includes/Parser.php

   1 <?php
   2
   3 /**
   4  * File for Parser and related classes
   5  *
   6  * @package MediaWiki
   7  * @version $Id$
   8  */
   9
  10 /**
  11  * Variable substitution O(N^2) attack
  12  *
  13  * Without countermeasures, it would be possible to attack the parser by saving
  14  * a page filled with a large number of inclusions of large pages. The size of
  15  * the generated page would be proportional to the square of the input size.
  16  * Hence, we limit the number of inclusions of any given page, thus bringing any
  17  * attack back to O(N).
  18  */
  19
  20 define( 'MAX_INCLUDE_REPEAT', 100 );
  21 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  22
  23 define( 'RLH_FOR_UPDATE', 1 );
  24
  25 # Allowed values for $mOutputType
  26 define( 'OT_HTML', 1 );
  27 define( 'OT_WIKI', 2 );
  28 define( 'OT_MSG' , 3 );
  29
  30 # string parameter for extractTags which will cause it
  31 # to strip HTML comments in addition to regular
  32 # <XML>-style tags. This should not be anything we
  33 # may want to use in wikisyntax
  34 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  35
  36 # prefix for escaping, used in two functions at least
  37 define( 'UNIQ_PREFIX', 'NaodW29');
  38
  39 # Constants needed for external link processing
  40 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  41 define( 'HTTP_PROTOCOLS', 'http|https' );
  42 # Everything except bracket, space, or control characters
  43 define( 'EXT_LINK_URL_CLASS', '[^]<>"\\x00-\\x20\\x7F]' );
  44 # Including space
  45 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  46 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  47 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  48 define( 'EXT_LINK_BRACKETED',  '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  49 define( 'EXT_IMAGE_REGEX',
  50         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  51         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  52         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  53 );
  54
  55 /**
  56  * PHP Parser
  57  *
  58  * Processes wiki markup
  59  *
  60  * <pre>
  61  * There are three main entry points into the Parser class:
  62  * parse()
  63  *   produces HTML output
  64  * preSaveTransform().
  65  *   produces altered wiki markup.
  66  * transformMsg()
  67  *   performs brace substitution on MediaWiki messages
  68  *
  69  * Globals used:
  70  *    objects:   $wgLang, $wgDateFormatter, $wgLinkCache
  71  *
  72  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  73  *
  74  * settings:
  75  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  76  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  77  *  $wgLocaltimezone
  78  *
  79  *  * only within ParserOptions
  80  * </pre>
  81  *
  82  * @package MediaWiki
  83  */
  84 class Parser
  85 {
  86         /**#@+
  87          * @access private
  88          */
  89         # Persistent:
  90         var $mTagHooks;
  91
  92         # Cleared with clearState():
  93         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  94         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  95
  96         # Temporary:
  97         var $mOptions, $mTitle, $mOutputType,
  98             $mTemplates,        // cache of already loaded templates, avoids
  99                                 // multiple SQL queries for the same string
 100             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 101                                 // in this path. Used for loop detection.
 102
 103         /**#@-*/
 104
 105         /**
 106          * Constructor
 107          *
 108          * @access public
 109          */
 110         function Parser() {
 111                 $this->mTemplates = array();
 112                 $this->mTemplatePath = array();
 113                 $this->mTagHooks = array();
 114                 $this->clearState();
 115         }
 116
 117         /**
 118          * Clear Parser state
 119          *
 120          * @access private
 121          */
 122         function clearState() {
 123                 $this->mOutput = new ParserOutput;
 124                 $this->mAutonumber = 0;
 125                 $this->mLastSection = "";
 126                 $this->mDTopen = false;
 127                 $this->mVariables = false;
 128                 $this->mIncludeCount = array();
 129                 $this->mStripState = array();
 130                 $this->mArgStack = array();
 131                 $this->mInPre = false;
 132         }
 133
 134         /**
 135          * First pass--just handle <nowiki> sections, pass the rest off
 136          * to internalParse() which does all the real work.
 137          *
 138          * @access private
 139          * @return ParserOutput a ParserOutput
 140          */
 141         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 142                 global $wgUseTidy, $wgContLang;
 143                 $fname = 'Parser::parse';
 144                 wfProfileIn( $fname );
 145
 146                 if ( $clearState ) {
 147                         $this->clearState();
 148                 }
 149
 150                 $this->mOptions = $options;
 151                 $this->mTitle =& $title;
 152                 $this->mOutputType = OT_HTML;
 153
 154                 $stripState = NULL;
 155                 $text = $this->strip( $text, $this->mStripState );
 156
 157                 $text = $this->internalParse( $text, $linestart );
 158                 $text = $this->unstrip( $text, $this->mStripState );
 159                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 160                 if(!$wgUseTidy) {
 161                         $fixtags = array(
 162                                 # french spaces, last one Guillemet-left
 163                                 # only if there is something before the space
 164                                 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 165                                 # french spaces, Guillemet-right
 166                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 167                                 '/<hr *>/i' => '<hr />',
 168                                 '/<br *>/i' => '<br />',
 169                                 '/<center *>/i' => '<div class="center">',
 170                                 '/<\\/center *>/i' => '</div>',
 171                                 # Clean up spare ampersands; note that we probably ought to be
 172                                 # more careful about named entities.
 173                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 174                         );
 175                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 176                 } else {
 177                         $fixtags = array(
 178                                 # french spaces, last one Guillemet-left
 179                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 180                                 # french spaces, Guillemet-right
 181                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 182                                 '/<center *>/i' => '<div class="center">',
 183                                 '/<\\/center *>/i' => '</div>'
 184                         );
 185                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 186                 }
 187                 # only once and last
 188                 $text = $this->doBlockLevels( $text, $linestart );
 189
 190                 $this->replaceLinkHolders( $text );
 191                 $text = $wgContLang->convert($text);
 192
 193                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 194                 global $wgUseTidy;
 195                 if ($wgUseTidy) {
 196                         $text = Parser::tidy($text);
 197                 }
 198
 199                 $this->mOutput->setText( $text );
 200                 wfProfileOut( $fname );
 201                 return $this->mOutput;
 202         }
 203
 204         /**
 205          * Get a random string
 206          *
 207          * @access private
 208          * @static
 209          */
 210         function getRandomString() {
 211                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 212         }
 213
 214         /**
 215          * Replaces all occurrences of <$tag>content</$tag> in the text
 216          * with a random marker and returns the new text. the output parameter
 217          * $content will be an associative array filled with data on the form
 218          * $unique_marker => content.
 219          *
 220          * If $content is already set, the additional entries will be appended
 221          * If $tag is set to STRIP_COMMENTS, the function will extract
 222          * <!-- HTML comments -->
 223          *
 224          * @access private
 225          * @static
 226          */
 227         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 228                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 229                 if ( !$content ) {
 230                         $content = array( );
 231                 }
 232                 $n = 1;
 233                 $stripped = '';
 234
 235                 while ( '' != $text ) {
 236                         if($tag==STRIP_COMMENTS) {
 237                                 $p = preg_split( '/<!--/i', $text, 2 );
 238                         } else {
 239                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 240                         }
 241                         $stripped .= $p[0];
 242                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 243                                 $text = '';
 244                         } else {
 245                                 if($tag==STRIP_COMMENTS) {
 246                                         $q = preg_split( '/-->/i', $p[1], 2 );
 247                                 } else {
 248                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 249                                 }
 250                                 $marker = $rnd . sprintf('%08X', $n++);
 251                                 $content[$marker] = $q[0];
 252                                 $stripped .= $marker;
 253                                 $text = $q[1];
 254                         }
 255                 }
 256                 return $stripped;
 257         }
 258
 259         /**
 260          * Strips and renders nowiki, pre, math, hiero
 261          * If $render is set, performs necessary rendering operations on plugins
 262          * Returns the text, and fills an array with data needed in unstrip()
 263          * If the $state is already a valid strip state, it adds to the state
 264          *
 265          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 266          *  will be stripped in addition to other tags. This is important
 267          *  for section editing, where these comments cause confusion when
 268          *  counting the sections in the wikisource
 269          *
 270          * @access private
 271          */
 272         function strip( $text, &$state, $stripcomments = false ) {
 273                 $render = ($this->mOutputType == OT_HTML);
 274                 $html_content = array();
 275                 $nowiki_content = array();
 276                 $math_content = array();
 277                 $pre_content = array();
 278                 $comment_content = array();
 279                 $ext_content = array();
 280                 $gallery_content = array();
 281
 282                 # Replace any instances of the placeholders
 283                 $uniq_prefix = UNIQ_PREFIX;
 284                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 285
 286                 # html
 287                 global $wgRawHtml, $wgWhitelistEdit;
 288                 if( $wgRawHtml && $wgWhitelistEdit ) {
 289                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 290                         foreach( $html_content as $marker => $content ) {
 291                                 if ($render ) {
 292                                         # Raw and unchecked for validity.
 293                                         $html_content[$marker] = $content;
 294                                 } else {
 295                                         $html_content[$marker] = '<html>'.$content.'</html>';
 296                                 }
 297                         }
 298                 }
 299
 300                 # nowiki
 301                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 302                 foreach( $nowiki_content as $marker => $content ) {
 303                         if( $render ){
 304                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 305                         } else {
 306                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 307                         }
 308                 }
 309
 310                 # math
 311                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 312                 foreach( $math_content as $marker => $content ){
 313                         if( $render ) {
 314                                 if( $this->mOptions->getUseTeX() ) {
 315                                         $math_content[$marker] = renderMath( $content );
 316                                 } else {
 317                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 318                                 }
 319                         } else {
 320                                 $math_content[$marker] = '<math>'.$content.'</math>';
 321                         }
 322                 }
 323
 324                 # pre
 325                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 326                 foreach( $pre_content as $marker => $content ){
 327                         if( $render ){
 328                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 329                         } else {
 330                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 331                         }
 332                 }
 333
 334                 # gallery
 335                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 336                 foreach( $gallery_content as $marker => $content ) {
 337                         require_once( 'ImageGallery.php' );
 338                         if ( $render ) {
 339                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 340                         } else {
 341                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 342                         }
 343                 }
 344
 345                 # Comments
 346                 if($stripcomments) {
 347                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 348                         foreach( $comment_content as $marker => $content ){
 349                                 $comment_content[$marker] = '<!--'.$content.'-->';
 350                         }
 351                 }
 352
 353                 # Extensions
 354                 foreach ( $this->mTagHooks as $tag => $callback ) {
 355                         $ext_contents[$tag] = array();
 356                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 357                         foreach( $ext_content[$tag] as $marker => $content ) {
 358                                 if ( $render ) {
 359                                         $ext_content[$tag][$marker] = $callback( $content );
 360                                 } else {
 361                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 362                                 }
 363                         }
 364                 }
 365
 366                 # Merge state with the pre-existing state, if there is one
 367                 if ( $state ) {
 368                         $state['html'] = $state['html'] + $html_content;
 369                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 370                         $state['math'] = $state['math'] + $math_content;
 371                         $state['pre'] = $state['pre'] + $pre_content;
 372                         $state['comment'] = $state['comment'] + $comment_content;
 373                         $state['gallery'] = $state['gallery'] + $gallery_content;
 374
 375                         foreach( $ext_content as $tag => $array ) {
 376                                 if ( array_key_exists( $tag, $state ) ) {
 377                                         $state[$tag] = $state[$tag] + $array;
 378                                 }
 379                         }
 380                 } else {
 381                         $state = array(
 382                           'html' => $html_content,
 383                           'nowiki' => $nowiki_content,
 384                           'math' => $math_content,
 385                           'pre' => $pre_content,
 386                           'comment' => $comment_content,
 387                           'gallery' => $gallery_content,
 388                         ) + $ext_content;
 389                 }
 390                 return $text;
 391         }
 392
 393         /**
 394          * restores pre, math, and hiero removed by strip()
 395          *
 396          * always call unstripNoWiki() after this one
 397          * @access private
 398          */
 399         function unstrip( $text, &$state ) {
 400                 # Must expand in reverse order, otherwise nested tags will be corrupted
 401                 $contentDict = end( $state );
 402                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 403                         if( key($state) != 'nowiki' && key($state) != 'html') {
 404                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 405                                         $text = str_replace( key( $contentDict ), $content, $text );
 406                                 }
 407                         }
 408                 }
 409
 410                 return $text;
 411         }
 412
 413         /**
 414          * always call this after unstrip() to preserve the order
 415          *
 416          * @access private
 417          */
 418         function unstripNoWiki( $text, &$state ) {
 419                 # Must expand in reverse order, otherwise nested tags will be corrupted
 420                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 421                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 422                 }
 423
 424                 global $wgRawHtml;
 425                 if ($wgRawHtml) {
 426                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 427                                 $text = str_replace( key( $state['html'] ), $content, $text );
 428                         }
 429                 }
 430
 431                 return $text;
 432         }
 433
 434         /**
 435          * Add an item to the strip state
 436          * Returns the unique tag which must be inserted into the stripped text
 437          * The tag will be replaced with the original text in unstrip()
 438          *
 439          * @access private
 440          */
 441         function insertStripItem( $text, &$state ) {
 442                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 443                 if ( !$state ) {
 444                         $state = array(
 445                           'html' => array(),
 446                           'nowiki' => array(),
 447                           'math' => array(),
 448                           'pre' => array()
 449                         );
 450                 }
 451                 $state['item'][$rnd] = $text;
 452                 return $rnd;
 453         }
 454
 455         /**
 456          * Return allowed HTML attributes
 457          *
 458          * @access private
 459          */
 460         function getHTMLattrs () {
 461                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 462                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 463                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 464                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 465                                 /* FONT */ 'type', 'start', 'value', 'compact',
 466                                 /* For various lists, mostly deprecated but safe */
 467                                 'summary', 'width', 'border', 'frame', 'rules',
 468                                 'cellspacing', 'cellpadding', 'valign', 'char',
 469                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 470                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 471                                 'id', 'class', 'name', 'style' /* For CSS */
 472                                 );
 473                 return $htmlattrs ;
 474         }
 475
 476         /**
 477          * Remove non approved attributes and javascript in css
 478          *
 479          * @access private
 480          */
 481         function fixTagAttributes ( $t ) {
 482                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 483                 $htmlattrs = $this->getHTMLattrs() ;
 484
 485                 # Strip non-approved attributes from the tag
 486                 $t = preg_replace(
 487                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 488                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 489                         $t);
 490
 491                 $t = str_replace ( '<></>' , '' , $t ) ; # This should fix bug 980557
 492
 493                 # Strip javascript "expression" from stylesheets. Brute force approach:
 494                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 495
 496                 if( preg_match(
 497                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 498                         wfMungeToUtf8( $t ) ) )
 499                 {
 500                         $t='';
 501                 }
 502
 503                 return trim ( $t ) ;
 504         }
 505
 506         /**
 507          * interface with html tidy, used if $wgUseTidy = true
 508          *
 509          * @access public
 510          * @static
 511          */
 512         function tidy ( $text ) {
 513                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 514                 global $wgInputEncoding, $wgOutputEncoding;
 515                 $fname = 'Parser::tidy';
 516                 wfProfileIn( $fname );
 517
 518                 $cleansource = '';
 519                 $opts = '';
 520                 switch(strtoupper($wgOutputEncoding)) {
 521                         case 'ISO-8859-1':
 522                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 523                                 break;
 524                         case 'UTF-8':
 525                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 526                                 break;
 527                         default:
 528                                 $opts .= ' -raw';
 529                         }
 530
 531                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 532 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 533 '<head><title>test</title></head><body>'.$text.'</body></html>';
 534                 $descriptorspec = array(
 535                         0 => array('pipe', 'r'),
 536                         1 => array('pipe', 'w'),
 537                         2 => array('file', '/dev/null', 'a')
 538                 );
 539                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 540                 if (is_resource($process)) {
 541                         fwrite($pipes[0], $wrappedtext);
 542                         fclose($pipes[0]);
 543                         while (!feof($pipes[1])) {
 544                                 $cleansource .= fgets($pipes[1], 1024);
 545                         }
 546                         fclose($pipes[1]);
 547                         $return_value = proc_close($process);
 548                 }
 549
 550                 wfProfileOut( $fname );
 551
 552                 if( $cleansource == '' && $text != '') {
 553                         wfDebug( "Tidy error detected!\n" );
 554                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 555                 } else {
 556                         return $cleansource;
 557                 }
 558         }
 559
 560         /**
 561          * parse the wiki syntax used to render tables
 562          *
 563          * @access private
 564          */
 565         function doTableStuff ( $t ) {
 566                 $fname = 'Parser::doTableStuff';
 567                 wfProfileIn( $fname );
 568
 569                 $t = explode ( "\n" , $t ) ;
 570                 $td = array () ; # Is currently a td tag open?
 571                 $ltd = array () ; # Was it TD or TH?
 572                 $tr = array () ; # Is currently a tr tag open?
 573                 $ltr = array () ; # tr attributes
 574                 $indent_level = 0; # indent level of the table
 575                 foreach ( $t AS $k => $x )
 576                 {
 577                         $x = trim ( $x ) ;
 578                         $fc = substr ( $x , 0 , 1 ) ;
 579                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 580                                 $indent_level = strlen( $matches[1] );
 581                                 $t[$k] = "\n" .
 582                                         str_repeat( '<dl><dd>', $indent_level ) .
 583                                         '<table ' . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 584                                 array_push ( $td , false ) ;
 585                                 array_push ( $ltd , '' ) ;
 586                                 array_push ( $tr , false ) ;
 587                                 array_push ( $ltr , '' ) ;
 588                         }
 589                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 590                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 591                                 $z = "</table>\n" ;
 592                                 $l = array_pop ( $ltd ) ;
 593                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 594                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 595                                 array_pop ( $ltr ) ;
 596                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 597                         }
 598                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 599                                 $x = substr ( $x , 1 ) ;
 600                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 601                                 $z = '' ;
 602                                 $l = array_pop ( $ltd ) ;
 603                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 604                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 605                                 array_pop ( $ltr ) ;
 606                                 $t[$k] = $z ;
 607                                 array_push ( $tr , false ) ;
 608                                 array_push ( $td , false ) ;
 609                                 array_push ( $ltd , '' ) ;
 610                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 611                         }
 612                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 613                                 # $x is a table row
 614                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 615                                         $fc = '+' ;
 616                                         $x = substr ( $x , 1 ) ;
 617                                 }
 618                                 $after = substr ( $x , 1 ) ;
 619                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 620                                 $after = explode ( '||' , $after ) ;
 621                                 $t[$k] = '' ;
 622
 623                                 # Loop through each table cell
 624                                 foreach ( $after AS $theline )
 625                                 {
 626                                         $z = '' ;
 627                                         if ( $fc != '+' )
 628                                         {
 629                                                 $tra = array_pop ( $ltr ) ;
 630                                                 if ( !array_pop ( $tr ) ) $z = '<tr '.$tra.">\n" ;
 631                                                 array_push ( $tr , true ) ;
 632                                                 array_push ( $ltr , '' ) ;
 633                                         }
 634
 635                                         $l = array_pop ( $ltd ) ;
 636                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 637                                         if ( $fc == '|' ) $l = 'td' ;
 638                                         else if ( $fc == '!' ) $l = 'th' ;
 639                                         else if ( $fc == '+' ) $l = 'caption' ;
 640                                         else $l = '' ;
 641                                         array_push ( $ltd , $l ) ;
 642
 643                                         # Cell parameters
 644                                         $y = explode ( '|' , $theline , 2 ) ;
 645                                         # Note that a '|' inside an invalid link should not
 646                                         # be mistaken as delimiting cell parameters
 647                                         if ( strpos( $y[0], '[[' ) !== false ) {
 648                                                 $y = array ($theline);
 649                                         }
 650                                         if ( count ( $y ) == 1 )
 651                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 652                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 653                                         $t[$k] .= $y ;
 654                                         array_push ( $td , true ) ;
 655                                 }
 656                         }
 657                 }
 658
 659                 # Closing open td, tr && table
 660                 while ( count ( $td ) > 0 )
 661                 {
 662                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 663                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 664                         $t[] = '</table>' ;
 665                 }
 666
 667                 $t = implode ( "\n" , $t ) ;
 668                 #               $t = $this->removeHTMLtags( $t );
 669                 wfProfileOut( $fname );
 670                 return $t ;
 671         }
 672
 673         /**
 674          * Helper function for parse() that transforms wiki markup into
 675          * HTML. Only called for $mOutputType == OT_HTML.
 676          *
 677          * @access private
 678          */
 679         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 680                 global $wgContLang;
 681
 682                 $fname = 'Parser::internalParse';
 683                 wfProfileIn( $fname );
 684
 685                 $text = $this->removeHTMLtags( $text );
 686                 $text = $this->replaceVariables( $text, $args );
 687
 688                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 689
 690                 $text = $this->doHeadings( $text );
 691                 if($this->mOptions->getUseDynamicDates()) {
 692                         global $wgDateFormatter;
 693                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 694                 }
 695                 $text = $this->doAllQuotes( $text );
 696                 $text = $this->replaceInternalLinks ( $text );
 697                 $text = $this->replaceExternalLinks( $text );
 698
 699                 # replaceInternalLinks may sometimes leave behind
 700                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 701                 $text = str_replace("http-noparse://","http://",$text);
 702
 703                 $text = $this->doMagicLinks( $text );
 704                 $text = $this->doTableStuff( $text );
 705                 $text = $this->formatHeadings( $text, $isMain );
 706                 $sk =& $this->mOptions->getSkin();
 707                 $text = $sk->transformContent( $text );
 708
 709                 wfProfileOut( $fname );
 710                 return $text;
 711         }
 712
 713         /**
 714          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 715          * magic external links.
 716          *
 717          * @access private
 718          */
 719         function &doMagicLinks( &$text ) {
 720                 global $wgUseGeoMode;
 721                 $text = $this->magicISBN( $text );
 722                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 723                         $text = $this->magicGEO( $text );
 724                 }
 725                 $text = $this->magicRFC( $text );
 726                 return $text;
 727         }
 728
 729         /**
 730          * Parse ^^ tokens and return html
 731          *
 732          * @access private
 733          */
 734         function doExponent ( $text ) {
 735                 $fname = 'Parser::doExponent';
 736                 wfProfileIn( $fname);
 737                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 738                 wfProfileOut( $fname);
 739                 return $text;
 740         }
 741
 742         /**
 743          * Parse headers and return html
 744          *
 745          * @access private
 746          */
 747         function doHeadings( $text ) {
 748                 $fname = 'Parser::doHeadings';
 749                 wfProfileIn( $fname );
 750                 for ( $i = 6; $i >= 1; --$i ) {
 751                         $h = substr( '======', 0, $i );
 752                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 753                           "<h{$i}>\\1</h{$i}>\\2", $text );
 754                 }
 755                 wfProfileOut( $fname );
 756                 return $text;
 757         }
 758
 759         /**
 760          * Replace single quotes with HTML markup
 761          * @access private
 762          * @return string the altered text
 763          */
 764         function doAllQuotes( $text ) {
 765                 $fname = 'Parser::doAllQuotes';
 766                 wfProfileIn( $fname );
 767                 $outtext = '';
 768                 $lines = explode( "\n", $text );
 769                 foreach ( $lines as $line ) {
 770                         $outtext .= $this->doQuotes ( $line ) . "\n";
 771                 }
 772                 $outtext = substr($outtext, 0,-1);
 773                 wfProfileOut( $fname );
 774                 return $outtext;
 775         }
 776
 777         /**
 778          * Helper function for doAllQuotes()
 779          * @access private
 780          */
 781         function doQuotes( $text ) {
 782                 $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
 783                 if (count ($arr) == 1)
 784                         return $text;
 785                 else
 786                 {
 787                         # First, do some preliminary work. This may shift some apostrophes from
 788                         # being mark-up to being text. It also counts the number of occurrences
 789                         # of bold and italics mark-ups.
 790                         $i = 0;
 791                         $numbold = 0;
 792                         $numitalics = 0;
 793                         foreach ($arr as $r)
 794                         {
 795                                 if (($i % 2) == 1)
 796                                 {
 797                                         # If there are ever four apostrophes, assume the first is supposed to
 798                                         # be text, and the remaining three constitute mark-up for bold text.
 799                                         if (strlen ($arr[$i]) == 4)
 800                                         {
 801                                                 $arr[$i-1] .= "'";
 802                                                 $arr[$i] = "'''";
 803                                         }
 804                                         # If there are more than 5 apostrophes in a row, assume they're all
 805                                         # text except for the last 5.
 806                                         else if (strlen ($arr[$i]) > 5)
 807                                         {
 808                                                 $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
 809                                                 $arr[$i] = "'''''";
 810                                         }
 811                                         # Count the number of occurrences of bold and italics mark-ups.
 812                                         # We are not counting sequences of five apostrophes.
 813                                         if (strlen ($arr[$i]) == 2) $numitalics++;  else
 814                                         if (strlen ($arr[$i]) == 3) $numbold++;     else
 815                                         if (strlen ($arr[$i]) == 5) { $numitalics++; $numbold++; }
 816                                 }
 817                                 $i++;
 818                         }
 819
 820                         # If there is an odd number of both bold and italics, it is likely
 821                         # that one of the bold ones was meant to be an apostrophe followed
 822                         # by italics. Which one we cannot know for certain, but it is more
 823                         # likely to be one that has a single-letter word before it.
 824                         if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
 825                         {
 826                                 $i = 0;
 827                                 $firstsingleletterword = -1;
 828                                 $firstmultiletterword = -1;
 829                                 $firstspace = -1;
 830                                 foreach ($arr as $r)
 831                                 {
 832                                         if (($i % 2 == 1) and (strlen ($r) == 3))
 833                                         {
 834                                                 $x1 = substr ($arr[$i-1], -1);
 835                                                 $x2 = substr ($arr[$i-1], -2, 1);
 836                                                 if ($x1 == ' ') {
 837                                                         if ($firstspace == -1) $firstspace = $i;
 838                                                 } else if ($x2 == ' ') {
 839                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 840                                                 } else {
 841                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 842                                                 }
 843                                         }
 844                                         $i++;
 845                                 }
 846
 847                                 # If there is a single-letter word, use it!
 848                                 if ($firstsingleletterword > -1)
 849                                 {
 850                                         $arr [ $firstsingleletterword ] = "''";
 851                                         $arr [ $firstsingleletterword-1 ] .= "'";
 852                                 }
 853                                 # If not, but there's a multi-letter word, use that one.
 854                                 else if ($firstmultiletterword > -1)
 855                                 {
 856                                         $arr [ $firstmultiletterword ] = "''";
 857                                         $arr [ $firstmultiletterword-1 ] .= "'";
 858                                 }
 859                                 # ... otherwise use the first one that has neither.
 860                                 # (notice that it is possible for all three to be -1 if, for example,
 861                                 # there is only one pentuple-apostrophe in the line)
 862                                 else if ($firstspace > -1)
 863                                 {
 864                                         $arr [ $firstspace ] = "''";
 865                                         $arr [ $firstspace-1 ] .= "'";
 866                                 }
 867                         }
 868
 869                         # Now let's actually convert our apostrophic mush to HTML!
 870                         $output = '';
 871                         $buffer = '';
 872                         $state = '';
 873                         $i = 0;
 874                         foreach ($arr as $r)
 875                         {
 876                                 if (($i % 2) == 0)
 877                                 {
 878                                         if ($state == 'both')
 879                                                 $buffer .= $r;
 880                                         else
 881                                                 $output .= $r;
 882                                 }
 883                                 else
 884                                 {
 885                                         if (strlen ($r) == 2)
 886                                         {
 887                                                 if ($state == 'i')
 888                                                 { $output .= '</i>'; $state = ''; }
 889                                                 else if ($state == 'bi')
 890                                                 { $output .= '</i>'; $state = 'b'; }
 891                                                 else if ($state == 'ib')
 892                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 893                                                 else if ($state == 'both')
 894                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 895                                                 else # $state can be 'b' or ''
 896                                                 { $output .= '<i>'; $state .= 'i'; }
 897                                         }
 898                                         else if (strlen ($r) == 3)
 899                                         {
 900                                                 if ($state == 'b')
 901                                                 { $output .= '</b>'; $state = ''; }
 902                                                 else if ($state == 'bi')
 903                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 904                                                 else if ($state == 'ib')
 905                                                 { $output .= '</b>'; $state = 'i'; }
 906                                                 else if ($state == 'both')
 907                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 908                                                 else # $state can be 'i' or ''
 909                                                 { $output .= '<b>'; $state .= 'b'; }
 910                                         }
 911                                         else if (strlen ($r) == 5)
 912                                         {
 913                                                 if ($state == 'b')
 914                                                 { $output .= '</b><i>'; $state = 'i'; }
 915                                                 else if ($state == 'i')
 916                                                 { $output .= '</i><b>'; $state = 'b'; }
 917                                                 else if ($state == 'bi')
 918                                                 { $output .= '</i></b>'; $state = ''; }
 919                                                 else if ($state == 'ib')
 920                                                 { $output .= '</b></i>'; $state = ''; }
 921                                                 else if ($state == 'both')
 922                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 923                                                 else # ($state == '')
 924                                                 { $buffer = ''; $state = 'both'; }
 925                                         }
 926                                 }
 927                                 $i++;
 928                         }
 929                         # Now close all remaining tags.  Notice that the order is important.
 930                         if ($state == 'b' || $state == 'ib')
 931                                 $output .= '</b>';
 932                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 933                                 $output .= '</i>';
 934                         if ($state == 'bi')
 935                                 $output .= '</b>';
 936                         if ($state == 'both')
 937                                 $output .= '<b><i>'.$buffer.'</i></b>';
 938                         return $output;
 939                 }
 940         }
 941
 942         /**
 943          * Replace external links
 944          *
 945          * Note: this is all very hackish and the order of execution matters a lot.
 946          * Make sure to run maintenance/parserTests.php if you change this code.
 947          *
 948          * @access private
 949          */
 950         function replaceExternalLinks( $text ) {
 951                 $fname = 'Parser::replaceExternalLinks';
 952                 wfProfileIn( $fname );
 953
 954                 $sk =& $this->mOptions->getSkin();
 955                 $linktrail = wfMsgForContent('linktrail');
 956                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 957
 958                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 959
 960                 $i = 0;
 961                 while ( $i<count( $bits ) ) {
 962                         $url = $bits[$i++];
 963                         $protocol = $bits[$i++];
 964                         $text = $bits[$i++];
 965                         $trail = $bits[$i++];
 966
 967                         # The characters '<' and '>' (which were escaped by
 968                         # removeHTMLtags()) should not be included in
 969                         # URLs, per RFC 2396.
 970                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 971                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 972                                 $url = substr($url, 0, $m2[0][1]);
 973                         }
 974
 975                         # If the link text is an image URL, replace it with an <img> tag
 976                         # This happened by accident in the original parser, but some people used it extensively
 977                         $img = $this->maybeMakeImageLink( $text );
 978                         if ( $img !== false ) {
 979                                 $text = $img;
 980                         }
 981
 982                         $dtrail = '';
 983
 984                         # No link text, e.g. [http://domain.tld/some.link]
 985                         if ( $text == '' ) {
 986                                 # Autonumber if allowed
 987                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 988                                         $text = '[' . ++$this->mAutonumber . ']';
 989                                 } else {
 990                                         # Otherwise just use the URL
 991                                         $text = htmlspecialchars( $url );
 992                                 }
 993                         } else {
 994                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 995                                 # Check for trail
 996                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 997                                         $dtrail = $m2[1];
 998                                         $trail = $m2[2];
 999                                 }
1000                         }
1001
1002                         $encUrl = htmlspecialchars( $url );
1003                         # Bit in parentheses showing the URL for the printable version
1004                         if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
1005                                 $paren = '';
1006                         } else {
1007                                 # Expand the URL for printable version
1008                                 if ( ! $sk->suppressUrlExpansion() ) {
1009                                         $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
1010                                 } else {
1011                                         $paren = '';
1012                                 }
1013                         }
1014
1015                         # Process the trail (i.e. everything after this link up until start of the next link),
1016                         # replacing any non-bracketed links
1017                         $trail = $this->replaceFreeExternalLinks( $trail );
1018
1019                         # Use the encoded URL
1020                         # This means that users can paste URLs directly into the text
1021                         # Funny characters like &ouml; aren't valid in URLs anyway
1022                         # This was changed in August 2004
1023                         $s .= $sk->makeExternalLink( $url, $text, false ) . $dtrail. $paren . $trail;
1024                 }
1025
1026                 wfProfileOut( $fname );
1027                 return $s;
1028         }
1029
1030         /**
1031          * Replace anything that looks like a URL with a link
1032          * @access private
1033          */
1034         function replaceFreeExternalLinks( $text ) {
1035                 $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1036                 $s = array_shift( $bits );
1037                 $i = 0;
1038
1039                 $sk =& $this->mOptions->getSkin();
1040
1041                 while ( $i < count( $bits ) ){
1042                         $protocol = $bits[$i++];
1043                         $remainder = $bits[$i++];
1044
1045                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1046                                 # Found some characters after the protocol that look promising
1047                                 $url = $protocol . $m[1];
1048                                 $trail = $m[2];
1049
1050                                 # The characters '<' and '>' (which were escaped by
1051                                 # removeHTMLtags()) should not be included in
1052                                 # URLs, per RFC 2396.
1053                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1054                                         $trail = substr($url, $m2[0][1]) . $trail;
1055                                         $url = substr($url, 0, $m2[0][1]);
1056                                 }
1057
1058                                 # Move trailing punctuation to $trail
1059                                 $sep = ',;\.:!?';
1060                                 # If there is no left bracket, then consider right brackets fair game too
1061                                 if ( strpos( $url, '(' ) === false ) {
1062                                         $sep .= ')';
1063                                 }
1064
1065                                 $numSepChars = strspn( strrev( $url ), $sep );
1066                                 if ( $numSepChars ) {
1067                                         $trail = substr( $url, -$numSepChars ) . $trail;
1068                                         $url = substr( $url, 0, -$numSepChars );
1069                                 }
1070
1071                                 # Replace &amp; from obsolete syntax with &.
1072                                 # All HTML entities will be escaped by makeExternalLink()
1073                                 # or maybeMakeImageLink()
1074                                 $url = str_replace( '&amp;', '&', $url );
1075
1076                                 # Is this an external image?
1077                                 $text = $this->maybeMakeImageLink( $url );
1078                                 if ( $text === false ) {
1079                                         # Not an image, make a link
1080                                         $text = $sk->makeExternalLink( $url, $url );
1081                                 }
1082                                 $s .= $text . $trail;
1083                         } else {
1084                                 $s .= $protocol . $remainder;
1085                         }
1086                 }
1087                 return $s;
1088         }
1089
1090         /**
1091          * make an image if it's allowed
1092          * @access private
1093          */
1094         function maybeMakeImageLink( $url ) {
1095                 $sk =& $this->mOptions->getSkin();
1096                 $text = false;
1097                 if ( $this->mOptions->getAllowExternalImages() ) {
1098                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1099                                 # Image found
1100                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1101                         }
1102                 }
1103                 return $text;
1104         }
1105
1106         /**
1107          * Process [[ ]] wikilinks
1108          *
1109          * @access private
1110          */
1111
1112         function replaceInternalLinks( $s ) {
1113                 global $wgLang, $wgContLang, $wgLinkCache;
1114                 global $wgDisableLangConversion;
1115                 static $fname = 'Parser::replaceInternalLinks' ;
1116
1117                 wfProfileIn( $fname );
1118
1119                 wfProfileIn( $fname.'-setup' );
1120                 static $tc = FALSE;
1121                 # the % is needed to support urlencoded titles as well
1122                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1123
1124                 $sk =& $this->mOptions->getSkin();
1125                 global $wgUseOldExistenceCheck;
1126                 # "Post-parse link colour check" works only on wiki text since it's now
1127                 # in Parser. Enable it, then disable it when we're done.
1128                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1129
1130                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1131
1132                 #split the entire text string on occurences of [[
1133                 $a = explode( '[[', ' ' . $s );
1134                 #get the first element (all text up to first [[), and remove the space we added
1135                 $s = array_shift( $a );
1136                 $s = substr( $s, 1 );
1137
1138                 # Match a link having the form [[namespace:link|alternate]]trail
1139                 static $e1 = FALSE;
1140                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1141                 # Match cases where there is no "]]", which might still be images
1142                 static $e1_img = FALSE;
1143                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1144                 # Match the end of a line for a word that's not followed by whitespace,
1145                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1146                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1147
1148                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1149
1150                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1151
1152                 if ( $useLinkPrefixExtension ) {
1153                         if ( preg_match( $e2, $s, $m ) ) {
1154                                 $first_prefix = $m[2];
1155                                 $s = $m[1];
1156                         } else {
1157                                 $first_prefix = false;
1158                         }
1159                 } else {
1160                         $prefix = '';
1161                 }
1162
1163                 wfProfileOut( $fname.'-setup' );
1164
1165                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1166                 # Loop for each link
1167                 for ($k = 0; isset( $a[$k] ); $k++) {
1168                         $line = $a[$k];
1169                         wfProfileIn( $fname.'-prefixhandling' );
1170                         if ( $useLinkPrefixExtension ) {
1171                                 if ( preg_match( $e2, $s, $m ) ) {
1172                                         $prefix = $m[2];
1173                                         $s = $m[1];
1174                                 } else {
1175                                         $prefix='';
1176                                 }
1177                                 # first link
1178                                 if($first_prefix) {
1179                                         $prefix = $first_prefix;
1180                                         $first_prefix = false;
1181                                 }
1182                         }
1183                         wfProfileOut( $fname.'-prefixhandling' );
1184
1185                         $might_be_img = false;
1186
1187                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1188                                 $text = $m[2];
1189                                 # fix up urlencoded title texts
1190                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1191                                 $trail = $m[3];
1192                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1193                                 $might_be_img = true;
1194                                 $text = $m[2];
1195                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1196                                 $trail = "";
1197                         } else { # Invalid form; output directly
1198                                 $s .= $prefix . '[[' . $line ;
1199                                 continue;
1200                         }
1201
1202                         # Don't allow internal links to pages containing
1203                         # PROTO: where PROTO is a valid URL protocol; these
1204                         # should be external links.
1205                         if (preg_match('/^((?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1206                                 $s .= $prefix . '[[' . $line ;
1207                                 continue;
1208                         }
1209
1210                         # Make subpage if necessary
1211                         $link = $this->maybeDoSubpageLink( $m[1], $text );
1212
1213                         $noforce = (substr($m[1], 0, 1) != ':');
1214                         if (!$noforce) {
1215                                 # Strip off leading ':'
1216                                 $link = substr($link, 1);
1217                         }
1218
1219                         $nt = Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1220                         if( !$nt ) {
1221                                 $s .= $prefix . '[[' . $line;
1222                                 continue;
1223                         }
1224
1225                         #check other language variants of the link
1226                         #if the article does not exist
1227                         if( $nt->getArticleID() == 0
1228                                 && $checkVariantLink ) {
1229                                 $wgContLang->findVariantLink($link, $nt);
1230                         }
1231
1232                         $ns = $nt->getNamespace();
1233                         $iw = $nt->getInterWiki();
1234
1235                         if ($might_be_img) { # if this is actually an invalid link
1236                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1237                                         $found = false;
1238                                         while (isset ($a[$k+1]) ) {
1239                                                 #look at the next 'line' to see if we can close it there
1240                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1241                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1242                                                 # the first ]] closes the inner link, the second the image
1243                                                         $found = true;
1244                                                         $text .= '[[' . $m[1];
1245                                                         $trail = $m[2];
1246                                                         break;
1247                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1248                                                         #if there's exactly one ]] that's fine, we'll keep looking
1249                                                         $text .= '[[' . $m[0];
1250                                                 } else {
1251                                                         #if $next_line is invalid too, we need look no further
1252                                                         $text .= '[[' . $next_line;
1253                                                         break;
1254                                                 }
1255                                         }
1256                                         if ( !$found ) {
1257                                                 # we couldn't find the end of this imageLink, so output it raw
1258                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1259                                                 $text = $this->replaceInternalLinks($text);
1260                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1261                                                 # note: no $trail, because without an end, there *is* no trail
1262                                                 continue;
1263                                         }
1264                                 } else { #it's not an image, so output it raw
1265                                         $s .= $prefix . '[[' . $link . '|' . $text;
1266                                         # note: no $trail, because without an end, there *is* no trail
1267                                         continue;
1268                                 }
1269                         }
1270
1271                         $wasblank = ( '' == $text );
1272                         if( $wasblank ) $text = $link;
1273
1274
1275                         # Link not escaped by : , create the various objects
1276                         if( $noforce ) {
1277
1278                                 # Interwikis
1279                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1280                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1281                                         $tmp = $prefix . $trail ;
1282                                         $s .= (trim($tmp) == '')? '': $tmp;
1283                                         continue;
1284                                 }
1285
1286                                 if ( $ns == NS_IMAGE ) {
1287                                         # recursively parse links inside the image caption
1288                                         # actually, this will parse them in any other parameters, too,
1289                                         # but it might be hard to fix that, and it doesn't matter ATM
1290                                         $text = $this->replaceExternalLinks($text);
1291                                         $text = $this->replaceInternalLinks($text);
1292
1293                                         # replace the image with a link-holder so that replaceExternalLinks() can't mess with it
1294                                         $s .= $prefix . $this->insertStripItem( $sk->makeImageLinkObj( $nt, $text ), $this->mStripState ) . $trail;
1295                                         $wgLinkCache->addImageLinkObj( $nt );
1296                                         continue;
1297                                 }
1298
1299                                 if ( $ns == NS_CATEGORY ) {
1300                                         $t = $nt->getText() ;
1301
1302                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1303                                         $pPLC=$sk->postParseLinkColour();
1304                                         $sk->postParseLinkColour( false );
1305                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1306                                         $sk->postParseLinkColour( $pPLC );
1307                                         $wgLinkCache->resume();
1308
1309                                         if ( $wasblank ) {
1310                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1311                                                         $sortkey = $this->mTitle->getText();
1312                                                 } else {
1313                                                         $sortkey = $this->mTitle->getPrefixedText();
1314                                                 }
1315                                         } else {
1316                                                 $sortkey = $text;
1317                                         }
1318                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1319                                         $this->mOutput->mCategoryLinks[] = $t ;
1320                                         $s .= $prefix . $trail ;
1321                                         continue;
1322                                 }
1323                         }
1324
1325                         if( ( $nt->getPrefixedText() === $this->mTitle->getPrefixedText() ) &&
1326                             ( $nt->getFragment() === '' ) ) {
1327                                 # Self-links are handled specially; generally de-link and change to bold.
1328                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1329                                 continue;
1330                         }
1331
1332                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1333                         if( $ns == NS_MEDIA ) {
1334                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1335                                 $wgLinkCache->addImageLinkObj( $nt );
1336                                 continue;
1337                         } elseif( $ns == NS_SPECIAL ) {
1338                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1339                                 continue;
1340                         }
1341                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1342                 }
1343                 $sk->postParseLinkColour( $saveParseColour );
1344                 wfProfileOut( $fname );
1345                 return $s;
1346         }
1347
1348         /**
1349          * Handle link to subpage if necessary
1350          * @param string $target the source of the link
1351          * @param string &$text the link text, modified as necessary
1352          * @return string the full name of the link
1353          * @access private
1354          */
1355         function maybeDoSubpageLink($target, &$text) {
1356                 # Valid link forms:
1357                 # Foobar -- normal
1358                 # :Foobar -- override special treatment of prefix (images, language links)
1359                 # /Foobar -- convert to CurrentPage/Foobar
1360                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1361                 global $wgNamespacesWithSubpages;
1362
1363                 $fname = 'Parser::maybeDoSubpageLink';
1364                 wfProfileIn( $fname );
1365                 # Look at the first character
1366                 if( $target{0} == '/' ) {
1367                         # / at end means we don't want the slash to be shown
1368                         if(substr($target,-1,1)=='/') {
1369                                 $target=substr($target,1,-1);
1370                                 $noslash=$target;
1371                         } else {
1372                                 $noslash=substr($target,1);
1373                         }
1374
1375                         # Some namespaces don't allow subpages
1376                         if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) {
1377                                 # subpages allowed here
1378                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1379                                 if( '' === $text ) {
1380                                         $text = $target;
1381                                 } # this might be changed for ugliness reasons
1382                         } else {
1383                                 # no subpage allowed, use standard link
1384                                 $ret = $target;
1385                         }
1386                 } else {
1387                         # no subpage
1388                         $ret = $target;
1389                 }
1390
1391                 wfProfileOut( $fname );
1392                 return $ret;
1393         }
1394
1395         /**#@+
1396          * Used by doBlockLevels()
1397          * @access private
1398          */
1399         /* private */ function closeParagraph() {
1400                 $result = '';
1401                 if ( '' != $this->mLastSection ) {
1402                         $result = '</' . $this->mLastSection  . ">\n";
1403                 }
1404                 $this->mInPre = false;
1405                 $this->mLastSection = '';
1406                 return $result;
1407         }
1408         # getCommon() returns the length of the longest common substring
1409         # of both arguments, starting at the beginning of both.
1410         #
1411         /* private */ function getCommon( $st1, $st2 ) {
1412                 $fl = strlen( $st1 );
1413                 $shorter = strlen( $st2 );
1414                 if ( $fl < $shorter ) { $shorter = $fl; }
1415
1416                 for ( $i = 0; $i < $shorter; ++$i ) {
1417                         if ( $st1{$i} != $st2{$i} ) { break; }
1418                 }
1419                 return $i;
1420         }
1421         # These next three functions open, continue, and close the list
1422         # element appropriate to the prefix character passed into them.
1423         #
1424         /* private */ function openList( $char ) {
1425                 $result = $this->closeParagraph();
1426
1427                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1428                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1429                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1430                 else if ( ';' == $char ) {
1431                         $result .= '<dl><dt>';
1432                         $this->mDTopen = true;
1433                 }
1434                 else { $result = '<!-- ERR 1 -->'; }
1435
1436                 return $result;
1437         }
1438
1439         /* private */ function nextItem( $char ) {
1440                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1441                 else if ( ':' == $char || ';' == $char ) {
1442                         $close = '</dd>';
1443                         if ( $this->mDTopen ) { $close = '</dt>'; }
1444                         if ( ';' == $char ) {
1445                                 $this->mDTopen = true;
1446                                 return $close . '<dt>';
1447                         } else {
1448                                 $this->mDTopen = false;
1449                                 return $close . '<dd>';
1450                         }
1451                 }
1452                 return '<!-- ERR 2 -->';
1453         }
1454
1455         /* private */ function closeList( $char ) {
1456                 if ( '*' == $char ) { $text = '</li></ul>'; }
1457                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1458                 else if ( ':' == $char ) {
1459                         if ( $this->mDTopen ) {
1460                                 $this->mDTopen = false;
1461                                 $text = '</dt></dl>';
1462                         } else {
1463                                 $text = '</dd></dl>';
1464                         }
1465                 }
1466                 else {  return '<!-- ERR 3 -->'; }
1467                 return $text."\n";
1468         }
1469         /**#@-*/
1470
1471         /**
1472          * Make lists from lines starting with ':', '*', '#', etc.
1473          *
1474          * @access private
1475          * @return string the lists rendered as HTML
1476          */
1477         function doBlockLevels( $text, $linestart ) {
1478                 $fname = 'Parser::doBlockLevels';
1479                 wfProfileIn( $fname );
1480
1481                 # Parsing through the text line by line.  The main thing
1482                 # happening here is handling of block-level elements p, pre,
1483                 # and making lists from lines starting with * # : etc.
1484                 #
1485                 $textLines = explode( "\n", $text );
1486
1487                 $lastPrefix = $output = $lastLine = '';
1488                 $this->mDTopen = $inBlockElem = false;
1489                 $prefixLength = 0;
1490                 $paragraphStack = false;
1491
1492                 if ( !$linestart ) {
1493                         $output .= array_shift( $textLines );
1494                 }
1495                 foreach ( $textLines as $oLine ) {
1496                         $lastPrefixLength = strlen( $lastPrefix );
1497                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1498                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1499                         if ( !$this->mInPre ) {
1500                                 # Multiple prefixes may abut each other for nested lists.
1501                                 $prefixLength = strspn( $oLine, '*#:;' );
1502                                 $pref = substr( $oLine, 0, $prefixLength );
1503
1504                                 # eh?
1505                                 $pref2 = str_replace( ';', ':', $pref );
1506                                 $t = substr( $oLine, $prefixLength );
1507                                 $this->mInPre = !empty($preOpenMatch);
1508                         } else {
1509                                 # Don't interpret any other prefixes in preformatted text
1510                                 $prefixLength = 0;
1511                                 $pref = $pref2 = '';
1512                                 $t = $oLine;
1513                         }
1514
1515                         # List generation
1516                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1517                                 # Same as the last item, so no need to deal with nesting or opening stuff
1518                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1519                                 $paragraphStack = false;
1520
1521                                 if ( substr( $pref, -1 ) == ';') {
1522                                         # The one nasty exception: definition lists work like this:
1523                                         # ; title : definition text
1524                                         # So we check for : in the remainder text to split up the
1525                                         # title and definition, without b0rking links.
1526                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1527                                                 $t = $t2;
1528                                                 $output .= $term . $this->nextItem( ':' );
1529                                         }
1530                                 }
1531                         } elseif( $prefixLength || $lastPrefixLength ) {
1532                                 # Either open or close a level...
1533                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1534                                 $paragraphStack = false;
1535
1536                                 while( $commonPrefixLength < $lastPrefixLength ) {
1537                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1538                                         --$lastPrefixLength;
1539                                 }
1540                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1541                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1542                                 }
1543                                 while ( $prefixLength > $commonPrefixLength ) {
1544                                         $char = substr( $pref, $commonPrefixLength, 1 );
1545                                         $output .= $this->openList( $char );
1546
1547                                         if ( ';' == $char ) {
1548                                                 # FIXME: This is dupe of code above
1549                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1550                                                         $t = $t2;
1551                                                         $output .= $term . $this->nextItem( ':' );
1552                                                 }
1553                                         }
1554                                         ++$commonPrefixLength;
1555                                 }
1556                                 $lastPrefix = $pref2;
1557                         }
1558                         if( 0 == $prefixLength ) {
1559                                 # No prefix (not in list)--go to paragraph mode
1560                                 $uniq_prefix = UNIQ_PREFIX;
1561                                 // XXX: use a stack for nestable elements like span, table and div
1562                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1563                                 $closematch = preg_match(
1564                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1565                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1566                                 if ( $openmatch or $closematch ) {
1567                                         $paragraphStack = false;
1568                                         $output .= $this->closeParagraph();
1569                                         if($preOpenMatch and !$preCloseMatch) {
1570                                                 $this->mInPre = true;
1571                                         }
1572                                         if ( $closematch ) {
1573                                                 $inBlockElem = false;
1574                                         } else {
1575                                                 $inBlockElem = true;
1576                                         }
1577                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1578                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1579                                                 // pre
1580                                                 if ($this->mLastSection != 'pre') {
1581                                                         $paragraphStack = false;
1582                                                         $output .= $this->closeParagraph().'<pre>';
1583                                                         $this->mLastSection = 'pre';
1584                                                 }
1585                                                 $t = substr( $t, 1 );
1586                                         } else {
1587                                                 // paragraph
1588                                                 if ( '' == trim($t) ) {
1589                                                         if ( $paragraphStack ) {
1590                                                                 $output .= $paragraphStack.'<br />';
1591                                                                 $paragraphStack = false;
1592                                                                 $this->mLastSection = 'p';
1593                                                         } else {
1594                                                                 if ($this->mLastSection != 'p' ) {
1595                                                                         $output .= $this->closeParagraph();
1596                                                                         $this->mLastSection = '';
1597                                                                         $paragraphStack = '<p>';
1598                                                                 } else {
1599                                                                         $paragraphStack = '</p><p>';
1600                                                                 }
1601                                                         }
1602                                                 } else {
1603                                                         if ( $paragraphStack ) {
1604                                                                 $output .= $paragraphStack;
1605                                                                 $paragraphStack = false;
1606                                                                 $this->mLastSection = 'p';
1607                                                         } else if ($this->mLastSection != 'p') {
1608                                                                 $output .= $this->closeParagraph().'<p>';
1609                                                                 $this->mLastSection = 'p';
1610                                                         }
1611                                                 }
1612                                         }
1613                                 }
1614                         }
1615                         if ($paragraphStack === false) {
1616                                 $output .= $t."\n";
1617                         }
1618                 }
1619                 while ( $prefixLength ) {
1620                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1621                         --$prefixLength;
1622                 }
1623                 if ( '' != $this->mLastSection ) {
1624                         $output .= '</' . $this->mLastSection . '>';
1625                         $this->mLastSection = '';
1626                 }
1627
1628                 wfProfileOut( $fname );
1629                 return $output;
1630         }
1631
1632         /**
1633          * Split up a string on ':', ignoring any occurences inside
1634          * <a>..</a> or <span>...</span>
1635          * @param string $str the string to split
1636          * @param string &$before set to everything before the ':'
1637          * @param string &$after set to everything after the ':'
1638          * return string the position of the ':', or false if none found
1639          */
1640         function findColonNoLinks($str, &$before, &$after) {
1641                 # I wonder if we should make this count all tags, not just <a>
1642                 # and <span>. That would prevent us from matching a ':' that
1643                 # comes in the middle of italics other such formatting....
1644                 # -- Wil
1645                 $fname = 'Parser::findColonNoLinks';
1646                 wfProfileIn( $fname );
1647                 $pos = 0;
1648                 do {
1649                         $colon = strpos($str, ':', $pos);
1650
1651                         if ($colon !== false) {
1652                                 $before = substr($str, 0, $colon);
1653                                 $after = substr($str, $colon + 1);
1654
1655                                 # Skip any ':' within <a> or <span> pairs
1656                                 $a = substr_count($before, '<a');
1657                                 $s = substr_count($before, '<span');
1658                                 $ca = substr_count($before, '</a>');
1659                                 $cs = substr_count($before, '</span>');
1660
1661                                 if ($a <= $ca and $s <= $cs) {
1662                                         # Tags are balanced before ':'; ok
1663                                         break;
1664                                 }
1665                                 $pos = $colon + 1;
1666                         }
1667                 } while ($colon !== false);
1668                 wfProfileOut( $fname );
1669                 return $colon;
1670         }
1671
1672         /**
1673          * Return value of a magic variable (like PAGENAME)
1674          *
1675          * @access private
1676          */
1677         function getVariableValue( $index ) {
1678                 global $wgContLang, $wgSitename, $wgServer;
1679
1680                 switch ( $index ) {
1681                         case MAG_CURRENTMONTH:
1682                                 return $wgContLang->formatNum( date( 'm' ) );
1683                         case MAG_CURRENTMONTHNAME:
1684                                 return $wgContLang->getMonthName( date('n') );
1685                         case MAG_CURRENTMONTHNAMEGEN:
1686                                 return $wgContLang->getMonthNameGen( date('n') );
1687                         case MAG_CURRENTDAY:
1688                                 return $wgContLang->formatNum( date('j') );
1689                         case MAG_PAGENAME:
1690                                 return $this->mTitle->getText();
1691                         case MAG_PAGENAMEE:
1692                                 return $this->mTitle->getPartialURL();
1693                         case MAG_NAMESPACE:
1694                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1695                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1696                         case MAG_CURRENTDAYNAME:
1697                                 return $wgContLang->getWeekdayName( date('w')+1 );
1698                         case MAG_CURRENTYEAR:
1699                                 return $wgContLang->formatNum( date( 'Y' ) );
1700                         case MAG_CURRENTTIME:
1701                                 return $wgContLang->time( wfTimestampNow(), false );
1702                         case MAG_NUMBEROFARTICLES:
1703                                 return $wgContLang->formatNum( wfNumberOfArticles() );
1704                         case MAG_SITENAME:
1705                                 return $wgSitename;
1706                         case MAG_SERVER:
1707                                 return $wgServer;
1708                         default:
1709                                 return NULL;
1710                 }
1711         }
1712
1713         /**
1714          * initialise the magic variables (like CURRENTMONTHNAME)
1715          *
1716          * @access private
1717          */
1718         function initialiseVariables() {
1719                 $fname = 'Parser::initialiseVariables';
1720                 wfProfileIn( $fname );
1721                 global $wgVariableIDs;
1722                 $this->mVariables = array();
1723                 foreach ( $wgVariableIDs as $id ) {
1724                         $mw =& MagicWord::get( $id );
1725                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1726                 }
1727                 wfProfileOut( $fname );
1728         }
1729
1730         /**
1731          * Replace magic variables, templates, and template arguments
1732          * with the appropriate text. Templates are substituted recursively,
1733          * taking care to avoid infinite loops.
1734          *
1735          * Note that the substitution depends on value of $mOutputType:
1736          *  OT_WIKI: only {{subst:}} templates
1737          *  OT_MSG: only magic variables
1738          *  OT_HTML: all templates and magic variables
1739          *
1740          * @param string $tex The text to transform
1741          * @param array $args Key-value pairs representing template parameters to substitute
1742          * @access private
1743          */
1744         function replaceVariables( $text, $args = array() ) {
1745                 global $wgLang, $wgScript, $wgArticlePath;
1746
1747                 # Prevent too big inclusions
1748                 if(strlen($text)> MAX_INCLUDE_SIZE)
1749                 return $text;
1750
1751                 $fname = 'Parser::replaceVariables';
1752                 wfProfileIn( $fname );
1753
1754                 $titleChars = Title::legalChars();
1755
1756                 # This function is called recursively. To keep track of arguments we need a stack:
1757                 array_push( $this->mArgStack, $args );
1758
1759                 # Variable substitution
1760                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1761
1762                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1763                         # Argument substitution
1764                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1765                 }
1766                 # Template substitution
1767                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1768                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1769
1770                 array_pop( $this->mArgStack );
1771
1772                 wfProfileOut( $fname );
1773                 return $text;
1774         }
1775
1776         /**
1777          * Replace magic variables
1778          * @access private
1779          */
1780         function variableSubstitution( $matches ) {
1781                 if ( !$this->mVariables ) {
1782                         $this->initialiseVariables();
1783                 }
1784                 $skip = false;
1785                 if ( $this->mOutputType == OT_WIKI ) {
1786                         # Do only magic variables prefixed by SUBST
1787                         $mwSubst =& MagicWord::get( MAG_SUBST );
1788                         if (!$mwSubst->matchStartAndRemove( $matches[1] ))
1789                                 $skip = true;
1790                         # Note that if we don't substitute the variable below,
1791                         # we don't remove the {{subst:}} magic word, in case
1792                         # it is a template rather than a magic variable.
1793                 }
1794                 if ( !$skip && array_key_exists( $matches[1], $this->mVariables ) ) {
1795                         $text = $this->mVariables[$matches[1]];
1796                         $this->mOutput->mContainsOldMagic = true;
1797                 } else {
1798                         $text = $matches[0];
1799                 }
1800                 return $text;
1801         }
1802
1803         # Split template arguments
1804         function getTemplateArgs( $argsString ) {
1805                 if ( $argsString === '' ) {
1806                         return array();
1807                 }
1808
1809                 $args = explode( '|', substr( $argsString, 1 ) );
1810
1811                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1812                 # merged with the next arg because the '|' character between belongs
1813                 # to the link syntax and not the template parameter syntax.
1814                 $argc = count($args);
1815                 $i = 0;
1816                 for ( $i = 0; $i < $argc-1; $i++ ) {
1817                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1818                                 $args[$i] .= '|'.$args[$i+1];
1819                                 array_splice($args, $i+1, 1);
1820                                 $i--;
1821                                 $argc--;
1822                         }
1823                 }
1824
1825                 return $args;
1826         }
1827
1828         /**
1829          * Return the text of a template, after recursively
1830          * replacing any variables or templates within the template.
1831          *
1832          * @param array $matches The parts of the template
1833          *  $matches[1]: the title, i.e. the part before the |
1834          *  $matches[2]: the parameters (including a leading |), if  any
1835          * @return string the text of the template
1836          * @access private
1837          */
1838         function braceSubstitution( $matches ) {
1839                 global $wgLinkCache, $wgContLang;
1840                 $fname = 'Parser::braceSubstitution';
1841                 $found = false;
1842                 $nowiki = false;
1843                 $noparse = false;
1844
1845                 $title = NULL;
1846
1847                 # Need to know if the template comes at the start of a line,
1848                 # to treat the beginning of the template like the beginning
1849                 # of a line for tables and block-level elements.
1850                 $linestart = $matches[1];
1851
1852                 # $part1 is the bit before the first |, and must contain only title characters
1853                 # $args is a list of arguments, starting from index 0, not including $part1
1854
1855                 $part1 = $matches[2];
1856                 # If the third subpattern matched anything, it will start with |
1857
1858                 $args = $this->getTemplateArgs($matches[3]);
1859                 $argc = count( $args );
1860
1861                 # Don't parse {{{}}} because that's only for template arguments
1862                 if ( $linestart === '{' ) {
1863                         $text = $matches[0];
1864                         $found = true;
1865                         $noparse = true;
1866                 }
1867
1868                 # SUBST
1869                 if ( !$found ) {
1870                         $mwSubst =& MagicWord::get( MAG_SUBST );
1871                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1872                                 # One of two possibilities is true:
1873                                 # 1) Found SUBST but not in the PST phase
1874                                 # 2) Didn't find SUBST and in the PST phase
1875                                 # In either case, return without further processing
1876                                 $text = $matches[0];
1877                                 $found = true;
1878                                 $noparse = true;
1879                         }
1880                 }
1881
1882                 # MSG, MSGNW and INT
1883                 if ( !$found ) {
1884                         # Check for MSGNW:
1885                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1886                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1887                                 $nowiki = true;
1888                         } else {
1889                                 # Remove obsolete MSG:
1890                                 $mwMsg =& MagicWord::get( MAG_MSG );
1891                                 $mwMsg->matchStartAndRemove( $part1 );
1892                         }
1893
1894                         # Check if it is an internal message
1895                         $mwInt =& MagicWord::get( MAG_INT );
1896                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1897                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1898                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1899                                         $found = true;
1900                                 }
1901                         }
1902                 }
1903
1904                 # NS
1905                 if ( !$found ) {
1906                         # Check for NS: (namespace expansion)
1907                         $mwNs = MagicWord::get( MAG_NS );
1908                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1909                                 if ( intval( $part1 ) ) {
1910                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1911                                         $found = true;
1912                                 } else {
1913                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1914                                         if ( !is_null( $index ) ) {
1915                                                 $text = $linestart . $wgContLang->getNsText( $index );
1916                                                 $found = true;
1917                                         }
1918                                 }
1919                         }
1920                 }
1921
1922                 # LOCALURL and LOCALURLE
1923                 if ( !$found ) {
1924                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1925                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1926
1927                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1928                                 $func = 'getLocalURL';
1929                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1930                                 $func = 'escapeLocalURL';
1931                         } else {
1932                                 $func = '';
1933                         }
1934
1935                         if ( $func !== '' ) {
1936                                 $title = Title::newFromText( $part1 );
1937                                 if ( !is_null( $title ) ) {
1938                                         if ( $argc > 0 ) {
1939                                                 $text = $linestart . $title->$func( $args[0] );
1940                                         } else {
1941                                                 $text = $linestart . $title->$func();
1942                                         }
1943                                         $found = true;
1944                                 }
1945                         }
1946                 }
1947
1948                 # GRAMMAR
1949                 if ( !$found && $argc == 1 ) {
1950                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1951                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1952                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
1953                                 $found = true;
1954                         }
1955                 }
1956
1957                 # Template table test
1958
1959                 # Did we encounter this template already? If yes, it is in the cache
1960                 # and we need to check for loops.
1961                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
1962                         # set $text to cached message.
1963                         $text = $linestart . $this->mTemplates[$part1];
1964                         $found = true;
1965
1966                         # Infinite loop test
1967                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1968                                 $noparse = true;
1969                                 $found = true;
1970                                 $text .= '<!-- WARNING: template loop detected -->';
1971                         }
1972                 }
1973
1974                 # Load from database
1975                 $itcamefromthedatabase = false;
1976                 if ( !$found ) {
1977                         $ns = NS_TEMPLATE;
1978                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
1979                         if ($subpage !== '') {
1980                                 $ns = $this->mTitle->getNamespace();
1981                         }
1982                         $title = Title::newFromText( $part1, $ns );
1983                         if ( !is_null( $title ) && !$title->isExternal() ) {
1984                                 # Check for excessive inclusion
1985                                 $dbk = $title->getPrefixedDBkey();
1986                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1987                                         # This should never be reached.
1988                                         $article = new Article( $title );
1989                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1990                                         if ( $articleContent !== false ) {
1991                                                 $found = true;
1992                                                 $text = $linestart . $articleContent;
1993                                                 $itcamefromthedatabase = true;
1994                                         }
1995                                 }
1996
1997                                 # If the title is valid but undisplayable, make a link to it
1998                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1999                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2000                                         $found = true;
2001                                 }
2002
2003                                 # Template cache array insertion
2004                                 $this->mTemplates[$part1] = $text;
2005                         }
2006                 }
2007
2008                 # Recursive parsing, escaping and link table handling
2009                 # Only for HTML output
2010                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2011                         $text = wfEscapeWikiText( $text );
2012                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2013                         # Clean up argument array
2014                         $assocArgs = array();
2015                         $index = 1;
2016                         foreach( $args as $arg ) {
2017                                 $eqpos = strpos( $arg, '=' );
2018                                 if ( $eqpos === false ) {
2019                                         $assocArgs[$index++] = $arg;
2020                                 } else {
2021                                         $name = trim( substr( $arg, 0, $eqpos ) );
2022                                         $value = trim( substr( $arg, $eqpos+1 ) );
2023                                         if ( $value === false ) {
2024                                                 $value = '';
2025                                         }
2026                                         if ( $name !== false ) {
2027                                                 $assocArgs[$name] = $value;
2028                                         }
2029                                 }
2030                         }
2031
2032                         # Add a new element to the templace recursion path
2033                         $this->mTemplatePath[$part1] = 1;
2034
2035                         $text = $this->strip( $text, $this->mStripState );
2036                         $text = $this->removeHTMLtags( $text );
2037                         $text = $this->replaceVariables( $text, $assocArgs );
2038
2039                         # Resume the link cache and register the inclusion as a link
2040                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2041                                 $wgLinkCache->addLinkObj( $title );
2042                         }
2043
2044                         # If the template begins with a table or block-level
2045                         # element, it should be treated as beginning a new line.
2046                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2047                                 $text = "\n" . $text;
2048                         }
2049                 }
2050
2051                 # Empties the template path
2052                 $this->mTemplatePath = array();
2053                 if ( !$found ) {
2054                         return $matches[0];
2055                 } else {
2056                         # replace ==section headers==
2057                         # XXX this needs to go away once we have a better parser.
2058                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2059                                 if( !is_null( $title ) )
2060                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2061                                 else
2062                                         $encodedname = base64_encode("");
2063                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2064                                         PREG_SPLIT_DELIM_CAPTURE);
2065                                 $text = '';
2066                                 $nsec = 0;
2067                                 for( $i = 0; $i < count($m); $i += 2 ) {
2068                                         $text .= $m[$i];
2069                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2070                                         $hl = $m[$i + 1];
2071                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2072                                                 $text .= $hl;
2073                                                 continue;
2074                                         }
2075                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2076                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2077                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2078
2079                                         $nsec++;
2080                                 }
2081                         }
2082                 }
2083
2084                 # Empties the template path
2085                 $this->mTemplatePath = array();
2086                 if ( !$found ) {
2087                         return $matches[0];
2088                 } else {
2089                         return $text;
2090                 }
2091         }
2092
2093         /**
2094          * Triple brace replacement -- used for template arguments
2095          * @access private
2096          */
2097         function argSubstitution( $matches ) {
2098                 $arg = trim( $matches[1] );
2099                 $text = $matches[0];
2100                 $inputArgs = end( $this->mArgStack );
2101
2102                 if ( array_key_exists( $arg, $inputArgs ) ) {
2103                         $text = $inputArgs[$arg];
2104                 }
2105
2106                 return $text;
2107         }
2108
2109         /**
2110          * Returns true if the function is allowed to include this entity
2111          * @access private
2112          */
2113         function incrementIncludeCount( $dbk ) {
2114                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2115                         $this->mIncludeCount[$dbk] = 0;
2116                 }
2117                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2118                         return true;
2119                 } else {
2120                         return false;
2121                 }
2122         }
2123
2124
2125         /**
2126          * Cleans up HTML, removes dangerous tags and attributes, and
2127          * removes HTML comments
2128          * @access private
2129          */
2130         function removeHTMLtags( $text ) {
2131                 global $wgUseTidy, $wgUserHtml;
2132                 $fname = 'Parser::removeHTMLtags';
2133                 wfProfileIn( $fname );
2134
2135                 if( $wgUserHtml ) {
2136                         $htmlpairs = array( # Tags that must be closed
2137                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
2138                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
2139                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
2140                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
2141                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
2142                         );
2143                         $htmlsingle = array(
2144                                 'br', 'hr', 'li', 'dt', 'dd'
2145                         );
2146                         $htmlnest = array( # Tags that can be nested--??
2147                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
2148                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
2149                         );
2150                         $tabletags = array( # Can only appear inside table
2151                                 'td', 'th', 'tr'
2152                         );
2153                 } else {
2154                         $htmlpairs = array();
2155                         $htmlsingle = array();
2156                         $htmlnest = array();
2157                         $tabletags = array();
2158                 }
2159
2160                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
2161                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
2162
2163                 $htmlattrs = $this->getHTMLattrs () ;
2164
2165                 # Remove HTML comments
2166                 $text = $this->removeHTMLcomments( $text );
2167
2168                 $bits = explode( '<', $text );
2169                 $text = array_shift( $bits );
2170                 if(!$wgUseTidy) {
2171                         $tagstack = array(); $tablestack = array();
2172                         foreach ( $bits as $x ) {
2173                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2174                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2175                                 $x, $regs );
2176                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2177                                 error_reporting( $prev );
2178
2179                                 $badtag = 0 ;
2180                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2181                                         # Check our stack
2182                                         if ( $slash ) {
2183                                                 # Closing a tag...
2184                                                 if ( ! in_array( $t, $htmlsingle ) &&
2185                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2186                                                         @array_push( $tagstack, $ot );
2187                                                         $badtag = 1;
2188                                                 } else {
2189                                                         if ( $t == 'table' ) {
2190                                                                 $tagstack = array_pop( $tablestack );
2191                                                         }
2192                                                         $newparams = '';
2193                                                 }
2194                                         } else {
2195                                                 # Keep track for later
2196                                                 if ( in_array( $t, $tabletags ) &&
2197                                                 ! in_array( 'table', $tagstack ) ) {
2198                                                         $badtag = 1;
2199                                                 } else if ( in_array( $t, $tagstack ) &&
2200                                                 ! in_array ( $t , $htmlnest ) ) {
2201                                                         $badtag = 1 ;
2202                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
2203                                                         if ( $t == 'table' ) {
2204                                                                 array_push( $tablestack, $tagstack );
2205                                                                 $tagstack = array();
2206                                                         }
2207                                                         array_push( $tagstack, $t );
2208                                                 }
2209                                                 # Strip non-approved attributes from the tag
2210                                                 $newparams = $this->fixTagAttributes($params);
2211
2212                                         }
2213                                         if ( ! $badtag ) {
2214                                                 $rest = str_replace( '>', '&gt;', $rest );
2215                                                 $text .= "<$slash$t $newparams$brace$rest";
2216                                                 continue;
2217                                         }
2218                                 }
2219                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2220                         }
2221                         # Close off any remaining tags
2222                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2223                                 $text .= "</$t>\n";
2224                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2225                         }
2226                 } else {
2227                         # this might be possible using tidy itself
2228                         foreach ( $bits as $x ) {
2229                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2230                                 $x, $regs );
2231                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2232                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2233                                         $newparams = $this->fixTagAttributes($params);
2234                                         $rest = str_replace( '>', '&gt;', $rest );
2235                                         $text .= "<$slash$t $newparams$brace$rest";
2236                                 } else {
2237                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2238                                 }
2239                         }
2240                 }
2241                 wfProfileOut( $fname );
2242                 return $text;
2243         }
2244
2245         /**
2246          * Remove '<!--', '-->', and everything between.
2247          * To avoid leaving blank lines, when a comment is both preceded
2248          * and followed by a newline (ignoring spaces), trim leading and
2249          * trailing spaces and one of the newlines.
2250          *
2251          * @access private
2252          */
2253         function removeHTMLcomments( $text ) {
2254                 $fname='Parser::removeHTMLcomments';
2255                 wfProfileIn( $fname );
2256                 while (($start = strpos($text, '<!--')) !== false) {
2257                         $end = strpos($text, '-->', $start + 4);
2258                         if ($end === false) {
2259                                 # Unterminated comment; bail out
2260                                 break;
2261                         }
2262
2263                         $end += 3;
2264
2265                         # Trim space and newline if the comment is both
2266                         # preceded and followed by a newline
2267                         $spaceStart = max($start - 1, 0);
2268                         $spaceLen = $end - $spaceStart;
2269                         while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2270                                 $spaceStart--;
2271                                 $spaceLen++;
2272                         }
2273                         while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2274                                 $spaceLen++;
2275                         if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2276                                 # Remove the comment, leading and trailing
2277                                 # spaces, and leave only one newline.
2278                                 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2279                         }
2280                         else {
2281                                 # Remove just the comment.
2282                                 $text = substr_replace($text, '', $start, $end - $start);
2283                         }
2284                 }
2285                 wfProfileOut( $fname );
2286                 return $text;
2287         }
2288
2289         /**
2290          * This function accomplishes several tasks:
2291          * 1) Auto-number headings if that option is enabled
2292          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2293          * 3) Add a Table of contents on the top for users who have enabled the option
2294          * 4) Auto-anchor headings
2295          *
2296          * It loops through all headlines, collects the necessary data, then splits up the
2297          * string and re-inserts the newly formatted headlines.
2298          * @access private
2299          */
2300         /* private */ function formatHeadings( $text, $isMain=true ) {
2301                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders;
2302
2303                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2304                 $doShowToc = $this->mOptions->getShowToc();
2305                 $forceTocHere = false;
2306                 if( !$this->mTitle->userCanEdit() ) {
2307                         $showEditLink = 0;
2308                         $rightClickHack = 0;
2309                 } else {
2310                         $showEditLink = $this->mOptions->getEditSection();
2311                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2312                 }
2313
2314                 # Inhibit editsection links if requested in the page
2315                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2316                 if( $esw->matchAndRemove( $text ) ) {
2317                         $showEditLink = 0;
2318                 }
2319                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2320                 # do not add TOC
2321                 $mw =& MagicWord::get( MAG_NOTOC );
2322                 if( $mw->matchAndRemove( $text ) ) {
2323                         $doShowToc = 0;
2324                 }
2325
2326                 # never add the TOC to the Main Page. This is an entry page that should not
2327                 # be more than 1-2 screens large anyway
2328                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2329                         $doShowToc = 0;
2330                 }
2331
2332                 # Get all headlines for numbering them and adding funky stuff like [edit]
2333                 # links - this is for later, but we need the number of headlines right now
2334                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
2335
2336                 # if there are fewer than 4 headlines in the article, do not show TOC
2337                 if( $numMatches < 4 ) {
2338                         $doShowToc = 0;
2339                 }
2340
2341                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2342                 # override above conditions and always show TOC at that place
2343                 $mw =& MagicWord::get( MAG_TOC );
2344                 if ($mw->match( $text ) ) {
2345                         $doShowToc = 1;
2346                         $forceTocHere = true;
2347                 } else {
2348                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2349                         # override above conditions and always show TOC above first header
2350                         $mw =& MagicWord::get( MAG_FORCETOC );
2351                         if ($mw->matchAndRemove( $text ) ) {
2352                                 $doShowToc = 1;
2353                         }
2354                 }
2355
2356
2357
2358                 # We need this to perform operations on the HTML
2359                 $sk =& $this->mOptions->getSkin();
2360
2361                 # headline counter
2362                 $headlineCount = 0;
2363                 $sectionCount = 0; # headlineCount excluding template sections
2364
2365                 # Ugh .. the TOC should have neat indentation levels which can be
2366                 # passed to the skin functions. These are determined here
2367                 $toclevel = 0;
2368                 $toc = '';
2369                 $full = '';
2370                 $head = array();
2371                 $sublevelCount = array();
2372                 $level = 0;
2373                 $prevlevel = 0;
2374                 foreach( $matches[3] as $headline ) {
2375                         $istemplate = 0;
2376                         $templatetitle = "";
2377                         $templatesection = 0;
2378
2379                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2380                                 $istemplate = 1;
2381                                 $templatetitle = base64_decode($mat[1]);
2382                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2383                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2384                         }
2385
2386                         $numbering = '';
2387                         if( $level ) {
2388                                 $prevlevel = $level;
2389                         }
2390                         $level = $matches[1][$headlineCount];
2391                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2392                                 # reset when we enter a new level
2393                                 $sublevelCount[$level] = 0;
2394                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2395                                 $toclevel += $level - $prevlevel;
2396                         }
2397                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2398                                 # reset when we step back a level
2399                                 $sublevelCount[$level+1]=0;
2400                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2401                                 $toclevel -= $prevlevel - $level;
2402                         }
2403                         # count number of headlines for each level
2404                         @$sublevelCount[$level]++;
2405                         if( $doNumberHeadings || $doShowToc ) {
2406                                 $dot = 0;
2407                                 for( $i = 1; $i <= $level; $i++ ) {
2408                                         if( !empty( $sublevelCount[$i] ) ) {
2409                                                 if( $dot ) {
2410                                                         $numbering .= '.';
2411                                                 }
2412                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2413                                                 $dot = 1;
2414                                         }
2415                                 }
2416                         }
2417
2418                         # The canonized header is a version of the header text safe to use for links
2419                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2420                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2421                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2422
2423                         # Remove link placeholders by the link text.
2424                         #     <!--LINK number-->
2425                         # turns into
2426                         #     link text with suffix
2427                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2428                                                             "\$wgLinkHolders['texts'][\$1]",
2429                                                             $canonized_headline );
2430
2431                         # strip out HTML
2432                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2433                         $tocline = trim( $canonized_headline );
2434                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2435                         $replacearray = array(
2436                                 '%3A' => ':',
2437                                 '%' => '.'
2438                         );
2439                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2440                         $refer[$headlineCount] = $canonized_headline;
2441
2442                         # count how many in assoc. array so we can track dupes in anchors
2443                         @$refers[$canonized_headline]++;
2444                         $refcount[$headlineCount]=$refers[$canonized_headline];
2445
2446                         # Prepend the number to the heading text
2447
2448                         if( $doNumberHeadings || $doShowToc ) {
2449                                 $tocline = $numbering . ' ' . $tocline;
2450
2451                                 # Don't number the heading if it is the only one (looks silly)
2452                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2453                                         # the two are different if the line contains a link
2454                                         $headline=$numbering . ' ' . $headline;
2455                                 }
2456                         }
2457
2458                         # Create the anchor for linking from the TOC to the section
2459                         $anchor = $canonized_headline;
2460                         if($refcount[$headlineCount] > 1 ) {
2461                                 $anchor .= '_' . $refcount[$headlineCount];
2462                         }
2463                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2464                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2465                         }
2466                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2467                                 if ( empty( $head[$headlineCount] ) ) {
2468                                         $head[$headlineCount] = '';
2469                                 }
2470                                 if( $istemplate )
2471                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2472                                 else
2473                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2474                         }
2475
2476                         # Add the edit section span
2477                         if( $rightClickHack ) {
2478                                 if( $istemplate )
2479                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2480                                 else
2481                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2482                         }
2483
2484                         # give headline the correct <h#> tag
2485                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2486
2487                         $headlineCount++;
2488                         if( !$istemplate )
2489                                 $sectionCount++;
2490                 }
2491
2492                 if( $doShowToc ) {
2493                         $toclines = $headlineCount;
2494                         $toc .= $sk->tocUnindent( $toclevel );
2495                         $toc = $sk->tocTable( $toc );
2496                 }
2497
2498                 # split up and insert constructed headlines
2499
2500                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2501                 $i = 0;
2502
2503                 foreach( $blocks as $block ) {
2504                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2505                                 # This is the [edit] link that appears for the top block of text when
2506                                 # section editing is enabled
2507
2508                                 # Disabled because it broke block formatting
2509                                 # For example, a bullet point in the top line
2510                                 # $full .= $sk->editSectionLink(0);
2511                         }
2512                         $full .= $block;
2513                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2514                         # Top anchor now in skin
2515                                 $full = $full.$toc;
2516                         }
2517
2518                         if( !empty( $head[$i] ) ) {
2519                                 $full .= $head[$i];
2520                         }
2521                         $i++;
2522                 }
2523                 if($forceTocHere) {
2524                         $mw =& MagicWord::get( MAG_TOC );
2525                         return $mw->replace( $toc, $full );
2526                 } else {
2527                         return $full;
2528                 }
2529         }
2530
2531         /**
2532          * Return an HTML link for the "ISBN 123456" text
2533          * @access private
2534          */
2535         function magicISBN( $text ) {
2536                 global $wgLang;
2537                 $fname = 'Parser::magicISBN';
2538                 wfProfileIn( $fname );
2539
2540                 $a = split( 'ISBN ', ' '.$text );
2541                 if ( count ( $a ) < 2 ) {
2542                         wfProfileOut( $fname );
2543                         return $text;
2544                 }
2545                 $text = substr( array_shift( $a ), 1);
2546                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2547
2548                 foreach ( $a as $x ) {
2549                         $isbn = $blank = '' ;
2550                         while ( ' ' == $x{0} ) {
2551                                 $blank .= ' ';
2552                                 $x = substr( $x, 1 );
2553                         }
2554                         if ( $x == '' ) { # blank isbn
2555                                 $text .= "ISBN $blank";
2556                                 continue;
2557                         }
2558                         while ( strstr( $valid, $x{0} ) != false ) {
2559                                 $isbn .= $x{0};
2560                                 $x = substr( $x, 1 );
2561                         }
2562                         $num = str_replace( '-', '', $isbn );
2563                         $num = str_replace( ' ', '', $num );
2564
2565                         if ( '' == $num ) {
2566                                 $text .= "ISBN $blank$x";
2567                         } else {
2568                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2569                                 $text .= '<a href="' .
2570                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2571                                         "\" class=\"internal\">ISBN $isbn</a>";
2572                                 $text .= $x;
2573                         }
2574                 }
2575                 wfProfileOut( $fname );
2576                 return $text;
2577         }
2578
2579         /**
2580          * Return an HTML link for the "GEO ..." text
2581          * @access private
2582          */
2583         function magicGEO( $text ) {
2584                 global $wgLang, $wgUseGeoMode;
2585                 $fname = 'Parser::magicGEO';
2586                 wfProfileIn( $fname );
2587
2588                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2589                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2590                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2591                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2592                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2593                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2594
2595                 $a = split( 'GEO ', ' '.$text );
2596                 if ( count ( $a ) < 2 ) {
2597                         wfProfileOut( $fname );
2598                         return $text;
2599                 }
2600                 $text = substr( array_shift( $a ), 1);
2601                 $valid = '0123456789.+-:';
2602
2603                 foreach ( $a as $x ) {
2604                         $geo = $blank = '' ;
2605                         while ( ' ' == $x{0} ) {
2606                                 $blank .= ' ';
2607                                 $x = substr( $x, 1 );
2608                         }
2609                         while ( strstr( $valid, $x{0} ) != false ) {
2610                                 $geo .= $x{0};
2611                                 $x = substr( $x, 1 );
2612                         }
2613                         $num = str_replace( '+', '', $geo );
2614                         $num = str_replace( ' ', '', $num );
2615
2616                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2617                                 $text .= "GEO $blank$x";
2618                         } else {
2619                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2620                                 $text .= '<a href="' .
2621                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2622                                         "\" class=\"internal\">GEO $geo</a>";
2623                                 $text .= $x;
2624                         }
2625                 }
2626                 wfProfileOut( $fname );
2627                 return $text;
2628         }
2629
2630         /**
2631          * Return an HTML link for the "RFC 1234" text
2632          * @access private
2633          * @param string $text text to be processed
2634          */
2635         function magicRFC( $text ) {
2636                 global $wgLang;
2637
2638                 $valid = '0123456789';
2639                 $internal = false;
2640
2641                 $a = split( 'RFC ', ' '.$text );
2642                 if ( count ( $a ) < 2 ) return $text;
2643                 $text = substr( array_shift( $a ), 1);
2644
2645                 /* Check if RFC keyword is preceed by [[.
2646                  * This test is made here cause of the array_shift above
2647                  * that prevent the test to be done in the foreach.
2648                  */
2649                 if(substr($text, -2) == '[[') { $internal = true; }
2650
2651                 foreach ( $a as $x ) {
2652                         /* token might be empty if we have RFC RFC 1234 */
2653                         if($x=='') {
2654                                 $text.='RFC ';
2655                                 continue;
2656                                 }
2657
2658                         $rfc = $blank = '' ;
2659
2660                         /** remove and save whitespaces in $blank */
2661                         while ( $x{0} == ' ' ) {
2662                                 $blank .= ' ';
2663                                 $x = substr( $x, 1 );
2664                         }
2665
2666                         /** remove and save the rfc number in $rfc */
2667                         while ( strstr( $valid, $x{0} ) != false ) {
2668                                 $rfc .= $x{0};
2669                                 $x = substr( $x, 1 );
2670                         }
2671
2672                         if ( $rfc == '') {
2673                                 /* call back stripped spaces*/
2674                                 $text .= "RFC $blank$x";
2675                         } elseif( $internal) {
2676                                 /* normal link */
2677                                 $text .= "RFC $rfc$x";
2678                         } else {
2679                                 /* build the external link*/
2680                                 $url = wfmsg( 'rfcurl' );
2681                                 $url = str_replace( '$1', $rfc, $url);
2682                                 $sk =& $this->mOptions->getSkin();
2683                                 $la = $sk->getExternalLinkAttributes( $url, 'RFC '.$rfc );
2684                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2685                         }
2686
2687                         /* Check if the next RFC keyword is preceed by [[ */
2688                         $internal = (substr($x,-2) == '[[');
2689                 }
2690                 return $text;
2691         }
2692
2693         /**
2694          * Transform wiki markup when saving a page by doing \r\n -> \n
2695          * conversion, substitting signatures, {{subst:}} templates, etc.
2696          *
2697          * @param string $text the text to transform
2698          * @param Title &$title the Title object for the current article
2699          * @param User &$user the User object describing the current user
2700          * @param ParserOptions $options parsing options
2701          * @param bool $clearState whether to clear the parser state first
2702          * @return string the altered wiki markup
2703          * @access public
2704          */
2705         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2706                 $this->mOptions = $options;
2707                 $this->mTitle =& $title;
2708                 $this->mOutputType = OT_WIKI;
2709
2710                 if ( $clearState ) {
2711                         $this->clearState();
2712                 }
2713
2714                 $stripState = false;
2715                 $pairs = array(
2716                         "\r\n" => "\n",
2717                         );
2718                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2719                 $text = $this->strip( $text, $stripState, false );
2720                 $text = $this->pstPass2( $text, $user );
2721                 $text = $this->unstrip( $text, $stripState );
2722                 $text = $this->unstripNoWiki( $text, $stripState );
2723                 return $text;
2724         }
2725
2726         /**
2727          * Pre-save transform helper function
2728          * @access private
2729          */
2730         function pstPass2( $text, &$user ) {
2731                 global $wgLang, $wgContLang, $wgLocaltimezone;
2732
2733                 # Variable replacement
2734                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2735                 $text = $this->replaceVariables( $text );
2736
2737                 # Signatures
2738                 #
2739                 $n = $user->getName();
2740                 $k = $user->getOption( 'nickname' );
2741                 if ( '' == $k ) { $k = $n; }
2742                 if(isset($wgLocaltimezone)) {
2743                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2744                 }
2745                 /* Note: this is an ugly timezone hack for the European wikis */
2746                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2747                   ' (' . date( 'T' ) . ')';
2748                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2749
2750                 $text = preg_replace( '/~~~~~/', $d, $text );
2751                 $text = preg_replace( '/~~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2752                 $text = preg_replace( '/~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2753
2754                 # Context links: [[|name]] and [[name (context)|]]
2755                 #
2756                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2757                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2758                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2759                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2760
2761                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2762                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2763                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2764                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2765                 $context = '';
2766                 $t = $this->mTitle->getText();
2767                 if ( preg_match( $conpat, $t, $m ) ) {
2768                         $context = $m[2];
2769                 }
2770                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2771                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2772                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2773
2774                 if ( '' == $context ) {
2775                         $text = preg_replace( $p2, '[[\\1]]', $text );
2776                 } else {
2777                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2778                 }
2779
2780                 # Trim trailing whitespace
2781                 # MAG_END (__END__) tag allows for trailing
2782                 # whitespace to be deliberately included
2783                 $text = rtrim( $text );
2784                 $mw =& MagicWord::get( MAG_END );
2785                 $mw->matchAndRemove( $text );
2786
2787                 return $text;
2788         }
2789
2790         /**
2791          * Set up some variables which are usually set up in parse()
2792          * so that an external function can call some class members with confidence
2793          * @access public
2794          */
2795         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2796                 $this->mTitle =& $title;
2797                 $this->mOptions = $options;
2798                 $this->mOutputType = $outputType;
2799                 if ( $clearState ) {
2800                         $this->clearState();
2801                 }
2802         }
2803
2804         /**
2805          * Transform a MediaWiki message by replacing magic variables.
2806          *
2807          * @param string $text the text to transform
2808          * @param ParserOptions $options  options
2809          * @return string the text with variables substituted
2810          * @access public
2811          */
2812         function transformMsg( $text, $options ) {
2813                 global $wgTitle;
2814                 static $executing = false;
2815
2816                 # Guard against infinite recursion
2817                 if ( $executing ) {
2818                         return $text;
2819                 }
2820                 $executing = true;
2821
2822                 $this->mTitle = $wgTitle;
2823                 $this->mOptions = $options;
2824                 $this->mOutputType = OT_MSG;
2825                 $this->clearState();
2826                 $text = $this->replaceVariables( $text );
2827
2828                 $executing = false;
2829                 return $text;
2830         }
2831
2832         /**
2833          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2834          * Callback will be called with the text within
2835          * Transform and return the text within
2836          * @access public
2837          */
2838         function setHook( $tag, $callback ) {
2839                 $oldVal = @$this->mTagHooks[$tag];
2840                 $this->mTagHooks[$tag] = $callback;
2841                 return $oldVal;
2842         }
2843
2844         /**
2845          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2846          * Placeholders created in Skin::makeLinkObj()
2847          * Returns an array of links found, indexed by PDBK:
2848          *  0 - broken
2849          *  1 - normal link
2850          *  2 - stub
2851          * $options is a bit field, RLH_FOR_UPDATE to select for update
2852          */
2853         function replaceLinkHolders( &$text, $options = 0 ) {
2854                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2855                 global $wgInterwikiLinkHolders;
2856                 global $outputReplace;
2857
2858                 if ( $wgUseOldExistenceCheck ) {
2859                         return array();
2860                 }
2861
2862                 $fname = 'Parser::replaceLinkHolders';
2863                 wfProfileIn( $fname );
2864
2865                 $pdbks = array();
2866                 $colours = array();
2867
2868                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2869                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2870                         wfProfileIn( $fname.'-check' );
2871                         $dbr =& wfGetDB( DB_SLAVE );
2872                         $cur = $dbr->tableName( 'cur' );
2873                         $sk = $wgUser->getSkin();
2874                         $threshold = $wgUser->getOption('stubthreshold');
2875
2876                         # Sort by namespace
2877                         asort( $wgLinkHolders['namespaces'] );
2878
2879                         # Generate query
2880                         $query = false;
2881                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2882                                 # Make title object
2883                                 $title = $wgLinkHolders['titles'][$key];
2884
2885                                 # Skip invalid entries.
2886                                 # Result will be ugly, but prevents crash.
2887                                 if ( is_null( $title ) ) {
2888                                         continue;
2889                                 }
2890                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2891
2892                                 # Check if it's in the link cache already
2893                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2894                                         $colours[$pdbk] = 1;
2895                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2896                                         $colours[$pdbk] = 0;
2897                                 } else {
2898                                         # Not in the link cache, add it to the query
2899                                         if ( !isset( $current ) ) {
2900                                                 $current = $val;
2901                                                 $query =  "SELECT cur_id, cur_namespace, cur_title";
2902                                                 if ( $threshold > 0 ) {
2903                                                         $query .= ", LENGTH(cur_text) AS cur_len, cur_is_redirect";
2904                                                 }
2905                                                 $query .= " FROM $cur WHERE (cur_namespace=$val AND cur_title IN(";
2906                                         } elseif ( $current != $val ) {
2907                                                 $current = $val;
2908                                                 $query .= ")) OR (cur_namespace=$val AND cur_title IN(";
2909                                         } else {
2910                                                 $query .= ', ';
2911                                         }
2912
2913                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
2914                                 }
2915                         }
2916                         if ( $query ) {
2917                                 $query .= '))';
2918                                 if ( $options & RLH_FOR_UPDATE ) {
2919                                         $query .= ' FOR UPDATE';
2920                                 }
2921
2922                                 $res = $dbr->query( $query, $fname );
2923
2924                                 # Fetch data and form into an associative array
2925                                 # non-existent = broken
2926                                 # 1 = known
2927                                 # 2 = stub
2928                                 while ( $s = $dbr->fetchObject($res) ) {
2929                                         $title = Title::makeTitle( $s->cur_namespace, $s->cur_title );
2930                                         $pdbk = $title->getPrefixedDBkey();
2931                                         $wgLinkCache->addGoodLink( $s->cur_id, $pdbk );
2932
2933                                         if ( $threshold >  0 ) {
2934                                                 $size = $s->cur_len;
2935                                                 if ( $s->cur_is_redirect || $s->cur_namespace != 0 || $length < $threshold ) {
2936                                                         $colours[$pdbk] = 1;
2937                                                 } else {
2938                                                         $colours[$pdbk] = 2;
2939                                                 }
2940                                         } else {
2941                                                 $colours[$pdbk] = 1;
2942                                         }
2943                                 }
2944                         }
2945                         wfProfileOut( $fname.'-check' );
2946
2947                         # Construct search and replace arrays
2948                         wfProfileIn( $fname.'-construct' );
2949                         $outputReplace = array();
2950                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
2951                                 $pdbk = $pdbks[$key];
2952                                 $searchkey = '<!--LINK '.$key.'-->';
2953                                 $title = $wgLinkHolders['titles'][$key];
2954                                 if ( empty( $colours[$pdbk] ) ) {
2955                                         $wgLinkCache->addBadLink( $pdbk );
2956                                         $colours[$pdbk] = 0;
2957                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
2958                                                                         $wgLinkHolders['texts'][$key],
2959                                                                         $wgLinkHolders['queries'][$key] );
2960                                 } elseif ( $colours[$pdbk] == 1 ) {
2961                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
2962                                                                         $wgLinkHolders['texts'][$key],
2963                                                                         $wgLinkHolders['queries'][$key] );
2964                                 } elseif ( $colours[$pdbk] == 2 ) {
2965                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
2966                                                                         $wgLinkHolders['texts'][$key],
2967                                                                         $wgLinkHolders['queries'][$key] );
2968                                 }
2969                         }
2970                         wfProfileOut( $fname.'-construct' );
2971
2972                         # Do the thing
2973                         wfProfileIn( $fname.'-replace' );
2974
2975                         $text = preg_replace_callback(
2976                                 '/(<!--LINK .*?-->)/',
2977                                 "outputReplaceMatches",
2978                                 $text);
2979                         wfProfileOut( $fname.'-replace' );
2980                 }
2981
2982                 if ( !empty( $wgInterwikiLinkHolders ) ) {
2983                         wfProfileIn( $fname.'-interwiki' );
2984                         $outputReplace = $wgInterwikiLinkHolders;
2985                         $text = preg_replace_callback(
2986                                 '/<!--IWLINK (.*?)-->/',
2987                                 "outputReplaceMatches",
2988                                 $text);
2989                         wfProfileOut( $fname.'-interwiki' );
2990                 }
2991
2992                 wfProfileOut( $fname );
2993                 return $colours;
2994         }
2995         /**
2996          * Renders an image gallery from a text with one line per image.
2997          * text labels may be given by using |-style alternative text. E.g.
2998          *   Image:one.jpg|The number "1"
2999          *   Image:tree.jpg|A tree
3000          * given as text will return the HTML of a gallery with two images,
3001          * labeled 'The number "1"' and
3002          * 'A tree'.
3003          */
3004         function renderImageGallery( $text ) {
3005                 global $wgLinkCache;
3006                 $ig = new ImageGallery();
3007                 $ig->setShowBytes( false );
3008                 $ig->setShowFilename( false );
3009                 $lines = explode( "\n", $text );
3010                 foreach ( $lines as $line ) {
3011                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3012                         # Skip empty lines
3013                         if ( count( $matches ) == 0 ) {
3014                                 continue;
3015                         }
3016                         $nt = Title::newFromURL( $matches[1] );
3017                         if ( isset( $matches[3] ) ) {
3018                                 $label = $matches[3];
3019                         } else {
3020                                 $label = '';
3021                         }
3022                         $ig->add( Image::newFromTitle( $nt ), $label );
3023                         $wgLinkCache->addImageLinkObj( $nt );
3024                 }
3025                 return $ig->toHTML();
3026         }
3027 }
3028
3029 /**
3030  * @todo document
3031  * @package MediaWiki
3032  */
3033 class ParserOutput
3034 {
3035         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3036         var $mCacheTime; # Used in ParserCache
3037
3038         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3039                 $containsOldMagic = false )
3040         {
3041                 $this->mText = $text;
3042                 $this->mLanguageLinks = $languageLinks;
3043                 $this->mCategoryLinks = $categoryLinks;
3044                 $this->mContainsOldMagic = $containsOldMagic;
3045                 $this->mCacheTime = '';
3046         }
3047
3048         function getText() { return $this->mText; }
3049         function getLanguageLinks() { return $this->mLanguageLinks; }
3050         function getCategoryLinks() { return $this->mCategoryLinks; }
3051         function getCacheTime() { return $this->mCacheTime; }
3052         function containsOldMagic() { return $this->mContainsOldMagic; }
3053         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
3054         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
3055         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
3056         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3057         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
3058
3059         function merge( $other ) {
3060                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3061                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3062                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3063         }
3064
3065 }
3066
3067 /**
3068  * Set options of the Parser
3069  * @todo document
3070  * @package MediaWiki
3071  */
3072 class ParserOptions
3073 {
3074         # All variables are private
3075         var $mUseTeX;                    # Use texvc to expand <math> tags
3076         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3077         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3078         var $mAllowExternalImages;       # Allow external images inline
3079         var $mSkin;                      # Reference to the preferred skin
3080         var $mDateFormat;                # Date format index
3081         var $mEditSection;               # Create "edit section" links
3082         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3083         var $mNumberHeadings;            # Automatically number headings
3084         var $mShowToc;                   # Show table of contents
3085
3086         function getUseTeX()                        { return $this->mUseTeX; }
3087         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3088         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3089         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3090         function getSkin()                          { return $this->mSkin; }
3091         function getDateFormat()                    { return $this->mDateFormat; }
3092         function getEditSection()                   { return $this->mEditSection; }
3093         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3094         function getNumberHeadings()                { return $this->mNumberHeadings; }
3095         function getShowToc()                       { return $this->mShowToc; }
3096
3097         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3098         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3099         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3100         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3101         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3102         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3103         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3104         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3105         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3106
3107         function setSkin( &$x ) { $this->mSkin =& $x; }
3108
3109         # Get parser options
3110         /* static */ function newFromUser( &$user ) {
3111                 $popts = new ParserOptions;
3112                 $popts->initialiseFromUser( $user );
3113                 return $popts;
3114         }
3115
3116         # Get user options
3117         function initialiseFromUser( &$userInput ) {
3118                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3119
3120                 $fname = 'ParserOptions::initialiseFromUser';
3121                 wfProfileIn( $fname );
3122                 if ( !$userInput ) {
3123                         $user = new User;
3124                         $user->setLoaded( true );
3125                 } else {
3126                         $user =& $userInput;
3127                 }
3128
3129                 $this->mUseTeX = $wgUseTeX;
3130                 $this->mUseDynamicDates = $wgUseDynamicDates;
3131                 $this->mInterwikiMagic = $wgInterwikiMagic;
3132                 $this->mAllowExternalImages = $wgAllowExternalImages;
3133                 wfProfileIn( $fname.'-skin' );
3134                 $this->mSkin =& $user->getSkin();
3135                 wfProfileOut( $fname.'-skin' );
3136                 $this->mDateFormat = $user->getOption( 'date' );
3137                 $this->mEditSection = $user->getOption( 'editsection' );
3138                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3139                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3140                 $this->mShowToc = $user->getOption( 'showtoc' );
3141                 wfProfileOut( $fname );
3142         }
3143
3144
3145 }
3146
3147 /**
3148  * Callback function used by Parser::replaceLinkHolders()
3149  * to substitute link placeholders.
3150  */
3151 function &outputReplaceMatches($matches) {
3152         global $outputReplace;
3153         return $outputReplace[$matches[1]];
3154 }
3155
3156 /**
3157  * Return the total number of articles
3158  */
3159 function wfNumberOfArticles() {
3160         global $wgNumberOfArticles;
3161
3162         wfLoadSiteStats();
3163         return $wgNumberOfArticles;
3164 }
3165
3166 /**
3167  * Get various statistics from the database
3168  * @private
3169  */
3170 function wfLoadSiteStats() {
3171         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3172         $fname = 'wfLoadSiteStats';
3173
3174         if ( -1 != $wgNumberOfArticles ) return;
3175         $dbr =& wfGetDB( DB_SLAVE );
3176         $s = $dbr->selectRow( 'site_stats',
3177                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3178                 array( 'ss_row_id' => 1 ), $fname
3179         );
3180
3181         if ( $s === false ) {
3182                 return;
3183         } else {
3184                 $wgTotalViews = $s->ss_total_views;
3185                 $wgTotalEdits = $s->ss_total_edits;
3186                 $wgNumberOfArticles = $s->ss_good_articles;
3187         }
3188 }
3189
3190 function wfEscapeHTMLTagsOnly( $in ) {
3191         return str_replace(
3192                 array( '"', '>', '<' ),
3193                 array( '&quot;', '&gt;', '&lt;' ),
3194                 $in );
3195 }
3196
3197 ?>