includes/Parser.php

   1 <?php
   2
   3 /**
   4  * File for Parser and related classes
   5  *
   6  * @package MediaWiki
   7  */
   8
   9 require_once( 'Sanitizer.php' );
  10
  11 /**
  12  * Update this version number when the ParserOutput format
  13  * changes in an incompatible way, so the parser cache
  14  * can automatically discard old data.
  15  */
  16 define( 'MW_PARSER_VERSION', '1.4.0' );
  17
  18 /**
  19  * Variable substitution O(N^2) attack
  20  *
  21  * Without countermeasures, it would be possible to attack the parser by saving
  22  * a page filled with a large number of inclusions of large pages. The size of
  23  * the generated page would be proportional to the square of the input size.
  24  * Hence, we limit the number of inclusions of any given page, thus bringing any
  25  * attack back to O(N).
  26  */
  27
  28 define( 'MAX_INCLUDE_REPEAT', 100 );
  29 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  30
  31 define( 'RLH_FOR_UPDATE', 1 );
  32
  33 # Allowed values for $mOutputType
  34 define( 'OT_HTML', 1 );
  35 define( 'OT_WIKI', 2 );
  36 define( 'OT_MSG' , 3 );
  37
  38 # string parameter for extractTags which will cause it
  39 # to strip HTML comments in addition to regular
  40 # <XML>-style tags. This should not be anything we
  41 # may want to use in wikisyntax
  42 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  43
  44 # prefix for escaping, used in two functions at least
  45 define( 'UNIQ_PREFIX', 'NaodW29');
  46
  47 # Constants needed for external link processing
  48 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  49 define( 'HTTP_PROTOCOLS', 'http|https' );
  50 # Everything except bracket, space, or control characters
  51 define( 'EXT_LINK_URL_CLASS', '[^]<>"\\x00-\\x20\\x7F]' );
  52 # Including space
  53 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  54 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  55 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  56 define( 'EXT_LINK_BRACKETED',  '/\[(\b('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  57 define( 'EXT_IMAGE_REGEX',
  58         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  59         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  60         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  61 );
  62
  63 /**
  64  * PHP Parser
  65  *
  66  * Processes wiki markup
  67  *
  68  * <pre>
  69  * There are three main entry points into the Parser class:
  70  * parse()
  71  *   produces HTML output
  72  * preSaveTransform().
  73  *   produces altered wiki markup.
  74  * transformMsg()
  75  *   performs brace substitution on MediaWiki messages
  76  *
  77  * Globals used:
  78  *    objects:   $wgLang, $wgDateFormatter, $wgLinkCache
  79  *
  80  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  81  *
  82  * settings:
  83  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  84  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  85  *  $wgLocaltimezone
  86  *
  87  *  * only within ParserOptions
  88  * </pre>
  89  *
  90  * @package MediaWiki
  91  */
  92 class Parser
  93 {
  94         /**#@+
  95          * @access private
  96          */
  97         # Persistent:
  98         var $mTagHooks;
  99
 100         # Cleared with clearState():
 101         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
 102         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
 103
 104         # Temporary:
 105         var $mOptions, $mTitle, $mOutputType,
 106             $mTemplates,        // cache of already loaded templates, avoids
 107                                 // multiple SQL queries for the same string
 108             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 109                                 // in this path. Used for loop detection.
 110
 111         /**#@-*/
 112
 113         /**
 114          * Constructor
 115          *
 116          * @access public
 117          */
 118         function Parser() {
 119                 $this->mTemplates = array();
 120                 $this->mTemplatePath = array();
 121                 $this->mTagHooks = array();
 122                 $this->clearState();
 123         }
 124
 125         /**
 126          * Clear Parser state
 127          *
 128          * @access private
 129          */
 130         function clearState() {
 131                 $this->mOutput = new ParserOutput;
 132                 $this->mAutonumber = 0;
 133                 $this->mLastSection = '';
 134                 $this->mDTopen = false;
 135                 $this->mVariables = false;
 136                 $this->mIncludeCount = array();
 137                 $this->mStripState = array();
 138                 $this->mArgStack = array();
 139                 $this->mInPre = false;
 140         }
 141
 142         /**
 143          * First pass--just handle <nowiki> sections, pass the rest off
 144          * to internalParse() which does all the real work.
 145          *
 146          * @access private
 147          * @param string $text Text we want to parse
 148          * @param Title &$title A title object
 149          * @param array $options
 150          * @param boolean $linestart
 151          * @param boolean $clearState
 152          * @return ParserOutput a ParserOutput
 153          */
 154         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 155                 global $wgUseTidy, $wgContLang;
 156                 $fname = 'Parser::parse';
 157                 wfProfileIn( $fname );
 158
 159                 if ( $clearState ) {
 160                         $this->clearState();
 161                 }
 162
 163                 $this->mOptions = $options;
 164                 $this->mTitle =& $title;
 165                 $this->mOutputType = OT_HTML;
 166
 167                 $stripState = NULL;
 168                 global $fnord; $fnord = 1;
 169                 //$text = $this->strip( $text, $this->mStripState );
 170                 // VOODOO MAGIC FIX! Sometimes the above segfaults in PHP5.
 171                 $x =& $this->mStripState;
 172                 $text = $this->strip( $text, $x );
 173
 174                 $text = $this->internalParse( $text, $linestart );
 175                 $text = $this->unstrip( $text, $this->mStripState );
 176                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 177                 if(!$wgUseTidy) {
 178                         $fixtags = array(
 179                                 # french spaces, last one Guillemet-left
 180                                 # only if there is something before the space
 181                                 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 182                                 # french spaces, Guillemet-right
 183                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 184                                 '/<hr *>/i' => '<hr />',
 185                                 '/<br *>/i' => '<br />',
 186                                 '/<center *>/i' => '<div class="center">',
 187                                 '/<\\/center *>/i' => '</div>',
 188                         );
 189                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 190                         $text = Sanitizer::normalizeCharReferences( $text );
 191                 } else {
 192                         $fixtags = array(
 193                                 # french spaces, last one Guillemet-left
 194                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 195                                 # french spaces, Guillemet-right
 196                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 197                                 '/<center *>/i' => '<div class="center">',
 198                                 '/<\\/center *>/i' => '</div>'
 199                         );
 200                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 201                 }
 202                 # only once and last
 203                 $text = $this->doBlockLevels( $text, $linestart );
 204
 205                 $this->replaceLinkHolders( $text );
 206                 $text = $wgContLang->convert($text);
 207
 208                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 209                 global $wgUseTidy;
 210                 if ($wgUseTidy) {
 211                         $text = Parser::tidy($text);
 212                 }
 213
 214                 $this->mOutput->setText( $text );
 215                 wfProfileOut( $fname );
 216                 return $this->mOutput;
 217         }
 218
 219         /**
 220          * Get a random string
 221          *
 222          * @access private
 223          * @static
 224          */
 225         function getRandomString() {
 226                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 227         }
 228
 229         /**
 230          * Replaces all occurrences of <$tag>content</$tag> in the text
 231          * with a random marker and returns the new text. the output parameter
 232          * $content will be an associative array filled with data on the form
 233          * $unique_marker => content.
 234          *
 235          * If $content is already set, the additional entries will be appended
 236          * If $tag is set to STRIP_COMMENTS, the function will extract
 237          * <!-- HTML comments -->
 238          *
 239          * @access private
 240          * @static
 241          */
 242         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 243                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 244                 if ( !$content ) {
 245                         $content = array( );
 246                 }
 247                 $n = 1;
 248                 $stripped = '';
 249
 250                 while ( '' != $text ) {
 251                         if($tag==STRIP_COMMENTS) {
 252                                 $p = preg_split( '/<!--/i', $text, 2 );
 253                         } else {
 254                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 255                         }
 256                         $stripped .= $p[0];
 257                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 258                                 $text = '';
 259                         } else {
 260                                 if($tag==STRIP_COMMENTS) {
 261                                         $q = preg_split( '/-->/i', $p[1], 2 );
 262                                 } else {
 263                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 264                                 }
 265                                 $marker = $rnd . sprintf('%08X', $n++);
 266                                 $content[$marker] = $q[0];
 267                                 $stripped .= $marker;
 268                                 $text = $q[1];
 269                         }
 270                 }
 271                 return $stripped;
 272         }
 273
 274         /**
 275          * Strips and renders nowiki, pre, math, hiero
 276          * If $render is set, performs necessary rendering operations on plugins
 277          * Returns the text, and fills an array with data needed in unstrip()
 278          * If the $state is already a valid strip state, it adds to the state
 279          *
 280          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 281          *  will be stripped in addition to other tags. This is important
 282          *  for section editing, where these comments cause confusion when
 283          *  counting the sections in the wikisource
 284          *
 285          * @access private
 286          */
 287         function strip( $text, &$state, $stripcomments = false ) {
 288                 $render = ($this->mOutputType == OT_HTML);
 289                 $html_content = array();
 290                 $nowiki_content = array();
 291                 $math_content = array();
 292                 $pre_content = array();
 293                 $comment_content = array();
 294                 $ext_content = array();
 295                 $gallery_content = array();
 296
 297                 # Replace any instances of the placeholders
 298                 $uniq_prefix = UNIQ_PREFIX;
 299                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 300
 301                 # html
 302                 global $wgRawHtml, $wgWhitelistEdit;
 303                 if( $wgRawHtml && $wgWhitelistEdit ) {
 304                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 305                         foreach( $html_content as $marker => $content ) {
 306                                 if ($render ) {
 307                                         # Raw and unchecked for validity.
 308                                         $html_content[$marker] = $content;
 309                                 } else {
 310                                         $html_content[$marker] = '<html>'.$content.'</html>';
 311                                 }
 312                         }
 313                 }
 314
 315                 # nowiki
 316                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 317                 foreach( $nowiki_content as $marker => $content ) {
 318                         if( $render ){
 319                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 320                         } else {
 321                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 322                         }
 323                 }
 324
 325                 # math
 326                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 327                 foreach( $math_content as $marker => $content ){
 328                         if( $render ) {
 329                                 if( $this->mOptions->getUseTeX() ) {
 330                                         $math_content[$marker] = renderMath( $content );
 331                                 } else {
 332                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 333                                 }
 334                         } else {
 335                                 $math_content[$marker] = '<math>'.$content.'</math>';
 336                         }
 337                 }
 338
 339                 # pre
 340                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 341                 foreach( $pre_content as $marker => $content ){
 342                         if( $render ){
 343                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 344                         } else {
 345                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 346                         }
 347                 }
 348
 349                 # gallery
 350                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 351                 foreach( $gallery_content as $marker => $content ) {
 352                         require_once( 'ImageGallery.php' );
 353                         if ( $render ) {
 354                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 355                         } else {
 356                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 357                         }
 358                 }
 359
 360                 # Comments
 361                 if($stripcomments) {
 362                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 363                         foreach( $comment_content as $marker => $content ){
 364                                 $comment_content[$marker] = '<!--'.$content.'-->';
 365                         }
 366                 }
 367
 368                 # Extensions
 369                 foreach ( $this->mTagHooks as $tag => $callback ) {
 370                         $ext_contents[$tag] = array();
 371                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 372                         foreach( $ext_content[$tag] as $marker => $content ) {
 373                                 if ( $render ) {
 374                                         $ext_content[$tag][$marker] = $callback( $content );
 375                                 } else {
 376                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 377                                 }
 378                         }
 379                 }
 380
 381                 # Merge state with the pre-existing state, if there is one
 382                 if ( $state ) {
 383                         $state['html'] = $state['html'] + $html_content;
 384                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 385                         $state['math'] = $state['math'] + $math_content;
 386                         $state['pre'] = $state['pre'] + $pre_content;
 387                         $state['comment'] = $state['comment'] + $comment_content;
 388                         $state['gallery'] = $state['gallery'] + $gallery_content;
 389
 390                         foreach( $ext_content as $tag => $array ) {
 391                                 if ( array_key_exists( $tag, $state ) ) {
 392                                         $state[$tag] = $state[$tag] + $array;
 393                                 }
 394                         }
 395                 } else {
 396                         $state = array(
 397                           'html' => $html_content,
 398                           'nowiki' => $nowiki_content,
 399                           'math' => $math_content,
 400                           'pre' => $pre_content,
 401                           'comment' => $comment_content,
 402                           'gallery' => $gallery_content,
 403                         ) + $ext_content;
 404                 }
 405                 return $text;
 406         }
 407
 408         /**
 409          * restores pre, math, and hiero removed by strip()
 410          *
 411          * always call unstripNoWiki() after this one
 412          * @access private
 413          */
 414         function unstrip( $text, &$state ) {
 415                 # Must expand in reverse order, otherwise nested tags will be corrupted
 416                 $contentDict = end( $state );
 417                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 418                         if( key($state) != 'nowiki' && key($state) != 'html') {
 419                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 420                                         $text = str_replace( key( $contentDict ), $content, $text );
 421                                 }
 422                         }
 423                 }
 424
 425                 return $text;
 426         }
 427
 428         /**
 429          * always call this after unstrip() to preserve the order
 430          *
 431          * @access private
 432          */
 433         function unstripNoWiki( $text, &$state ) {
 434                 # Must expand in reverse order, otherwise nested tags will be corrupted
 435                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 436                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 437                 }
 438
 439                 global $wgRawHtml;
 440                 if ($wgRawHtml) {
 441                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 442                                 $text = str_replace( key( $state['html'] ), $content, $text );
 443                         }
 444                 }
 445
 446                 return $text;
 447         }
 448
 449         /**
 450          * Add an item to the strip state
 451          * Returns the unique tag which must be inserted into the stripped text
 452          * The tag will be replaced with the original text in unstrip()
 453          *
 454          * @access private
 455          */
 456         function insertStripItem( $text, &$state ) {
 457                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 458                 if ( !$state ) {
 459                         $state = array(
 460                           'html' => array(),
 461                           'nowiki' => array(),
 462                           'math' => array(),
 463                           'pre' => array()
 464                         );
 465                 }
 466                 $state['item'][$rnd] = $text;
 467                 return $rnd;
 468         }
 469
 470         /**
 471          * interface with html tidy, used if $wgUseTidy = true
 472          *
 473          * @access public
 474          * @static
 475          */
 476         function tidy ( $text ) {
 477                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 478                 global $wgInputEncoding, $wgOutputEncoding;
 479                 $fname = 'Parser::tidy';
 480                 wfProfileIn( $fname );
 481
 482                 $cleansource = '';
 483                 $opts = '';
 484                 switch(strtoupper($wgOutputEncoding)) {
 485                         case 'ISO-8859-1':
 486                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 487                                 break;
 488                         case 'UTF-8':
 489                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 490                                 break;
 491                         default:
 492                                 $opts .= ' -raw';
 493                         }
 494
 495                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 496 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 497 '<head><title>test</title></head><body>'.$text.'</body></html>';
 498                 $descriptorspec = array(
 499                         0 => array('pipe', 'r'),
 500                         1 => array('pipe', 'w'),
 501                         2 => array('file', '/dev/null', 'a')
 502                 );
 503                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 504                 if (is_resource($process)) {
 505                         fwrite($pipes[0], $wrappedtext);
 506                         fclose($pipes[0]);
 507                         while (!feof($pipes[1])) {
 508                                 $cleansource .= fgets($pipes[1], 1024);
 509                         }
 510                         fclose($pipes[1]);
 511                         $return_value = proc_close($process);
 512                 }
 513
 514                 wfProfileOut( $fname );
 515
 516                 if( $cleansource == '' && $text != '') {
 517                         wfDebug( "Tidy error detected!\n" );
 518                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 519                 } else {
 520                         return $cleansource;
 521                 }
 522         }
 523
 524         /**
 525          * parse the wiki syntax used to render tables
 526          *
 527          * @access private
 528          */
 529         function doTableStuff ( $t ) {
 530                 $fname = 'Parser::doTableStuff';
 531                 wfProfileIn( $fname );
 532
 533                 $t = explode ( "\n" , $t ) ;
 534                 $td = array () ; # Is currently a td tag open?
 535                 $ltd = array () ; # Was it TD or TH?
 536                 $tr = array () ; # Is currently a tr tag open?
 537                 $ltr = array () ; # tr attributes
 538                 $indent_level = 0; # indent level of the table
 539                 foreach ( $t AS $k => $x )
 540                 {
 541                         $x = trim ( $x ) ;
 542                         $fc = substr ( $x , 0 , 1 ) ;
 543                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 544                                 $indent_level = strlen( $matches[1] );
 545                                 $t[$k] = "\n" .
 546                                         str_repeat( '<dl><dd>', $indent_level ) .
 547                                         '<table' . Sanitizer::fixTagAttributes ( $matches[2], 'table' ) . '>' ;
 548                                 array_push ( $td , false ) ;
 549                                 array_push ( $ltd , '' ) ;
 550                                 array_push ( $tr , false ) ;
 551                                 array_push ( $ltr , '' ) ;
 552                         }
 553                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 554                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 555                                 $z = "</table>\n" ;
 556                                 $l = array_pop ( $ltd ) ;
 557                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 558                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 559                                 array_pop ( $ltr ) ;
 560                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 561                         }
 562                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 563                                 $x = substr ( $x , 1 ) ;
 564                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 565                                 $z = '' ;
 566                                 $l = array_pop ( $ltd ) ;
 567                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 568                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 569                                 array_pop ( $ltr ) ;
 570                                 $t[$k] = $z ;
 571                                 array_push ( $tr , false ) ;
 572                                 array_push ( $td , false ) ;
 573                                 array_push ( $ltd , '' ) ;
 574                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $x, 'tr' ) ) ;
 575                         }
 576                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 577                                 # $x is a table row
 578                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 579                                         $fc = '+' ;
 580                                         $x = substr ( $x , 1 ) ;
 581                                 }
 582                                 $after = substr ( $x , 1 ) ;
 583                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 584                                 $after = explode ( '||' , $after ) ;
 585                                 $t[$k] = '' ;
 586
 587                                 # Loop through each table cell
 588                                 foreach ( $after AS $theline )
 589                                 {
 590                                         $z = '' ;
 591                                         if ( $fc != '+' )
 592                                         {
 593                                                 $tra = array_pop ( $ltr ) ;
 594                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 595                                                 array_push ( $tr , true ) ;
 596                                                 array_push ( $ltr , '' ) ;
 597                                         }
 598
 599                                         $l = array_pop ( $ltd ) ;
 600                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 601                                         if ( $fc == '|' ) $l = 'td' ;
 602                                         else if ( $fc == '!' ) $l = 'th' ;
 603                                         else if ( $fc == '+' ) $l = 'caption' ;
 604                                         else $l = '' ;
 605                                         array_push ( $ltd , $l ) ;
 606
 607                                         # Cell parameters
 608                                         $y = explode ( '|' , $theline , 2 ) ;
 609                                         # Note that a '|' inside an invalid link should not
 610                                         # be mistaken as delimiting cell parameters
 611                                         if ( strpos( $y[0], '[[' ) !== false ) {
 612                                                 $y = array ($theline);
 613                                         }
 614                                         if ( count ( $y ) == 1 )
 615                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 616                                         else $y = $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($y[0], $l).">{$y[1]}" ;
 617                                         $t[$k] .= $y ;
 618                                         array_push ( $td , true ) ;
 619                                 }
 620                         }
 621                 }
 622
 623                 # Closing open td, tr && table
 624                 while ( count ( $td ) > 0 )
 625                 {
 626                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 627                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 628                         $t[] = '</table>' ;
 629                 }
 630
 631                 $t = implode ( "\n" , $t ) ;
 632                 #               $t = Sanitizer::removeHTMLtags( $t );
 633                 wfProfileOut( $fname );
 634                 return $t ;
 635         }
 636
 637         /**
 638          * Helper function for parse() that transforms wiki markup into
 639          * HTML. Only called for $mOutputType == OT_HTML.
 640          *
 641          * @access private
 642          */
 643         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 644                 global $wgContLang;
 645
 646                 $fname = 'Parser::internalParse';
 647                 wfProfileIn( $fname );
 648
 649                 $text = Sanitizer::removeHTMLtags( $text );
 650                 $text = $this->replaceVariables( $text, $args );
 651
 652                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 653
 654                 $text = $this->doHeadings( $text );
 655                 if($this->mOptions->getUseDynamicDates()) {
 656                         global $wgDateFormatter;
 657                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 658                 }
 659                 $text = $this->doAllQuotes( $text );
 660                 $text = $this->replaceInternalLinks( $text );
 661                 $text = $this->replaceExternalLinks( $text );
 662
 663                 # replaceInternalLinks may sometimes leave behind
 664                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 665                 $text = str_replace("http-noparse://","http://",$text);
 666
 667                 $text = $this->doMagicLinks( $text );
 668                 $text = $this->doTableStuff( $text );
 669                 $text = $this->formatHeadings( $text, $isMain );
 670
 671                 wfProfileOut( $fname );
 672                 return $text;
 673         }
 674
 675         /**
 676          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 677          * magic external links.
 678          *
 679          * @access private
 680          */
 681         function &doMagicLinks( &$text ) {
 682                 $text = $this->magicISBN( $text );
 683                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 684                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 685                 return $text;
 686         }
 687
 688         /**
 689          * Parse ^^ tokens and return html
 690          *
 691          * @access private
 692          */
 693         function doExponent( $text ) {
 694                 $fname = 'Parser::doExponent';
 695                 wfProfileIn( $fname );
 696                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 697                 wfProfileOut( $fname );
 698                 return $text;
 699         }
 700
 701         /**
 702          * Parse headers and return html
 703          *
 704          * @access private
 705          */
 706         function doHeadings( $text ) {
 707                 $fname = 'Parser::doHeadings';
 708                 wfProfileIn( $fname );
 709                 for ( $i = 6; $i >= 1; --$i ) {
 710                         $h = substr( '======', 0, $i );
 711                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 712                           "<h{$i}>\\1</h{$i}>\\2", $text );
 713                 }
 714                 wfProfileOut( $fname );
 715                 return $text;
 716         }
 717
 718         /**
 719          * Replace single quotes with HTML markup
 720          * @access private
 721          * @return string the altered text
 722          */
 723         function doAllQuotes( $text ) {
 724                 $fname = 'Parser::doAllQuotes';
 725                 wfProfileIn( $fname );
 726                 $outtext = '';
 727                 $lines = explode( "\n", $text );
 728                 foreach ( $lines as $line ) {
 729                         $outtext .= $this->doQuotes ( $line ) . "\n";
 730                 }
 731                 $outtext = substr($outtext, 0,-1);
 732                 wfProfileOut( $fname );
 733                 return $outtext;
 734         }
 735
 736         /**
 737          * Helper function for doAllQuotes()
 738          * @access private
 739          */
 740         function doQuotes( $text ) {
 741                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 742                 if ( count( $arr ) == 1 )
 743                         return $text;
 744                 else
 745                 {
 746                         # First, do some preliminary work. This may shift some apostrophes from
 747                         # being mark-up to being text. It also counts the number of occurrences
 748                         # of bold and italics mark-ups.
 749                         $i = 0;
 750                         $numbold = 0;
 751                         $numitalics = 0;
 752                         foreach ( $arr as $r )
 753                         {
 754                                 if ( ( $i % 2 ) == 1 )
 755                                 {
 756                                         # If there are ever four apostrophes, assume the first is supposed to
 757                                         # be text, and the remaining three constitute mark-up for bold text.
 758                                         if ( strlen( $arr[$i] ) == 4 )
 759                                         {
 760                                                 $arr[$i-1] .= "'";
 761                                                 $arr[$i] = "'''";
 762                                         }
 763                                         # If there are more than 5 apostrophes in a row, assume they're all
 764                                         # text except for the last 5.
 765                                         else if ( strlen( $arr[$i] ) > 5 )
 766                                         {
 767                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 768                                                 $arr[$i] = "'''''";
 769                                         }
 770                                         # Count the number of occurrences of bold and italics mark-ups.
 771                                         # We are not counting sequences of five apostrophes.
 772                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 773                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 774                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 775                                 }
 776                                 $i++;
 777                         }
 778
 779                         # If there is an odd number of both bold and italics, it is likely
 780                         # that one of the bold ones was meant to be an apostrophe followed
 781                         # by italics. Which one we cannot know for certain, but it is more
 782                         # likely to be one that has a single-letter word before it.
 783                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 784                         {
 785                                 $i = 0;
 786                                 $firstsingleletterword = -1;
 787                                 $firstmultiletterword = -1;
 788                                 $firstspace = -1;
 789                                 foreach ( $arr as $r )
 790                                 {
 791                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 792                                         {
 793                                                 $x1 = substr ($arr[$i-1], -1);
 794                                                 $x2 = substr ($arr[$i-1], -2, 1);
 795                                                 if ($x1 == ' ') {
 796                                                         if ($firstspace == -1) $firstspace = $i;
 797                                                 } else if ($x2 == ' ') {
 798                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 799                                                 } else {
 800                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 801                                                 }
 802                                         }
 803                                         $i++;
 804                                 }
 805
 806                                 # If there is a single-letter word, use it!
 807                                 if ($firstsingleletterword > -1)
 808                                 {
 809                                         $arr [ $firstsingleletterword ] = "''";
 810                                         $arr [ $firstsingleletterword-1 ] .= "'";
 811                                 }
 812                                 # If not, but there's a multi-letter word, use that one.
 813                                 else if ($firstmultiletterword > -1)
 814                                 {
 815                                         $arr [ $firstmultiletterword ] = "''";
 816                                         $arr [ $firstmultiletterword-1 ] .= "'";
 817                                 }
 818                                 # ... otherwise use the first one that has neither.
 819                                 # (notice that it is possible for all three to be -1 if, for example,
 820                                 # there is only one pentuple-apostrophe in the line)
 821                                 else if ($firstspace > -1)
 822                                 {
 823                                         $arr [ $firstspace ] = "''";
 824                                         $arr [ $firstspace-1 ] .= "'";
 825                                 }
 826                         }
 827
 828                         # Now let's actually convert our apostrophic mush to HTML!
 829                         $output = '';
 830                         $buffer = '';
 831                         $state = '';
 832                         $i = 0;
 833                         foreach ($arr as $r)
 834                         {
 835                                 if (($i % 2) == 0)
 836                                 {
 837                                         if ($state == 'both')
 838                                                 $buffer .= $r;
 839                                         else
 840                                                 $output .= $r;
 841                                 }
 842                                 else
 843                                 {
 844                                         if (strlen ($r) == 2)
 845                                         {
 846                                                 if ($state == 'i')
 847                                                 { $output .= '</i>'; $state = ''; }
 848                                                 else if ($state == 'bi')
 849                                                 { $output .= '</i>'; $state = 'b'; }
 850                                                 else if ($state == 'ib')
 851                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 852                                                 else if ($state == 'both')
 853                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 854                                                 else # $state can be 'b' or ''
 855                                                 { $output .= '<i>'; $state .= 'i'; }
 856                                         }
 857                                         else if (strlen ($r) == 3)
 858                                         {
 859                                                 if ($state == 'b')
 860                                                 { $output .= '</b>'; $state = ''; }
 861                                                 else if ($state == 'bi')
 862                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 863                                                 else if ($state == 'ib')
 864                                                 { $output .= '</b>'; $state = 'i'; }
 865                                                 else if ($state == 'both')
 866                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 867                                                 else # $state can be 'i' or ''
 868                                                 { $output .= '<b>'; $state .= 'b'; }
 869                                         }
 870                                         else if (strlen ($r) == 5)
 871                                         {
 872                                                 if ($state == 'b')
 873                                                 { $output .= '</b><i>'; $state = 'i'; }
 874                                                 else if ($state == 'i')
 875                                                 { $output .= '</i><b>'; $state = 'b'; }
 876                                                 else if ($state == 'bi')
 877                                                 { $output .= '</i></b>'; $state = ''; }
 878                                                 else if ($state == 'ib')
 879                                                 { $output .= '</b></i>'; $state = ''; }
 880                                                 else if ($state == 'both')
 881                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 882                                                 else # ($state == '')
 883                                                 { $buffer = ''; $state = 'both'; }
 884                                         }
 885                                 }
 886                                 $i++;
 887                         }
 888                         # Now close all remaining tags.  Notice that the order is important.
 889                         if ($state == 'b' || $state == 'ib')
 890                                 $output .= '</b>';
 891                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 892                                 $output .= '</i>';
 893                         if ($state == 'bi')
 894                                 $output .= '</b>';
 895                         if ($state == 'both')
 896                                 $output .= '<b><i>'.$buffer.'</i></b>';
 897                         return $output;
 898                 }
 899         }
 900
 901         /**
 902          * Replace external links
 903          *
 904          * Note: this is all very hackish and the order of execution matters a lot.
 905          * Make sure to run maintenance/parserTests.php if you change this code.
 906          *
 907          * @access private
 908          */
 909         function replaceExternalLinks( $text ) {
 910                 $fname = 'Parser::replaceExternalLinks';
 911                 wfProfileIn( $fname );
 912
 913                 $sk =& $this->mOptions->getSkin();
 914                 global $wgContLang;
 915                 $linktrail = $wgContLang->linkTrail();
 916
 917                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 918
 919                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 920
 921                 $i = 0;
 922                 while ( $i<count( $bits ) ) {
 923                         $url = $bits[$i++];
 924                         $protocol = $bits[$i++];
 925                         $text = $bits[$i++];
 926                         $trail = $bits[$i++];
 927
 928                         # The characters '<' and '>' (which were escaped by
 929                         # removeHTMLtags()) should not be included in
 930                         # URLs, per RFC 2396.
 931                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 932                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 933                                 $url = substr($url, 0, $m2[0][1]);
 934                         }
 935
 936                         # If the link text is an image URL, replace it with an <img> tag
 937                         # This happened by accident in the original parser, but some people used it extensively
 938                         $img = $this->maybeMakeImageLink( $text );
 939                         if ( $img !== false ) {
 940                                 $text = $img;
 941                         }
 942
 943                         $dtrail = '';
 944
 945                         # Set linktype for CSS - if URL==text, link is essentially free
 946                         $linktype = ($text == $url) ? 'free' : 'text';
 947
 948                         # No link text, e.g. [http://domain.tld/some.link]
 949                         if ( $text == '' ) {
 950                                 # Autonumber if allowed
 951                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 952                                         $text = '[' . ++$this->mAutonumber . ']';
 953                                         $linktype = 'autonumber';
 954                                 } else {
 955                                         # Otherwise just use the URL
 956                                         $text = htmlspecialchars( $url );
 957                                         $linktype = 'free';
 958                                 }
 959                         } else {
 960                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 961                                 # Check for trail
 962                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 963                                         $dtrail = $m2[1];
 964                                         $trail = $m2[2];
 965                                 }
 966                         }
 967
 968                         # Replace &amp; from obsolete syntax with &.
 969                         # All HTML entities will be escaped by makeExternalLink()
 970                         # or maybeMakeImageLink()
 971                         $url = str_replace( '&amp;', '&', $url );
 972
 973                         # Process the trail (i.e. everything after this link up until start of the next link),
 974                         # replacing any non-bracketed links
 975                         $trail = $this->replaceFreeExternalLinks( $trail );
 976
 977                         # Use the encoded URL
 978                         # This means that users can paste URLs directly into the text
 979                         # Funny characters like &ouml; aren't valid in URLs anyway
 980                         # This was changed in August 2004
 981                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
 982                 }
 983
 984                 wfProfileOut( $fname );
 985                 return $s;
 986         }
 987
 988         /**
 989          * Replace anything that looks like a URL with a link
 990          * @access private
 991          */
 992         function replaceFreeExternalLinks( $text ) {
 993                 $fname = 'Parser::replaceFreeExternalLinks';
 994                 wfProfileIn( $fname );
 995
 996                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 997                 $s = array_shift( $bits );
 998                 $i = 0;
 999
1000                 $sk =& $this->mOptions->getSkin();
1001
1002                 while ( $i < count( $bits ) ){
1003                         $protocol = $bits[$i++];
1004                         $remainder = $bits[$i++];
1005
1006                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1007                                 # Found some characters after the protocol that look promising
1008                                 $url = $protocol . $m[1];
1009                                 $trail = $m[2];
1010
1011                                 # The characters '<' and '>' (which were escaped by
1012                                 # removeHTMLtags()) should not be included in
1013                                 # URLs, per RFC 2396.
1014                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1015                                         $trail = substr($url, $m2[0][1]) . $trail;
1016                                         $url = substr($url, 0, $m2[0][1]);
1017                                 }
1018
1019                                 # Move trailing punctuation to $trail
1020                                 $sep = ',;\.:!?';
1021                                 # If there is no left bracket, then consider right brackets fair game too
1022                                 if ( strpos( $url, '(' ) === false ) {
1023                                         $sep .= ')';
1024                                 }
1025
1026                                 $numSepChars = strspn( strrev( $url ), $sep );
1027                                 if ( $numSepChars ) {
1028                                         $trail = substr( $url, -$numSepChars ) . $trail;
1029                                         $url = substr( $url, 0, -$numSepChars );
1030                                 }
1031
1032                                 # Replace &amp; from obsolete syntax with &.
1033                                 # All HTML entities will be escaped by makeExternalLink()
1034                                 # or maybeMakeImageLink()
1035                                 $url = str_replace( '&amp;', '&', $url );
1036
1037                                 # Is this an external image?
1038                                 $text = $this->maybeMakeImageLink( $url );
1039                                 if ( $text === false ) {
1040                                         # Not an image, make a link
1041                                         $text = $sk->makeExternalLink( $url, $url, true, 'free' );
1042                                 }
1043                                 $s .= $text . $trail;
1044                         } else {
1045                                 $s .= $protocol . $remainder;
1046                         }
1047                 }
1048                 wfProfileOut();
1049                 return $s;
1050         }
1051
1052         /**
1053          * make an image if it's allowed
1054          * @access private
1055          */
1056         function maybeMakeImageLink( $url ) {
1057                 $sk =& $this->mOptions->getSkin();
1058                 $text = false;
1059                 if ( $this->mOptions->getAllowExternalImages() ) {
1060                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1061                                 # Image found
1062                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1063                         }
1064                 }
1065                 return $text;
1066         }
1067
1068         /**
1069          * Process [[ ]] wikilinks
1070          *
1071          * @access private
1072          */
1073         function replaceInternalLinks( $s ) {
1074                 global $wgLang, $wgContLang, $wgLinkCache;
1075                 global $wgDisableLangConversion;
1076                 static $fname = 'Parser::replaceInternalLinks' ;
1077
1078                 wfProfileIn( $fname );
1079
1080                 wfProfileIn( $fname.'-setup' );
1081                 static $tc = FALSE;
1082                 # the % is needed to support urlencoded titles as well
1083                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1084
1085                 $sk =& $this->mOptions->getSkin();
1086                 global $wgUseOldExistenceCheck;
1087                 # "Post-parse link colour check" works only on wiki text since it's now
1088                 # in Parser. Enable it, then disable it when we're done.
1089                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1090
1091                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1092
1093                 #split the entire text string on occurences of [[
1094                 $a = explode( '[[', ' ' . $s );
1095                 #get the first element (all text up to first [[), and remove the space we added
1096                 $s = array_shift( $a );
1097                 $s = substr( $s, 1 );
1098
1099                 # Match a link having the form [[namespace:link|alternate]]trail
1100                 static $e1 = FALSE;
1101                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1102                 # Match cases where there is no "]]", which might still be images
1103                 static $e1_img = FALSE;
1104                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1105                 # Match the end of a line for a word that's not followed by whitespace,
1106                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1107                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1108
1109                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1110
1111                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1112
1113                 if ( $useLinkPrefixExtension ) {
1114                         if ( preg_match( $e2, $s, $m ) ) {
1115                                 $first_prefix = $m[2];
1116                                 $s = $m[1];
1117                         } else {
1118                                 $first_prefix = false;
1119                         }
1120                 } else {
1121                         $prefix = '';
1122                 }
1123
1124                 $selflink = $this->mTitle->getPrefixedText();
1125                 wfProfileOut( $fname.'-setup' );
1126
1127                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1128                 $useSubpages = $this->areSubpagesAllowed();
1129
1130                 # Loop for each link
1131                 for ($k = 0; isset( $a[$k] ); $k++) {
1132                         $line = $a[$k];
1133                         if ( $useLinkPrefixExtension ) {
1134                                 wfProfileIn( $fname.'-prefixhandling' );
1135                                 if ( preg_match( $e2, $s, $m ) ) {
1136                                         $prefix = $m[2];
1137                                         $s = $m[1];
1138                                 } else {
1139                                         $prefix='';
1140                                 }
1141                                 # first link
1142                                 if($first_prefix) {
1143                                         $prefix = $first_prefix;
1144                                         $first_prefix = false;
1145                                 }
1146                                 wfProfileOut( $fname.'-prefixhandling' );
1147                         }
1148
1149                         $might_be_img = false;
1150
1151                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1152                                 $text = $m[2];
1153                                 # fix up urlencoded title texts
1154                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1155                                 $trail = $m[3];
1156                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1157                                 $might_be_img = true;
1158                                 $text = $m[2];
1159                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1160                                 $trail = "";
1161                         } else { # Invalid form; output directly
1162                                 $s .= $prefix . '[[' . $line ;
1163                                 continue;
1164                         }
1165
1166                         # Don't allow internal links to pages containing
1167                         # PROTO: where PROTO is a valid URL protocol; these
1168                         # should be external links.
1169                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1170                                 $s .= $prefix . '[[' . $line ;
1171                                 continue;
1172                         }
1173
1174                         # Make subpage if necessary
1175                         if( $useSubpages ) {
1176                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1177                         } else {
1178                                 $link = $m[1];
1179                         }
1180
1181                         $noforce = (substr($m[1], 0, 1) != ':');
1182                         if (!$noforce) {
1183                                 # Strip off leading ':'
1184                                 $link = substr($link, 1);
1185                         }
1186
1187                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1188                         if( !$nt ) {
1189                                 $s .= $prefix . '[[' . $line;
1190                                 continue;
1191                         }
1192
1193                         #check other language variants of the link
1194                         #if the article does not exist
1195                         if( $checkVariantLink
1196                             && $nt->getArticleID() == 0 ) {
1197                                 $wgContLang->findVariantLink($link, $nt);
1198                         }
1199
1200                         $ns = $nt->getNamespace();
1201                         $iw = $nt->getInterWiki();
1202
1203                         if ($might_be_img) { # if this is actually an invalid link
1204                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1205                                         $found = false;
1206                                         while (isset ($a[$k+1]) ) {
1207                                                 #look at the next 'line' to see if we can close it there
1208                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1209                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1210                                                 # the first ]] closes the inner link, the second the image
1211                                                         $found = true;
1212                                                         $text .= '[[' . $m[1];
1213                                                         $trail = $m[2];
1214                                                         break;
1215                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1216                                                         #if there's exactly one ]] that's fine, we'll keep looking
1217                                                         $text .= '[[' . $m[0];
1218                                                 } else {
1219                                                         #if $next_line is invalid too, we need look no further
1220                                                         $text .= '[[' . $next_line;
1221                                                         break;
1222                                                 }
1223                                         }
1224                                         if ( !$found ) {
1225                                                 # we couldn't find the end of this imageLink, so output it raw
1226                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1227                                                 $text = $this->replaceInternalLinks($text);
1228                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1229                                                 # note: no $trail, because without an end, there *is* no trail
1230                                                 continue;
1231                                         }
1232                                 } else { #it's not an image, so output it raw
1233                                         $s .= $prefix . '[[' . $link . '|' . $text;
1234                                         # note: no $trail, because without an end, there *is* no trail
1235                                         continue;
1236                                 }
1237                         }
1238
1239                         $wasblank = ( '' == $text );
1240                         if( $wasblank ) $text = $link;
1241
1242
1243                         # Link not escaped by : , create the various objects
1244                         if( $noforce ) {
1245
1246                                 # Interwikis
1247                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1248                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1249                                         $tmp = $prefix . $trail ;
1250                                         $s .= (trim($tmp) == '')? '': $tmp;
1251                                         continue;
1252                                 }
1253
1254                                 if ( $ns == NS_IMAGE ) {
1255                                         wfProfileIn( "$fname-image" );
1256
1257                                         # recursively parse links inside the image caption
1258                                         # actually, this will parse them in any other parameters, too,
1259                                         # but it might be hard to fix that, and it doesn't matter ATM
1260                                         $text = $this->replaceExternalLinks($text);
1261                                         $text = $this->replaceInternalLinks($text);
1262
1263                                         # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1264                                         $s .= $prefix . str_replace('http://', 'http-noparse://', $sk->makeImageLinkObj( $nt, $text ) ) . $trail;
1265                                         $wgLinkCache->addImageLinkObj( $nt );
1266
1267                                         wfProfileOut( "$fname-image" );
1268                                         continue;
1269                                 }
1270
1271                                 if ( $ns == NS_CATEGORY ) {
1272                                         wfProfileIn( "$fname-category" );
1273                                         $t = $nt->getText();
1274
1275                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1276                                         $pPLC=$sk->postParseLinkColour();
1277                                         $sk->postParseLinkColour( false );
1278                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1279                                         $sk->postParseLinkColour( $pPLC );
1280                                         $wgLinkCache->resume();
1281
1282                                         if ( $wasblank ) {
1283                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1284                                                         $sortkey = $this->mTitle->getText();
1285                                                 } else {
1286                                                         $sortkey = $this->mTitle->getPrefixedText();
1287                                                 }
1288                                         } else {
1289                                                 $sortkey = $text;
1290                                         }
1291                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1292                                         $this->mOutput->addCategoryLink( $t );
1293                                         $s .= $prefix . $trail ;
1294
1295                                         wfProfileOut( "$fname-category" );
1296                                         continue;
1297                                 }
1298                         }
1299
1300                         if( ( $nt->getPrefixedText() === $selflink ) &&
1301                             ( $nt->getFragment() === '' ) ) {
1302                                 # Self-links are handled specially; generally de-link and change to bold.
1303                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1304                                 continue;
1305                         }
1306
1307                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1308                         if( $ns == NS_MEDIA ) {
1309                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1310                                 $wgLinkCache->addImageLinkObj( $nt );
1311                                 continue;
1312                         } elseif( $ns == NS_SPECIAL ) {
1313                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1314                                 continue;
1315                         }
1316                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1317                 }
1318                 $sk->postParseLinkColour( $saveParseColour );
1319                 wfProfileOut( $fname );
1320                 return $s;
1321         }
1322
1323         /**
1324          * Return true if subpage links should be expanded on this page.
1325          * @return bool
1326          */
1327         function areSubpagesAllowed() {
1328                 # Some namespaces don't allow subpages
1329                 global $wgNamespacesWithSubpages;
1330                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1331         }
1332
1333         /**
1334          * Handle link to subpage if necessary
1335          * @param string $target the source of the link
1336          * @param string &$text the link text, modified as necessary
1337          * @return string the full name of the link
1338          * @access private
1339          */
1340         function maybeDoSubpageLink($target, &$text) {
1341                 # Valid link forms:
1342                 # Foobar -- normal
1343                 # :Foobar -- override special treatment of prefix (images, language links)
1344                 # /Foobar -- convert to CurrentPage/Foobar
1345                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1346                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1347                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1348
1349                 $fname = 'Parser::maybeDoSubpageLink';
1350                 wfProfileIn( $fname );
1351                 $ret = $target; # default return value is no change
1352
1353                 # Some namespaces don't allow subpages,
1354                 # so only perform processing if subpages are allowed
1355                 if( $this->areSubpagesAllowed() ) {
1356                         # Look at the first character
1357                         if( $target != '' && $target{0} == '/' ) {
1358                                 # / at end means we don't want the slash to be shown
1359                                 if( substr( $target, -1, 1 ) == '/' ) {
1360                                         $target = substr( $target, 1, -1 );
1361                                         $noslash = $target;
1362                                 } else {
1363                                         $noslash = substr( $target, 1 );
1364                                 }
1365
1366                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1367                                 if( '' === $text ) {
1368                                         $text = $target;
1369                                 } # this might be changed for ugliness reasons
1370                         } else {
1371                                 # check for .. subpage backlinks
1372                                 $dotdotcount = 0;
1373                                 $nodotdot = $target;
1374                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1375                                         ++$dotdotcount;
1376                                         $nodotdot = substr( $nodotdot, 3 );
1377                                 }
1378                                 if($dotdotcount > 0) {
1379                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1380                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1381                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1382                                                 # / at the end means don't show full path
1383                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1384                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1385                                                         if( '' === $text ) {
1386                                                                 $text = $nodotdot;
1387                                                         }
1388                                                 }
1389                                                 $nodotdot = trim( $nodotdot );
1390                                                 if( $nodotdot != '' ) {
1391                                                         $ret .= '/' . $nodotdot;
1392                                                 }
1393                                         }
1394                                 }
1395                         }
1396                 }
1397
1398                 wfProfileOut( $fname );
1399                 return $ret;
1400         }
1401
1402         /**#@+
1403          * Used by doBlockLevels()
1404          * @access private
1405          */
1406         /* private */ function closeParagraph() {
1407                 $result = '';
1408                 if ( '' != $this->mLastSection ) {
1409                         $result = '</' . $this->mLastSection  . ">\n";
1410                 }
1411                 $this->mInPre = false;
1412                 $this->mLastSection = '';
1413                 return $result;
1414         }
1415         # getCommon() returns the length of the longest common substring
1416         # of both arguments, starting at the beginning of both.
1417         #
1418         /* private */ function getCommon( $st1, $st2 ) {
1419                 $fl = strlen( $st1 );
1420                 $shorter = strlen( $st2 );
1421                 if ( $fl < $shorter ) { $shorter = $fl; }
1422
1423                 for ( $i = 0; $i < $shorter; ++$i ) {
1424                         if ( $st1{$i} != $st2{$i} ) { break; }
1425                 }
1426                 return $i;
1427         }
1428         # These next three functions open, continue, and close the list
1429         # element appropriate to the prefix character passed into them.
1430         #
1431         /* private */ function openList( $char ) {
1432                 $result = $this->closeParagraph();
1433
1434                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1435                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1436                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1437                 else if ( ';' == $char ) {
1438                         $result .= '<dl><dt>';
1439                         $this->mDTopen = true;
1440                 }
1441                 else { $result = '<!-- ERR 1 -->'; }
1442
1443                 return $result;
1444         }
1445
1446         /* private */ function nextItem( $char ) {
1447                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1448                 else if ( ':' == $char || ';' == $char ) {
1449                         $close = '</dd>';
1450                         if ( $this->mDTopen ) { $close = '</dt>'; }
1451                         if ( ';' == $char ) {
1452                                 $this->mDTopen = true;
1453                                 return $close . '<dt>';
1454                         } else {
1455                                 $this->mDTopen = false;
1456                                 return $close . '<dd>';
1457                         }
1458                 }
1459                 return '<!-- ERR 2 -->';
1460         }
1461
1462         /* private */ function closeList( $char ) {
1463                 if ( '*' == $char ) { $text = '</li></ul>'; }
1464                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1465                 else if ( ':' == $char ) {
1466                         if ( $this->mDTopen ) {
1467                                 $this->mDTopen = false;
1468                                 $text = '</dt></dl>';
1469                         } else {
1470                                 $text = '</dd></dl>';
1471                         }
1472                 }
1473                 else {  return '<!-- ERR 3 -->'; }
1474                 return $text."\n";
1475         }
1476         /**#@-*/
1477
1478         /**
1479          * Make lists from lines starting with ':', '*', '#', etc.
1480          *
1481          * @access private
1482          * @return string the lists rendered as HTML
1483          */
1484         function doBlockLevels( $text, $linestart ) {
1485                 $fname = 'Parser::doBlockLevels';
1486                 wfProfileIn( $fname );
1487
1488                 # Parsing through the text line by line.  The main thing
1489                 # happening here is handling of block-level elements p, pre,
1490                 # and making lists from lines starting with * # : etc.
1491                 #
1492                 $textLines = explode( "\n", $text );
1493
1494                 $lastPrefix = $output = $lastLine = '';
1495                 $this->mDTopen = $inBlockElem = false;
1496                 $prefixLength = 0;
1497                 $paragraphStack = false;
1498
1499                 if ( !$linestart ) {
1500                         $output .= array_shift( $textLines );
1501                 }
1502                 foreach ( $textLines as $oLine ) {
1503                         $lastPrefixLength = strlen( $lastPrefix );
1504                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1505                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1506                         if ( !$this->mInPre ) {
1507                                 # Multiple prefixes may abut each other for nested lists.
1508                                 $prefixLength = strspn( $oLine, '*#:;' );
1509                                 $pref = substr( $oLine, 0, $prefixLength );
1510
1511                                 # eh?
1512                                 $pref2 = str_replace( ';', ':', $pref );
1513                                 $t = substr( $oLine, $prefixLength );
1514                                 $this->mInPre = !empty($preOpenMatch);
1515                         } else {
1516                                 # Don't interpret any other prefixes in preformatted text
1517                                 $prefixLength = 0;
1518                                 $pref = $pref2 = '';
1519                                 $t = $oLine;
1520                         }
1521
1522                         # List generation
1523                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1524                                 # Same as the last item, so no need to deal with nesting or opening stuff
1525                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1526                                 $paragraphStack = false;
1527
1528                                 if ( substr( $pref, -1 ) == ';') {
1529                                         # The one nasty exception: definition lists work like this:
1530                                         # ; title : definition text
1531                                         # So we check for : in the remainder text to split up the
1532                                         # title and definition, without b0rking links.
1533                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1534                                                 $t = $t2;
1535                                                 $output .= $term . $this->nextItem( ':' );
1536                                         }
1537                                 }
1538                         } elseif( $prefixLength || $lastPrefixLength ) {
1539                                 # Either open or close a level...
1540                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1541                                 $paragraphStack = false;
1542
1543                                 while( $commonPrefixLength < $lastPrefixLength ) {
1544                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1545                                         --$lastPrefixLength;
1546                                 }
1547                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1548                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1549                                 }
1550                                 while ( $prefixLength > $commonPrefixLength ) {
1551                                         $char = substr( $pref, $commonPrefixLength, 1 );
1552                                         $output .= $this->openList( $char );
1553
1554                                         if ( ';' == $char ) {
1555                                                 # FIXME: This is dupe of code above
1556                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1557                                                         $t = $t2;
1558                                                         $output .= $term . $this->nextItem( ':' );
1559                                                 }
1560                                         }
1561                                         ++$commonPrefixLength;
1562                                 }
1563                                 $lastPrefix = $pref2;
1564                         }
1565                         if( 0 == $prefixLength ) {
1566                                 wfProfileIn( "$fname-paragraph" );
1567                                 # No prefix (not in list)--go to paragraph mode
1568                                 $uniq_prefix = UNIQ_PREFIX;
1569                                 // XXX: use a stack for nestable elements like span, table and div
1570                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1571                                 $closematch = preg_match(
1572                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1573                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1574                                 if ( $openmatch or $closematch ) {
1575                                         $paragraphStack = false;
1576                                         $output .= $this->closeParagraph();
1577                                         if($preOpenMatch and !$preCloseMatch) {
1578                                                 $this->mInPre = true;
1579                                         }
1580                                         if ( $closematch ) {
1581                                                 $inBlockElem = false;
1582                                         } else {
1583                                                 $inBlockElem = true;
1584                                         }
1585                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1586                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1587                                                 // pre
1588                                                 if ($this->mLastSection != 'pre') {
1589                                                         $paragraphStack = false;
1590                                                         $output .= $this->closeParagraph().'<pre>';
1591                                                         $this->mLastSection = 'pre';
1592                                                 }
1593                                                 $t = substr( $t, 1 );
1594                                         } else {
1595                                                 // paragraph
1596                                                 if ( '' == trim($t) ) {
1597                                                         if ( $paragraphStack ) {
1598                                                                 $output .= $paragraphStack.'<br />';
1599                                                                 $paragraphStack = false;
1600                                                                 $this->mLastSection = 'p';
1601                                                         } else {
1602                                                                 if ($this->mLastSection != 'p' ) {
1603                                                                         $output .= $this->closeParagraph();
1604                                                                         $this->mLastSection = '';
1605                                                                         $paragraphStack = '<p>';
1606                                                                 } else {
1607                                                                         $paragraphStack = '</p><p>';
1608                                                                 }
1609                                                         }
1610                                                 } else {
1611                                                         if ( $paragraphStack ) {
1612                                                                 $output .= $paragraphStack;
1613                                                                 $paragraphStack = false;
1614                                                                 $this->mLastSection = 'p';
1615                                                         } else if ($this->mLastSection != 'p') {
1616                                                                 $output .= $this->closeParagraph().'<p>';
1617                                                                 $this->mLastSection = 'p';
1618                                                         }
1619                                                 }
1620                                         }
1621                                 }
1622                                 wfProfileOut( "$fname-paragraph" );
1623                         }
1624                         if ($paragraphStack === false) {
1625                                 $output .= $t."\n";
1626                         }
1627                 }
1628                 while ( $prefixLength ) {
1629                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1630                         --$prefixLength;
1631                 }
1632                 if ( '' != $this->mLastSection ) {
1633                         $output .= '</' . $this->mLastSection . '>';
1634                         $this->mLastSection = '';
1635                 }
1636
1637                 wfProfileOut( $fname );
1638                 return $output;
1639         }
1640
1641         /**
1642          * Split up a string on ':', ignoring any occurences inside
1643          * <a>..</a> or <span>...</span>
1644          * @param string $str the string to split
1645          * @param string &$before set to everything before the ':'
1646          * @param string &$after set to everything after the ':'
1647          * return string the position of the ':', or false if none found
1648          */
1649         function findColonNoLinks($str, &$before, &$after) {
1650                 # I wonder if we should make this count all tags, not just <a>
1651                 # and <span>. That would prevent us from matching a ':' that
1652                 # comes in the middle of italics other such formatting....
1653                 # -- Wil
1654                 $fname = 'Parser::findColonNoLinks';
1655                 wfProfileIn( $fname );
1656                 $pos = 0;
1657                 do {
1658                         $colon = strpos($str, ':', $pos);
1659
1660                         if ($colon !== false) {
1661                                 $before = substr($str, 0, $colon);
1662                                 $after = substr($str, $colon + 1);
1663
1664                                 # Skip any ':' within <a> or <span> pairs
1665                                 $a = substr_count($before, '<a');
1666                                 $s = substr_count($before, '<span');
1667                                 $ca = substr_count($before, '</a>');
1668                                 $cs = substr_count($before, '</span>');
1669
1670                                 if ($a <= $ca and $s <= $cs) {
1671                                         # Tags are balanced before ':'; ok
1672                                         break;
1673                                 }
1674                                 $pos = $colon + 1;
1675                         }
1676                 } while ($colon !== false);
1677                 wfProfileOut( $fname );
1678                 return $colon;
1679         }
1680
1681         /**
1682          * Return value of a magic variable (like PAGENAME)
1683          *
1684          * @access private
1685          */
1686         function getVariableValue( $index ) {
1687                 global $wgContLang, $wgSitename, $wgServer;
1688
1689                 /**
1690                  * Some of these require message or data lookups and can be
1691                  * expensive to check many times.
1692                  */
1693                 static $varCache = array();
1694                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1695
1696                 switch ( $index ) {
1697                         case MAG_CURRENTMONTH:
1698                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1699                         case MAG_CURRENTMONTHNAME:
1700                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1701                         case MAG_CURRENTMONTHNAMEGEN:
1702                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1703                         case MAG_CURRENTDAY:
1704                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1705                         case MAG_PAGENAME:
1706                                 return $this->mTitle->getText();
1707                         case MAG_PAGENAMEE:
1708                                 return $this->mTitle->getPartialURL();
1709                         case MAG_NAMESPACE:
1710                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1711                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1712                         case MAG_CURRENTDAYNAME:
1713                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1714                         case MAG_CURRENTYEAR:
1715                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1716                         case MAG_CURRENTTIME:
1717                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1718                         case MAG_CURRENTWEEK:
1719                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1720                         case MAG_CURRENTDOW:
1721                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1722                         case MAG_NUMBEROFARTICLES:
1723                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1724                         case MAG_SITENAME:
1725                                 return $wgSitename;
1726                         case MAG_SERVER:
1727                                 return $wgServer;
1728                         default:
1729                                 return NULL;
1730                 }
1731         }
1732
1733         /**
1734          * initialise the magic variables (like CURRENTMONTHNAME)
1735          *
1736          * @access private
1737          */
1738         function initialiseVariables() {
1739                 $fname = 'Parser::initialiseVariables';
1740                 wfProfileIn( $fname );
1741                 global $wgVariableIDs;
1742                 $this->mVariables = array();
1743                 foreach ( $wgVariableIDs as $id ) {
1744                         $mw =& MagicWord::get( $id );
1745                         $mw->addToArray( $this->mVariables, $id );
1746                 }
1747                 wfProfileOut( $fname );
1748         }
1749
1750         /**
1751          * Replace magic variables, templates, and template arguments
1752          * with the appropriate text. Templates are substituted recursively,
1753          * taking care to avoid infinite loops.
1754          *
1755          * Note that the substitution depends on value of $mOutputType:
1756          *  OT_WIKI: only {{subst:}} templates
1757          *  OT_MSG: only magic variables
1758          *  OT_HTML: all templates and magic variables
1759          *
1760          * @param string $tex The text to transform
1761          * @param array $args Key-value pairs representing template parameters to substitute
1762          * @access private
1763          */
1764         function replaceVariables( $text, $args = array() ) {
1765                 global $wgLang, $wgScript, $wgArticlePath;
1766
1767                 # Prevent too big inclusions
1768                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1769                         return $text;
1770                 }
1771
1772                 $fname = 'Parser::replaceVariables';
1773                 wfProfileIn( $fname );
1774
1775                 $titleChars = Title::legalChars();
1776
1777                 # This function is called recursively. To keep track of arguments we need a stack:
1778                 array_push( $this->mArgStack, $args );
1779
1780                 # Variable substitution
1781                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1782
1783                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1784                         # Argument substitution
1785                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1786                 }
1787                 # Template substitution
1788                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1789                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1790
1791                 array_pop( $this->mArgStack );
1792
1793                 wfProfileOut( $fname );
1794                 return $text;
1795         }
1796
1797         /**
1798          * Replace magic variables
1799          * @access private
1800          */
1801         function variableSubstitution( $matches ) {
1802                 $fname = 'parser::variableSubstitution';
1803                 $varname = $matches[1];
1804                 wfProfileIn( $fname );
1805                 if ( !$this->mVariables ) {
1806                         $this->initialiseVariables();
1807                 }
1808                 $skip = false;
1809                 if ( $this->mOutputType == OT_WIKI ) {
1810                         # Do only magic variables prefixed by SUBST
1811                         $mwSubst =& MagicWord::get( MAG_SUBST );
1812                         if (!$mwSubst->matchStartAndRemove( $varname ))
1813                                 $skip = true;
1814                         # Note that if we don't substitute the variable below,
1815                         # we don't remove the {{subst:}} magic word, in case
1816                         # it is a template rather than a magic variable.
1817                 }
1818                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
1819                         $id = $this->mVariables[$varname];
1820                         $text = $this->getVariableValue( $id );
1821                         $this->mOutput->mContainsOldMagic = true;
1822                 } else {
1823                         $text = $matches[0];
1824                 }
1825                 wfProfileOut( $fname );
1826                 return $text;
1827         }
1828
1829         # Split template arguments
1830         function getTemplateArgs( $argsString ) {
1831                 if ( $argsString === '' ) {
1832                         return array();
1833                 }
1834
1835                 $args = explode( '|', substr( $argsString, 1 ) );
1836
1837                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1838                 # merged with the next arg because the '|' character between belongs
1839                 # to the link syntax and not the template parameter syntax.
1840                 $argc = count($args);
1841                 $i = 0;
1842                 for ( $i = 0; $i < $argc-1; $i++ ) {
1843                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1844                                 $args[$i] .= '|'.$args[$i+1];
1845                                 array_splice($args, $i+1, 1);
1846                                 $i--;
1847                                 $argc--;
1848                         }
1849                 }
1850
1851                 return $args;
1852         }
1853
1854         /**
1855          * Return the text of a template, after recursively
1856          * replacing any variables or templates within the template.
1857          *
1858          * @param array $matches The parts of the template
1859          *  $matches[1]: the title, i.e. the part before the |
1860          *  $matches[2]: the parameters (including a leading |), if  any
1861          * @return string the text of the template
1862          * @access private
1863          */
1864         function braceSubstitution( $matches ) {
1865                 global $wgLinkCache, $wgContLang;
1866                 $fname = 'Parser::braceSubstitution';
1867                 wfProfileIn( $fname );
1868
1869                 $found = false;
1870                 $nowiki = false;
1871                 $noparse = false;
1872
1873                 $title = NULL;
1874
1875                 # Need to know if the template comes at the start of a line,
1876                 # to treat the beginning of the template like the beginning
1877                 # of a line for tables and block-level elements.
1878                 $linestart = $matches[1];
1879
1880                 # $part1 is the bit before the first |, and must contain only title characters
1881                 # $args is a list of arguments, starting from index 0, not including $part1
1882
1883                 $part1 = $matches[2];
1884                 # If the third subpattern matched anything, it will start with |
1885
1886                 $args = $this->getTemplateArgs($matches[3]);
1887                 $argc = count( $args );
1888
1889                 # Don't parse {{{}}} because that's only for template arguments
1890                 if ( $linestart === '{' ) {
1891                         $text = $matches[0];
1892                         $found = true;
1893                         $noparse = true;
1894                 }
1895
1896                 # SUBST
1897                 if ( !$found ) {
1898                         $mwSubst =& MagicWord::get( MAG_SUBST );
1899                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1900                                 # One of two possibilities is true:
1901                                 # 1) Found SUBST but not in the PST phase
1902                                 # 2) Didn't find SUBST and in the PST phase
1903                                 # In either case, return without further processing
1904                                 $text = $matches[0];
1905                                 $found = true;
1906                                 $noparse = true;
1907                         }
1908                 }
1909
1910                 # MSG, MSGNW and INT
1911                 if ( !$found ) {
1912                         # Check for MSGNW:
1913                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1914                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1915                                 $nowiki = true;
1916                         } else {
1917                                 # Remove obsolete MSG:
1918                                 $mwMsg =& MagicWord::get( MAG_MSG );
1919                                 $mwMsg->matchStartAndRemove( $part1 );
1920                         }
1921
1922                         # Check if it is an internal message
1923                         $mwInt =& MagicWord::get( MAG_INT );
1924                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1925                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1926                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1927                                         $found = true;
1928                                 }
1929                         }
1930                 }
1931
1932                 # NS
1933                 if ( !$found ) {
1934                         # Check for NS: (namespace expansion)
1935                         $mwNs = MagicWord::get( MAG_NS );
1936                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1937                                 if ( intval( $part1 ) ) {
1938                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1939                                         $found = true;
1940                                 } else {
1941                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1942                                         if ( !is_null( $index ) ) {
1943                                                 $text = $linestart . $wgContLang->getNsText( $index );
1944                                                 $found = true;
1945                                         }
1946                                 }
1947                         }
1948                 }
1949
1950                 # LOCALURL and LOCALURLE
1951                 if ( !$found ) {
1952                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1953                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1954
1955                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1956                                 $func = 'getLocalURL';
1957                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1958                                 $func = 'escapeLocalURL';
1959                         } else {
1960                                 $func = '';
1961                         }
1962
1963                         if ( $func !== '' ) {
1964                                 $title = Title::newFromText( $part1 );
1965                                 if ( !is_null( $title ) ) {
1966                                         if ( $argc > 0 ) {
1967                                                 $text = $linestart . $title->$func( $args[0] );
1968                                         } else {
1969                                                 $text = $linestart . $title->$func();
1970                                         }
1971                                         $found = true;
1972                                 }
1973                         }
1974                 }
1975
1976                 # GRAMMAR
1977                 if ( !$found && $argc == 1 ) {
1978                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1979                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1980                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
1981                                 $found = true;
1982                         }
1983                 }
1984
1985                 # Template table test
1986
1987                 # Did we encounter this template already? If yes, it is in the cache
1988                 # and we need to check for loops.
1989                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
1990                         $found = true;
1991
1992                         # Infinite loop test
1993                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1994                                 $noparse = true;
1995                                 $found = true;
1996                                 $text = $linestart .
1997                                         "\{\{$part1}}" .
1998                                         '<!-- WARNING: template loop detected -->';
1999                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2000                         } else {
2001                                 # set $text to cached message.
2002                                 $text = $linestart . $this->mTemplates[$part1];
2003                         }
2004                 }
2005
2006                 # Load from database
2007                 $itcamefromthedatabase = false;
2008                 $lastPathLevel = $this->mTemplatePath;
2009                 if ( !$found ) {
2010                         $ns = NS_TEMPLATE;
2011                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2012                         if ($subpage !== '') {
2013                                 $ns = $this->mTitle->getNamespace();
2014                         }
2015                         $title = Title::newFromText( $part1, $ns );
2016                         if ( !is_null( $title ) && !$title->isExternal() ) {
2017                                 # Check for excessive inclusion
2018                                 $dbk = $title->getPrefixedDBkey();
2019                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2020                                         # This should never be reached.
2021                                         $article = new Article( $title );
2022                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2023                                         if ( $articleContent !== false ) {
2024                                                 $found = true;
2025                                                 $text = $linestart . $articleContent;
2026                                                 $itcamefromthedatabase = true;
2027                                         }
2028                                 }
2029
2030                                 # If the title is valid but undisplayable, make a link to it
2031                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2032                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2033                                         $found = true;
2034                                 }
2035
2036                                 # Template cache array insertion
2037                                 if( $found ) {
2038                                         $this->mTemplates[$part1] = $text;
2039                                 }
2040                         }
2041                 }
2042
2043                 # Recursive parsing, escaping and link table handling
2044                 # Only for HTML output
2045                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2046                         $text = wfEscapeWikiText( $text );
2047                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2048                         # Clean up argument array
2049                         $assocArgs = array();
2050                         $index = 1;
2051                         foreach( $args as $arg ) {
2052                                 $eqpos = strpos( $arg, '=' );
2053                                 if ( $eqpos === false ) {
2054                                         $assocArgs[$index++] = $arg;
2055                                 } else {
2056                                         $name = trim( substr( $arg, 0, $eqpos ) );
2057                                         $value = trim( substr( $arg, $eqpos+1 ) );
2058                                         if ( $value === false ) {
2059                                                 $value = '';
2060                                         }
2061                                         if ( $name !== false ) {
2062                                                 $assocArgs[$name] = $value;
2063                                         }
2064                                 }
2065                         }
2066
2067                         # Add a new element to the templace recursion path
2068                         $this->mTemplatePath[$part1] = 1;
2069
2070                         $text = $this->strip( $text, $this->mStripState );
2071                         $text = Sanitizer::removeHTMLtags( $text );
2072                         $text = $this->replaceVariables( $text, $assocArgs );
2073
2074                         # Resume the link cache and register the inclusion as a link
2075                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2076                                 $wgLinkCache->addLinkObj( $title );
2077                         }
2078
2079                         # If the template begins with a table or block-level
2080                         # element, it should be treated as beginning a new line.
2081                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2082                                 $text = "\n" . $text;
2083                         }
2084                 }
2085                 # Prune lower levels off the recursion check path
2086                 $this->mTemplatePath = $lastPathLevel;
2087
2088                 if ( !$found ) {
2089                         wfProfileOut( $fname );
2090                         return $matches[0];
2091                 } else {
2092                         # replace ==section headers==
2093                         # XXX this needs to go away once we have a better parser.
2094                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2095                                 if( !is_null( $title ) )
2096                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2097                                 else
2098                                         $encodedname = base64_encode("");
2099                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2100                                         PREG_SPLIT_DELIM_CAPTURE);
2101                                 $text = '';
2102                                 $nsec = 0;
2103                                 for( $i = 0; $i < count($m); $i += 2 ) {
2104                                         $text .= $m[$i];
2105                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2106                                         $hl = $m[$i + 1];
2107                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2108                                                 $text .= $hl;
2109                                                 continue;
2110                                         }
2111                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2112                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2113                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2114
2115                                         $nsec++;
2116                                 }
2117                         }
2118                 }
2119                 # Prune lower levels off the recursion check path
2120                 $this->mTemplatePath = $lastPathLevel;
2121
2122                 if ( !$found ) {
2123                         wfProfileOut( $fname );
2124                         return $matches[0];
2125                 } else {
2126                         wfProfileOut( $fname );
2127                         return $text;
2128                 }
2129         }
2130
2131         /**
2132          * Triple brace replacement -- used for template arguments
2133          * @access private
2134          */
2135         function argSubstitution( $matches ) {
2136                 $arg = trim( $matches[1] );
2137                 $text = $matches[0];
2138                 $inputArgs = end( $this->mArgStack );
2139
2140                 if ( array_key_exists( $arg, $inputArgs ) ) {
2141                         $text = $inputArgs[$arg];
2142                 }
2143
2144                 return $text;
2145         }
2146
2147         /**
2148          * Returns true if the function is allowed to include this entity
2149          * @access private
2150          */
2151         function incrementIncludeCount( $dbk ) {
2152                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2153                         $this->mIncludeCount[$dbk] = 0;
2154                 }
2155                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2156                         return true;
2157                 } else {
2158                         return false;
2159                 }
2160         }
2161
2162         /**
2163          * This function accomplishes several tasks:
2164          * 1) Auto-number headings if that option is enabled
2165          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2166          * 3) Add a Table of contents on the top for users who have enabled the option
2167          * 4) Auto-anchor headings
2168          *
2169          * It loops through all headlines, collects the necessary data, then splits up the
2170          * string and re-inserts the newly formatted headlines.
2171          *
2172          * @param string $text
2173          * @param boolean $isMain
2174          * @access private
2175          */
2176         function formatHeadings( $text, $isMain=true ) {
2177                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2178
2179                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2180                 $doShowToc = $this->mOptions->getShowToc();
2181                 $forceTocHere = false;
2182                 if( !$this->mTitle->userCanEdit() ) {
2183                         $showEditLink = 0;
2184                         $rightClickHack = 0;
2185                 } else {
2186                         $showEditLink = $this->mOptions->getEditSection();
2187                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2188                 }
2189
2190                 # Inhibit editsection links if requested in the page
2191                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2192                 if( $esw->matchAndRemove( $text ) ) {
2193                         $showEditLink = 0;
2194                 }
2195                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2196                 # do not add TOC
2197                 $mw =& MagicWord::get( MAG_NOTOC );
2198                 if( $mw->matchAndRemove( $text ) ) {
2199                         $doShowToc = 0;
2200                 }
2201
2202                 # never add the TOC to the Main Page. This is an entry page that should not
2203                 # be more than 1-2 screens large anyway
2204                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2205                         $doShowToc = 0;
2206                 }
2207
2208                 # Get all headlines for numbering them and adding funky stuff like [edit]
2209                 # links - this is for later, but we need the number of headlines right now
2210                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2211
2212                 # if there are fewer than 4 headlines in the article, do not show TOC
2213                 if( $numMatches < 4 ) {
2214                         $doShowToc = 0;
2215                 }
2216
2217                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2218                 # override above conditions and always show TOC at that place
2219
2220                 $mw =& MagicWord::get( MAG_TOC );
2221                 if($mw->match( $text ) ) {
2222                         $doShowToc = 1;
2223                         $forceTocHere = true;
2224                 } else {
2225                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2226                         # override above conditions and always show TOC above first header
2227                         $mw =& MagicWord::get( MAG_FORCETOC );
2228                         if ($mw->matchAndRemove( $text ) ) {
2229                                 $doShowToc = 1;
2230                         }
2231                 }
2232
2233                 # Never ever show TOC if no headers
2234                 if( $numMatches < 1 ) {
2235                         $doShowToc = 0;
2236                 }
2237
2238                 # We need this to perform operations on the HTML
2239                 $sk =& $this->mOptions->getSkin();
2240
2241                 # headline counter
2242                 $headlineCount = 0;
2243                 $sectionCount = 0; # headlineCount excluding template sections
2244
2245                 # Ugh .. the TOC should have neat indentation levels which can be
2246                 # passed to the skin functions. These are determined here
2247                 $toc = '';
2248                 $full = '';
2249                 $head = array();
2250                 $sublevelCount = array();
2251                 $levelCount = array();
2252                 $toclevel = 0;
2253                 $level = 0;
2254                 $prevlevel = 0;
2255                 $toclevel = 0;
2256                 $prevtoclevel = 0;
2257
2258                 foreach( $matches[3] as $headline ) {
2259                         $istemplate = 0;
2260                         $templatetitle = '';
2261                         $templatesection = 0;
2262                         $numbering = '';
2263
2264                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2265                                 $istemplate = 1;
2266                                 $templatetitle = base64_decode($mat[1]);
2267                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2268                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2269                         }
2270
2271                         if( $toclevel ) {
2272                                 $prevlevel = $level;
2273                                 $prevtoclevel = $toclevel;
2274                         }
2275                         $level = $matches[1][$headlineCount];
2276
2277                         if( $doNumberHeadings || $doShowToc ) {
2278
2279                                 if ( $level > $prevlevel ) {
2280                                         # Increase TOC level
2281                                         $toclevel++;
2282                                         $sublevelCount[$toclevel] = 0;
2283                                         $toc .= $sk->tocIndent();
2284                                 }
2285                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2286                                         # Decrease TOC level, find level to jump to
2287
2288                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2289                                                 # Can only go down to level 1
2290                                                 $toclevel = 1;
2291                                         } else {
2292                                                 for ($i = $toclevel; $i > 0; $i--) {
2293                                                         if ( $levelCount[$i] == $level ) {
2294                                                                 # Found last matching level
2295                                                                 $toclevel = $i;
2296                                                                 break;
2297                                                         }
2298                                                         elseif ( $levelCount[$i] < $level ) {
2299                                                                 # Found first matching level below current level
2300                                                                 $toclevel = $i + 1;
2301                                                                 break;
2302                                                         }
2303                                                 }
2304                                         }
2305
2306                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2307                                 }
2308                                 else {
2309                                         # No change in level, end TOC line
2310                                         $toc .= $sk->tocLineEnd();
2311                                 }
2312
2313                                 $levelCount[$toclevel] = $level;
2314
2315                                 # count number of headlines for each level
2316                                 @$sublevelCount[$toclevel]++;
2317                                 $dot = 0;
2318                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2319                                         if( !empty( $sublevelCount[$i] ) ) {
2320                                                 if( $dot ) {
2321                                                         $numbering .= '.';
2322                                                 }
2323                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2324                                                 $dot = 1;
2325                                         }
2326                                 }
2327                         }
2328
2329                         # The canonized header is a version of the header text safe to use for links
2330                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2331                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2332                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2333
2334                         # Remove link placeholders by the link text.
2335                         #     <!--LINK number-->
2336                         # turns into
2337                         #     link text with suffix
2338                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2339                                                             "\$wgLinkHolders['texts'][\$1]",
2340                                                             $canonized_headline );
2341                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2342                                                             "\$wgInterwikiLinkHolders[\$1]",
2343                                                             $canonized_headline );
2344
2345                         # strip out HTML
2346                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2347                         $tocline = trim( $canonized_headline );
2348                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2349                         $replacearray = array(
2350                                 '%3A' => ':',
2351                                 '%' => '.'
2352                         );
2353                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2354                         $refer[$headlineCount] = $canonized_headline;
2355
2356                         # count how many in assoc. array so we can track dupes in anchors
2357                         @$refers[$canonized_headline]++;
2358                         $refcount[$headlineCount]=$refers[$canonized_headline];
2359
2360                         # Don't number the heading if it is the only one (looks silly)
2361                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2362                                 # the two are different if the line contains a link
2363                                 $headline=$numbering . ' ' . $headline;
2364                         }
2365
2366                         # Create the anchor for linking from the TOC to the section
2367                         $anchor = $canonized_headline;
2368                         if($refcount[$headlineCount] > 1 ) {
2369                                 $anchor .= '_' . $refcount[$headlineCount];
2370                         }
2371                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2372                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2373                         }
2374                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2375                                 if ( empty( $head[$headlineCount] ) ) {
2376                                         $head[$headlineCount] = '';
2377                                 }
2378                                 if( $istemplate )
2379                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2380                                 else
2381                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2382                         }
2383
2384                         # Add the edit section span
2385                         if( $rightClickHack ) {
2386                                 if( $istemplate )
2387                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2388                                 else
2389                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2390                         }
2391
2392                         # give headline the correct <h#> tag
2393                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2394
2395                         $headlineCount++;
2396                         if( !$istemplate )
2397                                 $sectionCount++;
2398                 }
2399
2400                 if( $doShowToc ) {
2401                         $toclines = $headlineCount;
2402                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2403                         $toc = $sk->tocList( $toc );
2404                 }
2405
2406                 # split up and insert constructed headlines
2407
2408                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2409                 $i = 0;
2410
2411                 foreach( $blocks as $block ) {
2412                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2413                                 # This is the [edit] link that appears for the top block of text when
2414                                 # section editing is enabled
2415
2416                                 # Disabled because it broke block formatting
2417                                 # For example, a bullet point in the top line
2418                                 # $full .= $sk->editSectionLink(0);
2419                         }
2420                         $full .= $block;
2421                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2422                         # Top anchor now in skin
2423                                 $full = $full.$toc;
2424                         }
2425
2426                         if( !empty( $head[$i] ) ) {
2427                                 $full .= $head[$i];
2428                         }
2429                         $i++;
2430                 }
2431                 if($forceTocHere) {
2432                         $mw =& MagicWord::get( MAG_TOC );
2433                         return $mw->replace( $toc, $full );
2434                 } else {
2435                         return $full;
2436                 }
2437         }
2438
2439         /**
2440          * Return an HTML link for the "ISBN 123456" text
2441          * @access private
2442          */
2443         function magicISBN( $text ) {
2444                 global $wgLang;
2445                 $fname = 'Parser::magicISBN';
2446                 wfProfileIn( $fname );
2447
2448                 $a = split( 'ISBN ', ' '.$text );
2449                 if ( count ( $a ) < 2 ) {
2450                         wfProfileOut( $fname );
2451                         return $text;
2452                 }
2453                 $text = substr( array_shift( $a ), 1);
2454                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2455
2456                 foreach ( $a as $x ) {
2457                         $isbn = $blank = '' ;
2458                         while ( ' ' == $x{0} ) {
2459                                 $blank .= ' ';
2460                                 $x = substr( $x, 1 );
2461                         }
2462                         if ( $x == '' ) { # blank isbn
2463                                 $text .= "ISBN $blank";
2464                                 continue;
2465                         }
2466                         while ( strstr( $valid, $x{0} ) != false ) {
2467                                 $isbn .= $x{0};
2468                                 $x = substr( $x, 1 );
2469                         }
2470                         $num = str_replace( '-', '', $isbn );
2471                         $num = str_replace( ' ', '', $num );
2472
2473                         if ( '' == $num ) {
2474                                 $text .= "ISBN $blank$x";
2475                         } else {
2476                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2477                                 $text .= '<a href="' .
2478                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2479                                         "\" class=\"internal\">ISBN $isbn</a>";
2480                                 $text .= $x;
2481                         }
2482                 }
2483                 wfProfileOut( $fname );
2484                 return $text;
2485         }
2486
2487         /**
2488          * Return an HTML link for the "RFC 1234" text
2489          * @access private
2490          * @param string $text text to be processed
2491          */
2492         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2493                 global $wgLang;
2494
2495                 $valid = '0123456789';
2496                 $internal = false;
2497
2498                 $a = split( $keyword, ' '.$text );
2499                 if ( count ( $a ) < 2 ) {
2500                         return $text;
2501                 }
2502                 $text = substr( array_shift( $a ), 1);
2503
2504                 /* Check if keyword is preceed by [[.
2505                  * This test is made here cause of the array_shift above
2506                  * that prevent the test to be done in the foreach.
2507                  */
2508                 if ( substr( $text, -2 ) == '[[' ) {
2509                         $internal = true;
2510                 }
2511
2512                 foreach ( $a as $x ) {
2513                         /* token might be empty if we have RFC RFC 1234 */
2514                         if ( $x=='' ) {
2515                                 $text.=$keyword;
2516                                 continue;
2517                                 }
2518
2519                         $id = $blank = '' ;
2520
2521                         /** remove and save whitespaces in $blank */
2522                         while ( $x{0} == ' ' ) {
2523                                 $blank .= ' ';
2524                                 $x = substr( $x, 1 );
2525                         }
2526
2527                         /** remove and save the rfc number in $id */
2528                         while ( strstr( $valid, $x{0} ) != false ) {
2529                                 $id .= $x{0};
2530                                 $x = substr( $x, 1 );
2531                         }
2532
2533                         if ( $id == '' ) {
2534                                 /* call back stripped spaces*/
2535                                 $text .= $keyword.$blank.$x;
2536                         } elseif( $internal ) {
2537                                 /* normal link */
2538                                 $text .= $keyword.$id.$x;
2539                         } else {
2540                                 /* build the external link*/
2541                                 $url = wfmsg( $urlmsg );
2542                                 $url = str_replace( '$1', $id, $url);
2543                                 $sk =& $this->mOptions->getSkin();
2544                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2545                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2546                         }
2547
2548                         /* Check if the next RFC keyword is preceed by [[ */
2549                         $internal = ( substr($x,-2) == '[[' );
2550                 }
2551                 return $text;
2552         }
2553
2554         /**
2555          * Transform wiki markup when saving a page by doing \r\n -> \n
2556          * conversion, substitting signatures, {{subst:}} templates, etc.
2557          *
2558          * @param string $text the text to transform
2559          * @param Title &$title the Title object for the current article
2560          * @param User &$user the User object describing the current user
2561          * @param ParserOptions $options parsing options
2562          * @param bool $clearState whether to clear the parser state first
2563          * @return string the altered wiki markup
2564          * @access public
2565          */
2566         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2567                 $this->mOptions = $options;
2568                 $this->mTitle =& $title;
2569                 $this->mOutputType = OT_WIKI;
2570
2571                 if ( $clearState ) {
2572                         $this->clearState();
2573                 }
2574
2575                 $stripState = false;
2576                 $pairs = array(
2577                         "\r\n" => "\n",
2578                         );
2579                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2580                 $text = $this->strip( $text, $stripState, false );
2581                 $text = $this->pstPass2( $text, $user );
2582                 $text = $this->unstrip( $text, $stripState );
2583                 $text = $this->unstripNoWiki( $text, $stripState );
2584                 return $text;
2585         }
2586
2587         /**
2588          * Pre-save transform helper function
2589          * @access private
2590          */
2591         function pstPass2( $text, &$user ) {
2592                 global $wgLang, $wgContLang, $wgLocaltimezone;
2593
2594                 # Variable replacement
2595                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2596                 $text = $this->replaceVariables( $text );
2597
2598                 # Signatures
2599                 #
2600                 $n = $user->getName();
2601                 $k = $user->getOption( 'nickname' );
2602                 if ( '' == $k ) { $k = $n; }
2603                 if ( isset( $wgLocaltimezone ) ) {
2604                         $oldtz = getenv( 'TZ' );
2605                         putenv( 'TZ='.$wgLocaltimezone );
2606                 }
2607                 /* Note: this is an ugly timezone hack for the European wikis */
2608                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2609                   ' (' . date( 'T' ) . ')';
2610                 if ( isset( $wgLocaltimezone ) ) {
2611                         putenv( 'TZ='.$oldtzs );
2612                 }
2613
2614                 if( $user->getOption( 'fancysig' ) ) {
2615                         $sigText = $k;
2616                 } else {
2617                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2618                 }
2619                 $text = preg_replace( '/~~~~~/', $d, $text );
2620                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2621                 $text = preg_replace( '/~~~/', $sigText, $text );
2622
2623                 # Context links: [[|name]] and [[name (context)|]]
2624                 #
2625                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2626                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2627                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2628                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2629
2630                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2631                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2632                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2633                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2634                 $context = '';
2635                 $t = $this->mTitle->getText();
2636                 if ( preg_match( $conpat, $t, $m ) ) {
2637                         $context = $m[2];
2638                 }
2639                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2640                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2641                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2642
2643                 if ( '' == $context ) {
2644                         $text = preg_replace( $p2, '[[\\1]]', $text );
2645                 } else {
2646                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2647                 }
2648
2649                 # Trim trailing whitespace
2650                 # MAG_END (__END__) tag allows for trailing
2651                 # whitespace to be deliberately included
2652                 $text = rtrim( $text );
2653                 $mw =& MagicWord::get( MAG_END );
2654                 $mw->matchAndRemove( $text );
2655
2656                 return $text;
2657         }
2658
2659         /**
2660          * Set up some variables which are usually set up in parse()
2661          * so that an external function can call some class members with confidence
2662          * @access public
2663          */
2664         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2665                 $this->mTitle =& $title;
2666                 $this->mOptions = $options;
2667                 $this->mOutputType = $outputType;
2668                 if ( $clearState ) {
2669                         $this->clearState();
2670                 }
2671         }
2672
2673         /**
2674          * Transform a MediaWiki message by replacing magic variables.
2675          *
2676          * @param string $text the text to transform
2677          * @param ParserOptions $options  options
2678          * @return string the text with variables substituted
2679          * @access public
2680          */
2681         function transformMsg( $text, $options ) {
2682                 global $wgTitle;
2683                 static $executing = false;
2684
2685                 # Guard against infinite recursion
2686                 if ( $executing ) {
2687                         return $text;
2688                 }
2689                 $executing = true;
2690
2691                 $this->mTitle = $wgTitle;
2692                 $this->mOptions = $options;
2693                 $this->mOutputType = OT_MSG;
2694                 $this->clearState();
2695                 $text = $this->replaceVariables( $text );
2696
2697                 $executing = false;
2698                 return $text;
2699         }
2700
2701         /**
2702          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2703          * Callback will be called with the text within
2704          * Transform and return the text within
2705          * @access public
2706          */
2707         function setHook( $tag, $callback ) {
2708                 $oldVal = @$this->mTagHooks[$tag];
2709                 $this->mTagHooks[$tag] = $callback;
2710                 return $oldVal;
2711         }
2712
2713         /**
2714          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2715          * Placeholders created in Skin::makeLinkObj()
2716          * Returns an array of links found, indexed by PDBK:
2717          *  0 - broken
2718          *  1 - normal link
2719          *  2 - stub
2720          * $options is a bit field, RLH_FOR_UPDATE to select for update
2721          */
2722         function replaceLinkHolders( &$text, $options = 0 ) {
2723                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2724                 global $wgInterwikiLinkHolders;
2725                 global $outputReplace;
2726
2727                 if ( $wgUseOldExistenceCheck ) {
2728                         return array();
2729                 }
2730
2731                 $fname = 'Parser::replaceLinkHolders';
2732                 wfProfileIn( $fname );
2733
2734                 $pdbks = array();
2735                 $colours = array();
2736
2737                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2738                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2739                         wfProfileIn( $fname.'-check' );
2740                         $dbr =& wfGetDB( DB_SLAVE );
2741                         $page = $dbr->tableName( 'page' );
2742                         $sk = $wgUser->getSkin();
2743                         $threshold = $wgUser->getOption('stubthreshold');
2744
2745                         # Sort by namespace
2746                         asort( $wgLinkHolders['namespaces'] );
2747
2748                         # Generate query
2749                         $query = false;
2750                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2751                                 # Make title object
2752                                 $title = $wgLinkHolders['titles'][$key];
2753
2754                                 # Skip invalid entries.
2755                                 # Result will be ugly, but prevents crash.
2756                                 if ( is_null( $title ) ) {
2757                                         continue;
2758                                 }
2759                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2760
2761                                 # Check if it's in the link cache already
2762                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2763                                         $colours[$pdbk] = 1;
2764                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2765                                         $colours[$pdbk] = 0;
2766                                 } else {
2767                                         # Not in the link cache, add it to the query
2768                                         if ( !isset( $current ) ) {
2769                                                 $current = $val;
2770                                                 $tables = $page;
2771                                                 $join = '';
2772                                                 $query =  "SELECT page_id, page_namespace, page_title";
2773                                                 if ( $threshold > 0 ) {
2774                                                         $textTable = $dbr->tableName( 'text' );
2775                                                         $query .= ', LENGTH(old_text) AS page_len, page_is_redirect';
2776                                                         $tables .= ", $textTable";
2777                                                         $join = 'page_latest=old_id AND';
2778                                                 }
2779                                                 $query .= " FROM $tables WHERE $join (page_namespace=$val AND page_title IN(";
2780                                         } elseif ( $current != $val ) {
2781                                                 $current = $val;
2782                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
2783                                         } else {
2784                                                 $query .= ', ';
2785                                         }
2786
2787                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2788                                 }
2789                         }
2790                         if ( $query ) {
2791                                 $query .= '))';
2792                                 if ( $options & RLH_FOR_UPDATE ) {
2793                                         $query .= ' FOR UPDATE';
2794                                 }
2795
2796                                 $res = $dbr->query( $query, $fname );
2797
2798                                 # Fetch data and form into an associative array
2799                                 # non-existent = broken
2800                                 # 1 = known
2801                                 # 2 = stub
2802                                 while ( $s = $dbr->fetchObject($res) ) {
2803                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
2804                                         $pdbk = $title->getPrefixedDBkey();
2805                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
2806
2807                                         if ( $threshold >  0 ) {
2808                                                 $size = $s->page_len;
2809                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $length < $threshold ) {
2810                                                         $colours[$pdbk] = 1;
2811                                                 } else {
2812                                                         $colours[$pdbk] = 2;
2813                                                 }
2814                                         } else {
2815                                                 $colours[$pdbk] = 1;
2816                                         }
2817                                 }
2818                         }
2819                         wfProfileOut( $fname.'-check' );
2820
2821                         # Construct search and replace arrays
2822                         wfProfileIn( $fname.'-construct' );
2823                         $outputReplace = array();
2824                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
2825                                 $pdbk = $pdbks[$key];
2826                                 $searchkey = '<!--LINK '.$key.'-->';
2827                                 $title = $wgLinkHolders['titles'][$key];
2828                                 if ( empty( $colours[$pdbk] ) ) {
2829                                         $wgLinkCache->addBadLink( $pdbk );
2830                                         $colours[$pdbk] = 0;
2831                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
2832                                                                         $wgLinkHolders['texts'][$key],
2833                                                                         $wgLinkHolders['queries'][$key] );
2834                                 } elseif ( $colours[$pdbk] == 1 ) {
2835                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
2836                                                                         $wgLinkHolders['texts'][$key],
2837                                                                         $wgLinkHolders['queries'][$key] );
2838                                 } elseif ( $colours[$pdbk] == 2 ) {
2839                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
2840                                                                         $wgLinkHolders['texts'][$key],
2841                                                                         $wgLinkHolders['queries'][$key] );
2842                                 }
2843                         }
2844                         wfProfileOut( $fname.'-construct' );
2845
2846                         # Do the thing
2847                         wfProfileIn( $fname.'-replace' );
2848
2849                         $text = preg_replace_callback(
2850                                 '/(<!--LINK .*?-->)/',
2851                                 "outputReplaceMatches",
2852                                 $text);
2853                         wfProfileOut( $fname.'-replace' );
2854                 }
2855
2856                 if ( !empty( $wgInterwikiLinkHolders ) ) {
2857                         wfProfileIn( $fname.'-interwiki' );
2858                         $outputReplace = $wgInterwikiLinkHolders;
2859                         $text = preg_replace_callback(
2860                                 '/<!--IWLINK (.*?)-->/',
2861                                 "outputReplaceMatches",
2862                                 $text );
2863                         wfProfileOut( $fname.'-interwiki' );
2864                 }
2865
2866                 wfProfileOut( $fname );
2867                 return $colours;
2868         }
2869
2870         /**
2871          * Renders an image gallery from a text with one line per image.
2872          * text labels may be given by using |-style alternative text. E.g.
2873          *   Image:one.jpg|The number "1"
2874          *   Image:tree.jpg|A tree
2875          * given as text will return the HTML of a gallery with two images,
2876          * labeled 'The number "1"' and
2877          * 'A tree'.
2878          */
2879         function renderImageGallery( $text ) {
2880                 global $wgLinkCache;
2881                 $ig = new ImageGallery();
2882                 $ig->setShowBytes( false );
2883                 $ig->setShowFilename( false );
2884                 $lines = explode( "\n", $text );
2885
2886                 foreach ( $lines as $line ) {
2887                         # match lines like these:
2888                         # Image:someimage.jpg|This is some image
2889                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
2890                         # Skip empty lines
2891                         if ( count( $matches ) == 0 ) {
2892                                 continue;
2893                         }
2894                         $nt = Title::newFromURL( $matches[1] );
2895                         if( is_null( $nt ) ) {
2896                                 # Bogus title. Ignore these so we don't bomb out later.
2897                                 continue;
2898                         }
2899                         if ( isset( $matches[3] ) ) {
2900                                 $label = $matches[3];
2901                         } else {
2902                                 $label = '';
2903                         }
2904
2905                         # FIXME: Use the full wiki parser and add its links
2906                         # to the page's links.
2907                         $html = $this->mOptions->mSkin->formatComment( $label );
2908
2909                         $ig->add( Image::newFromTitle( $nt ), $html );
2910                         $wgLinkCache->addImageLinkObj( $nt );
2911                 }
2912                 return $ig->toHTML();
2913         }
2914 }
2915
2916 /**
2917  * @todo document
2918  * @package MediaWiki
2919  */
2920 class ParserOutput
2921 {
2922         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2923         var $mCacheTime; # Used in ParserCache
2924         var $mVersion;   # Compatibility check
2925
2926         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2927                 $containsOldMagic = false )
2928         {
2929                 $this->mText = $text;
2930                 $this->mLanguageLinks = $languageLinks;
2931                 $this->mCategoryLinks = $categoryLinks;
2932                 $this->mContainsOldMagic = $containsOldMagic;
2933                 $this->mCacheTime = '';
2934                 $this->mVersion = MW_PARSER_VERSION;
2935         }
2936
2937         function getText()                   { return $this->mText; }
2938         function getLanguageLinks()          { return $this->mLanguageLinks; }
2939         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
2940         function getCacheTime()              { return $this->mCacheTime; }
2941         function containsOldMagic()          { return $this->mContainsOldMagic; }
2942         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
2943         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
2944         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
2945         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2946         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
2947         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
2948
2949         function merge( $other ) {
2950                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2951                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2952                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2953         }
2954
2955         /**
2956          * Return true if this cached output object predates the global or
2957          * per-article cache invalidation timestamps, or if it comes from
2958          * an incompatible older version.
2959          *
2960          * @param string $touched the affected article's last touched timestamp
2961          * @return bool
2962          * @access public
2963          */
2964         function expired( $touched ) {
2965                 global $wgCacheEpoch;
2966                 return $this->getCacheTime() <= $touched ||
2967                        $this->getCacheTime() <= $wgCacheEpoch ||
2968                        !isset( $this->mVersion ) ||
2969                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
2970         }
2971 }
2972
2973 /**
2974  * Set options of the Parser
2975  * @todo document
2976  * @package MediaWiki
2977  */
2978 class ParserOptions
2979 {
2980         # All variables are private
2981         var $mUseTeX;                    # Use texvc to expand <math> tags
2982         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2983         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2984         var $mAllowExternalImages;       # Allow external images inline
2985         var $mSkin;                      # Reference to the preferred skin
2986         var $mDateFormat;                # Date format index
2987         var $mEditSection;               # Create "edit section" links
2988         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2989         var $mNumberHeadings;            # Automatically number headings
2990         var $mShowToc;                   # Show table of contents
2991
2992         function getUseTeX()                        { return $this->mUseTeX; }
2993         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2994         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2995         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2996         function getSkin()                          { return $this->mSkin; }
2997         function getDateFormat()                    { return $this->mDateFormat; }
2998         function getEditSection()                   { return $this->mEditSection; }
2999         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3000         function getNumberHeadings()                { return $this->mNumberHeadings; }
3001         function getShowToc()                       { return $this->mShowToc; }
3002
3003         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3004         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3005         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3006         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3007         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3008         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3009         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3010         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3011         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3012
3013         function setSkin( &$x ) { $this->mSkin =& $x; }
3014
3015         /**
3016          * Get parser options
3017          * @static
3018          */
3019         function newFromUser( &$user ) {
3020                 $popts = new ParserOptions;
3021                 $popts->initialiseFromUser( $user );
3022                 return $popts;
3023         }
3024
3025         /** Get user options */
3026         function initialiseFromUser( &$userInput ) {
3027                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3028                 $fname = 'ParserOptions::initialiseFromUser';
3029                 wfProfileIn( $fname );
3030                 if ( !$userInput ) {
3031                         $user = new User;
3032                         $user->setLoaded( true );
3033                 } else {
3034                         $user =& $userInput;
3035                 }
3036
3037                 $this->mUseTeX = $wgUseTeX;
3038                 $this->mUseDynamicDates = $wgUseDynamicDates;
3039                 $this->mInterwikiMagic = $wgInterwikiMagic;
3040                 $this->mAllowExternalImages = $wgAllowExternalImages;
3041                 wfProfileIn( $fname.'-skin' );
3042                 $this->mSkin =& $user->getSkin();
3043                 wfProfileOut( $fname.'-skin' );
3044                 $this->mDateFormat = $user->getOption( 'date' );
3045                 $this->mEditSection = $user->getOption( 'editsection' );
3046                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3047                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3048                 $this->mShowToc = $user->getOption( 'showtoc' );
3049                 wfProfileOut( $fname );
3050         }
3051 }
3052
3053 /**
3054  * Callback function used by Parser::replaceLinkHolders()
3055  * to substitute link placeholders.
3056  */
3057 function &outputReplaceMatches( $matches ) {
3058         global $outputReplace;
3059         return $outputReplace[$matches[1]];
3060 }
3061
3062 /**
3063  * Return the total number of articles
3064  */
3065 function wfNumberOfArticles() {
3066         global $wgNumberOfArticles;
3067
3068         wfLoadSiteStats();
3069         return $wgNumberOfArticles;
3070 }
3071
3072 /**
3073  * Get various statistics from the database
3074  * @private
3075  */
3076 function wfLoadSiteStats() {
3077         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3078         $fname = 'wfLoadSiteStats';
3079
3080         if ( -1 != $wgNumberOfArticles ) return;
3081         $dbr =& wfGetDB( DB_SLAVE );
3082         $s = $dbr->selectRow( 'site_stats',
3083                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3084                 array( 'ss_row_id' => 1 ), $fname
3085         );
3086
3087         if ( $s === false ) {
3088                 return;
3089         } else {
3090                 $wgTotalViews = $s->ss_total_views;
3091                 $wgTotalEdits = $s->ss_total_edits;
3092                 $wgNumberOfArticles = $s->ss_good_articles;
3093         }
3094 }
3095
3096 /**
3097  * Escape html tags
3098  * Basicly replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3099  *
3100  * @param string $in Text that might contain HTML tags
3101  * @return string Escaped string
3102  */
3103 function wfEscapeHTMLTagsOnly( $in ) {
3104         return str_replace(
3105                 array( '"', '>', '<' ),
3106                 array( '&quot;', '&gt;', '&lt;' ),
3107                 $in );
3108 }
3109
3110 ?>