( $tablestack, $tagstack );
2258                                                                 $tagstack = array();
2259                                                         }
2260                                                         array_push( $tagstack, $t );
2261                                                 }
2262                                                 # Strip non-approved attributes from the tag
2263                                                 $newparams = $this->fixTagAttributes($params);
2264
2265                                         }
2266                                         if ( ! $badtag ) {
2267                                                 $rest = str_replace( '>', '&gt;', $rest );
2268                                                 $text .= "<$slash$t $newparams$brace$rest";
2269                                                 continue;
2270                                         }
2271                                 }
2272                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2273                         }
2274                         # Close off any remaining tags
2275                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2276                                 $text .= "</$t>\n";
2277                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2278                         }
2279                 } else {
2280                         # this might be possible using tidy itself
2281                         foreach ( $bits as $x ) {
2282                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2283                                 $x, $regs );
2284                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2285                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2286                                         $newparams = $this->fixTagAttributes($params);
2287                                         $rest = str_replace( '>', '&gt;', $rest );
2288                                         $text .= "<$slash$t $newparams$brace$rest";
2289                                 } else {
2290                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2291                                 }
2292                         }
2293                 }
2294                 wfProfileOut( $fname );
2295                 return $text;
2296         }
2297
2298         /**
2299          * Remove '<!--', '-->', and everything between.
2300          * To avoid leaving blank lines, when a comment is both preceded
2301          * and followed by a newline (ignoring spaces), trim leading and
2302          * trailing spaces and one of the newlines.
2303          *
2304          * @access private
2305          */
2306         function removeHTMLcomments( $text ) {
2307                 $fname='Parser::removeHTMLcomments';
2308                 wfProfileIn( $fname );
2309                 while (($start = strpos($text, '<!--')) !== false) {
2310                         $end = strpos($text, '-->', $start + 4);
2311                         if ($end === false) {
2312                                 # Unterminated comment; bail out
2313                                 break;
2314                         }
2315
2316                         $end += 3;
2317
2318                         # Trim space and newline if the comment is both
2319                         # preceded and followed by a newline
2320                         $spaceStart = max($start - 1, 0);
2321                         $spaceLen = $end - $spaceStart;
2322                         while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2323                                 $spaceStart--;
2324                                 $spaceLen++;
2325                         }
2326                         while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2327                                 $spaceLen++;
2328                         if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2329                                 # Remove the comment, leading and trailing
2330                                 # spaces, and leave only one newline.
2331                                 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2332                         }
2333                         else {
2334                                 # Remove just the comment.
2335                                 $text = substr_replace($text, '', $start, $end - $start);
2336                         }
2337                 }
2338                 wfProfileOut( $fname );
2339                 return $text;
2340         }
2341
2342         /**
2343          * This function accomplishes several tasks:
2344          * 1) Auto-number headings if that option is enabled
2345          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2346          * 3) Add a Table of contents on the top for users who have enabled the option
2347          * 4) Auto-anchor headings
2348          *
2349          * It loops through all headlines, collects the necessary data, then splits up the
2350          * string and re-inserts the newly formatted headlines.
2351          * @access private
2352          */
2353         /* private */ function formatHeadings( $text, $isMain=true ) {
2354                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders;
2355
2356                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2357                 $doShowToc = $this->mOptions->getShowToc();
2358                 $forceTocHere = false;
2359                 if( !$this->mTitle->userCanEdit() ) {
2360                         $showEditLink = 0;
2361                         $rightClickHack = 0;
2362                 } else {
2363                         $showEditLink = $this->mOptions->getEditSection();
2364                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2365                 }
2366
2367                 # Inhibit editsection links if requested in the page
2368                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2369                 if( $esw->matchAndRemove( $text ) ) {
2370                         $showEditLink = 0;
2371                 }
2372                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2373                 # do not add TOC
2374                 $mw =& MagicWord::get( MAG_NOTOC );
2375                 if( $mw->matchAndRemove( $text ) ) {
2376                         $doShowToc = 0;
2377                 }
2378
2379                 # never add the TOC to the Main Page. This is an entry page that should not
2380                 # be more than 1-2 screens large anyway
2381                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2382                         $doShowToc = 0;
2383                 }
2384
2385                 # Get all headlines for numbering them and adding funky stuff like [edit]
2386                 # links - this is for later, but we need the number of headlines right now
2387                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
2388
2389                 # if there are fewer than 4 headlines in the article, do not show TOC
2390                 if( $numMatches < 4 ) {
2391                         $doShowToc = 0;
2392                 }
2393
2394                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2395                 # override above conditions and always show TOC at that place
2396                 $mw =& MagicWord::get( MAG_TOC );
2397                 if ($mw->match( $text ) ) {
2398                         $doShowToc = 1;
2399                         $forceTocHere = true;
2400                 } else {
2401                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2402                         # override above conditions and always show TOC above first header
2403                         $mw =& MagicWord::get( MAG_FORCETOC );
2404                         if ($mw->matchAndRemove( $text ) ) {
2405                                 $doShowToc = 1;
2406                         }
2407                 }
2408
2409
2410
2411                 # We need this to perform operations on the HTML
2412                 $sk =& $this->mOptions->getSkin();
2413
2414                 # headline counter
2415                 $headlineCount = 0;
2416                 $sectionCount = 0; # headlineCount excluding template sections
2417
2418                 # Ugh .. the TOC should have neat indentation levels which can be
2419                 # passed to the skin functions. These are determined here
2420                 $toclevel = 0;
2421                 $toc = '';
2422                 $full = '';
2423                 $head = array();
2424                 $sublevelCount = array();
2425                 $level = 0;
2426                 $prevlevel = 0;
2427                 foreach( $matches[3] as $headline ) {
2428                         $istemplate = 0;
2429                         $templatetitle = "";
2430                         $templatesection = 0;
2431
2432                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2433                                 $istemplate = 1;
2434                                 $templatetitle = base64_decode($mat[1]);
2435                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2436                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2437                         }
2438
2439                         $numbering = '';
2440                         if( $level ) {
2441                                 $prevlevel = $level;
2442                         }
2443                         $level = $matches[1][$headlineCount];
2444                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2445                                 # reset when we enter a new level
2446                                 $sublevelCount[$level] = 0;
2447                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2448                                 $toclevel += $level - $prevlevel;
2449                         }
2450                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2451                                 # reset when we step back a level
2452                                 $sublevelCount[$level+1]=0;
2453                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2454                                 $toclevel -= $prevlevel - $level;
2455                         }
2456                         # count number of headlines for each level
2457                         @$sublevelCount[$level]++;
2458                         if( $doNumberHeadings || $doShowToc ) {
2459                                 $dot = 0;
2460                                 for( $i = 1; $i <= $level; $i++ ) {
2461                                         if( !empty( $sublevelCount[$i] ) ) {
2462                                                 if( $dot ) {
2463                                                         $numbering .= '.';
2464                                                 }
2465                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2466                                                 $dot = 1;
2467                                         }
2468                                 }
2469                         }
2470
2471                         # The canonized header is a version of the header text safe to use for links
2472                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2473                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2474                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2475
2476                         # Remove link placeholders by the link text.
2477                         #     <!--LINK number-->
2478                         # turns into
2479                         #     link text with suffix
2480                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2481                                                             "\$wgLinkHolders['texts'][\$1]",
2482                                                             $canonized_headline );
2483
2484                         # strip out HTML
2485                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2486                         $tocline = trim( $canonized_headline );
2487                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2488                         $replacearray = array(
2489                                 '%3A' => ':',
2490                                 '%' => '.'
2491                         );
2492                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2493                         $refer[$headlineCount] = $canonized_headline;
2494
2495                         # count how many in assoc. array so we can track dupes in anchors
2496                         @$refers[$canonized_headline]++;
2497                         $refcount[$headlineCount]=$refers[$canonized_headline];
2498
2499                         # Prepend the number to the heading text
2500
2501                         if( $doNumberHeadings || $doShowToc ) {
2502                                 $tocline = $numbering . ' ' . $tocline;
2503
2504                                 # Don't number the heading if it is the only one (looks silly)
2505                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2506                                         # the two are different if the line contains a link
2507                                         $headline=$numbering . ' ' . $headline;
2508                                 }
2509                         }
2510
2511                         # Create the anchor for linking from the TOC to the section
2512                         $anchor = $canonized_headline;
2513                         if($refcount[$headlineCount] > 1 ) {
2514                                 $anchor .= '_' . $refcount[$headlineCount];
2515                         }
2516                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2517                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2518                         }
2519                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2520                                 if ( empty( $head[$headlineCount] ) ) {
2521                                         $head[$headlineCount] = '';
2522                                 }
2523                                 if( $istemplate )
2524                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2525                                 else
2526                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2527                         }
2528
2529                         # Add the edit section span
2530                         if( $rightClickHack ) {
2531                                 if( $istemplate )
2532                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2533                                 else
2534                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2535                         }
2536
2537                         # give headline the correct <h#> tag
2538                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2539
2540                         $headlineCount++;
2541                         if( !$istemplate )
2542                                 $sectionCount++;
2543                 }
2544
2545                 if( $doShowToc ) {
2546                         $toclines = $headlineCount;
2547                         $toc .= $sk->tocUnindent( $toclevel );
2548                         $toc = $sk->tocTable( $toc );
2549                 }
2550
2551                 # split up and insert constructed headlines
2552
2553                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2554                 $i = 0;
2555
2556                 foreach( $blocks as $block ) {
2557                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2558                                 # This is the [edit] link that appears for the top block of text when
2559                                 # section editing is enabled
2560
2561                                 # Disabled because it broke block formatting
2562                                 # For example, a bullet point in the top line
2563                                 # $full .= $sk->editSectionLink(0);
2564                         }
2565                         $full .= $block;
2566                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2567                         # Top anchor now in skin
2568                                 $full = $full.$toc;
2569                         }
2570
2571                         if( !empty( $head[$i] ) ) {
2572                                 $full .= $head[$i];
2573                         }
2574                         $i++;
2575                 }
2576                 if($forceTocHere) {
2577                         $mw =& MagicWord::get( MAG_TOC );
2578                         return $mw->replace( $toc, $full );
2579                 } else {
2580                         return $full;
2581                 }
2582         }
2583
2584         /**
2585          * Return an HTML link for the "ISBN 123456" text
2586          * @access private
2587          */
2588         function magicISBN( $text ) {
2589                 global $wgLang;
2590                 $fname = 'Parser::magicISBN';
2591                 wfProfileIn( $fname );
2592
2593                 $a = split( 'ISBN ', ' '.$text );
2594                 if ( count ( $a ) < 2 ) {
2595                         wfProfileOut( $fname );
2596                         return $text;
2597                 }
2598                 $text = substr( array_shift( $a ), 1);
2599                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2600
2601                 foreach ( $a as $x ) {
2602                         $isbn = $blank = '' ;
2603                         while ( ' ' == $x{0} ) {
2604                                 $blank .= ' ';
2605                                 $x = substr( $x, 1 );
2606                         }
2607                         if ( $x == '' ) { # blank isbn
2608                                 $text .= "ISBN $blank";
2609                                 continue;
2610                         }
2611                         while ( strstr( $valid, $x{0} ) != false ) {
2612                                 $isbn .= $x{0};
2613                                 $x = substr( $x, 1 );
2614                         }
2615                         $num = str_replace( '-', '', $isbn );
2616                         $num = str_replace( ' ', '', $num );
2617
2618                         if ( '' == $num ) {
2619                                 $text .= "ISBN $blank$x";
2620                         } else {
2621                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2622                                 $text .= '<a href="' .
2623                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2624                                         "\" class=\"internal\">ISBN $isbn</a>";
2625                                 $text .= $x;
2626                         }
2627                 }
2628                 wfProfileOut( $fname );
2629                 return $text;
2630         }
2631
2632         /**
2633          * Return an HTML link for the "GEO ..." text
2634          * @access private
2635          */
2636         function magicGEO( $text ) {
2637                 global $wgLang, $wgUseGeoMode;
2638                 $fname = 'Parser::magicGEO';
2639                 wfProfileIn( $fname );
2640
2641                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2642                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2643                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2644                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2645                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2646                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2647
2648                 $a = split( 'GEO ', ' '.$text );
2649                 if ( count ( $a ) < 2 ) {
2650                         wfProfileOut( $fname );
2651                         return $text;
2652                 }
2653                 $text = substr( array_shift( $a ), 1);
2654                 $valid = '0123456789.+-:';
2655
2656                 foreach ( $a as $x ) {
2657                         $geo = $blank = '' ;
2658                         while ( ' ' == $x{0} ) {
2659                                 $blank .= ' ';
2660                                 $x = substr( $x, 1 );
2661                         }
2662                         while ( strstr( $valid, $x{0} ) != false ) {
2663                                 $geo .= $x{0};
2664                                 $x = substr( $x, 1 );
2665                         }
2666                         $num = str_replace( '+', '', $geo );
2667                         $num = str_replace( ' ', '', $num );
2668
2669                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2670                                 $text .= "GEO $blank$x";
2671                         } else {
2672                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2673                                 $text .= '<a href="' .
2674                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2675                                         "\" class=\"internal\">GEO $geo</a>";
2676                                 $text .= $x;
2677                         }
2678                 }
2679                 wfProfileOut( $fname );
2680                 return $text;
2681         }
2682
2683         /**
2684          * Return an HTML link for the "RFC 1234" text
2685          * @access private
2686          * @param string $text text to be processed
2687          */
2688         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2689                 global $wgLang;
2690
2691                 $valid = '0123456789';
2692                 $internal = false;
2693
2694                 $a = split( $keyword, ' '.$text );
2695                 if ( count ( $a ) < 2 ) {
2696                         return $text;
2697                 }
2698                 $text = substr( array_shift( $a ), 1);
2699
2700                 /* Check if keyword is preceed by [[.
2701                  * This test is made here cause of the array_shift above
2702                  * that prevent the test to be done in the foreach.
2703                  */
2704                 if ( substr( $text, -2 ) == '[[' ) {
2705                         $internal = true;
2706                 }
2707
2708                 foreach ( $a as $x ) {
2709                         /* token might be empty if we have RFC RFC 1234 */
2710                         if ( $x=='' ) {
2711                                 $text.=$keyword;
2712                                 continue;
2713                                 }
2714
2715                         $id = $blank = '' ;
2716
2717                         /** remove and save whitespaces in $blank */
2718                         while ( $x{0} == ' ' ) {
2719                                 $blank .= ' ';
2720                                 $x = substr( $x, 1 );
2721                         }
2722
2723                         /** remove and save the rfc number in $id */
2724                         while ( strstr( $valid, $x{0} ) != false ) {
2725                                 $id .= $x{0};
2726                                 $x = substr( $x, 1 );
2727                         }
2728
2729                         if ( $id == '' ) {
2730                                 /* call back stripped spaces*/
2731                                 $text .= $keyword.$blank.$x;
2732                         } elseif( $internal ) {
2733                                 /* normal link */
2734                                 $text .= $keyword.$id.$x;
2735                         } else {
2736                                 /* build the external link*/
2737                                 $url = wfmsg( $urlmsg );
2738                                 $url = str_replace( '$1', $id, $url);
2739                                 $sk =& $this->mOptions->getSkin();
2740                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2741                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2742                         }
2743
2744                         /* Check if the next RFC keyword is preceed by [[ */
2745                         $internal = ( substr($x,-2) == '[[' );
2746                 }
2747                 return $text;
2748         }
2749
2750         /**
2751          * Transform wiki markup when saving a page by doing \r\n -> \n
2752          * conversion, substitting signatures, {{subst:}} templates, etc.
2753          *
2754          * @param string $text the text to transform
2755          * @param Title &$title the Title object for the current article
2756          * @param User &$user the User object describing the current user
2757          * @param ParserOptions $options parsing options
2758          * @param bool $clearState whether to clear the parser state first
2759          * @return string the altered wiki markup
2760          * @access public
2761          */
2762         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2763                 $this->mOptions = $options;
2764                 $this->mTitle =& $title;
2765                 $this->mOutputType = OT_WIKI;
2766
2767                 if ( $clearState ) {
2768                         $this->clearState();
2769                 }
2770
2771                 $stripState = false;
2772                 $pairs = array(
2773                         "\r\n" => "\n",
2774                         );
2775                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2776                 $text = $this->strip( $text, $stripState, false );
2777                 $text = $this->pstPass2( $text, $user );
2778                 $text = $this->unstrip( $text, $stripState );
2779                 $text = $this->unstripNoWiki( $text, $stripState );
2780                 return $text;
2781         }
2782
2783         /**
2784          * Pre-save transform helper function
2785          * @access private
2786          */
2787         function pstPass2( $text, &$user ) {
2788                 global $wgLang, $wgContLang, $wgLocaltimezone;
2789
2790                 # Variable replacement
2791                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2792                 $text = $this->replaceVariables( $text );
2793
2794                 # Signatures
2795                 #
2796                 $n = $user->getName();
2797                 $k = $user->getOption( 'nickname' );
2798                 if ( '' == $k ) { $k = $n; }
2799                 if ( isset( $wgLocaltimezone ) ) {
2800                         $oldtz = getenv( 'TZ' );
2801                         putenv( 'TZ='.$wgLocaltimezone );
2802                 }
2803                 /* Note: this is an ugly timezone hack for the European wikis */
2804                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2805                   ' (' . date( 'T' ) . ')';
2806                 if ( isset( $wgLocaltimezone ) ) {
2807                         putenv( 'TZ='.$oldtzs );
2808                 }
2809
2810                 $text = preg_replace( '/~~~~~~/', $d, $text );
2811                 $text = preg_replace( '/~~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2812                 $text = preg_replace( '/~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2813
2814                 # Context links: [[|name]] and [[name (context)|]]
2815                 #
2816                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2817                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2818                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2819                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2820
2821                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2822                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2823                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2824                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2825                 $context = '';
2826                 $t = $this->mTitle->getText();
2827                 if ( preg_match( $conpat, $t, $m ) ) {
2828                         $context = $m[2];
2829                 }
2830                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2831                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2832                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2833
2834                 if ( '' == $context ) {
2835                         $text = preg_replace( $p2, '[[\\1]]', $text );
2836                 } else {
2837                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2838                 }
2839
2840                 # Trim trailing whitespace
2841                 # MAG_END (__END__) tag allows for trailing
2842                 # whitespace to be deliberately included
2843                 $text = rtrim( $text );
2844                 $mw =& MagicWord::get( MAG_END );
2845                 $mw->matchAndRemove( $text );
2846
2847                 return $text;
2848         }
2849
2850         /**
2851          * Set up some variables which are usually set up in parse()
2852          * so that an external function can call some class members with confidence
2853          * @access public
2854          */
2855         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2856                 $this->mTitle =& $title;
2857                 $this->mOptions = $options;
2858                 $this->mOutputType = $outputType;
2859                 if ( $clearState ) {
2860                         $this->clearState();
2861                 }
2862         }
2863
2864         /**
2865          * Transform a MediaWiki message by replacing magic variables.
2866          *
2867          * @param string $text the text to transform
2868          * @param ParserOptions $options  options
2869          * @return string the text with variables substituted
2870          * @access public
2871          */
2872         function transformMsg( $text, $options ) {
2873                 global $wgTitle;
2874                 static $executing = false;
2875
2876                 # Guard against infinite recursion
2877                 if ( $executing ) {
2878                         return $text;
2879                 }
2880                 $executing = true;
2881
2882                 $this->mTitle = $wgTitle;
2883                 $this->mOptions = $options;
2884                 $this->mOutputType = OT_MSG;
2885                 $this->clearState();
2886                 $text = $this->replaceVariables( $text );
2887
2888                 $executing = false;
2889                 return $text;
2890         }
2891
2892         /**
2893          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2894          * Callback will be called with the text within
2895          * Transform and return the text within
2896          * @access public
2897          */
2898         function setHook( $tag, $callback ) {
2899                 $oldVal = @$this->mTagHooks[$tag];
2900                 $this->mTagHooks[$tag] = $callback;
2901                 return $oldVal;
2902         }
2903
2904         /**
2905          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2906          * Placeholders created in Skin::makeLinkObj()
2907          * Returns an array of links found, indexed by PDBK:
2908          *  0 - broken
2909          *  1 - normal link
2910          *  2 - stub
2911          * $options is a bit field, RLH_FOR_UPDATE to select for update
2912          */
2913         function replaceLinkHolders( &$text, $options = 0 ) {
2914                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2915                 global $wgInterwikiLinkHolders;
2916                 global $outputReplace;
2917
2918                 if ( $wgUseOldExistenceCheck ) {
2919                         return array();
2920                 }
2921
2922                 $fname = 'Parser::replaceLinkHolders';
2923                 wfProfileIn( $fname );
2924
2925                 $pdbks = array();
2926                 $colours = array();
2927
2928                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2929                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2930                         wfProfileIn( $fname.'-check' );
2931                         $dbr =& wfGetDB( DB_SLAVE );
2932                         $cur = $dbr->tableName( 'cur' );
2933                         $sk = $wgUser->getSkin();
2934                         $threshold = $wgUser->getOption('stubthreshold');
2935
2936                         # Sort by namespace
2937                         asort( $wgLinkHolders['namespaces'] );
2938
2939                         # Generate query
2940                         $query = false;
2941                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2942                                 # Make title object
2943                                 $title = $wgLinkHolders['titles'][$key];
2944
2945                                 # Skip invalid entries.
2946                                 # Result will be ugly, but prevents crash.
2947                                 if ( is_null( $title ) ) {
2948                                         continue;
2949                                 }
2950                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2951
2952                                 # Check if it's in the link cache already
2953                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2954                                         $colours[$pdbk] = 1;
2955                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2956                                         $colours[$pdbk] = 0;
2957                                 } else {
2958                                         # Not in the link cache, add it to the query
2959                                         if ( !isset( $current ) ) {
2960                                                 $current = $val;
2961                                                 $query =  "SELECT cur_id, cur_namespace, cur_title";
2962                                                 if ( $threshold > 0 ) {
2963                                                         $query .= ", LENGTH(cur_text) AS cur_len, cur_is_redirect";
2964                                                 }
2965                                                 $query .= " FROM $cur WHERE (cur_namespace=$val AND cur_title IN(";
2966                                         } elseif ( $current != $val ) {
2967                                                 $current = $val;
2968                                                 $query .= ")) OR (cur_namespace=$val AND cur_title IN(";
2969                                         } else {
2970                                                 $query .= ', ';
2971                                         }
2972
2973                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2974                                 }
2975                         }
2976                         if ( $query ) {
2977                                 $query .= '))';
2978                                 if ( $options & RLH_FOR_UPDATE ) {
2979                                         $query .= ' FOR UPDATE';
2980                                 }
2981
2982                                 $res = $dbr->query( $query, $fname );
2983
2984                                 # Fetch data and form into an associative array
2985                                 # non-existent = broken
2986                                 # 1 = known
2987                                 # 2 = stub
2988                                 while ( $s = $dbr->fetchObject($res) ) {
2989                                         $title = Title::makeTitle( $s->cur_namespace, $s->cur_title );
2990                                         $pdbk = $title->getPrefixedDBkey();
2991                                         $wgLinkCache->addGoodLink( $s->cur_id, $pdbk );
2992
2993                                         if ( $threshold >  0 ) {
2994                                                 $size = $s->cur_len;
2995                                                 if ( $s->cur_is_redirect || $s->cur_namespace != 0 || $length < $threshold ) {
2996                                                         $colours[$pdbk] = 1;
2997                                                 } else {
2998                                                         $colours[$pdbk] = 2;
2999                                                 }
3000                                         } else {
3001                                                 $colours[$pdbk] = 1;
3002                                         }
3003                                 }
3004                         }
3005                         wfProfileOut( $fname.'-check' );
3006
3007                         # Construct search and replace arrays
3008                         wfProfileIn( $fname.'-construct' );
3009                         $outputReplace = array();
3010                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
3011                                 $pdbk = $pdbks[$key];
3012                                 $searchkey = '<!--LINK '.$key.'-->';
3013                                 $title = $wgLinkHolders['titles'][$key];
3014                                 if ( empty( $colours[$pdbk] ) ) {
3015                                         $wgLinkCache->addBadLink( $pdbk );
3016                                         $colours[$pdbk] = 0;
3017                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3018                                                                         $wgLinkHolders['texts'][$key],
3019                                                                         $wgLinkHolders['queries'][$key] );
3020                                 } elseif ( $colours[$pdbk] == 1 ) {
3021                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3022                                                                         $wgLinkHolders['texts'][$key],
3023                                                                         $wgLinkHolders['queries'][$key] );
3024                                 } elseif ( $colours[$pdbk] == 2 ) {
3025                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3026                                                                         $wgLinkHolders['texts'][$key],
3027                                                                         $wgLinkHolders['queries'][$key] );
3028                                 }
3029                         }
3030                         wfProfileOut( $fname.'-construct' );
3031
3032                         # Do the thing
3033                         wfProfileIn( $fname.'-replace' );
3034
3035                         $text = preg_replace_callback(
3036                                 '/(<!--LINK .*?-->)/',
3037                                 "outputReplaceMatches",
3038                                 $text);
3039                         wfProfileOut( $fname.'-replace' );
3040                 }
3041
3042                 if ( !empty( $wgInterwikiLinkHolders ) ) {
3043                         wfProfileIn( $fname.'-interwiki' );
3044                         $outputReplace = $wgInterwikiLinkHolders;
3045                         $text = preg_replace_callback(
3046                                 '/<!--IWLINK (.*?)-->/',
3047                                 "outputReplaceMatches",
3048                                 $text );
3049                         wfProfileOut( $fname.'-interwiki' );
3050                 }
3051
3052                 wfProfileOut( $fname );
3053                 return $colours;
3054         }
3055
3056         /**
3057          * Renders an image gallery from a text with one line per image.
3058          * text labels may be given by using |-style alternative text. E.g.
3059          *   Image:one.jpg|The number "1"
3060          *   Image:tree.jpg|A tree
3061          * given as text will return the HTML of a gallery with two images,
3062          * labeled 'The number "1"' and
3063          * 'A tree'.
3064          */
3065         function renderImageGallery( $text ) {
3066                 global $wgLinkCache;
3067                 $ig = new ImageGallery();
3068                 $ig->setShowBytes( false );
3069                 $ig->setShowFilename( false );
3070                 $lines = explode( "\n", $text );
3071
3072                 foreach ( $lines as $line ) {
3073                         # match lines like these:
3074                         # Image:someimage.jpg|This is some image
3075                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3076                         # Skip empty lines
3077                         if ( count( $matches ) == 0 ) {
3078                                 continue;
3079                         }
3080                         $nt = Title::newFromURL( $matches[1] );
3081                         if ( isset( $matches[3] ) ) {
3082                                 $label = $matches[3];
3083                         } else {
3084                                 $label = '';
3085                         }
3086                         $ig->add( Image::newFromTitle( $nt ), $label );
3087                         $wgLinkCache->addImageLinkObj( $nt );
3088                 }
3089                 return $ig->toHTML();
3090         }
3091 }
3092
3093 /**
3094  * @todo document
3095  * @package MediaWiki
3096  */
3097 class ParserOutput
3098 {
3099         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3100         var $mCacheTime; # Used in ParserCache
3101         var $mVersion;   # Compatibility check
3102
3103         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3104                 $containsOldMagic = false )
3105         {
3106                 $this->mText = $text;
3107                 $this->mLanguageLinks = $languageLinks;
3108                 $this->mCategoryLinks = $categoryLinks;
3109                 $this->mContainsOldMagic = $containsOldMagic;
3110                 $this->mCacheTime = '';
3111                 $this->mVersion = MW_PARSER_VERSION;
3112         }
3113
3114         function getText()                   { return $this->mText; }
3115         function getLanguageLinks()          { return $this->mLanguageLinks; }
3116         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
3117         function getCacheTime()              { return $this->mCacheTime; }
3118         function containsOldMagic()          { return $this->mContainsOldMagic; }
3119         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3120         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3121         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3122         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3123         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3124         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3125
3126         function merge( $other ) {
3127                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3128                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3129                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3130         }
3131
3132         /**
3133          * Return true if this cached output object predates the global or
3134          * per-article cache invalidation timestamps, or if it comes from
3135          * an incompatible older version.
3136          *
3137          * @param string $touched the affected article's last touched timestamp
3138          * @return bool
3139          * @access public
3140          */
3141         function expired( $touched ) {
3142                 global $wgCacheEpoch;
3143                 return $this->getCacheTime() <= $touched ||
3144                        $this->getCacheTime() <= $wgCacheEpoch ||
3145                        !isset( $this->mVersion ) ||
3146                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3147         }
3148 }
3149
3150 /**
3151  * Set options of the Parser
3152  * @todo document
3153  * @package MediaWiki
3154  */
3155 class ParserOptions
3156 {
3157         # All variables are private
3158         var $mUseTeX;                    # Use texvc to expand <math> tags
3159         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3160         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3161         var $mAllowExternalImages;       # Allow external images inline
3162         var $mSkin;                      # Reference to the preferred skin
3163         var $mDateFormat;                # Date format index
3164         var $mEditSection;               # Create "edit section" links
3165         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3166         var $mNumberHeadings;            # Automatically number headings
3167         var $mShowToc;                   # Show table of contents
3168
3169         function getUseTeX()                        { return $this->mUseTeX; }
3170         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3171         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3172         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3173         function getSkin()                          { return $this->mSkin; }
3174         function getDateFormat()                    { return $this->mDateFormat; }
3175         function getEditSection()                   { return $this->mEditSection; }
3176         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3177         function getNumberHeadings()                { return $this->mNumberHeadings; }
3178         function getShowToc()                       { return $this->mShowToc; }
3179
3180         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3181         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3182         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3183         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3184         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3185         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3186         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3187         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3188         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3189
3190         function setSkin( &$x ) { $this->mSkin =& $x; }
3191
3192         # Get parser options
3193         /* static */ function newFromUser( &$user ) {
3194                 $popts = new ParserOptions;
3195                 $popts->initialiseFromUser( $user );
3196                 return $popts;
3197         }
3198
3199         # Get user options
3200         function initialiseFromUser( &$userInput ) {
3201                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3202                 $fname = 'ParserOptions::initialiseFromUser';
3203                 wfProfileIn( $fname );
3204                 if ( !$userInput ) {
3205                         $user = new User;
3206                         $user->setLoaded( true );
3207                 } else {
3208                         $user =& $userInput;
3209                 }
3210
3211                 $this->mUseTeX = $wgUseTeX;
3212                 $this->mUseDynamicDates = $wgUseDynamicDates;
3213                 $this->mInterwikiMagic = $wgInterwikiMagic;
3214                 $this->mAllowExternalImages = $wgAllowExternalImages;
3215                 wfProfileIn( $fname.'-skin' );
3216                 $this->mSkin =& $user->getSkin();
3217                 wfProfileOut( $fname.'-skin' );
3218                 $this->mDateFormat = $user->getOption( 'date' );
3219                 $this->mEditSection = $user->getOption( 'editsection' );
3220                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3221                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3222                 $this->mShowToc = $user->getOption( 'showtoc' );
3223                 wfProfileOut( $fname );
3224         }
3225
3226
3227 }
3228
3229 /**
3230  * Callback function used by Parser::replaceLinkHolders()
3231  * to substitute link placeholders.
3232  */
3233 function &outputReplaceMatches( $matches ) {
3234         global $outputReplace;
3235         return $outputReplace[$matches[1]];
3236 }
3237
3238 /**
3239  * Return the total number of articles
3240  */
3241 function wfNumberOfArticles() {
3242         global $wgNumberOfArticles;
3243
3244         wfLoadSiteStats();
3245         return $wgNumberOfArticles;
3246 }
3247
3248 /**
3249  * Get various statistics from the database
3250  * @private
3251  */
3252 function wfLoadSiteStats() {
3253         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3254         $fname = 'wfLoadSiteStats';
3255
3256         if ( -1 != $wgNumberOfArticles ) return;
3257         $dbr =& wfGetDB( DB_SLAVE );
3258         $s = $dbr->selectRow( 'site_stats',
3259                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3260                 array( 'ss_row_id' => 1 ), $fname
3261         );
3262
3263         if ( $s === false ) {
3264                 return;
3265         } else {
3266                 $wgTotalViews = $s->ss_total_views;
3267                 $wgTotalEdits = $s->ss_total_edits;
3268                 $wgNumberOfArticles = $s->ss_good_articles;
3269         }
3270 }
3271
3272 function wfEscapeHTMLTagsOnly( $in ) {
3273         return str_replace(
3274                 array( '"', '>', '<' ),
3275                 array( '&quot;', '&gt;', '&lt;' ),
3276                 $in );
3277 }
3278
3279 ?>