includes/Parser.php

   1 <?php
   2
   3 /**
   4  * File for Parser and related classes
   5  *
   6  * @package MediaWiki
   7  */
   8
   9 /**
  10  * Update this version number when the ParserOutput format
  11  * changes in an incompatible way, so the parser cache
  12  * can automatically discard old data.
  13  */
  14 define( 'MW_PARSER_VERSION', '1.4.0' );
  15
  16 /**
  17  * Variable substitution O(N^2) attack
  18  *
  19  * Without countermeasures, it would be possible to attack the parser by saving
  20  * a page filled with a large number of inclusions of large pages. The size of
  21  * the generated page would be proportional to the square of the input size.
  22  * Hence, we limit the number of inclusions of any given page, thus bringing any
  23  * attack back to O(N).
  24  */
  25
  26 define( 'MAX_INCLUDE_REPEAT', 100 );
  27 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  28
  29 define( 'RLH_FOR_UPDATE', 1 );
  30
  31 # Allowed values for $mOutputType
  32 define( 'OT_HTML', 1 );
  33 define( 'OT_WIKI', 2 );
  34 define( 'OT_MSG' , 3 );
  35
  36 # string parameter for extractTags which will cause it
  37 # to strip HTML comments in addition to regular
  38 # <XML>-style tags. This should not be anything we
  39 # may want to use in wikisyntax
  40 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  41
  42 # prefix for escaping, used in two functions at least
  43 define( 'UNIQ_PREFIX', 'NaodW29');
  44
  45 # Constants needed for external link processing
  46 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  47 define( 'HTTP_PROTOCOLS', 'http|https' );
  48 # Everything except bracket, space, or control characters
  49 define( 'EXT_LINK_URL_CLASS', '[^]<>"\\x00-\\x20\\x7F]' );
  50 # Including space
  51 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  52 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  53 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  54 define( 'EXT_LINK_BRACKETED',  '/\[(\b('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  55 define( 'EXT_IMAGE_REGEX',
  56         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  57         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  58         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  59 );
  60
  61 /**
  62  * PHP Parser
  63  *
  64  * Processes wiki markup
  65  *
  66  * <pre>
  67  * There are three main entry points into the Parser class:
  68  * parse()
  69  *   produces HTML output
  70  * preSaveTransform().
  71  *   produces altered wiki markup.
  72  * transformMsg()
  73  *   performs brace substitution on MediaWiki messages
  74  *
  75  * Globals used:
  76  *    objects:   $wgLang, $wgDateFormatter, $wgLinkCache
  77  *
  78  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  79  *
  80  * settings:
  81  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  82  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  83  *  $wgLocaltimezone
  84  *
  85  *  * only within ParserOptions
  86  * </pre>
  87  *
  88  * @package MediaWiki
  89  */
  90 class Parser
  91 {
  92         /**#@+
  93          * @access private
  94          */
  95         # Persistent:
  96         var $mTagHooks;
  97
  98         # Cleared with clearState():
  99         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
 100         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
 101
 102         # Temporary:
 103         var $mOptions, $mTitle, $mOutputType,
 104             $mTemplates,        // cache of already loaded templates, avoids
 105                                 // multiple SQL queries for the same string
 106             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 107                                 // in this path. Used for loop detection.
 108
 109         /**#@-*/
 110
 111         /**
 112          * Constructor
 113          *
 114          * @access public
 115          */
 116         function Parser() {
 117                 $this->mTemplates = array();
 118                 $this->mTemplatePath = array();
 119                 $this->mTagHooks = array();
 120                 $this->clearState();
 121         }
 122
 123         /**
 124          * Clear Parser state
 125          *
 126          * @access private
 127          */
 128         function clearState() {
 129                 $this->mOutput = new ParserOutput;
 130                 $this->mAutonumber = 0;
 131                 $this->mLastSection = "";
 132                 $this->mDTopen = false;
 133                 $this->mVariables = false;
 134                 $this->mIncludeCount = array();
 135                 $this->mStripState = array();
 136                 $this->mArgStack = array();
 137                 $this->mInPre = false;
 138         }
 139
 140         /**
 141          * First pass--just handle <nowiki> sections, pass the rest off
 142          * to internalParse() which does all the real work.
 143          *
 144          * @access private
 145          * @return ParserOutput a ParserOutput
 146          */
 147         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 148                 global $wgUseTidy, $wgContLang;
 149                 $fname = 'Parser::parse';
 150                 wfProfileIn( $fname );
 151
 152                 if ( $clearState ) {
 153                         $this->clearState();
 154                 }
 155
 156                 $this->mOptions = $options;
 157                 $this->mTitle =& $title;
 158                 $this->mOutputType = OT_HTML;
 159
 160                 $stripState = NULL;
 161                 global $fnord; $fnord = 1;
 162                 //$text = $this->strip( $text, $this->mStripState );
 163                 // VOODOO MAGIC FIX! Sometimes the above segfaults in PHP5.
 164                 $x =& $this->mStripState;
 165                 $text = $this->strip( $text, $x );
 166
 167                 $text = $this->internalParse( $text, $linestart );
 168                 $text = $this->unstrip( $text, $this->mStripState );
 169                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 170                 if(!$wgUseTidy) {
 171                         $fixtags = array(
 172                                 # french spaces, last one Guillemet-left
 173                                 # only if there is something before the space
 174                                 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 175                                 # french spaces, Guillemet-right
 176                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 177                                 '/<hr *>/i' => '<hr />',
 178                                 '/<br *>/i' => '<br />',
 179                                 '/<center *>/i' => '<div class="center">',
 180                                 '/<\\/center *>/i' => '</div>',
 181                                 # Clean up spare ampersands; note that we probably ought to be
 182                                 # more careful about named entities.
 183                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 184                         );
 185                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 186                 } else {
 187                         $fixtags = array(
 188                                 # french spaces, last one Guillemet-left
 189                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 190                                 # french spaces, Guillemet-right
 191                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 192                                 '/<center *>/i' => '<div class="center">',
 193                                 '/<\\/center *>/i' => '</div>'
 194                         );
 195                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 196                 }
 197                 # only once and last
 198                 $text = $this->doBlockLevels( $text, $linestart );
 199
 200                 $this->replaceLinkHolders( $text );
 201                 $text = $wgContLang->convert($text);
 202
 203                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 204                 global $wgUseTidy;
 205                 if ($wgUseTidy) {
 206                         $text = Parser::tidy($text);
 207                 }
 208
 209                 $this->mOutput->setText( $text );
 210                 wfProfileOut( $fname );
 211                 return $this->mOutput;
 212         }
 213
 214         /**
 215          * Get a random string
 216          *
 217          * @access private
 218          * @static
 219          */
 220         function getRandomString() {
 221                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 222         }
 223
 224         /**
 225          * Replaces all occurrences of <$tag>content</$tag> in the text
 226          * with a random marker and returns the new text. the output parameter
 227          * $content will be an associative array filled with data on the form
 228          * $unique_marker => content.
 229          *
 230          * If $content is already set, the additional entries will be appended
 231          * If $tag is set to STRIP_COMMENTS, the function will extract
 232          * <!-- HTML comments -->
 233          *
 234          * @access private
 235          * @static
 236          */
 237         function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 238                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 239                 if ( !$content ) {
 240                         $content = array( );
 241                 }
 242                 $n = 1;
 243                 $stripped = '';
 244
 245                 while ( '' != $text ) {
 246                         if($tag==STRIP_COMMENTS) {
 247                                 $p = preg_split( '/<!--/i', $text, 2 );
 248                         } else {
 249                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 250                         }
 251                         $stripped .= $p[0];
 252                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 253                                 $text = '';
 254                         } else {
 255                                 if($tag==STRIP_COMMENTS) {
 256                                         $q = preg_split( '/-->/i', $p[1], 2 );
 257                                 } else {
 258                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 259                                 }
 260                                 $marker = $rnd . sprintf('%08X', $n++);
 261                                 $content[$marker] = $q[0];
 262                                 $stripped .= $marker;
 263                                 $text = $q[1];
 264                         }
 265                 }
 266                 return $stripped;
 267         }
 268
 269         /**
 270          * Strips and renders nowiki, pre, math, hiero
 271          * If $render is set, performs necessary rendering operations on plugins
 272          * Returns the text, and fills an array with data needed in unstrip()
 273          * If the $state is already a valid strip state, it adds to the state
 274          *
 275          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 276          *  will be stripped in addition to other tags. This is important
 277          *  for section editing, where these comments cause confusion when
 278          *  counting the sections in the wikisource
 279          *
 280          * @access private
 281          */
 282         function strip( $text, &$state, $stripcomments = false ) {
 283                 $render = ($this->mOutputType == OT_HTML);
 284                 $html_content = array();
 285                 $nowiki_content = array();
 286                 $math_content = array();
 287                 $pre_content = array();
 288                 $comment_content = array();
 289                 $ext_content = array();
 290                 $gallery_content = array();
 291
 292                 # Replace any instances of the placeholders
 293                 $uniq_prefix = UNIQ_PREFIX;
 294                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 295
 296                 # html
 297                 global $wgRawHtml, $wgWhitelistEdit;
 298                 if( $wgRawHtml && $wgWhitelistEdit ) {
 299                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 300                         foreach( $html_content as $marker => $content ) {
 301                                 if ($render ) {
 302                                         # Raw and unchecked for validity.
 303                                         $html_content[$marker] = $content;
 304                                 } else {
 305                                         $html_content[$marker] = '<html>'.$content.'</html>';
 306                                 }
 307                         }
 308                 }
 309
 310                 # nowiki
 311                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 312                 foreach( $nowiki_content as $marker => $content ) {
 313                         if( $render ){
 314                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 315                         } else {
 316                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 317                         }
 318                 }
 319
 320                 # math
 321                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 322                 foreach( $math_content as $marker => $content ){
 323                         if( $render ) {
 324                                 if( $this->mOptions->getUseTeX() ) {
 325                                         $math_content[$marker] = renderMath( $content );
 326                                 } else {
 327                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 328                                 }
 329                         } else {
 330                                 $math_content[$marker] = '<math>'.$content.'</math>';
 331                         }
 332                 }
 333
 334                 # pre
 335                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 336                 foreach( $pre_content as $marker => $content ){
 337                         if( $render ){
 338                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 339                         } else {
 340                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 341                         }
 342                 }
 343
 344                 # gallery
 345                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 346                 foreach( $gallery_content as $marker => $content ) {
 347                         require_once( 'ImageGallery.php' );
 348                         if ( $render ) {
 349                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 350                         } else {
 351                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 352                         }
 353                 }
 354
 355                 # Comments
 356                 if($stripcomments) {
 357                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 358                         foreach( $comment_content as $marker => $content ){
 359                                 $comment_content[$marker] = '<!--'.$content.'-->';
 360                         }
 361                 }
 362
 363                 # Extensions
 364                 foreach ( $this->mTagHooks as $tag => $callback ) {
 365                         $ext_contents[$tag] = array();
 366                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 367                         foreach( $ext_content[$tag] as $marker => $content ) {
 368                                 if ( $render ) {
 369                                         $ext_content[$tag][$marker] = $callback( $content );
 370                                 } else {
 371                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 372                                 }
 373                         }
 374                 }
 375
 376                 # Merge state with the pre-existing state, if there is one
 377                 if ( $state ) {
 378                         $state['html'] = $state['html'] + $html_content;
 379                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 380                         $state['math'] = $state['math'] + $math_content;
 381                         $state['pre'] = $state['pre'] + $pre_content;
 382                         $state['comment'] = $state['comment'] + $comment_content;
 383                         $state['gallery'] = $state['gallery'] + $gallery_content;
 384
 385                         foreach( $ext_content as $tag => $array ) {
 386                                 if ( array_key_exists( $tag, $state ) ) {
 387                                         $state[$tag] = $state[$tag] + $array;
 388                                 }
 389                         }
 390                 } else {
 391                         $state = array(
 392                           'html' => $html_content,
 393                           'nowiki' => $nowiki_content,
 394                           'math' => $math_content,
 395                           'pre' => $pre_content,
 396                           'comment' => $comment_content,
 397                           'gallery' => $gallery_content,
 398                         ) + $ext_content;
 399                 }
 400                 return $text;
 401         }
 402
 403         /**
 404          * restores pre, math, and hiero removed by strip()
 405          *
 406          * always call unstripNoWiki() after this one
 407          * @access private
 408          */
 409         function unstrip( $text, &$state ) {
 410                 # Must expand in reverse order, otherwise nested tags will be corrupted
 411                 $contentDict = end( $state );
 412                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 413                         if( key($state) != 'nowiki' && key($state) != 'html') {
 414                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 415                                         $text = str_replace( key( $contentDict ), $content, $text );
 416                                 }
 417                         }
 418                 }
 419
 420                 return $text;
 421         }
 422
 423         /**
 424          * always call this after unstrip() to preserve the order
 425          *
 426          * @access private
 427          */
 428         function unstripNoWiki( $text, &$state ) {
 429                 # Must expand in reverse order, otherwise nested tags will be corrupted
 430                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 431                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 432                 }
 433
 434                 global $wgRawHtml;
 435                 if ($wgRawHtml) {
 436                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 437                                 $text = str_replace( key( $state['html'] ), $content, $text );
 438                         }
 439                 }
 440
 441                 return $text;
 442         }
 443
 444         /**
 445          * Add an item to the strip state
 446          * Returns the unique tag which must be inserted into the stripped text
 447          * The tag will be replaced with the original text in unstrip()
 448          *
 449          * @access private
 450          */
 451         function insertStripItem( $text, &$state ) {
 452                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 453                 if ( !$state ) {
 454                         $state = array(
 455                           'html' => array(),
 456                           'nowiki' => array(),
 457                           'math' => array(),
 458                           'pre' => array()
 459                         );
 460                 }
 461                 $state['item'][$rnd] = $text;
 462                 return $rnd;
 463         }
 464
 465         /**
 466          * Return allowed HTML attributes
 467          *
 468          * @access private
 469          */
 470         function getHTMLattrs () {
 471                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 472                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 473                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 474                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 475                                 /* FONT */ 'type', 'start', 'value', 'compact',
 476                                 /* For various lists, mostly deprecated but safe */
 477                                 'summary', 'width', 'border', 'frame', 'rules',
 478                                 'cellspacing', 'cellpadding', 'valign', 'char',
 479                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 480                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 481                                 'id', 'class', 'name', 'style' /* For CSS */
 482                                 );
 483                 return $htmlattrs ;
 484         }
 485
 486         /**
 487          * Remove non approved attributes and javascript in css
 488          *
 489          * @access private
 490          */
 491         function fixTagAttributes ( $t ) {
 492                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 493                 $htmlattrs = $this->getHTMLattrs() ;
 494
 495                 # Strip non-approved attributes from the tag
 496                 $t = preg_replace(
 497                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 498                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 499                         $t);
 500
 501                 $t = str_replace ( '<></>' , '' , $t ) ; # This should fix bug 980557
 502
 503                 # Strip javascript "expression" from stylesheets. Brute force approach:
 504                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 505
 506                 if( preg_match(
 507                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 508                         wfMungeToUtf8( $t ) ) )
 509                 {
 510                         $t='';
 511                 }
 512
 513                 return trim ( $t ) ;
 514         }
 515
 516         /**
 517          * interface with html tidy, used if $wgUseTidy = true
 518          *
 519          * @access public
 520          * @static
 521          */
 522         function tidy ( $text ) {
 523                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 524                 global $wgInputEncoding, $wgOutputEncoding;
 525                 $fname = 'Parser::tidy';
 526                 wfProfileIn( $fname );
 527
 528                 $cleansource = '';
 529                 $opts = '';
 530                 switch(strtoupper($wgOutputEncoding)) {
 531                         case 'ISO-8859-1':
 532                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 533                                 break;
 534                         case 'UTF-8':
 535                                 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 536                                 break;
 537                         default:
 538                                 $opts .= ' -raw';
 539                         }
 540
 541                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 542 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 543 '<head><title>test</title></head><body>'.$text.'</body></html>';
 544                 $descriptorspec = array(
 545                         0 => array('pipe', 'r'),
 546                         1 => array('pipe', 'w'),
 547                         2 => array('file', '/dev/null', 'a')
 548                 );
 549                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 550                 if (is_resource($process)) {
 551                         fwrite($pipes[0], $wrappedtext);
 552                         fclose($pipes[0]);
 553                         while (!feof($pipes[1])) {
 554                                 $cleansource .= fgets($pipes[1], 1024);
 555                         }
 556                         fclose($pipes[1]);
 557                         $return_value = proc_close($process);
 558                 }
 559
 560                 wfProfileOut( $fname );
 561
 562                 if( $cleansource == '' && $text != '') {
 563                         wfDebug( "Tidy error detected!\n" );
 564                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 565                 } else {
 566                         return $cleansource;
 567                 }
 568         }
 569
 570         /**
 571          * parse the wiki syntax used to render tables
 572          *
 573          * @access private
 574          */
 575         function doTableStuff ( $t ) {
 576                 $fname = 'Parser::doTableStuff';
 577                 wfProfileIn( $fname );
 578
 579                 $t = explode ( "\n" , $t ) ;
 580                 $td = array () ; # Is currently a td tag open?
 581                 $ltd = array () ; # Was it TD or TH?
 582                 $tr = array () ; # Is currently a tr tag open?
 583                 $ltr = array () ; # tr attributes
 584                 $indent_level = 0; # indent level of the table
 585                 foreach ( $t AS $k => $x )
 586                 {
 587                         $x = trim ( $x ) ;
 588                         $fc = substr ( $x , 0 , 1 ) ;
 589                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 590                                 $indent_level = strlen( $matches[1] );
 591                                 $t[$k] = "\n" .
 592                                         str_repeat( '<dl><dd>', $indent_level ) .
 593                                         '<table ' . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 594                                 array_push ( $td , false ) ;
 595                                 array_push ( $ltd , '' ) ;
 596                                 array_push ( $tr , false ) ;
 597                                 array_push ( $ltr , '' ) ;
 598                         }
 599                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 600                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 601                                 $z = "</table>\n" ;
 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 . str_repeat( '</dd></dl>', $indent_level );
 607                         }
 608                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 609                                 $x = substr ( $x , 1 ) ;
 610                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 611                                 $z = '' ;
 612                                 $l = array_pop ( $ltd ) ;
 613                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 614                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 615                                 array_pop ( $ltr ) ;
 616                                 $t[$k] = $z ;
 617                                 array_push ( $tr , false ) ;
 618                                 array_push ( $td , false ) ;
 619                                 array_push ( $ltd , '' ) ;
 620                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 621                         }
 622                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 623                                 # $x is a table row
 624                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 625                                         $fc = '+' ;
 626                                         $x = substr ( $x , 1 ) ;
 627                                 }
 628                                 $after = substr ( $x , 1 ) ;
 629                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 630                                 $after = explode ( '||' , $after ) ;
 631                                 $t[$k] = '' ;
 632
 633                                 # Loop through each table cell
 634                                 foreach ( $after AS $theline )
 635                                 {
 636                                         $z = '' ;
 637                                         if ( $fc != '+' )
 638                                         {
 639                                                 $tra = array_pop ( $ltr ) ;
 640                                                 if ( !array_pop ( $tr ) ) $z = '<tr '.$tra.">\n" ;
 641                                                 array_push ( $tr , true ) ;
 642                                                 array_push ( $ltr , '' ) ;
 643                                         }
 644
 645                                         $l = array_pop ( $ltd ) ;
 646                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 647                                         if ( $fc == '|' ) $l = 'td' ;
 648                                         else if ( $fc == '!' ) $l = 'th' ;
 649                                         else if ( $fc == '+' ) $l = 'caption' ;
 650                                         else $l = '' ;
 651                                         array_push ( $ltd , $l ) ;
 652
 653                                         # Cell parameters
 654                                         $y = explode ( '|' , $theline , 2 ) ;
 655                                         # Note that a '|' inside an invalid link should not
 656                                         # be mistaken as delimiting cell parameters
 657                                         if ( strpos( $y[0], '[[' ) !== false ) {
 658                                                 $y = array ($theline);
 659                                         }
 660                                         if ( count ( $y ) == 1 )
 661                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 662                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 663                                         $t[$k] .= $y ;
 664                                         array_push ( $td , true ) ;
 665                                 }
 666                         }
 667                 }
 668
 669                 # Closing open td, tr && table
 670                 while ( count ( $td ) > 0 )
 671                 {
 672                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 673                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 674                         $t[] = '</table>' ;
 675                 }
 676
 677                 $t = implode ( "\n" , $t ) ;
 678                 #               $t = $this->removeHTMLtags( $t );
 679                 wfProfileOut( $fname );
 680                 return $t ;
 681         }
 682
 683         /**
 684          * Helper function for parse() that transforms wiki markup into
 685          * HTML. Only called for $mOutputType == OT_HTML.
 686          *
 687          * @access private
 688          */
 689         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 690                 global $wgContLang;
 691
 692                 $fname = 'Parser::internalParse';
 693                 wfProfileIn( $fname );
 694
 695                 $text = $this->removeHTMLtags( $text );
 696                 $text = $this->replaceVariables( $text, $args );
 697
 698                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 699
 700                 $text = $this->doHeadings( $text );
 701                 if($this->mOptions->getUseDynamicDates()) {
 702                         global $wgDateFormatter;
 703                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 704                 }
 705                 $text = $this->doAllQuotes( $text );
 706                 $text = $this->replaceInternalLinks( $text );
 707                 $text = $this->replaceExternalLinks( $text );
 708
 709                 # replaceInternalLinks may sometimes leave behind
 710                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 711                 $text = str_replace("http-noparse://","http://",$text);
 712
 713                 $text = $this->doMagicLinks( $text );
 714                 $text = $this->doTableStuff( $text );
 715                 $text = $this->formatHeadings( $text, $isMain );
 716
 717                 wfProfileOut( $fname );
 718                 return $text;
 719         }
 720
 721         /**
 722          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 723          * magic external links.
 724          *
 725          * @access private
 726          */
 727         function &doMagicLinks( &$text ) {
 728                 global $wgUseGeoMode;
 729                 $text = $this->magicISBN( $text );
 730                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 731                         $text = $this->magicGEO( $text );
 732                 }
 733                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 734                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 735                 return $text;
 736         }
 737
 738         /**
 739          * Parse ^^ tokens and return html
 740          *
 741          * @access private
 742          */
 743         function doExponent( $text ) {
 744                 $fname = 'Parser::doExponent';
 745                 wfProfileIn( $fname );
 746                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 747                 wfProfileOut( $fname );
 748                 return $text;
 749         }
 750
 751         /**
 752          * Parse headers and return html
 753          *
 754          * @access private
 755          */
 756         function doHeadings( $text ) {
 757                 $fname = 'Parser::doHeadings';
 758                 wfProfileIn( $fname );
 759                 for ( $i = 6; $i >= 1; --$i ) {
 760                         $h = substr( '======', 0, $i );
 761                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 762                           "<h{$i}>\\1</h{$i}>\\2", $text );
 763                 }
 764                 wfProfileOut( $fname );
 765                 return $text;
 766         }
 767
 768         /**
 769          * Replace single quotes with HTML markup
 770          * @access private
 771          * @return string the altered text
 772          */
 773         function doAllQuotes( $text ) {
 774                 $fname = 'Parser::doAllQuotes';
 775                 wfProfileIn( $fname );
 776                 $outtext = '';
 777                 $lines = explode( "\n", $text );
 778                 foreach ( $lines as $line ) {
 779                         $outtext .= $this->doQuotes ( $line ) . "\n";
 780                 }
 781                 $outtext = substr($outtext, 0,-1);
 782                 wfProfileOut( $fname );
 783                 return $outtext;
 784         }
 785
 786         /**
 787          * Helper function for doAllQuotes()
 788          * @access private
 789          */
 790         function doQuotes( $text ) {
 791                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 792                 if ( count( $arr ) == 1 )
 793                         return $text;
 794                 else
 795                 {
 796                         # First, do some preliminary work. This may shift some apostrophes from
 797                         # being mark-up to being text. It also counts the number of occurrences
 798                         # of bold and italics mark-ups.
 799                         $i = 0;
 800                         $numbold = 0;
 801                         $numitalics = 0;
 802                         foreach ( $arr as $r )
 803                         {
 804                                 if ( ( $i % 2 ) == 1 )
 805                                 {
 806                                         # If there are ever four apostrophes, assume the first is supposed to
 807                                         # be text, and the remaining three constitute mark-up for bold text.
 808                                         if ( strlen( $arr[$i] ) == 4 )
 809                                         {
 810                                                 $arr[$i-1] .= "'";
 811                                                 $arr[$i] = "'''";
 812                                         }
 813                                         # If there are more than 5 apostrophes in a row, assume they're all
 814                                         # text except for the last 5.
 815                                         else if ( strlen( $arr[$i] ) > 5 )
 816                                         {
 817                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 818                                                 $arr[$i] = "'''''";
 819                                         }
 820                                         # Count the number of occurrences of bold and italics mark-ups.
 821                                         # We are not counting sequences of five apostrophes.
 822                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 823                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 824                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 825                                 }
 826                                 $i++;
 827                         }
 828
 829                         # If there is an odd number of both bold and italics, it is likely
 830                         # that one of the bold ones was meant to be an apostrophe followed
 831                         # by italics. Which one we cannot know for certain, but it is more
 832                         # likely to be one that has a single-letter word before it.
 833                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 834                         {
 835                                 $i = 0;
 836                                 $firstsingleletterword = -1;
 837                                 $firstmultiletterword = -1;
 838                                 $firstspace = -1;
 839                                 foreach ( $arr as $r )
 840                                 {
 841                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 842                                         {
 843                                                 $x1 = substr ($arr[$i-1], -1);
 844                                                 $x2 = substr ($arr[$i-1], -2, 1);
 845                                                 if ($x1 == ' ') {
 846                                                         if ($firstspace == -1) $firstspace = $i;
 847                                                 } else if ($x2 == ' ') {
 848                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 849                                                 } else {
 850                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 851                                                 }
 852                                         }
 853                                         $i++;
 854                                 }
 855
 856                                 # If there is a single-letter word, use it!
 857                                 if ($firstsingleletterword > -1)
 858                                 {
 859                                         $arr [ $firstsingleletterword ] = "''";
 860                                         $arr [ $firstsingleletterword-1 ] .= "'";
 861                                 }
 862                                 # If not, but there's a multi-letter word, use that one.
 863                                 else if ($firstmultiletterword > -1)
 864                                 {
 865                                         $arr [ $firstmultiletterword ] = "''";
 866                                         $arr [ $firstmultiletterword-1 ] .= "'";
 867                                 }
 868                                 # ... otherwise use the first one that has neither.
 869                                 # (notice that it is possible for all three to be -1 if, for example,
 870                                 # there is only one pentuple-apostrophe in the line)
 871                                 else if ($firstspace > -1)
 872                                 {
 873                                         $arr [ $firstspace ] = "''";
 874                                         $arr [ $firstspace-1 ] .= "'";
 875                                 }
 876                         }
 877
 878                         # Now let's actually convert our apostrophic mush to HTML!
 879                         $output = '';
 880                         $buffer = '';
 881                         $state = '';
 882                         $i = 0;
 883                         foreach ($arr as $r)
 884                         {
 885                                 if (($i % 2) == 0)
 886                                 {
 887                                         if ($state == 'both')
 888                                                 $buffer .= $r;
 889                                         else
 890                                                 $output .= $r;
 891                                 }
 892                                 else
 893                                 {
 894                                         if (strlen ($r) == 2)
 895                                         {
 896                                                 if ($state == 'i')
 897                                                 { $output .= '</i>'; $state = ''; }
 898                                                 else if ($state == 'bi')
 899                                                 { $output .= '</i>'; $state = 'b'; }
 900                                                 else if ($state == 'ib')
 901                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 902                                                 else if ($state == 'both')
 903                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 904                                                 else # $state can be 'b' or ''
 905                                                 { $output .= '<i>'; $state .= 'i'; }
 906                                         }
 907                                         else if (strlen ($r) == 3)
 908                                         {
 909                                                 if ($state == 'b')
 910                                                 { $output .= '</b>'; $state = ''; }
 911                                                 else if ($state == 'bi')
 912                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 913                                                 else if ($state == 'ib')
 914                                                 { $output .= '</b>'; $state = 'i'; }
 915                                                 else if ($state == 'both')
 916                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 917                                                 else # $state can be 'i' or ''
 918                                                 { $output .= '<b>'; $state .= 'b'; }
 919                                         }
 920                                         else if (strlen ($r) == 5)
 921                                         {
 922                                                 if ($state == 'b')
 923                                                 { $output .= '</b><i>'; $state = 'i'; }
 924                                                 else if ($state == 'i')
 925                                                 { $output .= '</i><b>'; $state = 'b'; }
 926                                                 else if ($state == 'bi')
 927                                                 { $output .= '</i></b>'; $state = ''; }
 928                                                 else if ($state == 'ib')
 929                                                 { $output .= '</b></i>'; $state = ''; }
 930                                                 else if ($state == 'both')
 931                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 932                                                 else # ($state == '')
 933                                                 { $buffer = ''; $state = 'both'; }
 934                                         }
 935                                 }
 936                                 $i++;
 937                         }
 938                         # Now close all remaining tags.  Notice that the order is important.
 939                         if ($state == 'b' || $state == 'ib')
 940                                 $output .= '</b>';
 941                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 942                                 $output .= '</i>';
 943                         if ($state == 'bi')
 944                                 $output .= '</b>';
 945                         if ($state == 'both')
 946                                 $output .= '<b><i>'.$buffer.'</i></b>';
 947                         return $output;
 948                 }
 949         }
 950
 951         /**
 952          * Replace external links
 953          *
 954          * Note: this is all very hackish and the order of execution matters a lot.
 955          * Make sure to run maintenance/parserTests.php if you change this code.
 956          *
 957          * @access private
 958          */
 959         function replaceExternalLinks( $text ) {
 960                 $fname = 'Parser::replaceExternalLinks';
 961                 wfProfileIn( $fname );
 962
 963                 $sk =& $this->mOptions->getSkin();
 964                 global $wgContLang;
 965                 $linktrail = $wgContLang->linkTrail();
 966
 967                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 968
 969                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 970
 971                 $i = 0;
 972                 while ( $i<count( $bits ) ) {
 973                         $url = $bits[$i++];
 974                         $protocol = $bits[$i++];
 975                         $text = $bits[$i++];
 976                         $trail = $bits[$i++];
 977
 978                         # The characters '<' and '>' (which were escaped by
 979                         # removeHTMLtags()) should not be included in
 980                         # URLs, per RFC 2396.
 981                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
 982                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
 983                                 $url = substr($url, 0, $m2[0][1]);
 984                         }
 985
 986                         # If the link text is an image URL, replace it with an <img> tag
 987                         # This happened by accident in the original parser, but some people used it extensively
 988                         $img = $this->maybeMakeImageLink( $text );
 989                         if ( $img !== false ) {
 990                                 $text = $img;
 991                         }
 992
 993                         $dtrail = '';
 994
 995                         # Set linktype for CSS - if URL==text, link is essentially free
 996                         $linktype = ($text == $url) ? 'free' : 'text';
 997
 998                         # No link text, e.g. [http://domain.tld/some.link]
 999                         if ( $text == '' ) {
1000                                 # Autonumber if allowed
1001                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
1002                                         $text = '[' . ++$this->mAutonumber . ']';
1003                                         $linktype = 'autonumber';
1004                                 } else {
1005                                         # Otherwise just use the URL
1006                                         $text = htmlspecialchars( $url );
1007                                         $linktype = 'free';
1008                                 }
1009                         } else {
1010                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1011                                 # Check for trail
1012                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
1013                                         $dtrail = $m2[1];
1014                                         $trail = $m2[2];
1015                                 }
1016                         }
1017
1018                         # Process the trail (i.e. everything after this link up until start of the next link),
1019                         # replacing any non-bracketed links
1020                         $trail = $this->replaceFreeExternalLinks( $trail );
1021
1022                         # Use the encoded URL
1023                         # This means that users can paste URLs directly into the text
1024                         # Funny characters like &ouml; aren't valid in URLs anyway
1025                         # This was changed in August 2004
1026                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
1027                 }
1028
1029                 wfProfileOut( $fname );
1030                 return $s;
1031         }
1032
1033         /**
1034          * Replace anything that looks like a URL with a link
1035          * @access private
1036          */
1037         function replaceFreeExternalLinks( $text ) {
1038                 $fname = 'Parser::replaceFreeExternalLinks';
1039                 wfProfileIn( $fname );
1040
1041                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1042                 $s = array_shift( $bits );
1043                 $i = 0;
1044
1045                 $sk =& $this->mOptions->getSkin();
1046
1047                 while ( $i < count( $bits ) ){
1048                         $protocol = $bits[$i++];
1049                         $remainder = $bits[$i++];
1050
1051                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1052                                 # Found some characters after the protocol that look promising
1053                                 $url = $protocol . $m[1];
1054                                 $trail = $m[2];
1055
1056                                 # The characters '<' and '>' (which were escaped by
1057                                 # removeHTMLtags()) should not be included in
1058                                 # URLs, per RFC 2396.
1059                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1060                                         $trail = substr($url, $m2[0][1]) . $trail;
1061                                         $url = substr($url, 0, $m2[0][1]);
1062                                 }
1063
1064                                 # Move trailing punctuation to $trail
1065                                 $sep = ',;\.:!?';
1066                                 # If there is no left bracket, then consider right brackets fair game too
1067                                 if ( strpos( $url, '(' ) === false ) {
1068                                         $sep .= ')';
1069                                 }
1070
1071                                 $numSepChars = strspn( strrev( $url ), $sep );
1072                                 if ( $numSepChars ) {
1073                                         $trail = substr( $url, -$numSepChars ) . $trail;
1074                                         $url = substr( $url, 0, -$numSepChars );
1075                                 }
1076
1077                                 # Replace &amp; from obsolete syntax with &.
1078                                 # All HTML entities will be escaped by makeExternalLink()
1079                                 # or maybeMakeImageLink()
1080                                 $url = str_replace( '&amp;', '&', $url );
1081
1082                                 # Is this an external image?
1083                                 $text = $this->maybeMakeImageLink( $url );
1084                                 if ( $text === false ) {
1085                                         # Not an image, make a link
1086                                         $text = $sk->makeExternalLink( $url, $url, true, 'free' );
1087                                 }
1088                                 $s .= $text . $trail;
1089                         } else {
1090                                 $s .= $protocol . $remainder;
1091                         }
1092                 }
1093                 wfProfileOut();
1094                 return $s;
1095         }
1096
1097         /**
1098          * make an image if it's allowed
1099          * @access private
1100          */
1101         function maybeMakeImageLink( $url ) {
1102                 $sk =& $this->mOptions->getSkin();
1103                 $text = false;
1104                 if ( $this->mOptions->getAllowExternalImages() ) {
1105                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1106                                 # Image found
1107                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1108                         }
1109                 }
1110                 return $text;
1111         }
1112
1113         /**
1114          * Process [[ ]] wikilinks
1115          *
1116          * @access private
1117          */
1118
1119         function replaceInternalLinks( $s ) {
1120                 global $wgLang, $wgContLang, $wgLinkCache;
1121                 global $wgDisableLangConversion;
1122                 static $fname = 'Parser::replaceInternalLinks' ;
1123
1124                 wfProfileIn( $fname );
1125
1126                 wfProfileIn( $fname.'-setup' );
1127                 static $tc = FALSE;
1128                 # the % is needed to support urlencoded titles as well
1129                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1130
1131                 $sk =& $this->mOptions->getSkin();
1132                 global $wgUseOldExistenceCheck;
1133                 # "Post-parse link colour check" works only on wiki text since it's now
1134                 # in Parser. Enable it, then disable it when we're done.
1135                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1136
1137                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1138
1139                 #split the entire text string on occurences of [[
1140                 $a = explode( '[[', ' ' . $s );
1141                 #get the first element (all text up to first [[), and remove the space we added
1142                 $s = array_shift( $a );
1143                 $s = substr( $s, 1 );
1144
1145                 # Match a link having the form [[namespace:link|alternate]]trail
1146                 static $e1 = FALSE;
1147                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1148                 # Match cases where there is no "]]", which might still be images
1149                 static $e1_img = FALSE;
1150                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1151                 # Match the end of a line for a word that's not followed by whitespace,
1152                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1153                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1154
1155                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1156
1157                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1158
1159                 if ( $useLinkPrefixExtension ) {
1160                         if ( preg_match( $e2, $s, $m ) ) {
1161                                 $first_prefix = $m[2];
1162                                 $s = $m[1];
1163                         } else {
1164                                 $first_prefix = false;
1165                         }
1166                 } else {
1167                         $prefix = '';
1168                 }
1169
1170                 $selflink = $this->mTitle->getPrefixedText();
1171                 wfProfileOut( $fname.'-setup' );
1172
1173                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1174                 $useSubpages = $this->areSubpagesAllowed();
1175
1176                 # Loop for each link
1177                 for ($k = 0; isset( $a[$k] ); $k++) {
1178                         $line = $a[$k];
1179                         if ( $useLinkPrefixExtension ) {
1180                                 wfProfileIn( $fname.'-prefixhandling' );
1181                                 if ( preg_match( $e2, $s, $m ) ) {
1182                                         $prefix = $m[2];
1183                                         $s = $m[1];
1184                                 } else {
1185                                         $prefix='';
1186                                 }
1187                                 # first link
1188                                 if($first_prefix) {
1189                                         $prefix = $first_prefix;
1190                                         $first_prefix = false;
1191                                 }
1192                                 wfProfileOut( $fname.'-prefixhandling' );
1193                         }
1194
1195                         $might_be_img = false;
1196
1197                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1198                                 $text = $m[2];
1199                                 # fix up urlencoded title texts
1200                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1201                                 $trail = $m[3];
1202                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1203                                 $might_be_img = true;
1204                                 $text = $m[2];
1205                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1206                                 $trail = "";
1207                         } else { # Invalid form; output directly
1208                                 $s .= $prefix . '[[' . $line ;
1209                                 continue;
1210                         }
1211
1212                         # Don't allow internal links to pages containing
1213                         # PROTO: where PROTO is a valid URL protocol; these
1214                         # should be external links.
1215                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1216                                 $s .= $prefix . '[[' . $line ;
1217                                 continue;
1218                         }
1219
1220                         # Make subpage if necessary
1221                         if( $useSubpages ) {
1222                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1223                         } else {
1224                                 $link = $m[1];
1225                         }
1226
1227                         $noforce = (substr($m[1], 0, 1) != ':');
1228                         if (!$noforce) {
1229                                 # Strip off leading ':'
1230                                 $link = substr($link, 1);
1231                         }
1232
1233                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1234                         if( !$nt ) {
1235                                 $s .= $prefix . '[[' . $line;
1236                                 continue;
1237                         }
1238
1239                         #check other language variants of the link
1240                         #if the article does not exist
1241                         if( $checkVariantLink
1242                             && $nt->getArticleID() == 0 ) {
1243                                 $wgContLang->findVariantLink($link, $nt);
1244                         }
1245
1246                         $ns = $nt->getNamespace();
1247                         $iw = $nt->getInterWiki();
1248
1249                         if ($might_be_img) { # if this is actually an invalid link
1250                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1251                                         $found = false;
1252                                         while (isset ($a[$k+1]) ) {
1253                                                 #look at the next 'line' to see if we can close it there
1254                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1255                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1256                                                 # the first ]] closes the inner link, the second the image
1257                                                         $found = true;
1258                                                         $text .= '[[' . $m[1];
1259                                                         $trail = $m[2];
1260                                                         break;
1261                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1262                                                         #if there's exactly one ]] that's fine, we'll keep looking
1263                                                         $text .= '[[' . $m[0];
1264                                                 } else {
1265                                                         #if $next_line is invalid too, we need look no further
1266                                                         $text .= '[[' . $next_line;
1267                                                         break;
1268                                                 }
1269                                         }
1270                                         if ( !$found ) {
1271                                                 # we couldn't find the end of this imageLink, so output it raw
1272                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1273                                                 $text = $this->replaceInternalLinks($text);
1274                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1275                                                 # note: no $trail, because without an end, there *is* no trail
1276                                                 continue;
1277                                         }
1278                                 } else { #it's not an image, so output it raw
1279                                         $s .= $prefix . '[[' . $link . '|' . $text;
1280                                         # note: no $trail, because without an end, there *is* no trail
1281                                         continue;
1282                                 }
1283                         }
1284
1285                         $wasblank = ( '' == $text );
1286                         if( $wasblank ) $text = $link;
1287
1288
1289                         # Link not escaped by : , create the various objects
1290                         if( $noforce ) {
1291
1292                                 # Interwikis
1293                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1294                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1295                                         $tmp = $prefix . $trail ;
1296                                         $s .= (trim($tmp) == '')? '': $tmp;
1297                                         continue;
1298                                 }
1299
1300                                 if ( $ns == NS_IMAGE ) {
1301                                         wfProfileIn( "$fname-image" );
1302
1303                                         # recursively parse links inside the image caption
1304                                         # actually, this will parse them in any other parameters, too,
1305                                         # but it might be hard to fix that, and it doesn't matter ATM
1306                                         $text = $this->replaceExternalLinks($text);
1307                                         $text = $this->replaceInternalLinks($text);
1308
1309                                         # replace the image with a link-holder so that replaceExternalLinks() can't mess with it
1310                                         $s .= $prefix . $this->insertStripItem( $sk->makeImageLinkObj( $nt, $text ), $this->mStripState ) . $trail;
1311                                         $wgLinkCache->addImageLinkObj( $nt );
1312
1313                                         wfProfileOut( "$fname-image" );
1314                                         continue;
1315                                 }
1316
1317                                 if ( $ns == NS_CATEGORY ) {
1318                                         wfProfileIn( "$fname-category" );
1319                                         $t = $nt->getText();
1320
1321                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1322                                         $pPLC=$sk->postParseLinkColour();
1323                                         $sk->postParseLinkColour( false );
1324                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1325                                         $sk->postParseLinkColour( $pPLC );
1326                                         $wgLinkCache->resume();
1327
1328                                         if ( $wasblank ) {
1329                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1330                                                         $sortkey = $this->mTitle->getText();
1331                                                 } else {
1332                                                         $sortkey = $this->mTitle->getPrefixedText();
1333                                                 }
1334                                         } else {
1335                                                 $sortkey = $text;
1336                                         }
1337                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1338                                         $this->mOutput->addCategoryLink( $t );
1339                                         $s .= $prefix . $trail ;
1340
1341                                         wfProfileOut( "$fname-category" );
1342                                         continue;
1343                                 }
1344                         }
1345
1346                         if( ( $nt->getPrefixedText() === $selflink ) &&
1347                             ( $nt->getFragment() === '' ) ) {
1348                                 # Self-links are handled specially; generally de-link and change to bold.
1349                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1350                                 continue;
1351                         }
1352
1353                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1354                         if( $ns == NS_MEDIA ) {
1355                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1356                                 $wgLinkCache->addImageLinkObj( $nt );
1357                                 continue;
1358                         } elseif( $ns == NS_SPECIAL ) {
1359                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1360                                 continue;
1361                         }
1362                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1363                 }
1364                 $sk->postParseLinkColour( $saveParseColour );
1365                 wfProfileOut( $fname );
1366                 return $s;
1367         }
1368
1369         /**
1370          * Return true if subpage links should be expanded on this page.
1371          * @return bool
1372          */
1373         function areSubpagesAllowed() {
1374                 # Some namespaces don't allow subpages
1375                 global $wgNamespacesWithSubpages;
1376                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1377         }
1378
1379         /**
1380          * Handle link to subpage if necessary
1381          * @param string $target the source of the link
1382          * @param string &$text the link text, modified as necessary
1383          * @return string the full name of the link
1384          * @access private
1385          */
1386         function maybeDoSubpageLink($target, &$text) {
1387                 # Valid link forms:
1388                 # Foobar -- normal
1389                 # :Foobar -- override special treatment of prefix (images, language links)
1390                 # /Foobar -- convert to CurrentPage/Foobar
1391                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1392                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1393                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1394
1395                 $fname = 'Parser::maybeDoSubpageLink';
1396                 wfProfileIn( $fname );
1397                 $ret = $target; # default return value is no change
1398
1399                 # Some namespaces don't allow subpages,
1400                 # so only perform processing if subpages are allowed
1401                 if( $this->areSubpagesAllowed() ) {
1402                         # Look at the first character
1403                         if( $target != '' && $target{0} == '/' ) {
1404                                 # / at end means we don't want the slash to be shown
1405                                 if( substr( $target, -1, 1 ) == '/' ) {
1406                                         $target = substr( $target, 1, -1 );
1407                                         $noslash = $target;
1408                                 } else {
1409                                         $noslash = substr( $target, 1 );
1410                                 }
1411
1412                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1413                                 if( '' === $text ) {
1414                                         $text = $target;
1415                                 } # this might be changed for ugliness reasons
1416                         } else {
1417                                 # check for .. subpage backlinks
1418                                 $dotdotcount = 0;
1419                                 $nodotdot = $target;
1420                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1421                                         ++$dotdotcount;
1422                                         $nodotdot = substr( $nodotdot, 3 );
1423                                 }
1424                                 if($dotdotcount > 0) {
1425                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1426                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1427                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1428                                                 # / at the end means don't show full path
1429                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1430                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1431                                                         if( '' === $text ) {
1432                                                                 $text = $nodotdot;
1433                                                         }
1434                                                 }
1435                                                 $nodotdot = trim( $nodotdot );
1436                                                 if( $nodotdot != '' ) {
1437                                                         $ret .= '/' . $nodotdot;
1438                                                 }
1439                                         }
1440                                 }
1441                         }
1442                 }
1443
1444                 wfProfileOut( $fname );
1445                 return $ret;
1446         }
1447
1448         /**#@+
1449          * Used by doBlockLevels()
1450          * @access private
1451          */
1452         /* private */ function closeParagraph() {
1453                 $result = '';
1454                 if ( '' != $this->mLastSection ) {
1455                         $result = '</' . $this->mLastSection  . ">\n";
1456                 }
1457                 $this->mInPre = false;
1458                 $this->mLastSection = '';
1459                 return $result;
1460         }
1461         # getCommon() returns the length of the longest common substring
1462         # of both arguments, starting at the beginning of both.
1463         #
1464         /* private */ function getCommon( $st1, $st2 ) {
1465                 $fl = strlen( $st1 );
1466                 $shorter = strlen( $st2 );
1467                 if ( $fl < $shorter ) { $shorter = $fl; }
1468
1469                 for ( $i = 0; $i < $shorter; ++$i ) {
1470                         if ( $st1{$i} != $st2{$i} ) { break; }
1471                 }
1472                 return $i;
1473         }
1474         # These next three functions open, continue, and close the list
1475         # element appropriate to the prefix character passed into them.
1476         #
1477         /* private */ function openList( $char ) {
1478                 $result = $this->closeParagraph();
1479
1480                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1481                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1482                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1483                 else if ( ';' == $char ) {
1484                         $result .= '<dl><dt>';
1485                         $this->mDTopen = true;
1486                 }
1487                 else { $result = '<!-- ERR 1 -->'; }
1488
1489                 return $result;
1490         }
1491
1492         /* private */ function nextItem( $char ) {
1493                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1494                 else if ( ':' == $char || ';' == $char ) {
1495                         $close = '</dd>';
1496                         if ( $this->mDTopen ) { $close = '</dt>'; }
1497                         if ( ';' == $char ) {
1498                                 $this->mDTopen = true;
1499                                 return $close . '<dt>';
1500                         } else {
1501                                 $this->mDTopen = false;
1502                                 return $close . '<dd>';
1503                         }
1504                 }
1505                 return '<!-- ERR 2 -->';
1506         }
1507
1508         /* private */ function closeList( $char ) {
1509                 if ( '*' == $char ) { $text = '</li></ul>'; }
1510                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1511                 else if ( ':' == $char ) {
1512                         if ( $this->mDTopen ) {
1513                                 $this->mDTopen = false;
1514                                 $text = '</dt></dl>';
1515                         } else {
1516                                 $text = '</dd></dl>';
1517                         }
1518                 }
1519                 else {  return '<!-- ERR 3 -->'; }
1520                 return $text."\n";
1521         }
1522         /**#@-*/
1523
1524         /**
1525          * Make lists from lines starting with ':', '*', '#', etc.
1526          *
1527          * @access private
1528          * @return string the lists rendered as HTML
1529          */
1530         function doBlockLevels( $text, $linestart ) {
1531                 $fname = 'Parser::doBlockLevels';
1532                 wfProfileIn( $fname );
1533
1534                 # Parsing through the text line by line.  The main thing
1535                 # happening here is handling of block-level elements p, pre,
1536                 # and making lists from lines starting with * # : etc.
1537                 #
1538                 $textLines = explode( "\n", $text );
1539
1540                 $lastPrefix = $output = $lastLine = '';
1541                 $this->mDTopen = $inBlockElem = false;
1542                 $prefixLength = 0;
1543                 $paragraphStack = false;
1544
1545                 if ( !$linestart ) {
1546                         $output .= array_shift( $textLines );
1547                 }
1548                 foreach ( $textLines as $oLine ) {
1549                         $lastPrefixLength = strlen( $lastPrefix );
1550                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1551                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1552                         if ( !$this->mInPre ) {
1553                                 # Multiple prefixes may abut each other for nested lists.
1554                                 $prefixLength = strspn( $oLine, '*#:;' );
1555                                 $pref = substr( $oLine, 0, $prefixLength );
1556
1557                                 # eh?
1558                                 $pref2 = str_replace( ';', ':', $pref );
1559                                 $t = substr( $oLine, $prefixLength );
1560                                 $this->mInPre = !empty($preOpenMatch);
1561                         } else {
1562                                 # Don't interpret any other prefixes in preformatted text
1563                                 $prefixLength = 0;
1564                                 $pref = $pref2 = '';
1565                                 $t = $oLine;
1566                         }
1567
1568                         # List generation
1569                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1570                                 # Same as the last item, so no need to deal with nesting or opening stuff
1571                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1572                                 $paragraphStack = false;
1573
1574                                 if ( substr( $pref, -1 ) == ';') {
1575                                         # The one nasty exception: definition lists work like this:
1576                                         # ; title : definition text
1577                                         # So we check for : in the remainder text to split up the
1578                                         # title and definition, without b0rking links.
1579                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1580                                                 $t = $t2;
1581                                                 $output .= $term . $this->nextItem( ':' );
1582                                         }
1583                                 }
1584                         } elseif( $prefixLength || $lastPrefixLength ) {
1585                                 # Either open or close a level...
1586                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1587                                 $paragraphStack = false;
1588
1589                                 while( $commonPrefixLength < $lastPrefixLength ) {
1590                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1591                                         --$lastPrefixLength;
1592                                 }
1593                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1594                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1595                                 }
1596                                 while ( $prefixLength > $commonPrefixLength ) {
1597                                         $char = substr( $pref, $commonPrefixLength, 1 );
1598                                         $output .= $this->openList( $char );
1599
1600                                         if ( ';' == $char ) {
1601                                                 # FIXME: This is dupe of code above
1602                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1603                                                         $t = $t2;
1604                                                         $output .= $term . $this->nextItem( ':' );
1605                                                 }
1606                                         }
1607                                         ++$commonPrefixLength;
1608                                 }
1609                                 $lastPrefix = $pref2;
1610                         }
1611                         if( 0 == $prefixLength ) {
1612                                 wfProfileIn( "$fname-paragraph" );
1613                                 # No prefix (not in list)--go to paragraph mode
1614                                 $uniq_prefix = UNIQ_PREFIX;
1615                                 // XXX: use a stack for nestable elements like span, table and div
1616                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1617                                 $closematch = preg_match(
1618                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1619                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1620                                 if ( $openmatch or $closematch ) {
1621                                         $paragraphStack = false;
1622                                         $output .= $this->closeParagraph();
1623                                         if($preOpenMatch and !$preCloseMatch) {
1624                                                 $this->mInPre = true;
1625                                         }
1626                                         if ( $closematch ) {
1627                                                 $inBlockElem = false;
1628                                         } else {
1629                                                 $inBlockElem = true;
1630                                         }
1631                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1632                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1633                                                 // pre
1634                                                 if ($this->mLastSection != 'pre') {
1635                                                         $paragraphStack = false;
1636                                                         $output .= $this->closeParagraph().'<pre>';
1637                                                         $this->mLastSection = 'pre';
1638                                                 }
1639                                                 $t = substr( $t, 1 );
1640                                         } else {
1641                                                 // paragraph
1642                                                 if ( '' == trim($t) ) {
1643                                                         if ( $paragraphStack ) {
1644                                                                 $output .= $paragraphStack.'<br />';
1645                                                                 $paragraphStack = false;
1646                                                                 $this->mLastSection = 'p';
1647                                                         } else {
1648                                                                 if ($this->mLastSection != 'p' ) {
1649                                                                         $output .= $this->closeParagraph();
1650                                                                         $this->mLastSection = '';
1651                                                                         $paragraphStack = '<p>';
1652                                                                 } else {
1653                                                                         $paragraphStack = '</p><p>';
1654                                                                 }
1655                                                         }
1656                                                 } else {
1657                                                         if ( $paragraphStack ) {
1658                                                                 $output .= $paragraphStack;
1659                                                                 $paragraphStack = false;
1660                                                                 $this->mLastSection = 'p';
1661                                                         } else if ($this->mLastSection != 'p') {
1662                                                                 $output .= $this->closeParagraph().'<p>';
1663                                                                 $this->mLastSection = 'p';
1664                                                         }
1665                                                 }
1666                                         }
1667                                 }
1668                                 wfProfileOut( "$fname-paragraph" );
1669                         }
1670                         if ($paragraphStack === false) {
1671                                 $output .= $t."\n";
1672                         }
1673                 }
1674                 while ( $prefixLength ) {
1675                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1676                         --$prefixLength;
1677                 }
1678                 if ( '' != $this->mLastSection ) {
1679                         $output .= '</' . $this->mLastSection . '>';
1680                         $this->mLastSection = '';
1681                 }
1682
1683                 wfProfileOut( $fname );
1684                 return $output;
1685         }
1686
1687         /**
1688          * Split up a string on ':', ignoring any occurences inside
1689          * <a>..</a> or <span>...</span>
1690          * @param string $str the string to split
1691          * @param string &$before set to everything before the ':'
1692          * @param string &$after set to everything after the ':'
1693          * return string the position of the ':', or false if none found
1694          */
1695         function findColonNoLinks($str, &$before, &$after) {
1696                 # I wonder if we should make this count all tags, not just <a>
1697                 # and <span>. That would prevent us from matching a ':' that
1698                 # comes in the middle of italics other such formatting....
1699                 # -- Wil
1700                 $fname = 'Parser::findColonNoLinks';
1701                 wfProfileIn( $fname );
1702                 $pos = 0;
1703                 do {
1704                         $colon = strpos($str, ':', $pos);
1705
1706                         if ($colon !== false) {
1707                                 $before = substr($str, 0, $colon);
1708                                 $after = substr($str, $colon + 1);
1709
1710                                 # Skip any ':' within <a> or <span> pairs
1711                                 $a = substr_count($before, '<a');
1712                                 $s = substr_count($before, '<span');
1713                                 $ca = substr_count($before, '</a>');
1714                                 $cs = substr_count($before, '</span>');
1715
1716                                 if ($a <= $ca and $s <= $cs) {
1717                                         # Tags are balanced before ':'; ok
1718                                         break;
1719                                 }
1720                                 $pos = $colon + 1;
1721                         }
1722                 } while ($colon !== false);
1723                 wfProfileOut( $fname );
1724                 return $colon;
1725         }
1726
1727         /**
1728          * Return value of a magic variable (like PAGENAME)
1729          *
1730          * @access private
1731          */
1732         function getVariableValue( $index ) {
1733                 global $wgContLang, $wgSitename, $wgServer;
1734
1735                 /**
1736                  * Some of these require message or data lookups and can be
1737                  * expensive to check many times.
1738                  */
1739                 static $varCache = array();
1740                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1741
1742                 switch ( $index ) {
1743                         case MAG_CURRENTMONTH:
1744                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1745                         case MAG_CURRENTMONTHNAME:
1746                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1747                         case MAG_CURRENTMONTHNAMEGEN:
1748                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1749                         case MAG_CURRENTDAY:
1750                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1751                         case MAG_PAGENAME:
1752                                 return $this->mTitle->getText();
1753                         case MAG_PAGENAMEE:
1754                                 return $this->mTitle->getPartialURL();
1755                         case MAG_NAMESPACE:
1756                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1757                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1758                         case MAG_CURRENTDAYNAME:
1759                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1760                         case MAG_CURRENTYEAR:
1761                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1762                         case MAG_CURRENTTIME:
1763                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1764                         case MAG_CURRENTWEEK:
1765                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1766                         case MAG_CURRENTDOW:
1767                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1768                         case MAG_NUMBEROFARTICLES:
1769                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1770                         case MAG_SITENAME:
1771                                 return $wgSitename;
1772                         case MAG_SERVER:
1773                                 return $wgServer;
1774                         default:
1775                                 return NULL;
1776                 }
1777         }
1778
1779         /**
1780          * initialise the magic variables (like CURRENTMONTHNAME)
1781          *
1782          * @access private
1783          */
1784         function initialiseVariables() {
1785                 $fname = 'Parser::initialiseVariables';
1786                 wfProfileIn( $fname );
1787                 global $wgVariableIDs;
1788                 $this->mVariables = array();
1789                 foreach ( $wgVariableIDs as $id ) {
1790                         $mw =& MagicWord::get( $id );
1791                         $mw->addToArray( $this->mVariables, $id );
1792                 }
1793                 wfProfileOut( $fname );
1794         }
1795
1796         /**
1797          * Replace magic variables, templates, and template arguments
1798          * with the appropriate text. Templates are substituted recursively,
1799          * taking care to avoid infinite loops.
1800          *
1801          * Note that the substitution depends on value of $mOutputType:
1802          *  OT_WIKI: only {{subst:}} templates
1803          *  OT_MSG: only magic variables
1804          *  OT_HTML: all templates and magic variables
1805          *
1806          * @param string $tex The text to transform
1807          * @param array $args Key-value pairs representing template parameters to substitute
1808          * @access private
1809          */
1810         function replaceVariables( $text, $args = array() ) {
1811                 global $wgLang, $wgScript, $wgArticlePath;
1812
1813                 # Prevent too big inclusions
1814                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1815                         return $text;
1816                 }
1817
1818                 $fname = 'Parser::replaceVariables';
1819                 wfProfileIn( $fname );
1820
1821                 $titleChars = Title::legalChars();
1822
1823                 # This function is called recursively. To keep track of arguments we need a stack:
1824                 array_push( $this->mArgStack, $args );
1825
1826                 # Variable substitution
1827                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1828
1829                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1830                         # Argument substitution
1831                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1832                 }
1833                 # Template substitution
1834                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1835                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1836
1837                 array_pop( $this->mArgStack );
1838
1839                 wfProfileOut( $fname );
1840                 return $text;
1841         }
1842
1843         /**
1844          * Replace magic variables
1845          * @access private
1846          */
1847         function variableSubstitution( $matches ) {
1848                 $fname = 'parser::variableSubstitution';
1849                 $varname = $matches[1];
1850                 wfProfileIn( $fname );
1851                 if ( !$this->mVariables ) {
1852                         $this->initialiseVariables();
1853                 }
1854                 $skip = false;
1855                 if ( $this->mOutputType == OT_WIKI ) {
1856                         # Do only magic variables prefixed by SUBST
1857                         $mwSubst =& MagicWord::get( MAG_SUBST );
1858                         if (!$mwSubst->matchStartAndRemove( $varname ))
1859                                 $skip = true;
1860                         # Note that if we don't substitute the variable below,
1861                         # we don't remove the {{subst:}} magic word, in case
1862                         # it is a template rather than a magic variable.
1863                 }
1864                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
1865                         $id = $this->mVariables[$varname];
1866                         $text = $this->getVariableValue( $id );
1867                         $this->mOutput->mContainsOldMagic = true;
1868                 } else {
1869                         $text = $matches[0];
1870                 }
1871                 wfProfileOut( $fname );
1872                 return $text;
1873         }
1874
1875         # Split template arguments
1876         function getTemplateArgs( $argsString ) {
1877                 if ( $argsString === '' ) {
1878                         return array();
1879                 }
1880
1881                 $args = explode( '|', substr( $argsString, 1 ) );
1882
1883                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1884                 # merged with the next arg because the '|' character between belongs
1885                 # to the link syntax and not the template parameter syntax.
1886                 $argc = count($args);
1887                 $i = 0;
1888                 for ( $i = 0; $i < $argc-1; $i++ ) {
1889                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1890                                 $args[$i] .= '|'.$args[$i+1];
1891                                 array_splice($args, $i+1, 1);
1892                                 $i--;
1893                                 $argc--;
1894                         }
1895                 }
1896
1897                 return $args;
1898         }
1899
1900         /**
1901          * Return the text of a template, after recursively
1902          * replacing any variables or templates within the template.
1903          *
1904          * @param array $matches The parts of the template
1905          *  $matches[1]: the title, i.e. the part before the |
1906          *  $matches[2]: the parameters (including a leading |), if  any
1907          * @return string the text of the template
1908          * @access private
1909          */
1910         function braceSubstitution( $matches ) {
1911                 global $wgLinkCache, $wgContLang;
1912                 $fname = 'Parser::braceSubstitution';
1913                 wfProfileIn( $fname );
1914
1915                 $found = false;
1916                 $nowiki = false;
1917                 $noparse = false;
1918
1919                 $title = NULL;
1920
1921                 # Need to know if the template comes at the start of a line,
1922                 # to treat the beginning of the template like the beginning
1923                 # of a line for tables and block-level elements.
1924                 $linestart = $matches[1];
1925
1926                 # $part1 is the bit before the first |, and must contain only title characters
1927                 # $args is a list of arguments, starting from index 0, not including $part1
1928
1929                 $part1 = $matches[2];
1930                 # If the third subpattern matched anything, it will start with |
1931
1932                 $args = $this->getTemplateArgs($matches[3]);
1933                 $argc = count( $args );
1934
1935                 # Don't parse {{{}}} because that's only for template arguments
1936                 if ( $linestart === '{' ) {
1937                         $text = $matches[0];
1938                         $found = true;
1939                         $noparse = true;
1940                 }
1941
1942                 # SUBST
1943                 if ( !$found ) {
1944                         $mwSubst =& MagicWord::get( MAG_SUBST );
1945                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1946                                 # One of two possibilities is true:
1947                                 # 1) Found SUBST but not in the PST phase
1948                                 # 2) Didn't find SUBST and in the PST phase
1949                                 # In either case, return without further processing
1950                                 $text = $matches[0];
1951                                 $found = true;
1952                                 $noparse = true;
1953                         }
1954                 }
1955
1956                 # MSG, MSGNW and INT
1957                 if ( !$found ) {
1958                         # Check for MSGNW:
1959                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1960                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1961                                 $nowiki = true;
1962                         } else {
1963                                 # Remove obsolete MSG:
1964                                 $mwMsg =& MagicWord::get( MAG_MSG );
1965                                 $mwMsg->matchStartAndRemove( $part1 );
1966                         }
1967
1968                         # Check if it is an internal message
1969                         $mwInt =& MagicWord::get( MAG_INT );
1970                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1971                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1972                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1973                                         $found = true;
1974                                 }
1975                         }
1976                 }
1977
1978                 # NS
1979                 if ( !$found ) {
1980                         # Check for NS: (namespace expansion)
1981                         $mwNs = MagicWord::get( MAG_NS );
1982                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1983                                 if ( intval( $part1 ) ) {
1984                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1985                                         $found = true;
1986                                 } else {
1987                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1988                                         if ( !is_null( $index ) ) {
1989                                                 $text = $linestart . $wgContLang->getNsText( $index );
1990                                                 $found = true;
1991                                         }
1992                                 }
1993                         }
1994                 }
1995
1996                 # LOCALURL and LOCALURLE
1997                 if ( !$found ) {
1998                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1999                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
2000
2001                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2002                                 $func = 'getLocalURL';
2003                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2004                                 $func = 'escapeLocalURL';
2005                         } else {
2006                                 $func = '';
2007                         }
2008
2009                         if ( $func !== '' ) {
2010                                 $title = Title::newFromText( $part1 );
2011                                 if ( !is_null( $title ) ) {
2012                                         if ( $argc > 0 ) {
2013                                                 $text = $linestart . $title->$func( $args[0] );
2014                                         } else {
2015                                                 $text = $linestart . $title->$func();
2016                                         }
2017                                         $found = true;
2018                                 }
2019                         }
2020                 }
2021
2022                 # GRAMMAR
2023                 if ( !$found && $argc == 1 ) {
2024                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2025                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2026                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2027                                 $found = true;
2028                         }
2029                 }
2030
2031                 # Template table test
2032
2033                 # Did we encounter this template already? If yes, it is in the cache
2034                 # and we need to check for loops.
2035                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2036                         # set $text to cached message.
2037                         $text = $linestart . $this->mTemplates[$part1];
2038                         $found = true;
2039
2040                         # Infinite loop test
2041                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2042                                 $noparse = true;
2043                                 $found = true;
2044                                 $text .= '<!-- WARNING: template loop detected -->';
2045                         }
2046                 }
2047
2048                 # Load from database
2049                 $itcamefromthedatabase = false;
2050                 if ( !$found ) {
2051                         $ns = NS_TEMPLATE;
2052                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2053                         if ($subpage !== '') {
2054                                 $ns = $this->mTitle->getNamespace();
2055                         }
2056                         $title = Title::newFromText( $part1, $ns );
2057                         if ( !is_null( $title ) && !$title->isExternal() ) {
2058                                 # Check for excessive inclusion
2059                                 $dbk = $title->getPrefixedDBkey();
2060                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2061                                         # This should never be reached.
2062                                         $article = new Article( $title );
2063                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2064                                         if ( $articleContent !== false ) {
2065                                                 $found = true;
2066                                                 $text = $linestart . $articleContent;
2067                                                 $itcamefromthedatabase = true;
2068                                         }
2069                                 }
2070
2071                                 # If the title is valid but undisplayable, make a link to it
2072                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2073                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2074                                         $found = true;
2075                                 }
2076
2077                                 # Template cache array insertion
2078                                 $this->mTemplates[$part1] = $text;
2079                         }
2080                 }
2081
2082                 # Recursive parsing, escaping and link table handling
2083                 # Only for HTML output
2084                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2085                         $text = wfEscapeWikiText( $text );
2086                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2087                         # Clean up argument array
2088                         $assocArgs = array();
2089                         $index = 1;
2090                         foreach( $args as $arg ) {
2091                                 $eqpos = strpos( $arg, '=' );
2092                                 if ( $eqpos === false ) {
2093                                         $assocArgs[$index++] = $arg;
2094                                 } else {
2095                                         $name = trim( substr( $arg, 0, $eqpos ) );
2096                                         $value = trim( substr( $arg, $eqpos+1 ) );
2097                                         if ( $value === false ) {
2098                                                 $value = '';
2099                                         }
2100                                         if ( $name !== false ) {
2101                                                 $assocArgs[$name] = $value;
2102                                         }
2103                                 }
2104                         }
2105
2106                         # Add a new element to the templace recursion path
2107                         $this->mTemplatePath[$part1] = 1;
2108
2109                         $text = $this->strip( $text, $this->mStripState );
2110                         $text = $this->removeHTMLtags( $text );
2111                         $text = $this->replaceVariables( $text, $assocArgs );
2112
2113                         # Resume the link cache and register the inclusion as a link
2114                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2115                                 $wgLinkCache->addLinkObj( $title );
2116                         }
2117
2118                         # If the template begins with a table or block-level
2119                         # element, it should be treated as beginning a new line.
2120                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2121                                 $text = "\n" . $text;
2122                         }
2123                 }
2124
2125                 # Empties the template path
2126                 $this->mTemplatePath = array();
2127                 if ( !$found ) {
2128                         wfProfileOut( $fname );
2129                         return $matches[0];
2130                 } else {
2131                         # replace ==section headers==
2132                         # XXX this needs to go away once we have a better parser.
2133                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2134                                 if( !is_null( $title ) )
2135                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2136                                 else
2137                                         $encodedname = base64_encode("");
2138                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2139                                         PREG_SPLIT_DELIM_CAPTURE);
2140                                 $text = '';
2141                                 $nsec = 0;
2142                                 for( $i = 0; $i < count($m); $i += 2 ) {
2143                                         $text .= $m[$i];
2144                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2145                                         $hl = $m[$i + 1];
2146                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2147                                                 $text .= $hl;
2148                                                 continue;
2149                                         }
2150                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2151                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2152                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2153
2154                                         $nsec++;
2155                                 }
2156                         }
2157                 }
2158
2159                 # Empties the template path
2160                 $this->mTemplatePath = array();
2161
2162                 if ( !$found ) {
2163                         wfProfileOut( $fname );
2164                         return $matches[0];
2165                 } else {
2166                         wfProfileOut( $fname );
2167                         return $text;
2168                 }
2169         }
2170
2171         /**
2172          * Triple brace replacement -- used for template arguments
2173          * @access private
2174          */
2175         function argSubstitution( $matches ) {
2176                 $arg = trim( $matches[1] );
2177                 $text = $matches[0];
2178                 $inputArgs = end( $this->mArgStack );
2179
2180                 if ( array_key_exists( $arg, $inputArgs ) ) {
2181                         $text = $inputArgs[$arg];
2182                 }
2183
2184                 return $text;
2185         }
2186
2187         /**
2188          * Returns true if the function is allowed to include this entity
2189          * @access private
2190          */
2191         function incrementIncludeCount( $dbk ) {
2192                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2193                         $this->mIncludeCount[$dbk] = 0;
2194                 }
2195                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2196                         return true;
2197                 } else {
2198                         return false;
2199                 }
2200         }
2201
2202
2203         /**
2204          * Cleans up HTML, removes dangerous tags and attributes, and
2205          * removes HTML comments
2206          * @access private
2207          */
2208         function removeHTMLtags( $text ) {
2209                 global $wgUseTidy, $wgUserHtml;
2210                 $fname = 'Parser::removeHTMLtags';
2211                 wfProfileIn( $fname );
2212
2213                 if( $wgUserHtml ) {
2214                         $htmlpairs = array( # Tags that must be closed
2215                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
2216                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
2217                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
2218                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
2219                                 'ruby', 'rt' , 'rb' , 'rp', 'p', 'span'
2220                         );
2221                         $htmlsingle = array(
2222                                 'br', 'hr', 'li', 'dt', 'dd'
2223                         );
2224                         $htmlnest = array( # Tags that can be nested--??
2225                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
2226                                 'dl', 'font', 'big', 'small', 'sub', 'sup', 'span'
2227                         );
2228                         $tabletags = array( # Can only appear inside table
2229                                 'td', 'th', 'tr'
2230                         );
2231                 } else {
2232                         $htmlpairs = array();
2233                         $htmlsingle = array();
2234                         $htmlnest = array();
2235                         $tabletags = array();
2236                 }
2237
2238                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
2239                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
2240
2241                 $htmlattrs = $this->getHTMLattrs () ;
2242
2243                 # Remove HTML comments
2244                 $text = $this->removeHTMLcomments( $text );
2245
2246                 $bits = explode( '<', $text );
2247                 $text = array_shift( $bits );
2248                 if(!$wgUseTidy) {
2249                         $tagstack = array(); $tablestack = array();
2250                         foreach ( $bits as $x ) {
2251                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2252                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2253                                 $x, $regs );
2254                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2255                                 error_reporting( $prev );
2256
2257                                 $badtag = 0 ;
2258                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2259                                         # Check our stack
2260                                         if ( $slash ) {
2261                                                 # Closing a tag...
2262                                                 if ( ! in_array( $t, $htmlsingle ) &&
2263                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2264                                                         @array_push( $tagstack, $ot );
2265                                                         $badtag = 1;
2266                                                 } else {
2267                                                         if ( $t == 'table' ) {
2268                                                                 $tagstack = array_pop( $tablestack );
2269                                                         }
2270                                                         $newparams = '';
2271                                                 }
2272                                         } else {
2273                                                 # Keep track for later
2274                                                 if ( in_array( $t, $tabletags ) &&
2275                                                 ! in_array( 'table', $tagstack ) ) {
2276                                                         $badtag = 1;
2277                                                 } else if ( in_array( $t, $tagstack ) &&
2278                                                 ! in_array ( $t , $htmlnest ) ) {
2279                                                         $badtag = 1 ;
2280                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
2281                                                         if ( $t == 'table' ) {
2282                                                                 array_push( $tablestack, $tagstack );
2283                                                                 $tagstack = array();
2284                                                         }
2285                                                         array_push( $tagstack, $t );
2286                                                 }
2287                                                 # Strip non-approved attributes from the tag
2288                                                 $newparams = $this->fixTagAttributes($params);
2289
2290                                         }
2291                                         if ( ! $badtag ) {
2292                                                 $rest = str_replace( '>', '&gt;', $rest );
2293                                                 $text .= "<$slash$t $newparams$brace$rest";
2294                                                 continue;
2295                                         }
2296                                 }
2297                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2298                         }
2299                         # Close off any remaining tags
2300                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2301                                 $text .= "</$t>\n";
2302                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2303                         }
2304                 } else {
2305                         # this might be possible using tidy itself
2306                         foreach ( $bits as $x ) {
2307                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2308                                 $x, $regs );
2309                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2310                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2311                                         $newparams = $this->fixTagAttributes($params);
2312                                         $rest = str_replace( '>', '&gt;', $rest );
2313                                         $text .= "<$slash$t $newparams$brace$rest";
2314                                 } else {
2315                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2316                                 }
2317                         }
2318                 }
2319                 wfProfileOut( $fname );
2320                 return $text;
2321         }
2322
2323         /**
2324          * Remove '<!--', '-->', and everything between.
2325          * To avoid leaving blank lines, when a comment is both preceded
2326          * and followed by a newline (ignoring spaces), trim leading and
2327          * trailing spaces and one of the newlines.
2328          *
2329          * @access private
2330          */
2331         function removeHTMLcomments( $text ) {
2332                 $fname='Parser::removeHTMLcomments';
2333                 wfProfileIn( $fname );
2334                 while (($start = strpos($text, '<!--')) !== false) {
2335                         $end = strpos($text, '-->', $start + 4);
2336                         if ($end === false) {
2337                                 # Unterminated comment; bail out
2338                                 break;
2339                         }
2340
2341                         $end += 3;
2342
2343                         # Trim space and newline if the comment is both
2344                         # preceded and followed by a newline
2345                         $spaceStart = max($start - 1, 0);
2346                         $spaceLen = $end - $spaceStart;
2347                         while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2348                                 $spaceStart--;
2349                                 $spaceLen++;
2350                         }
2351                         while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2352                                 $spaceLen++;
2353                         if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2354                                 # Remove the comment, leading and trailing
2355                                 # spaces, and leave only one newline.
2356                                 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2357                         }
2358                         else {
2359                                 # Remove just the comment.
2360                                 $text = substr_replace($text, '', $start, $end - $start);
2361                         }
2362                 }
2363                 wfProfileOut( $fname );
2364                 return $text;
2365         }
2366
2367         /**
2368          * This function accomplishes several tasks:
2369          * 1) Auto-number headings if that option is enabled
2370          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2371          * 3) Add a Table of contents on the top for users who have enabled the option
2372          * 4) Auto-anchor headings
2373          *
2374          * It loops through all headlines, collects the necessary data, then splits up the
2375          * string and re-inserts the newly formatted headlines.
2376          * @access private
2377          */
2378         /* private */ function formatHeadings( $text, $isMain=true ) {
2379                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2380
2381                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2382                 $doShowToc = $this->mOptions->getShowToc();
2383                 $forceTocHere = false;
2384                 if( !$this->mTitle->userCanEdit() ) {
2385                         $showEditLink = 0;
2386                         $rightClickHack = 0;
2387                 } else {
2388                         $showEditLink = $this->mOptions->getEditSection();
2389                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2390                 }
2391
2392                 # Inhibit editsection links if requested in the page
2393                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2394                 if( $esw->matchAndRemove( $text ) ) {
2395                         $showEditLink = 0;
2396                 }
2397                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2398                 # do not add TOC
2399                 $mw =& MagicWord::get( MAG_NOTOC );
2400                 if( $mw->matchAndRemove( $text ) ) {
2401                         $doShowToc = 0;
2402                 }
2403
2404                 # never add the TOC to the Main Page. This is an entry page that should not
2405                 # be more than 1-2 screens large anyway
2406                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2407                         $doShowToc = 0;
2408                 }
2409
2410                 # Get all headlines for numbering them and adding funky stuff like [edit]
2411                 # links - this is for later, but we need the number of headlines right now
2412                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2413
2414                 # if there are fewer than 4 headlines in the article, do not show TOC
2415                 if( $numMatches < 4 ) {
2416                         $doShowToc = 0;
2417                 }
2418
2419                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2420                 # override above conditions and always show TOC at that place
2421
2422                 $mw =& MagicWord::get( MAG_TOC );
2423                 if($mw->match( $text ) ) {
2424                         $doShowToc = 1;
2425                         $forceTocHere = true;
2426                 } else {
2427                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2428                         # override above conditions and always show TOC above first header
2429                         $mw =& MagicWord::get( MAG_FORCETOC );
2430                         if ($mw->matchAndRemove( $text ) ) {
2431                                 $doShowToc = 1;
2432                         }
2433                 }
2434
2435                 # Never ever show TOC if no headers
2436                 if( $numMatches < 1 ) {
2437                         $doShowToc = 0;
2438                 }
2439
2440                 # We need this to perform operations on the HTML
2441                 $sk =& $this->mOptions->getSkin();
2442
2443                 # headline counter
2444                 $headlineCount = 0;
2445                 $sectionCount = 0; # headlineCount excluding template sections
2446
2447                 # Ugh .. the TOC should have neat indentation levels which can be
2448                 # passed to the skin functions. These are determined here
2449                 $toc = '';
2450                 $full = '';
2451                 $head = array();
2452                 $sublevelCount = array();
2453                 $levelCount = array();
2454                 $toclevel = 0;
2455                 $level = 0;
2456                 $prevlevel = 0;
2457                 $toclevel = 0;
2458                 $prevtoclevel = 0;
2459
2460                 foreach( $matches[3] as $headline ) {
2461                         $istemplate = 0;
2462                         $templatetitle = '';
2463                         $templatesection = 0;
2464                         $numbering = '';
2465
2466                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2467                                 $istemplate = 1;
2468                                 $templatetitle = base64_decode($mat[1]);
2469                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2470                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2471                         }
2472
2473                         if( $toclevel ) {
2474                                 $prevlevel = $level;
2475                                 $prevtoclevel = $toclevel;
2476                         }
2477                         $level = $matches[1][$headlineCount];
2478
2479                         if( $doNumberHeadings || $doShowToc ) {
2480
2481                                 if ( $level > $prevlevel ) {
2482                                         # Increase TOC level
2483                                         $toclevel++;
2484                                         $sublevelCount[$toclevel] = 0;
2485                                         $toc .= $sk->tocIndent();
2486                                 }
2487                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2488                                         # Decrease TOC level, find level to jump to
2489
2490                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2491                                                 # Can only go down to level 1
2492                                                 $toclevel = 1;
2493                                         } else {
2494                                                 for ($i = $toclevel; $i > 0; $i--) {
2495                                                         if ( $levelCount[$i] == $level ) {
2496                                                                 # Found last matching level
2497                                                                 $toclevel = $i;
2498                                                                 break;
2499                                                         }
2500                                                         elseif ( $levelCount[$i] < $level ) {
2501                                                                 # Found first matching level below current level
2502                                                                 $toclevel = $i + 1;
2503                                                                 break;
2504                                                         }
2505                                                 }
2506                                         }
2507
2508                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2509                                 }
2510                                 else {
2511                                         # No change in level, end TOC line
2512                                         $toc .= $sk->tocLineEnd();
2513                                 }
2514
2515                                 $levelCount[$toclevel] = $level;
2516
2517                                 # count number of headlines for each level
2518                                 @$sublevelCount[$toclevel]++;
2519                                 $dot = 0;
2520                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2521                                         if( !empty( $sublevelCount[$i] ) ) {
2522                                                 if( $dot ) {
2523                                                         $numbering .= '.';
2524                                                 }
2525                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2526                                                 $dot = 1;
2527                                         }
2528                                 }
2529                         }
2530
2531                         # The canonized header is a version of the header text safe to use for links
2532                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2533                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2534                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2535
2536                         # Remove link placeholders by the link text.
2537                         #     <!--LINK number-->
2538                         # turns into
2539                         #     link text with suffix
2540                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2541                                                             "\$wgLinkHolders['texts'][\$1]",
2542                                                             $canonized_headline );
2543                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2544                                                             "\$wgInterwikiLinkHolders[\$1]",
2545                                                             $canonized_headline );
2546
2547                         # strip out HTML
2548                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2549                         $tocline = trim( $canonized_headline );
2550                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2551                         $replacearray = array(
2552                                 '%3A' => ':',
2553                                 '%' => '.'
2554                         );
2555                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2556                         $refer[$headlineCount] = $canonized_headline;
2557
2558                         # count how many in assoc. array so we can track dupes in anchors
2559                         @$refers[$canonized_headline]++;
2560                         $refcount[$headlineCount]=$refers[$canonized_headline];
2561
2562                         # Don't number the heading if it is the only one (looks silly)
2563                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2564                                 # the two are different if the line contains a link
2565                                 $headline=$numbering . ' ' . $headline;
2566                         }
2567
2568                         # Create the anchor for linking from the TOC to the section
2569                         $anchor = $canonized_headline;
2570                         if($refcount[$headlineCount] > 1 ) {
2571                                 $anchor .= '_' . $refcount[$headlineCount];
2572                         }
2573                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2574                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2575                         }
2576                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2577                                 if ( empty( $head[$headlineCount] ) ) {
2578                                         $head[$headlineCount] = '';
2579                                 }
2580                                 if( $istemplate )
2581                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2582                                 else
2583                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2584                         }
2585
2586                         # Add the edit section span
2587                         if( $rightClickHack ) {
2588                                 if( $istemplate )
2589                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2590                                 else
2591                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2592                         }
2593
2594                         # give headline the correct <h#> tag
2595                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2596
2597                         $headlineCount++;
2598                         if( !$istemplate )
2599                                 $sectionCount++;
2600                 }
2601
2602                 if( $doShowToc ) {
2603                         $toclines = $headlineCount;
2604                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2605                         $toc = $sk->tocList( $toc );
2606                 }
2607
2608                 # split up and insert constructed headlines
2609
2610                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2611                 $i = 0;
2612
2613                 foreach( $blocks as $block ) {
2614                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2615                                 # This is the [edit] link that appears for the top block of text when
2616                                 # section editing is enabled
2617
2618                                 # Disabled because it broke block formatting
2619                                 # For example, a bullet point in the top line
2620                                 # $full .= $sk->editSectionLink(0);
2621                         }
2622                         $full .= $block;
2623                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2624                         # Top anchor now in skin
2625                                 $full = $full.$toc;
2626                         }
2627
2628                         if( !empty( $head[$i] ) ) {
2629                                 $full .= $head[$i];
2630                         }
2631                         $i++;
2632                 }
2633                 if($forceTocHere) {
2634                         $mw =& MagicWord::get( MAG_TOC );
2635                         return $mw->replace( $toc, $full );
2636                 } else {
2637                         return $full;
2638                 }
2639         }
2640
2641         /**
2642          * Return an HTML link for the "ISBN 123456" text
2643          * @access private
2644          */
2645         function magicISBN( $text ) {
2646                 global $wgLang;
2647                 $fname = 'Parser::magicISBN';
2648                 wfProfileIn( $fname );
2649
2650                 $a = split( 'ISBN ', ' '.$text );
2651                 if ( count ( $a ) < 2 ) {
2652                         wfProfileOut( $fname );
2653                         return $text;
2654                 }
2655                 $text = substr( array_shift( $a ), 1);
2656                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2657
2658                 foreach ( $a as $x ) {
2659                         $isbn = $blank = '' ;
2660                         while ( ' ' == $x{0} ) {
2661                                 $blank .= ' ';
2662                                 $x = substr( $x, 1 );
2663                         }
2664                         if ( $x == '' ) { # blank isbn
2665                                 $text .= "ISBN $blank";
2666                                 continue;
2667                         }
2668                         while ( strstr( $valid, $x{0} ) != false ) {
2669                                 $isbn .= $x{0};
2670                                 $x = substr( $x, 1 );
2671                         }
2672                         $num = str_replace( '-', '', $isbn );
2673                         $num = str_replace( ' ', '', $num );
2674
2675                         if ( '' == $num ) {
2676                                 $text .= "ISBN $blank$x";
2677                         } else {
2678                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2679                                 $text .= '<a href="' .
2680                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2681                                         "\" class=\"internal\">ISBN $isbn</a>";
2682                                 $text .= $x;
2683                         }
2684                 }
2685                 wfProfileOut( $fname );
2686                 return $text;
2687         }
2688
2689         /**
2690          * Return an HTML link for the "GEO ..." text
2691          * @access private
2692          */
2693         function magicGEO( $text ) {
2694                 global $wgLang, $wgUseGeoMode;
2695                 $fname = 'Parser::magicGEO';
2696                 wfProfileIn( $fname );
2697
2698                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2699                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2700                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2701                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2702                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2703                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2704
2705                 $a = split( 'GEO ', ' '.$text );
2706                 if ( count ( $a ) < 2 ) {
2707                         wfProfileOut( $fname );
2708                         return $text;
2709                 }
2710                 $text = substr( array_shift( $a ), 1);
2711                 $valid = '0123456789.+-:';
2712
2713                 foreach ( $a as $x ) {
2714                         $geo = $blank = '' ;
2715                         while ( ' ' == $x{0} ) {
2716                                 $blank .= ' ';
2717                                 $x = substr( $x, 1 );
2718                         }
2719                         while ( strstr( $valid, $x{0} ) != false ) {
2720                                 $geo .= $x{0};
2721                                 $x = substr( $x, 1 );
2722                         }
2723                         $num = str_replace( '+', '', $geo );
2724                         $num = str_replace( ' ', '', $num );
2725
2726                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2727                                 $text .= "GEO $blank$x";
2728                         } else {
2729                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2730                                 $text .= '<a href="' .
2731                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2732                                         "\" class=\"internal\">GEO $geo</a>";
2733                                 $text .= $x;
2734                         }
2735                 }
2736                 wfProfileOut( $fname );
2737                 return $text;
2738         }
2739
2740         /**
2741          * Return an HTML link for the "RFC 1234" text
2742          * @access private
2743          * @param string $text text to be processed
2744          */
2745         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2746                 global $wgLang;
2747
2748                 $valid = '0123456789';
2749                 $internal = false;
2750
2751                 $a = split( $keyword, ' '.$text );
2752                 if ( count ( $a ) < 2 ) {
2753                         return $text;
2754                 }
2755                 $text = substr( array_shift( $a ), 1);
2756
2757                 /* Check if keyword is preceed by [[.
2758                  * This test is made here cause of the array_shift above
2759                  * that prevent the test to be done in the foreach.
2760                  */
2761                 if ( substr( $text, -2 ) == '[[' ) {
2762                         $internal = true;
2763                 }
2764
2765                 foreach ( $a as $x ) {
2766                         /* token might be empty if we have RFC RFC 1234 */
2767                         if ( $x=='' ) {
2768                                 $text.=$keyword;
2769                                 continue;
2770                                 }
2771
2772                         $id = $blank = '' ;
2773
2774                         /** remove and save whitespaces in $blank */
2775                         while ( $x{0} == ' ' ) {
2776                                 $blank .= ' ';
2777                                 $x = substr( $x, 1 );
2778                         }
2779
2780                         /** remove and save the rfc number in $id */
2781                         while ( strstr( $valid, $x{0} ) != false ) {
2782                                 $id .= $x{0};
2783                                 $x = substr( $x, 1 );
2784                         }
2785
2786                         if ( $id == '' ) {
2787                                 /* call back stripped spaces*/
2788                                 $text .= $keyword.$blank.$x;
2789                         } elseif( $internal ) {
2790                                 /* normal link */
2791                                 $text .= $keyword.$id.$x;
2792                         } else {
2793                                 /* build the external link*/
2794                                 $url = wfmsg( $urlmsg );
2795                                 $url = str_replace( '$1', $id, $url);
2796                                 $sk =& $this->mOptions->getSkin();
2797                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2798                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2799                         }
2800
2801                         /* Check if the next RFC keyword is preceed by [[ */
2802                         $internal = ( substr($x,-2) == '[[' );
2803                 }
2804                 return $text;
2805         }
2806
2807         /**
2808          * Transform wiki markup when saving a page by doing \r\n -> \n
2809          * conversion, substitting signatures, {{subst:}} templates, etc.
2810          *
2811          * @param string $text the text to transform
2812          * @param Title &$title the Title object for the current article
2813          * @param User &$user the User object describing the current user
2814          * @param ParserOptions $options parsing options
2815          * @param bool $clearState whether to clear the parser state first
2816          * @return string the altered wiki markup
2817          * @access public
2818          */
2819         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2820                 $this->mOptions = $options;
2821                 $this->mTitle =& $title;
2822                 $this->mOutputType = OT_WIKI;
2823
2824                 if ( $clearState ) {
2825                         $this->clearState();
2826                 }
2827
2828                 $stripState = false;
2829                 $pairs = array(
2830                         "\r\n" => "\n",
2831                         );
2832                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2833                 $text = $this->strip( $text, $stripState, false );
2834                 $text = $this->pstPass2( $text, $user );
2835                 $text = $this->unstrip( $text, $stripState );
2836                 $text = $this->unstripNoWiki( $text, $stripState );
2837                 return $text;
2838         }
2839
2840         /**
2841          * Pre-save transform helper function
2842          * @access private
2843          */
2844         function pstPass2( $text, &$user ) {
2845                 global $wgLang, $wgContLang, $wgLocaltimezone;
2846
2847                 # Variable replacement
2848                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2849                 $text = $this->replaceVariables( $text );
2850
2851                 # Signatures
2852                 #
2853                 $n = $user->getName();
2854                 $k = $user->getOption( 'nickname' );
2855                 if ( '' == $k ) { $k = $n; }
2856                 if ( isset( $wgLocaltimezone ) ) {
2857                         $oldtz = getenv( 'TZ' );
2858                         putenv( 'TZ='.$wgLocaltimezone );
2859                 }
2860                 /* Note: this is an ugly timezone hack for the European wikis */
2861                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2862                   ' (' . date( 'T' ) . ')';
2863                 if ( isset( $wgLocaltimezone ) ) {
2864                         putenv( 'TZ='.$oldtzs );
2865                 }
2866
2867                 if( $user->getOption( 'fancysig' ) ) {
2868                         $sigText = $k;
2869                 } else {
2870                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2871                 }
2872                 $text = preg_replace( '/~~~~~/', $d, $text );
2873                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2874                 $text = preg_replace( '/~~~/', $sigText, $text );
2875
2876                 # Context links: [[|name]] and [[name (context)|]]
2877                 #
2878                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2879                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2880                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2881                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2882
2883                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2884                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2885                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2886                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2887                 $context = '';
2888                 $t = $this->mTitle->getText();
2889                 if ( preg_match( $conpat, $t, $m ) ) {
2890                         $context = $m[2];
2891                 }
2892                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2893                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2894                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2895
2896                 if ( '' == $context ) {
2897                         $text = preg_replace( $p2, '[[\\1]]', $text );
2898                 } else {
2899                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2900                 }
2901
2902                 # Trim trailing whitespace
2903                 # MAG_END (__END__) tag allows for trailing
2904                 # whitespace to be deliberately included
2905                 $text = rtrim( $text );
2906                 $mw =& MagicWord::get( MAG_END );
2907                 $mw->matchAndRemove( $text );
2908
2909                 return $text;
2910         }
2911
2912         /**
2913          * Set up some variables which are usually set up in parse()
2914          * so that an external function can call some class members with confidence
2915          * @access public
2916          */
2917         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2918                 $this->mTitle =& $title;
2919                 $this->mOptions = $options;
2920                 $this->mOutputType = $outputType;
2921                 if ( $clearState ) {
2922                         $this->clearState();
2923                 }
2924         }
2925
2926         /**
2927          * Transform a MediaWiki message by replacing magic variables.
2928          *
2929          * @param string $text the text to transform
2930          * @param ParserOptions $options  options
2931          * @return string the text with variables substituted
2932          * @access public
2933          */
2934         function transformMsg( $text, $options ) {
2935                 global $wgTitle;
2936                 static $executing = false;
2937
2938                 # Guard against infinite recursion
2939                 if ( $executing ) {
2940                         return $text;
2941                 }
2942                 $executing = true;
2943
2944                 $this->mTitle = $wgTitle;
2945                 $this->mOptions = $options;
2946                 $this->mOutputType = OT_MSG;
2947                 $this->clearState();
2948                 $text = $this->replaceVariables( $text );
2949
2950                 $executing = false;
2951                 return $text;
2952         }
2953
2954         /**
2955          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2956          * Callback will be called with the text within
2957          * Transform and return the text within
2958          * @access public
2959          */
2960         function setHook( $tag, $callback ) {
2961                 $oldVal = @$this->mTagHooks[$tag];
2962                 $this->mTagHooks[$tag] = $callback;
2963                 return $oldVal;
2964         }
2965
2966         /**
2967          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2968          * Placeholders created in Skin::makeLinkObj()
2969          * Returns an array of links found, indexed by PDBK:
2970          *  0 - broken
2971          *  1 - normal link
2972          *  2 - stub
2973          * $options is a bit field, RLH_FOR_UPDATE to select for update
2974          */
2975         function replaceLinkHolders( &$text, $options = 0 ) {
2976                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2977                 global $wgInterwikiLinkHolders;
2978                 global $outputReplace;
2979
2980                 if ( $wgUseOldExistenceCheck ) {
2981                         return array();
2982                 }
2983
2984                 $fname = 'Parser::replaceLinkHolders';
2985                 wfProfileIn( $fname );
2986
2987                 $pdbks = array();
2988                 $colours = array();
2989
2990                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2991                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2992                         wfProfileIn( $fname.'-check' );
2993                         $dbr =& wfGetDB( DB_SLAVE );
2994                         $page = $dbr->tableName( 'page' );
2995                         $sk = $wgUser->getSkin();
2996                         $threshold = $wgUser->getOption('stubthreshold');
2997
2998                         # Sort by namespace
2999                         asort( $wgLinkHolders['namespaces'] );
3000
3001                         # Generate query
3002                         $query = false;
3003                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
3004                                 # Make title object
3005                                 $title = $wgLinkHolders['titles'][$key];
3006
3007                                 # Skip invalid entries.
3008                                 # Result will be ugly, but prevents crash.
3009                                 if ( is_null( $title ) ) {
3010                                         continue;
3011                                 }
3012                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
3013
3014                                 # Check if it's in the link cache already
3015                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
3016                                         $colours[$pdbk] = 1;
3017                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
3018                                         $colours[$pdbk] = 0;
3019                                 } else {
3020                                         # Not in the link cache, add it to the query
3021                                         if ( !isset( $current ) ) {
3022                                                 $current = $val;
3023                                                 $tables = $page;
3024                                                 $join = '';
3025                                                 $query =  "SELECT page_id, page_namespace, page_title";
3026                                                 if ( $threshold > 0 ) {
3027                                                         $textTable = $dbr->tableName( 'text' );
3028                                                         $query .= ', LENGTH(old_text) AS page_len, page_is_redirect';
3029                                                         $tables .= ", $textTable";
3030                                                         $join = 'page_latest=old_id AND';
3031                                                 }
3032                                                 $query .= " FROM $tables WHERE $join (page_namespace=$val AND page_title IN(";
3033                                         } elseif ( $current != $val ) {
3034                                                 $current = $val;
3035                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
3036                                         } else {
3037                                                 $query .= ', ';
3038                                         }
3039
3040                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
3041                                 }
3042                         }
3043                         if ( $query ) {
3044                                 $query .= '))';
3045                                 if ( $options & RLH_FOR_UPDATE ) {
3046                                         $query .= ' FOR UPDATE';
3047                                 }
3048
3049                                 $res = $dbr->query( $query, $fname );
3050
3051                                 # Fetch data and form into an associative array
3052                                 # non-existent = broken
3053                                 # 1 = known
3054                                 # 2 = stub
3055                                 while ( $s = $dbr->fetchObject($res) ) {
3056                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
3057                                         $pdbk = $title->getPrefixedDBkey();
3058                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
3059
3060                                         if ( $threshold >  0 ) {
3061                                                 $size = $s->page_len;
3062                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $length < $threshold ) {
3063                                                         $colours[$pdbk] = 1;
3064                                                 } else {
3065                                                         $colours[$pdbk] = 2;
3066                                                 }
3067                                         } else {
3068                                                 $colours[$pdbk] = 1;
3069                                         }
3070                                 }
3071                         }
3072                         wfProfileOut( $fname.'-check' );
3073
3074                         # Construct search and replace arrays
3075                         wfProfileIn( $fname.'-construct' );
3076                         $outputReplace = array();
3077                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
3078                                 $pdbk = $pdbks[$key];
3079                                 $searchkey = '<!--LINK '.$key.'-->';
3080                                 $title = $wgLinkHolders['titles'][$key];
3081                                 if ( empty( $colours[$pdbk] ) ) {
3082                                         $wgLinkCache->addBadLink( $pdbk );
3083                                         $colours[$pdbk] = 0;
3084                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3085                                                                         $wgLinkHolders['texts'][$key],
3086                                                                         $wgLinkHolders['queries'][$key] );
3087                                 } elseif ( $colours[$pdbk] == 1 ) {
3088                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3089                                                                         $wgLinkHolders['texts'][$key],
3090                                                                         $wgLinkHolders['queries'][$key] );
3091                                 } elseif ( $colours[$pdbk] == 2 ) {
3092                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3093                                                                         $wgLinkHolders['texts'][$key],
3094                                                                         $wgLinkHolders['queries'][$key] );
3095                                 }
3096                         }
3097                         wfProfileOut( $fname.'-construct' );
3098
3099                         # Do the thing
3100                         wfProfileIn( $fname.'-replace' );
3101
3102                         $text = preg_replace_callback(
3103                                 '/(<!--LINK .*?-->)/',
3104                                 "outputReplaceMatches",
3105                                 $text);
3106                         wfProfileOut( $fname.'-replace' );
3107                 }
3108
3109                 if ( !empty( $wgInterwikiLinkHolders ) ) {
3110                         wfProfileIn( $fname.'-interwiki' );
3111                         $outputReplace = $wgInterwikiLinkHolders;
3112                         $text = preg_replace_callback(
3113                                 '/<!--IWLINK (.*?)-->/',
3114                                 "outputReplaceMatches",
3115                                 $text );
3116                         wfProfileOut( $fname.'-interwiki' );
3117                 }
3118
3119                 wfProfileOut( $fname );
3120                 return $colours;
3121         }
3122
3123         /**
3124          * Renders an image gallery from a text with one line per image.
3125          * text labels may be given by using |-style alternative text. E.g.
3126          *   Image:one.jpg|The number "1"
3127          *   Image:tree.jpg|A tree
3128          * given as text will return the HTML of a gallery with two images,
3129          * labeled 'The number "1"' and
3130          * 'A tree'.
3131          */
3132         function renderImageGallery( $text ) {
3133                 global $wgLinkCache;
3134                 $ig = new ImageGallery();
3135                 $ig->setShowBytes( false );
3136                 $ig->setShowFilename( false );
3137                 $lines = explode( "\n", $text );
3138
3139                 foreach ( $lines as $line ) {
3140                         # match lines like these:
3141                         # Image:someimage.jpg|This is some image
3142                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3143                         # Skip empty lines
3144                         if ( count( $matches ) == 0 ) {
3145                                 continue;
3146                         }
3147                         $nt = Title::newFromURL( $matches[1] );
3148                         if( is_null( $nt ) ) {
3149                                 # Bogus title. Ignore these so we don't bomb out later.
3150                                 continue;
3151                         }
3152                         if ( isset( $matches[3] ) ) {
3153                                 $label = $matches[3];
3154                         } else {
3155                                 $label = '';
3156                         }
3157
3158                         # FIXME: Use the full wiki parser and add its links
3159                         # to the page's links.
3160                         $html = $this->mOptions->mSkin->formatComment( $label );
3161
3162                         $ig->add( Image::newFromTitle( $nt ), $html );
3163                         $wgLinkCache->addImageLinkObj( $nt );
3164                 }
3165                 return $ig->toHTML();
3166         }
3167 }
3168
3169 /**
3170  * @todo document
3171  * @package MediaWiki
3172  */
3173 class ParserOutput
3174 {
3175         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3176         var $mCacheTime; # Used in ParserCache
3177         var $mVersion;   # Compatibility check
3178
3179         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3180                 $containsOldMagic = false )
3181         {
3182                 $this->mText = $text;
3183                 $this->mLanguageLinks = $languageLinks;
3184                 $this->mCategoryLinks = $categoryLinks;
3185                 $this->mContainsOldMagic = $containsOldMagic;
3186                 $this->mCacheTime = '';
3187                 $this->mVersion = MW_PARSER_VERSION;
3188         }
3189
3190         function getText()                   { return $this->mText; }
3191         function getLanguageLinks()          { return $this->mLanguageLinks; }
3192         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
3193         function getCacheTime()              { return $this->mCacheTime; }
3194         function containsOldMagic()          { return $this->mContainsOldMagic; }
3195         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3196         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3197         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3198         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3199         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3200         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3201
3202         function merge( $other ) {
3203                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3204                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3205                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3206         }
3207
3208         /**
3209          * Return true if this cached output object predates the global or
3210          * per-article cache invalidation timestamps, or if it comes from
3211          * an incompatible older version.
3212          *
3213          * @param string $touched the affected article's last touched timestamp
3214          * @return bool
3215          * @access public
3216          */
3217         function expired( $touched ) {
3218                 global $wgCacheEpoch;
3219                 return $this->getCacheTime() <= $touched ||
3220                        $this->getCacheTime() <= $wgCacheEpoch ||
3221                        !isset( $this->mVersion ) ||
3222                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3223         }
3224 }
3225
3226 /**
3227  * Set options of the Parser
3228  * @todo document
3229  * @package MediaWiki
3230  */
3231 class ParserOptions
3232 {
3233         # All variables are private
3234         var $mUseTeX;                    # Use texvc to expand <math> tags
3235         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3236         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3237         var $mAllowExternalImages;       # Allow external images inline
3238         var $mSkin;                      # Reference to the preferred skin
3239         var $mDateFormat;                # Date format index
3240         var $mEditSection;               # Create "edit section" links
3241         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3242         var $mNumberHeadings;            # Automatically number headings
3243         var $mShowToc;                   # Show table of contents
3244
3245         function getUseTeX()                        { return $this->mUseTeX; }
3246         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3247         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3248         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3249         function getSkin()                          { return $this->mSkin; }
3250         function getDateFormat()                    { return $this->mDateFormat; }
3251         function getEditSection()                   { return $this->mEditSection; }
3252         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3253         function getNumberHeadings()                { return $this->mNumberHeadings; }
3254         function getShowToc()                       { return $this->mShowToc; }
3255
3256         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3257         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3258         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3259         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3260         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3261         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3262         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3263         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3264         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3265
3266         function setSkin( &$x ) { $this->mSkin =& $x; }
3267
3268         # Get parser options
3269         /* static */ function newFromUser( &$user ) {
3270                 $popts = new ParserOptions;
3271                 $popts->initialiseFromUser( $user );
3272                 return $popts;
3273         }
3274
3275         # Get user options
3276         function initialiseFromUser( &$userInput ) {
3277                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3278                 $fname = 'ParserOptions::initialiseFromUser';
3279                 wfProfileIn( $fname );
3280                 if ( !$userInput ) {
3281                         $user = new User;
3282                         $user->setLoaded( true );
3283                 } else {
3284                         $user =& $userInput;
3285                 }
3286
3287                 $this->mUseTeX = $wgUseTeX;
3288                 $this->mUseDynamicDates = $wgUseDynamicDates;
3289                 $this->mInterwikiMagic = $wgInterwikiMagic;
3290                 $this->mAllowExternalImages = $wgAllowExternalImages;
3291                 wfProfileIn( $fname.'-skin' );
3292                 $this->mSkin =& $user->getSkin();
3293                 wfProfileOut( $fname.'-skin' );
3294                 $this->mDateFormat = $user->getOption( 'date' );
3295                 $this->mEditSection = $user->getOption( 'editsection' );
3296                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3297                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3298                 $this->mShowToc = $user->getOption( 'showtoc' );
3299                 wfProfileOut( $fname );
3300         }
3301
3302
3303 }
3304
3305 /**
3306  * Callback function used by Parser::replaceLinkHolders()
3307  * to substitute link placeholders.
3308  */
3309 function &outputReplaceMatches( $matches ) {
3310         global $outputReplace;
3311         return $outputReplace[$matches[1]];
3312 }
3313
3314 /**
3315  * Return the total number of articles
3316  */
3317 function wfNumberOfArticles() {
3318         global $wgNumberOfArticles;
3319
3320         wfLoadSiteStats();
3321         return $wgNumberOfArticles;
3322 }
3323
3324 /**
3325  * Get various statistics from the database
3326  * @private
3327  */
3328 function wfLoadSiteStats() {
3329         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3330         $fname = 'wfLoadSiteStats';
3331
3332         if ( -1 != $wgNumberOfArticles ) return;
3333         $dbr =& wfGetDB( DB_SLAVE );
3334         $s = $dbr->selectRow( 'site_stats',
3335                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3336                 array( 'ss_row_id' => 1 ), $fname
3337         );
3338
3339         if ( $s === false ) {
3340                 return;
3341         } else {
3342                 $wgTotalViews = $s->ss_total_views;
3343                 $wgTotalEdits = $s->ss_total_edits;
3344                 $wgNumberOfArticles = $s->ss_good_articles;
3345         }
3346 }
3347
3348 function wfEscapeHTMLTagsOnly( $in ) {
3349         return str_replace(
3350                 array( '"', '>', '<' ),
3351                 array( '&quot;', '&gt;', '&lt;' ),
3352                 $in );
3353 }
3354
3355 ?>