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                         # No link text, e.g. [http://domain.tld/some.link]
 996                         if ( $text == '' ) {
 997                                 # Autonumber if allowed
 998                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 999                                         $text = '[' . ++$this->mAutonumber . ']';
1000                                 } else {
1001                                         # Otherwise just use the URL
1002                                         $text = htmlspecialchars( $url );
1003                                 }
1004                         } else {
1005                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1006                                 # Check for trail
1007                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
1008                                         $dtrail = $m2[1];
1009                                         $trail = $m2[2];
1010                                 }
1011                         }
1012
1013                         $encUrl = htmlspecialchars( $url );
1014                         # Bit in parentheses showing the URL for the printable version
1015                         if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
1016                                 $paren = '';
1017                         } else {
1018                                 # Expand the URL for printable version
1019                                 if ( ! $sk->suppressUrlExpansion() ) {
1020                                         $paren = "<span class='urlexpansion'>&nbsp;(<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
1021                                 } else {
1022                                         $paren = '';
1023                                 }
1024                         }
1025
1026                         # Process the trail (i.e. everything after this link up until start of the next link),
1027                         # replacing any non-bracketed links
1028                         $trail = $this->replaceFreeExternalLinks( $trail );
1029
1030                         # Use the encoded URL
1031                         # This means that users can paste URLs directly into the text
1032                         # Funny characters like &ouml; aren't valid in URLs anyway
1033                         # This was changed in August 2004
1034                         $s .= $sk->makeExternalLink( $url, $text, false ) . $dtrail. $paren . $trail;
1035                 }
1036
1037                 wfProfileOut( $fname );
1038                 return $s;
1039         }
1040
1041         /**
1042          * Replace anything that looks like a URL with a link
1043          * @access private
1044          */
1045         function replaceFreeExternalLinks( $text ) {
1046                 $fname = 'Parser::replaceFreeExternalLinks';
1047                 wfProfileIn( $fname );
1048
1049                 $bits = preg_split( '/(\b(?:'.URL_PROTOCOLS.'):)/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1050                 $s = array_shift( $bits );
1051                 $i = 0;
1052
1053                 $sk =& $this->mOptions->getSkin();
1054
1055                 while ( $i < count( $bits ) ){
1056                         $protocol = $bits[$i++];
1057                         $remainder = $bits[$i++];
1058
1059                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1060                                 # Found some characters after the protocol that look promising
1061                                 $url = $protocol . $m[1];
1062                                 $trail = $m[2];
1063
1064                                 # The characters '<' and '>' (which were escaped by
1065                                 # removeHTMLtags()) should not be included in
1066                                 # URLs, per RFC 2396.
1067                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1068                                         $trail = substr($url, $m2[0][1]) . $trail;
1069                                         $url = substr($url, 0, $m2[0][1]);
1070                                 }
1071
1072                                 # Move trailing punctuation to $trail
1073                                 $sep = ',;\.:!?';
1074                                 # If there is no left bracket, then consider right brackets fair game too
1075                                 if ( strpos( $url, '(' ) === false ) {
1076                                         $sep .= ')';
1077                                 }
1078
1079                                 $numSepChars = strspn( strrev( $url ), $sep );
1080                                 if ( $numSepChars ) {
1081                                         $trail = substr( $url, -$numSepChars ) . $trail;
1082                                         $url = substr( $url, 0, -$numSepChars );
1083                                 }
1084
1085                                 # Replace &amp; from obsolete syntax with &.
1086                                 # All HTML entities will be escaped by makeExternalLink()
1087                                 # or maybeMakeImageLink()
1088                                 $url = str_replace( '&amp;', '&', $url );
1089
1090                                 # Is this an external image?
1091                                 $text = $this->maybeMakeImageLink( $url );
1092                                 if ( $text === false ) {
1093                                         # Not an image, make a link
1094                                         $text = $sk->makeExternalLink( $url, $url );
1095                                 }
1096                                 $s .= $text . $trail;
1097                         } else {
1098                                 $s .= $protocol . $remainder;
1099                         }
1100                 }
1101                 wfProfileOut();
1102                 return $s;
1103         }
1104
1105         /**
1106          * make an image if it's allowed
1107          * @access private
1108          */
1109         function maybeMakeImageLink( $url ) {
1110                 $sk =& $this->mOptions->getSkin();
1111                 $text = false;
1112                 if ( $this->mOptions->getAllowExternalImages() ) {
1113                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1114                                 # Image found
1115                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
1116                         }
1117                 }
1118                 return $text;
1119         }
1120
1121         /**
1122          * Process [[ ]] wikilinks
1123          *
1124          * @access private
1125          */
1126
1127         function replaceInternalLinks( $s ) {
1128                 global $wgLang, $wgContLang, $wgLinkCache;
1129                 global $wgDisableLangConversion;
1130                 static $fname = 'Parser::replaceInternalLinks' ;
1131
1132                 wfProfileIn( $fname );
1133
1134                 wfProfileIn( $fname.'-setup' );
1135                 static $tc = FALSE;
1136                 # the % is needed to support urlencoded titles as well
1137                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1138
1139                 $sk =& $this->mOptions->getSkin();
1140                 global $wgUseOldExistenceCheck;
1141                 # "Post-parse link colour check" works only on wiki text since it's now
1142                 # in Parser. Enable it, then disable it when we're done.
1143                 $saveParseColour = $sk->postParseLinkColour( !$wgUseOldExistenceCheck );
1144
1145                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1146
1147                 #split the entire text string on occurences of [[
1148                 $a = explode( '[[', ' ' . $s );
1149                 #get the first element (all text up to first [[), and remove the space we added
1150                 $s = array_shift( $a );
1151                 $s = substr( $s, 1 );
1152
1153                 # Match a link having the form [[namespace:link|alternate]]trail
1154                 static $e1 = FALSE;
1155                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1156                 # Match cases where there is no "]]", which might still be images
1157                 static $e1_img = FALSE;
1158                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1159                 # Match the end of a line for a word that's not followed by whitespace,
1160                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1161                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1162
1163                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1164
1165                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1166
1167                 if ( $useLinkPrefixExtension ) {
1168                         if ( preg_match( $e2, $s, $m ) ) {
1169                                 $first_prefix = $m[2];
1170                                 $s = $m[1];
1171                         } else {
1172                                 $first_prefix = false;
1173                         }
1174                 } else {
1175                         $prefix = '';
1176                 }
1177
1178                 $selflink = $this->mTitle->getPrefixedText();
1179                 wfProfileOut( $fname.'-setup' );
1180
1181                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1182                 $useSubpages = $this->areSubpagesAllowed();
1183
1184                 # Loop for each link
1185                 for ($k = 0; isset( $a[$k] ); $k++) {
1186                         $line = $a[$k];
1187                         if ( $useLinkPrefixExtension ) {
1188                                 wfProfileIn( $fname.'-prefixhandling' );
1189                                 if ( preg_match( $e2, $s, $m ) ) {
1190                                         $prefix = $m[2];
1191                                         $s = $m[1];
1192                                 } else {
1193                                         $prefix='';
1194                                 }
1195                                 # first link
1196                                 if($first_prefix) {
1197                                         $prefix = $first_prefix;
1198                                         $first_prefix = false;
1199                                 }
1200                                 wfProfileOut( $fname.'-prefixhandling' );
1201                         }
1202
1203                         $might_be_img = false;
1204
1205                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1206                                 $text = $m[2];
1207                                 # fix up urlencoded title texts
1208                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1209                                 $trail = $m[3];
1210                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1211                                 $might_be_img = true;
1212                                 $text = $m[2];
1213                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1214                                 $trail = "";
1215                         } else { # Invalid form; output directly
1216                                 $s .= $prefix . '[[' . $line ;
1217                                 continue;
1218                         }
1219
1220                         # Don't allow internal links to pages containing
1221                         # PROTO: where PROTO is a valid URL protocol; these
1222                         # should be external links.
1223                         if (preg_match('/^(\b(?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1224                                 $s .= $prefix . '[[' . $line ;
1225                                 continue;
1226                         }
1227
1228                         # Make subpage if necessary
1229                         if( $useSubpages ) {
1230                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1231                         } else {
1232                                 $link = $m[1];
1233                         }
1234
1235                         $noforce = (substr($m[1], 0, 1) != ':');
1236                         if (!$noforce) {
1237                                 # Strip off leading ':'
1238                                 $link = substr($link, 1);
1239                         }
1240
1241                         $nt =& Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1242                         if( !$nt ) {
1243                                 $s .= $prefix . '[[' . $line;
1244                                 continue;
1245                         }
1246
1247                         #check other language variants of the link
1248                         #if the article does not exist
1249                         if( $checkVariantLink
1250                             && $nt->getArticleID() == 0 ) {
1251                                 $wgContLang->findVariantLink($link, $nt);
1252                         }
1253
1254                         $ns = $nt->getNamespace();
1255                         $iw = $nt->getInterWiki();
1256
1257                         if ($might_be_img) { # if this is actually an invalid link
1258                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1259                                         $found = false;
1260                                         while (isset ($a[$k+1]) ) {
1261                                                 #look at the next 'line' to see if we can close it there
1262                                                 $next_line =  array_shift(array_splice( $a, $k + 1, 1) );
1263                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1264                                                 # the first ]] closes the inner link, the second the image
1265                                                         $found = true;
1266                                                         $text .= '[[' . $m[1];
1267                                                         $trail = $m[2];
1268                                                         break;
1269                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1270                                                         #if there's exactly one ]] that's fine, we'll keep looking
1271                                                         $text .= '[[' . $m[0];
1272                                                 } else {
1273                                                         #if $next_line is invalid too, we need look no further
1274                                                         $text .= '[[' . $next_line;
1275                                                         break;
1276                                                 }
1277                                         }
1278                                         if ( !$found ) {
1279                                                 # we couldn't find the end of this imageLink, so output it raw
1280                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1281                                                 $text = $this->replaceInternalLinks($text);
1282                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1283                                                 # note: no $trail, because without an end, there *is* no trail
1284                                                 continue;
1285                                         }
1286                                 } else { #it's not an image, so output it raw
1287                                         $s .= $prefix . '[[' . $link . '|' . $text;
1288                                         # note: no $trail, because without an end, there *is* no trail
1289                                         continue;
1290                                 }
1291                         }
1292
1293                         $wasblank = ( '' == $text );
1294                         if( $wasblank ) $text = $link;
1295
1296
1297                         # Link not escaped by : , create the various objects
1298                         if( $noforce ) {
1299
1300                                 # Interwikis
1301                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1302                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1303                                         $tmp = $prefix . $trail ;
1304                                         $s .= (trim($tmp) == '')? '': $tmp;
1305                                         continue;
1306                                 }
1307
1308                                 if ( $ns == NS_IMAGE ) {
1309                                         wfProfileIn( "$fname-image" );
1310
1311                                         # recursively parse links inside the image caption
1312                                         # actually, this will parse them in any other parameters, too,
1313                                         # but it might be hard to fix that, and it doesn't matter ATM
1314                                         $text = $this->replaceExternalLinks($text);
1315                                         $text = $this->replaceInternalLinks($text);
1316
1317                                         # replace the image with a link-holder so that replaceExternalLinks() can't mess with it
1318                                         $s .= $prefix . $this->insertStripItem( $sk->makeImageLinkObj( $nt, $text ), $this->mStripState ) . $trail;
1319                                         $wgLinkCache->addImageLinkObj( $nt );
1320
1321                                         wfProfileOut( "$fname-image" );
1322                                         continue;
1323                                 }
1324
1325                                 if ( $ns == NS_CATEGORY ) {
1326                                         wfProfileIn( "$fname-category" );
1327                                         $t = $nt->getText();
1328
1329                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1330                                         $pPLC=$sk->postParseLinkColour();
1331                                         $sk->postParseLinkColour( false );
1332                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1333                                         $sk->postParseLinkColour( $pPLC );
1334                                         $wgLinkCache->resume();
1335
1336                                         if ( $wasblank ) {
1337                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1338                                                         $sortkey = $this->mTitle->getText();
1339                                                 } else {
1340                                                         $sortkey = $this->mTitle->getPrefixedText();
1341                                                 }
1342                                         } else {
1343                                                 $sortkey = $text;
1344                                         }
1345                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1346                                         $this->mOutput->addCategoryLink( $t );
1347                                         $s .= $prefix . $trail ;
1348
1349                                         wfProfileOut( "$fname-category" );
1350                                         continue;
1351                                 }
1352                         }
1353
1354                         if( ( $nt->getPrefixedText() === $selflink ) &&
1355                             ( $nt->getFragment() === '' ) ) {
1356                                 # Self-links are handled specially; generally de-link and change to bold.
1357                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1358                                 continue;
1359                         }
1360
1361                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1362                         if( $ns == NS_MEDIA ) {
1363                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1364                                 $wgLinkCache->addImageLinkObj( $nt );
1365                                 continue;
1366                         } elseif( $ns == NS_SPECIAL ) {
1367                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1368                                 continue;
1369                         }
1370                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1371                 }
1372                 $sk->postParseLinkColour( $saveParseColour );
1373                 wfProfileOut( $fname );
1374                 return $s;
1375         }
1376
1377         /**
1378          * Return true if subpage links should be expanded on this page.
1379          * @return bool
1380          */
1381         function areSubpagesAllowed() {
1382                 # Some namespaces don't allow subpages
1383                 global $wgNamespacesWithSubpages;
1384                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1385         }
1386
1387         /**
1388          * Handle link to subpage if necessary
1389          * @param string $target the source of the link
1390          * @param string &$text the link text, modified as necessary
1391          * @return string the full name of the link
1392          * @access private
1393          */
1394         function maybeDoSubpageLink($target, &$text) {
1395                 # Valid link forms:
1396                 # Foobar -- normal
1397                 # :Foobar -- override special treatment of prefix (images, language links)
1398                 # /Foobar -- convert to CurrentPage/Foobar
1399                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1400                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1401                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1402
1403                 $fname = 'Parser::maybeDoSubpageLink';
1404                 wfProfileIn( $fname );
1405                 $ret = $target; # default return value is no change
1406
1407                 # Some namespaces don't allow subpages,
1408                 # so only perform processing if subpages are allowed
1409                 if( $this->areSubpagesAllowed() ) {
1410                         # Look at the first character
1411                         if( $target != '' && $target{0} == '/' ) {
1412                                 # / at end means we don't want the slash to be shown
1413                                 if( substr( $target, -1, 1 ) == '/' ) {
1414                                         $target = substr( $target, 1, -1 );
1415                                         $noslash = $target;
1416                                 } else {
1417                                         $noslash = substr( $target, 1 );
1418                                 }
1419
1420                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1421                                 if( '' === $text ) {
1422                                         $text = $target;
1423                                 } # this might be changed for ugliness reasons
1424                         } else {
1425                                 # check for .. subpage backlinks
1426                                 $dotdotcount = 0;
1427                                 $nodotdot = $target;
1428                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1429                                         ++$dotdotcount;
1430                                         $nodotdot = substr( $nodotdot, 3 );
1431                                 }
1432                                 if($dotdotcount > 0) {
1433                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1434                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1435                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1436                                                 # / at the end means don't show full path
1437                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1438                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1439                                                         if( '' === $text ) {
1440                                                                 $text = $nodotdot;
1441                                                         }
1442                                                 }
1443                                                 $nodotdot = trim( $nodotdot );
1444                                                 if( $nodotdot != '' ) {
1445                                                         $ret .= '/' . $nodotdot;
1446                                                 }
1447                                         }
1448                                 }
1449                         }
1450                 }
1451
1452                 wfProfileOut( $fname );
1453                 return $ret;
1454         }
1455
1456         /**#@+
1457          * Used by doBlockLevels()
1458          * @access private
1459          */
1460         /* private */ function closeParagraph() {
1461                 $result = '';
1462                 if ( '' != $this->mLastSection ) {
1463                         $result = '</' . $this->mLastSection  . ">\n";
1464                 }
1465                 $this->mInPre = false;
1466                 $this->mLastSection = '';
1467                 return $result;
1468         }
1469         # getCommon() returns the length of the longest common substring
1470         # of both arguments, starting at the beginning of both.
1471         #
1472         /* private */ function getCommon( $st1, $st2 ) {
1473                 $fl = strlen( $st1 );
1474                 $shorter = strlen( $st2 );
1475                 if ( $fl < $shorter ) { $shorter = $fl; }
1476
1477                 for ( $i = 0; $i < $shorter; ++$i ) {
1478                         if ( $st1{$i} != $st2{$i} ) { break; }
1479                 }
1480                 return $i;
1481         }
1482         # These next three functions open, continue, and close the list
1483         # element appropriate to the prefix character passed into them.
1484         #
1485         /* private */ function openList( $char ) {
1486                 $result = $this->closeParagraph();
1487
1488                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1489                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1490                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1491                 else if ( ';' == $char ) {
1492                         $result .= '<dl><dt>';
1493                         $this->mDTopen = true;
1494                 }
1495                 else { $result = '<!-- ERR 1 -->'; }
1496
1497                 return $result;
1498         }
1499
1500         /* private */ function nextItem( $char ) {
1501                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1502                 else if ( ':' == $char || ';' == $char ) {
1503                         $close = '</dd>';
1504                         if ( $this->mDTopen ) { $close = '</dt>'; }
1505                         if ( ';' == $char ) {
1506                                 $this->mDTopen = true;
1507                                 return $close . '<dt>';
1508                         } else {
1509                                 $this->mDTopen = false;
1510                                 return $close . '<dd>';
1511                         }
1512                 }
1513                 return '<!-- ERR 2 -->';
1514         }
1515
1516         /* private */ function closeList( $char ) {
1517                 if ( '*' == $char ) { $text = '</li></ul>'; }
1518                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1519                 else if ( ':' == $char ) {
1520                         if ( $this->mDTopen ) {
1521                                 $this->mDTopen = false;
1522                                 $text = '</dt></dl>';
1523                         } else {
1524                                 $text = '</dd></dl>';
1525                         }
1526                 }
1527                 else {  return '<!-- ERR 3 -->'; }
1528                 return $text."\n";
1529         }
1530         /**#@-*/
1531
1532         /**
1533          * Make lists from lines starting with ':', '*', '#', etc.
1534          *
1535          * @access private
1536          * @return string the lists rendered as HTML
1537          */
1538         function doBlockLevels( $text, $linestart ) {
1539                 $fname = 'Parser::doBlockLevels';
1540                 wfProfileIn( $fname );
1541
1542                 # Parsing through the text line by line.  The main thing
1543                 # happening here is handling of block-level elements p, pre,
1544                 # and making lists from lines starting with * # : etc.
1545                 #
1546                 $textLines = explode( "\n", $text );
1547
1548                 $lastPrefix = $output = $lastLine = '';
1549                 $this->mDTopen = $inBlockElem = false;
1550                 $prefixLength = 0;
1551                 $paragraphStack = false;
1552
1553                 if ( !$linestart ) {
1554                         $output .= array_shift( $textLines );
1555                 }
1556                 foreach ( $textLines as $oLine ) {
1557                         $lastPrefixLength = strlen( $lastPrefix );
1558                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1559                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1560                         if ( !$this->mInPre ) {
1561                                 # Multiple prefixes may abut each other for nested lists.
1562                                 $prefixLength = strspn( $oLine, '*#:;' );
1563                                 $pref = substr( $oLine, 0, $prefixLength );
1564
1565                                 # eh?
1566                                 $pref2 = str_replace( ';', ':', $pref );
1567                                 $t = substr( $oLine, $prefixLength );
1568                                 $this->mInPre = !empty($preOpenMatch);
1569                         } else {
1570                                 # Don't interpret any other prefixes in preformatted text
1571                                 $prefixLength = 0;
1572                                 $pref = $pref2 = '';
1573                                 $t = $oLine;
1574                         }
1575
1576                         # List generation
1577                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1578                                 # Same as the last item, so no need to deal with nesting or opening stuff
1579                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1580                                 $paragraphStack = false;
1581
1582                                 if ( substr( $pref, -1 ) == ';') {
1583                                         # The one nasty exception: definition lists work like this:
1584                                         # ; title : definition text
1585                                         # So we check for : in the remainder text to split up the
1586                                         # title and definition, without b0rking links.
1587                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1588                                                 $t = $t2;
1589                                                 $output .= $term . $this->nextItem( ':' );
1590                                         }
1591                                 }
1592                         } elseif( $prefixLength || $lastPrefixLength ) {
1593                                 # Either open or close a level...
1594                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1595                                 $paragraphStack = false;
1596
1597                                 while( $commonPrefixLength < $lastPrefixLength ) {
1598                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1599                                         --$lastPrefixLength;
1600                                 }
1601                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1602                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1603                                 }
1604                                 while ( $prefixLength > $commonPrefixLength ) {
1605                                         $char = substr( $pref, $commonPrefixLength, 1 );
1606                                         $output .= $this->openList( $char );
1607
1608                                         if ( ';' == $char ) {
1609                                                 # FIXME: This is dupe of code above
1610                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1611                                                         $t = $t2;
1612                                                         $output .= $term . $this->nextItem( ':' );
1613                                                 }
1614                                         }
1615                                         ++$commonPrefixLength;
1616                                 }
1617                                 $lastPrefix = $pref2;
1618                         }
1619                         if( 0 == $prefixLength ) {
1620                                 wfProfileIn( "$fname-paragraph" );
1621                                 # No prefix (not in list)--go to paragraph mode
1622                                 $uniq_prefix = UNIQ_PREFIX;
1623                                 // XXX: use a stack for nestable elements like span, table and div
1624                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1625                                 $closematch = preg_match(
1626                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1627                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1628                                 if ( $openmatch or $closematch ) {
1629                                         $paragraphStack = false;
1630                                         $output .= $this->closeParagraph();
1631                                         if($preOpenMatch and !$preCloseMatch) {
1632                                                 $this->mInPre = true;
1633                                         }
1634                                         if ( $closematch ) {
1635                                                 $inBlockElem = false;
1636                                         } else {
1637                                                 $inBlockElem = true;
1638                                         }
1639                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1640                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1641                                                 // pre
1642                                                 if ($this->mLastSection != 'pre') {
1643                                                         $paragraphStack = false;
1644                                                         $output .= $this->closeParagraph().'<pre>';
1645                                                         $this->mLastSection = 'pre';
1646                                                 }
1647                                                 $t = substr( $t, 1 );
1648                                         } else {
1649                                                 // paragraph
1650                                                 if ( '' == trim($t) ) {
1651                                                         if ( $paragraphStack ) {
1652                                                                 $output .= $paragraphStack.'<br />';
1653                                                                 $paragraphStack = false;
1654                                                                 $this->mLastSection = 'p';
1655                                                         } else {
1656                                                                 if ($this->mLastSection != 'p' ) {
1657                                                                         $output .= $this->closeParagraph();
1658                                                                         $this->mLastSection = '';
1659                                                                         $paragraphStack = '<p>';
1660                                                                 } else {
1661                                                                         $paragraphStack = '</p><p>';
1662                                                                 }
1663                                                         }
1664                                                 } else {
1665                                                         if ( $paragraphStack ) {
1666                                                                 $output .= $paragraphStack;
1667                                                                 $paragraphStack = false;
1668                                                                 $this->mLastSection = 'p';
1669                                                         } else if ($this->mLastSection != 'p') {
1670                                                                 $output .= $this->closeParagraph().'<p>';
1671                                                                 $this->mLastSection = 'p';
1672                                                         }
1673                                                 }
1674                                         }
1675                                 }
1676                                 wfProfileOut( "$fname-paragraph" );
1677                         }
1678                         if ($paragraphStack === false) {
1679                                 $output .= $t."\n";
1680                         }
1681                 }
1682                 while ( $prefixLength ) {
1683                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1684                         --$prefixLength;
1685                 }
1686                 if ( '' != $this->mLastSection ) {
1687                         $output .= '</' . $this->mLastSection . '>';
1688                         $this->mLastSection = '';
1689                 }
1690
1691                 wfProfileOut( $fname );
1692                 return $output;
1693         }
1694
1695         /**
1696          * Split up a string on ':', ignoring any occurences inside
1697          * <a>..</a> or <span>...</span>
1698          * @param string $str the string to split
1699          * @param string &$before set to everything before the ':'
1700          * @param string &$after set to everything after the ':'
1701          * return string the position of the ':', or false if none found
1702          */
1703         function findColonNoLinks($str, &$before, &$after) {
1704                 # I wonder if we should make this count all tags, not just <a>
1705                 # and <span>. That would prevent us from matching a ':' that
1706                 # comes in the middle of italics other such formatting....
1707                 # -- Wil
1708                 $fname = 'Parser::findColonNoLinks';
1709                 wfProfileIn( $fname );
1710                 $pos = 0;
1711                 do {
1712                         $colon = strpos($str, ':', $pos);
1713
1714                         if ($colon !== false) {
1715                                 $before = substr($str, 0, $colon);
1716                                 $after = substr($str, $colon + 1);
1717
1718                                 # Skip any ':' within <a> or <span> pairs
1719                                 $a = substr_count($before, '<a');
1720                                 $s = substr_count($before, '<span');
1721                                 $ca = substr_count($before, '</a>');
1722                                 $cs = substr_count($before, '</span>');
1723
1724                                 if ($a <= $ca and $s <= $cs) {
1725                                         # Tags are balanced before ':'; ok
1726                                         break;
1727                                 }
1728                                 $pos = $colon + 1;
1729                         }
1730                 } while ($colon !== false);
1731                 wfProfileOut( $fname );
1732                 return $colon;
1733         }
1734
1735         /**
1736          * Return value of a magic variable (like PAGENAME)
1737          *
1738          * @access private
1739          */
1740         function getVariableValue( $index ) {
1741                 global $wgContLang, $wgSitename, $wgServer;
1742
1743                 /**
1744                  * Some of these require message or data lookups and can be
1745                  * expensive to check many times.
1746                  */
1747                 static $varCache = array();
1748                 if( isset( $varCache[$index] ) ) return $varCache[$index];
1749
1750                 switch ( $index ) {
1751                         case MAG_CURRENTMONTH:
1752                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm' ) );
1753                         case MAG_CURRENTMONTHNAME:
1754                                 return $varCache[$index] = $wgContLang->getMonthName( date('n') );
1755                         case MAG_CURRENTMONTHNAMEGEN:
1756                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date('n') );
1757                         case MAG_CURRENTDAY:
1758                                 return $varCache[$index] = $wgContLang->formatNum( date('j') );
1759                         case MAG_PAGENAME:
1760                                 return $this->mTitle->getText();
1761                         case MAG_PAGENAMEE:
1762                                 return $this->mTitle->getPartialURL();
1763                         case MAG_NAMESPACE:
1764                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1765                                 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1766                         case MAG_CURRENTDAYNAME:
1767                                 return $varCache[$index] = $wgContLang->getWeekdayName( date('w')+1 );
1768                         case MAG_CURRENTYEAR:
1769                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y' ) );
1770                         case MAG_CURRENTTIME:
1771                                 return $varCache[$index] = $wgContLang->time( wfTimestampNow(), false );
1772                         case MAG_CURRENTWEEK:
1773                                 return $varCache[$index] = $wgContLang->formatNum( date('W') );
1774                         case MAG_CURRENTDOW:
1775                                 return $varCache[$index] = $wgContLang->formatNum( date('w') );
1776                         case MAG_NUMBEROFARTICLES:
1777                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1778                         case MAG_SITENAME:
1779                                 return $wgSitename;
1780                         case MAG_SERVER:
1781                                 return $wgServer;
1782                         default:
1783                                 return NULL;
1784                 }
1785         }
1786
1787         /**
1788          * initialise the magic variables (like CURRENTMONTHNAME)
1789          *
1790          * @access private
1791          */
1792         function initialiseVariables() {
1793                 $fname = 'Parser::initialiseVariables';
1794                 wfProfileIn( $fname );
1795                 global $wgVariableIDs;
1796                 $this->mVariables = array();
1797                 foreach ( $wgVariableIDs as $id ) {
1798                         $mw =& MagicWord::get( $id );
1799                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1800                 }
1801                 wfProfileOut( $fname );
1802         }
1803
1804         /**
1805          * Replace magic variables, templates, and template arguments
1806          * with the appropriate text. Templates are substituted recursively,
1807          * taking care to avoid infinite loops.
1808          *
1809          * Note that the substitution depends on value of $mOutputType:
1810          *  OT_WIKI: only {{subst:}} templates
1811          *  OT_MSG: only magic variables
1812          *  OT_HTML: all templates and magic variables
1813          *
1814          * @param string $tex The text to transform
1815          * @param array $args Key-value pairs representing template parameters to substitute
1816          * @access private
1817          */
1818         function replaceVariables( $text, $args = array() ) {
1819                 global $wgLang, $wgScript, $wgArticlePath;
1820
1821                 # Prevent too big inclusions
1822                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
1823                         return $text;
1824                 }
1825
1826                 $fname = 'Parser::replaceVariables';
1827                 wfProfileIn( $fname );
1828
1829                 $titleChars = Title::legalChars();
1830
1831                 # This function is called recursively. To keep track of arguments we need a stack:
1832                 array_push( $this->mArgStack, $args );
1833
1834                 # Variable substitution
1835                 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", array( &$this, 'variableSubstitution' ), $text );
1836
1837                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1838                         # Argument substitution
1839                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", array( &$this, 'argSubstitution' ), $text );
1840                 }
1841                 # Template substitution
1842                 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1843                 $text = preg_replace_callback( $regex, array( &$this, 'braceSubstitution' ), $text );
1844
1845                 array_pop( $this->mArgStack );
1846
1847                 wfProfileOut( $fname );
1848                 return $text;
1849         }
1850
1851         /**
1852          * Replace magic variables
1853          * @access private
1854          */
1855         function variableSubstitution( $matches ) {
1856                 $fname = 'parser::variableSubstitution';
1857                 wfProfileIn( $fname );
1858                 if ( !$this->mVariables ) {
1859                         $this->initialiseVariables();
1860                 }
1861                 $skip = false;
1862                 if ( $this->mOutputType == OT_WIKI ) {
1863                         # Do only magic variables prefixed by SUBST
1864                         $mwSubst =& MagicWord::get( MAG_SUBST );
1865                         if (!$mwSubst->matchStartAndRemove( $matches[1] ))
1866                                 $skip = true;
1867                         # Note that if we don't substitute the variable below,
1868                         # we don't remove the {{subst:}} magic word, in case
1869                         # it is a template rather than a magic variable.
1870                 }
1871                 if ( !$skip && array_key_exists( $matches[1], $this->mVariables ) ) {
1872                         $text = $this->mVariables[$matches[1]];
1873                         $this->mOutput->mContainsOldMagic = true;
1874                 } else {
1875                         $text = $matches[0];
1876                 }
1877                 wfProfileOut( $fname );
1878                 return $text;
1879         }
1880
1881         # Split template arguments
1882         function getTemplateArgs( $argsString ) {
1883                 if ( $argsString === '' ) {
1884                         return array();
1885                 }
1886
1887                 $args = explode( '|', substr( $argsString, 1 ) );
1888
1889                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1890                 # merged with the next arg because the '|' character between belongs
1891                 # to the link syntax and not the template parameter syntax.
1892                 $argc = count($args);
1893                 $i = 0;
1894                 for ( $i = 0; $i < $argc-1; $i++ ) {
1895                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1896                                 $args[$i] .= '|'.$args[$i+1];
1897                                 array_splice($args, $i+1, 1);
1898                                 $i--;
1899                                 $argc--;
1900                         }
1901                 }
1902
1903                 return $args;
1904         }
1905
1906         /**
1907          * Return the text of a template, after recursively
1908          * replacing any variables or templates within the template.
1909          *
1910          * @param array $matches The parts of the template
1911          *  $matches[1]: the title, i.e. the part before the |
1912          *  $matches[2]: the parameters (including a leading |), if  any
1913          * @return string the text of the template
1914          * @access private
1915          */
1916         function braceSubstitution( $matches ) {
1917                 global $wgLinkCache, $wgContLang;
1918                 $fname = 'Parser::braceSubstitution';
1919                 wfProfileIn( $fname );
1920
1921                 $found = false;
1922                 $nowiki = false;
1923                 $noparse = false;
1924
1925                 $title = NULL;
1926
1927                 # Need to know if the template comes at the start of a line,
1928                 # to treat the beginning of the template like the beginning
1929                 # of a line for tables and block-level elements.
1930                 $linestart = $matches[1];
1931
1932                 # $part1 is the bit before the first |, and must contain only title characters
1933                 # $args is a list of arguments, starting from index 0, not including $part1
1934
1935                 $part1 = $matches[2];
1936                 # If the third subpattern matched anything, it will start with |
1937
1938                 $args = $this->getTemplateArgs($matches[3]);
1939                 $argc = count( $args );
1940
1941                 # Don't parse {{{}}} because that's only for template arguments
1942                 if ( $linestart === '{' ) {
1943                         $text = $matches[0];
1944                         $found = true;
1945                         $noparse = true;
1946                 }
1947
1948                 # SUBST
1949                 if ( !$found ) {
1950                         $mwSubst =& MagicWord::get( MAG_SUBST );
1951                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1952                                 # One of two possibilities is true:
1953                                 # 1) Found SUBST but not in the PST phase
1954                                 # 2) Didn't find SUBST and in the PST phase
1955                                 # In either case, return without further processing
1956                                 $text = $matches[0];
1957                                 $found = true;
1958                                 $noparse = true;
1959                         }
1960                 }
1961
1962                 # MSG, MSGNW and INT
1963                 if ( !$found ) {
1964                         # Check for MSGNW:
1965                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1966                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1967                                 $nowiki = true;
1968                         } else {
1969                                 # Remove obsolete MSG:
1970                                 $mwMsg =& MagicWord::get( MAG_MSG );
1971                                 $mwMsg->matchStartAndRemove( $part1 );
1972                         }
1973
1974                         # Check if it is an internal message
1975                         $mwInt =& MagicWord::get( MAG_INT );
1976                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1977                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1978                                         $text = $linestart . wfMsgReal( $part1, $args, true );
1979                                         $found = true;
1980                                 }
1981                         }
1982                 }
1983
1984                 # NS
1985                 if ( !$found ) {
1986                         # Check for NS: (namespace expansion)
1987                         $mwNs = MagicWord::get( MAG_NS );
1988                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1989                                 if ( intval( $part1 ) ) {
1990                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1991                                         $found = true;
1992                                 } else {
1993                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1994                                         if ( !is_null( $index ) ) {
1995                                                 $text = $linestart . $wgContLang->getNsText( $index );
1996                                                 $found = true;
1997                                         }
1998                                 }
1999                         }
2000                 }
2001
2002                 # LOCALURL and LOCALURLE
2003                 if ( !$found ) {
2004                         $mwLocal = MagicWord::get( MAG_LOCALURL );
2005                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
2006
2007                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2008                                 $func = 'getLocalURL';
2009                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2010                                 $func = 'escapeLocalURL';
2011                         } else {
2012                                 $func = '';
2013                         }
2014
2015                         if ( $func !== '' ) {
2016                                 $title = Title::newFromText( $part1 );
2017                                 if ( !is_null( $title ) ) {
2018                                         if ( $argc > 0 ) {
2019                                                 $text = $linestart . $title->$func( $args[0] );
2020                                         } else {
2021                                                 $text = $linestart . $title->$func();
2022                                         }
2023                                         $found = true;
2024                                 }
2025                         }
2026                 }
2027
2028                 # GRAMMAR
2029                 if ( !$found && $argc == 1 ) {
2030                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2031                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2032                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2033                                 $found = true;
2034                         }
2035                 }
2036
2037                 # Template table test
2038
2039                 # Did we encounter this template already? If yes, it is in the cache
2040                 # and we need to check for loops.
2041                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2042                         # set $text to cached message.
2043                         $text = $linestart . $this->mTemplates[$part1];
2044                         $found = true;
2045
2046                         # Infinite loop test
2047                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2048                                 $noparse = true;
2049                                 $found = true;
2050                                 $text .= '<!-- WARNING: template loop detected -->';
2051                         }
2052                 }
2053
2054                 # Load from database
2055                 $itcamefromthedatabase = false;
2056                 if ( !$found ) {
2057                         $ns = NS_TEMPLATE;
2058                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2059                         if ($subpage !== '') {
2060                                 $ns = $this->mTitle->getNamespace();
2061                         }
2062                         $title = Title::newFromText( $part1, $ns );
2063                         if ( !is_null( $title ) && !$title->isExternal() ) {
2064                                 # Check for excessive inclusion
2065                                 $dbk = $title->getPrefixedDBkey();
2066                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2067                                         # This should never be reached.
2068                                         $article = new Article( $title );
2069                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
2070                                         if ( $articleContent !== false ) {
2071                                                 $found = true;
2072                                                 $text = $linestart . $articleContent;
2073                                                 $itcamefromthedatabase = true;
2074                                         }
2075                                 }
2076
2077                                 # If the title is valid but undisplayable, make a link to it
2078                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2079                                         $text = $linestart . '[['.$title->getPrefixedText().']]';
2080                                         $found = true;
2081                                 }
2082
2083                                 # Template cache array insertion
2084                                 $this->mTemplates[$part1] = $text;
2085                         }
2086                 }
2087
2088                 # Recursive parsing, escaping and link table handling
2089                 # Only for HTML output
2090                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2091                         $text = wfEscapeWikiText( $text );
2092                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2093                         # Clean up argument array
2094                         $assocArgs = array();
2095                         $index = 1;
2096                         foreach( $args as $arg ) {
2097                                 $eqpos = strpos( $arg, '=' );
2098                                 if ( $eqpos === false ) {
2099                                         $assocArgs[$index++] = $arg;
2100                                 } else {
2101                                         $name = trim( substr( $arg, 0, $eqpos ) );
2102                                         $value = trim( substr( $arg, $eqpos+1 ) );
2103                                         if ( $value === false ) {
2104                                                 $value = '';
2105                                         }
2106                                         if ( $name !== false ) {
2107                                                 $assocArgs[$name] = $value;
2108                                         }
2109                                 }
2110                         }
2111
2112                         # Add a new element to the templace recursion path
2113                         $this->mTemplatePath[$part1] = 1;
2114
2115                         $text = $this->strip( $text, $this->mStripState );
2116                         $text = $this->removeHTMLtags( $text );
2117                         $text = $this->replaceVariables( $text, $assocArgs );
2118
2119                         # Resume the link cache and register the inclusion as a link
2120                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2121                                 $wgLinkCache->addLinkObj( $title );
2122                         }
2123
2124                         # If the template begins with a table or block-level
2125                         # element, it should be treated as beginning a new line.
2126                         if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2127                                 $text = "\n" . $text;
2128                         }
2129                 }
2130
2131                 # Empties the template path
2132                 $this->mTemplatePath = array();
2133                 if ( !$found ) {
2134                         wfProfileOut( $fname );
2135                         return $matches[0];
2136                 } else {
2137                         # replace ==section headers==
2138                         # XXX this needs to go away once we have a better parser.
2139                         if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
2140                                 if( !is_null( $title ) )
2141                                         $encodedname = base64_encode($title->getPrefixedDBkey());
2142                                 else
2143                                         $encodedname = base64_encode("");
2144                                 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2145                                         PREG_SPLIT_DELIM_CAPTURE);
2146                                 $text = '';
2147                                 $nsec = 0;
2148                                 for( $i = 0; $i < count($m); $i += 2 ) {
2149                                         $text .= $m[$i];
2150                                         if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2151                                         $hl = $m[$i + 1];
2152                                         if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2153                                                 $text .= $hl;
2154                                                 continue;
2155                                         }
2156                                         preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2157                                         $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2158                                                 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2159
2160                                         $nsec++;
2161                                 }
2162                         }
2163                 }
2164
2165                 # Empties the template path
2166                 $this->mTemplatePath = array();
2167
2168                 if ( !$found ) {
2169                         wfProfileOut( $fname );
2170                         return $matches[0];
2171                 } else {
2172                         wfProfileOut( $fname );
2173                         return $text;
2174                 }
2175         }
2176
2177         /**
2178          * Triple brace replacement -- used for template arguments
2179          * @access private
2180          */
2181         function argSubstitution( $matches ) {
2182                 $arg = trim( $matches[1] );
2183                 $text = $matches[0];
2184                 $inputArgs = end( $this->mArgStack );
2185
2186                 if ( array_key_exists( $arg, $inputArgs ) ) {
2187                         $text = $inputArgs[$arg];
2188                 }
2189
2190                 return $text;
2191         }
2192
2193         /**
2194          * Returns true if the function is allowed to include this entity
2195          * @access private
2196          */
2197         function incrementIncludeCount( $dbk ) {
2198                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2199                         $this->mIncludeCount[$dbk] = 0;
2200                 }
2201                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2202                         return true;
2203                 } else {
2204                         return false;
2205                 }
2206         }
2207
2208
2209         /**
2210          * Cleans up HTML, removes dangerous tags and attributes, and
2211          * removes HTML comments
2212          * @access private
2213          */
2214         function removeHTMLtags( $text ) {
2215                 global $wgUseTidy, $wgUserHtml;
2216                 $fname = 'Parser::removeHTMLtags';
2217                 wfProfileIn( $fname );
2218
2219                 if( $wgUserHtml ) {
2220                         $htmlpairs = array( # Tags that must be closed
2221                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
2222                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
2223                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
2224                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
2225                                 'ruby', 'rt' , 'rb' , 'rp', 'p', 'span'
2226                         );
2227                         $htmlsingle = array(
2228                                 'br', 'hr', 'li', 'dt', 'dd'
2229                         );
2230                         $htmlnest = array( # Tags that can be nested--??
2231                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
2232                                 'dl', 'font', 'big', 'small', 'sub', 'sup', 'span'
2233                         );
2234                         $tabletags = array( # Can only appear inside table
2235                                 'td', 'th', 'tr'
2236                         );
2237                 } else {
2238                         $htmlpairs = array();
2239                         $htmlsingle = array();
2240                         $htmlnest = array();
2241                         $tabletags = array();
2242                 }
2243
2244                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
2245                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
2246
2247                 $htmlattrs = $this->getHTMLattrs () ;
2248
2249                 # Remove HTML comments
2250                 $text = $this->removeHTMLcomments( $text );
2251
2252                 $bits = explode( '<', $text );
2253                 $text = array_shift( $bits );
2254                 if(!$wgUseTidy) {
2255                         $tagstack = array(); $tablestack = array();
2256                         foreach ( $bits as $x ) {
2257                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2258                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2259                                 $x, $regs );
2260                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2261                                 error_reporting( $prev );
2262
2263                                 $badtag = 0 ;
2264                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2265                                         # Check our stack
2266                                         if ( $slash ) {
2267                                                 # Closing a tag...
2268                                                 if ( ! in_array( $t, $htmlsingle ) &&
2269                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2270                                                         @array_push( $tagstack, $ot );
2271                                                         $badtag = 1;
2272                                                 } else {
2273                                                         if ( $t == 'table' ) {
2274                                                                 $tagstack = array_pop( $tablestack );
2275                                                         }
2276                                                         $newparams = '';
2277                                                 }
2278                                         } else {
2279                                                 # Keep track for later
2280                                                 if ( in_array( $t, $tabletags ) &&
2281                                                 ! in_array( 'table', $tagstack ) ) {
2282                                                         $badtag = 1;
2283                                                 } else if ( in_array( $t, $tagstack ) &&
2284                                                 ! in_array ( $t , $htmlnest ) ) {
2285                                                         $badtag = 1 ;
2286                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
2287                                                         if ( $t == 'table' ) {
2288                                                                 array_push( $tablestack, $tagstack );
2289                                                                 $tagstack = array();
2290                                                         }
2291                                                         array_push( $tagstack, $t );
2292                                                 }
2293                                                 # Strip non-approved attributes from the tag
2294                                                 $newparams = $this->fixTagAttributes($params);
2295
2296                                         }
2297                                         if ( ! $badtag ) {
2298                                                 $rest = str_replace( '>', '&gt;', $rest );
2299                                                 $text .= "<$slash$t $newparams$brace$rest";
2300                                                 continue;
2301                                         }
2302                                 }
2303                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2304                         }
2305                         # Close off any remaining tags
2306                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2307                                 $text .= "</$t>\n";
2308                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2309                         }
2310                 } else {
2311                         # this might be possible using tidy itself
2312                         foreach ( $bits as $x ) {
2313                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2314                                 $x, $regs );
2315                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2316                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2317                                         $newparams = $this->fixTagAttributes($params);
2318                                         $rest = str_replace( '>', '&gt;', $rest );
2319                                         $text .= "<$slash$t $newparams$brace$rest";
2320                                 } else {
2321                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2322                                 }
2323                         }
2324                 }
2325                 wfProfileOut( $fname );
2326                 return $text;
2327         }
2328
2329         /**
2330          * Remove '<!--', '-->', and everything between.
2331          * To avoid leaving blank lines, when a comment is both preceded
2332          * and followed by a newline (ignoring spaces), trim leading and
2333          * trailing spaces and one of the newlines.
2334          *
2335          * @access private
2336          */
2337         function removeHTMLcomments( $text ) {
2338                 $fname='Parser::removeHTMLcomments';
2339                 wfProfileIn( $fname );
2340                 while (($start = strpos($text, '<!--')) !== false) {
2341                         $end = strpos($text, '-->', $start + 4);
2342                         if ($end === false) {
2343                                 # Unterminated comment; bail out
2344                                 break;
2345                         }
2346
2347                         $end += 3;
2348
2349                         # Trim space and newline if the comment is both
2350                         # preceded and followed by a newline
2351                         $spaceStart = max($start - 1, 0);
2352                         $spaceLen = $end - $spaceStart;
2353                         while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2354                                 $spaceStart--;
2355                                 $spaceLen++;
2356                         }
2357                         while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2358                                 $spaceLen++;
2359                         if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2360                                 # Remove the comment, leading and trailing
2361                                 # spaces, and leave only one newline.
2362                                 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2363                         }
2364                         else {
2365                                 # Remove just the comment.
2366                                 $text = substr_replace($text, '', $start, $end - $start);
2367                         }
2368                 }
2369                 wfProfileOut( $fname );
2370                 return $text;
2371         }
2372
2373         /**
2374          * This function accomplishes several tasks:
2375          * 1) Auto-number headings if that option is enabled
2376          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2377          * 3) Add a Table of contents on the top for users who have enabled the option
2378          * 4) Auto-anchor headings
2379          *
2380          * It loops through all headlines, collects the necessary data, then splits up the
2381          * string and re-inserts the newly formatted headlines.
2382          * @access private
2383          */
2384         /* private */ function formatHeadings( $text, $isMain=true ) {
2385                 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders;
2386
2387                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2388                 $doShowToc = $this->mOptions->getShowToc();
2389                 $forceTocHere = false;
2390                 if( !$this->mTitle->userCanEdit() ) {
2391                         $showEditLink = 0;
2392                         $rightClickHack = 0;
2393                 } else {
2394                         $showEditLink = $this->mOptions->getEditSection();
2395                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2396                 }
2397
2398                 # Inhibit editsection links if requested in the page
2399                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2400                 if( $esw->matchAndRemove( $text ) ) {
2401                         $showEditLink = 0;
2402                 }
2403                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2404                 # do not add TOC
2405                 $mw =& MagicWord::get( MAG_NOTOC );
2406                 if( $mw->matchAndRemove( $text ) ) {
2407                         $doShowToc = 0;
2408                 }
2409
2410                 # never add the TOC to the Main Page. This is an entry page that should not
2411                 # be more than 1-2 screens large anyway
2412                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2413                         $doShowToc = 0;
2414                 }
2415
2416                 # Get all headlines for numbering them and adding funky stuff like [edit]
2417                 # links - this is for later, but we need the number of headlines right now
2418                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2419
2420                 # if there are fewer than 4 headlines in the article, do not show TOC
2421                 if( $numMatches < 4 ) {
2422                         $doShowToc = 0;
2423                 }
2424
2425                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2426                 # override above conditions and always show TOC at that place
2427                 $mw =& MagicWord::get( MAG_TOC );
2428                 if ($mw->match( $text ) ) {
2429                         $doShowToc = 1;
2430                         $forceTocHere = true;
2431                 } else {
2432                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2433                         # override above conditions and always show TOC above first header
2434                         $mw =& MagicWord::get( MAG_FORCETOC );
2435                         if ($mw->matchAndRemove( $text ) ) {
2436                                 $doShowToc = 1;
2437                         }
2438                 }
2439
2440
2441
2442                 # We need this to perform operations on the HTML
2443                 $sk =& $this->mOptions->getSkin();
2444
2445                 # headline counter
2446                 $headlineCount = 0;
2447                 $sectionCount = 0; # headlineCount excluding template sections
2448
2449                 # Ugh .. the TOC should have neat indentation levels which can be
2450                 # passed to the skin functions. These are determined here
2451                 $toclevel = 0;
2452                 $toc = '';
2453                 $full = '';
2454                 $head = array();
2455                 $sublevelCount = array();
2456                 $level = 0;
2457                 $prevlevel = 0;
2458                 foreach( $matches[3] as $headline ) {
2459                         $istemplate = 0;
2460                         $templatetitle = "";
2461                         $templatesection = 0;
2462
2463                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2464                                 $istemplate = 1;
2465                                 $templatetitle = base64_decode($mat[1]);
2466                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2467                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2468                         }
2469
2470                         $numbering = '';
2471                         if( $level ) {
2472                                 $prevlevel = $level;
2473                         }
2474                         $level = $matches[1][$headlineCount];
2475                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2476                                 # reset when we enter a new level
2477                                 $sublevelCount[$level] = 0;
2478                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2479                                 $toclevel += $level - $prevlevel;
2480                         }
2481                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2482                                 # reset when we step back a level
2483                                 $sublevelCount[$level+1]=0;
2484                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2485                                 $toclevel -= $prevlevel - $level;
2486                         }
2487                         # count number of headlines for each level
2488                         @$sublevelCount[$level]++;
2489                         if( $doNumberHeadings || $doShowToc ) {
2490                                 $dot = 0;
2491                                 for( $i = 1; $i <= $level; $i++ ) {
2492                                         if( !empty( $sublevelCount[$i] ) ) {
2493                                                 if( $dot ) {
2494                                                         $numbering .= '.';
2495                                                 }
2496                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2497                                                 $dot = 1;
2498                                         }
2499                                 }
2500                         }
2501
2502                         # The canonized header is a version of the header text safe to use for links
2503                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2504                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2505                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2506
2507                         # Remove link placeholders by the link text.
2508                         #     <!--LINK number-->
2509                         # turns into
2510                         #     link text with suffix
2511                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2512                                                             "\$wgLinkHolders['texts'][\$1]",
2513                                                             $canonized_headline );
2514
2515                         # strip out HTML
2516                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2517                         $tocline = trim( $canonized_headline );
2518                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2519                         $replacearray = array(
2520                                 '%3A' => ':',
2521                                 '%' => '.'
2522                         );
2523                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2524                         $refer[$headlineCount] = $canonized_headline;
2525
2526                         # count how many in assoc. array so we can track dupes in anchors
2527                         @$refers[$canonized_headline]++;
2528                         $refcount[$headlineCount]=$refers[$canonized_headline];
2529
2530                         # Prepend the number to the heading text
2531
2532                         if( $doNumberHeadings || $doShowToc ) {
2533                                 $tocline = $numbering . ' ' . $tocline;
2534
2535                                 # Don't number the heading if it is the only one (looks silly)
2536                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2537                                         # the two are different if the line contains a link
2538                                         $headline=$numbering . ' ' . $headline;
2539                                 }
2540                         }
2541
2542                         # Create the anchor for linking from the TOC to the section
2543                         $anchor = $canonized_headline;
2544                         if($refcount[$headlineCount] > 1 ) {
2545                                 $anchor .= '_' . $refcount[$headlineCount];
2546                         }
2547                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2548                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2549                         }
2550                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2551                                 if ( empty( $head[$headlineCount] ) ) {
2552                                         $head[$headlineCount] = '';
2553                                 }
2554                                 if( $istemplate )
2555                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2556                                 else
2557                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2558                         }
2559
2560                         # Add the edit section span
2561                         if( $rightClickHack ) {
2562                                 if( $istemplate )
2563                                         $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2564                                 else
2565                                         $headline = $sk->editSectionScript($this->mTitle, $sectionCount+1,$headline);
2566                         }
2567
2568                         # give headline the correct <h#> tag
2569                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2570
2571                         $headlineCount++;
2572                         if( !$istemplate )
2573                                 $sectionCount++;
2574                 }
2575
2576                 if( $doShowToc ) {
2577                         $toclines = $headlineCount;
2578                         $toc .= $sk->tocUnindent( $toclevel );
2579                         $toc = $sk->tocTable( $toc );
2580                 }
2581
2582                 # split up and insert constructed headlines
2583
2584                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2585                 $i = 0;
2586
2587                 foreach( $blocks as $block ) {
2588                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2589                                 # This is the [edit] link that appears for the top block of text when
2590                                 # section editing is enabled
2591
2592                                 # Disabled because it broke block formatting
2593                                 # For example, a bullet point in the top line
2594                                 # $full .= $sk->editSectionLink(0);
2595                         }
2596                         $full .= $block;
2597                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2598                         # Top anchor now in skin
2599                                 $full = $full.$toc;
2600                         }
2601
2602                         if( !empty( $head[$i] ) ) {
2603                                 $full .= $head[$i];
2604                         }
2605                         $i++;
2606                 }
2607                 if($forceTocHere) {
2608                         $mw =& MagicWord::get( MAG_TOC );
2609                         return $mw->replace( $toc, $full );
2610                 } else {
2611                         return $full;
2612                 }
2613         }
2614
2615         /**
2616          * Return an HTML link for the "ISBN 123456" text
2617          * @access private
2618          */
2619         function magicISBN( $text ) {
2620                 global $wgLang;
2621                 $fname = 'Parser::magicISBN';
2622                 wfProfileIn( $fname );
2623
2624                 $a = split( 'ISBN ', ' '.$text );
2625                 if ( count ( $a ) < 2 ) {
2626                         wfProfileOut( $fname );
2627                         return $text;
2628                 }
2629                 $text = substr( array_shift( $a ), 1);
2630                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2631
2632                 foreach ( $a as $x ) {
2633                         $isbn = $blank = '' ;
2634                         while ( ' ' == $x{0} ) {
2635                                 $blank .= ' ';
2636                                 $x = substr( $x, 1 );
2637                         }
2638                         if ( $x == '' ) { # blank isbn
2639                                 $text .= "ISBN $blank";
2640                                 continue;
2641                         }
2642                         while ( strstr( $valid, $x{0} ) != false ) {
2643                                 $isbn .= $x{0};
2644                                 $x = substr( $x, 1 );
2645                         }
2646                         $num = str_replace( '-', '', $isbn );
2647                         $num = str_replace( ' ', '', $num );
2648
2649                         if ( '' == $num ) {
2650                                 $text .= "ISBN $blank$x";
2651                         } else {
2652                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2653                                 $text .= '<a href="' .
2654                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2655                                         "\" class=\"internal\">ISBN $isbn</a>";
2656                                 $text .= $x;
2657                         }
2658                 }
2659                 wfProfileOut( $fname );
2660                 return $text;
2661         }
2662
2663         /**
2664          * Return an HTML link for the "GEO ..." text
2665          * @access private
2666          */
2667         function magicGEO( $text ) {
2668                 global $wgLang, $wgUseGeoMode;
2669                 $fname = 'Parser::magicGEO';
2670                 wfProfileIn( $fname );
2671
2672                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2673                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2674                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2675                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2676                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2677                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2678
2679                 $a = split( 'GEO ', ' '.$text );
2680                 if ( count ( $a ) < 2 ) {
2681                         wfProfileOut( $fname );
2682                         return $text;
2683                 }
2684                 $text = substr( array_shift( $a ), 1);
2685                 $valid = '0123456789.+-:';
2686
2687                 foreach ( $a as $x ) {
2688                         $geo = $blank = '' ;
2689                         while ( ' ' == $x{0} ) {
2690                                 $blank .= ' ';
2691                                 $x = substr( $x, 1 );
2692                         }
2693                         while ( strstr( $valid, $x{0} ) != false ) {
2694                                 $geo .= $x{0};
2695                                 $x = substr( $x, 1 );
2696                         }
2697                         $num = str_replace( '+', '', $geo );
2698                         $num = str_replace( ' ', '', $num );
2699
2700                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2701                                 $text .= "GEO $blank$x";
2702                         } else {
2703                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2704                                 $text .= '<a href="' .
2705                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2706                                         "\" class=\"internal\">GEO $geo</a>";
2707                                 $text .= $x;
2708                         }
2709                 }
2710                 wfProfileOut( $fname );
2711                 return $text;
2712         }
2713
2714         /**
2715          * Return an HTML link for the "RFC 1234" text
2716          * @access private
2717          * @param string $text text to be processed
2718          */
2719         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
2720                 global $wgLang;
2721
2722                 $valid = '0123456789';
2723                 $internal = false;
2724
2725                 $a = split( $keyword, ' '.$text );
2726                 if ( count ( $a ) < 2 ) {
2727                         return $text;
2728                 }
2729                 $text = substr( array_shift( $a ), 1);
2730
2731                 /* Check if keyword is preceed by [[.
2732                  * This test is made here cause of the array_shift above
2733                  * that prevent the test to be done in the foreach.
2734                  */
2735                 if ( substr( $text, -2 ) == '[[' ) {
2736                         $internal = true;
2737                 }
2738
2739                 foreach ( $a as $x ) {
2740                         /* token might be empty if we have RFC RFC 1234 */
2741                         if ( $x=='' ) {
2742                                 $text.=$keyword;
2743                                 continue;
2744                                 }
2745
2746                         $id = $blank = '' ;
2747
2748                         /** remove and save whitespaces in $blank */
2749                         while ( $x{0} == ' ' ) {
2750                                 $blank .= ' ';
2751                                 $x = substr( $x, 1 );
2752                         }
2753
2754                         /** remove and save the rfc number in $id */
2755                         while ( strstr( $valid, $x{0} ) != false ) {
2756                                 $id .= $x{0};
2757                                 $x = substr( $x, 1 );
2758                         }
2759
2760                         if ( $id == '' ) {
2761                                 /* call back stripped spaces*/
2762                                 $text .= $keyword.$blank.$x;
2763                         } elseif( $internal ) {
2764                                 /* normal link */
2765                                 $text .= $keyword.$id.$x;
2766                         } else {
2767                                 /* build the external link*/
2768                                 $url = wfmsg( $urlmsg );
2769                                 $url = str_replace( '$1', $id, $url);
2770                                 $sk =& $this->mOptions->getSkin();
2771                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
2772                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
2773                         }
2774
2775                         /* Check if the next RFC keyword is preceed by [[ */
2776                         $internal = ( substr($x,-2) == '[[' );
2777                 }
2778                 return $text;
2779         }
2780
2781         /**
2782          * Transform wiki markup when saving a page by doing \r\n -> \n
2783          * conversion, substitting signatures, {{subst:}} templates, etc.
2784          *
2785          * @param string $text the text to transform
2786          * @param Title &$title the Title object for the current article
2787          * @param User &$user the User object describing the current user
2788          * @param ParserOptions $options parsing options
2789          * @param bool $clearState whether to clear the parser state first
2790          * @return string the altered wiki markup
2791          * @access public
2792          */
2793         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2794                 $this->mOptions = $options;
2795                 $this->mTitle =& $title;
2796                 $this->mOutputType = OT_WIKI;
2797
2798                 if ( $clearState ) {
2799                         $this->clearState();
2800                 }
2801
2802                 $stripState = false;
2803                 $pairs = array(
2804                         "\r\n" => "\n",
2805                         );
2806                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
2807                 $text = $this->strip( $text, $stripState, false );
2808                 $text = $this->pstPass2( $text, $user );
2809                 $text = $this->unstrip( $text, $stripState );
2810                 $text = $this->unstripNoWiki( $text, $stripState );
2811                 return $text;
2812         }
2813
2814         /**
2815          * Pre-save transform helper function
2816          * @access private
2817          */
2818         function pstPass2( $text, &$user ) {
2819                 global $wgLang, $wgContLang, $wgLocaltimezone;
2820
2821                 # Variable replacement
2822                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2823                 $text = $this->replaceVariables( $text );
2824
2825                 # Signatures
2826                 #
2827                 $n = $user->getName();
2828                 $k = $user->getOption( 'nickname' );
2829                 if ( '' == $k ) { $k = $n; }
2830                 if ( isset( $wgLocaltimezone ) ) {
2831                         $oldtz = getenv( 'TZ' );
2832                         putenv( 'TZ='.$wgLocaltimezone );
2833                 }
2834                 /* Note: this is an ugly timezone hack for the European wikis */
2835                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2836                   ' (' . date( 'T' ) . ')';
2837                 if ( isset( $wgLocaltimezone ) ) {
2838                         putenv( 'TZ='.$oldtzs );
2839                 }
2840
2841                 if( $user->getOption( 'fancysig' ) ) {
2842                         $sigText = $k;
2843                 } else {
2844                         $sigText = '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]";
2845                 }
2846                 $text = preg_replace( '/~~~~~/', $d, $text );
2847                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
2848                 $text = preg_replace( '/~~~/', $sigText, $text );
2849
2850                 # Context links: [[|name]] and [[name (context)|]]
2851                 #
2852                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2853                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2854                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2855                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2856
2857                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2858                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2859                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2860                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2861                 $context = '';
2862                 $t = $this->mTitle->getText();
2863                 if ( preg_match( $conpat, $t, $m ) ) {
2864                         $context = $m[2];
2865                 }
2866                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2867                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2868                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2869
2870                 if ( '' == $context ) {
2871                         $text = preg_replace( $p2, '[[\\1]]', $text );
2872                 } else {
2873                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2874                 }
2875
2876                 # Trim trailing whitespace
2877                 # MAG_END (__END__) tag allows for trailing
2878                 # whitespace to be deliberately included
2879                 $text = rtrim( $text );
2880                 $mw =& MagicWord::get( MAG_END );
2881                 $mw->matchAndRemove( $text );
2882
2883                 return $text;
2884         }
2885
2886         /**
2887          * Set up some variables which are usually set up in parse()
2888          * so that an external function can call some class members with confidence
2889          * @access public
2890          */
2891         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2892                 $this->mTitle =& $title;
2893                 $this->mOptions = $options;
2894                 $this->mOutputType = $outputType;
2895                 if ( $clearState ) {
2896                         $this->clearState();
2897                 }
2898         }
2899
2900         /**
2901          * Transform a MediaWiki message by replacing magic variables.
2902          *
2903          * @param string $text the text to transform
2904          * @param ParserOptions $options  options
2905          * @return string the text with variables substituted
2906          * @access public
2907          */
2908         function transformMsg( $text, $options ) {
2909                 global $wgTitle;
2910                 static $executing = false;
2911
2912                 # Guard against infinite recursion
2913                 if ( $executing ) {
2914                         return $text;
2915                 }
2916                 $executing = true;
2917
2918                 $this->mTitle = $wgTitle;
2919                 $this->mOptions = $options;
2920                 $this->mOutputType = OT_MSG;
2921                 $this->clearState();
2922                 $text = $this->replaceVariables( $text );
2923
2924                 $executing = false;
2925                 return $text;
2926         }
2927
2928         /**
2929          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2930          * Callback will be called with the text within
2931          * Transform and return the text within
2932          * @access public
2933          */
2934         function setHook( $tag, $callback ) {
2935                 $oldVal = @$this->mTagHooks[$tag];
2936                 $this->mTagHooks[$tag] = $callback;
2937                 return $oldVal;
2938         }
2939
2940         /**
2941          * Replace <!--LINK--> link placeholders with actual links, in the buffer
2942          * Placeholders created in Skin::makeLinkObj()
2943          * Returns an array of links found, indexed by PDBK:
2944          *  0 - broken
2945          *  1 - normal link
2946          *  2 - stub
2947          * $options is a bit field, RLH_FOR_UPDATE to select for update
2948          */
2949         function replaceLinkHolders( &$text, $options = 0 ) {
2950                 global $wgUser, $wgLinkCache, $wgUseOldExistenceCheck, $wgLinkHolders;
2951                 global $wgInterwikiLinkHolders;
2952                 global $outputReplace;
2953
2954                 if ( $wgUseOldExistenceCheck ) {
2955                         return array();
2956                 }
2957
2958                 $fname = 'Parser::replaceLinkHolders';
2959                 wfProfileIn( $fname );
2960
2961                 $pdbks = array();
2962                 $colours = array();
2963
2964                 #if ( !empty( $tmpLinks[0] ) ) { #TODO
2965                 if ( !empty( $wgLinkHolders['namespaces'] ) ) {
2966                         wfProfileIn( $fname.'-check' );
2967                         $dbr =& wfGetDB( DB_SLAVE );
2968                         $page = $dbr->tableName( 'page' );
2969                         $sk = $wgUser->getSkin();
2970                         $threshold = $wgUser->getOption('stubthreshold');
2971
2972                         # Sort by namespace
2973                         asort( $wgLinkHolders['namespaces'] );
2974
2975                         # Generate query
2976                         $query = false;
2977                         foreach ( $wgLinkHolders['namespaces'] as $key => $val ) {
2978                                 # Make title object
2979                                 $title = $wgLinkHolders['titles'][$key];
2980
2981                                 # Skip invalid entries.
2982                                 # Result will be ugly, but prevents crash.
2983                                 if ( is_null( $title ) ) {
2984                                         continue;
2985                                 }
2986                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
2987
2988                                 # Check if it's in the link cache already
2989                                 if ( $wgLinkCache->getGoodLinkID( $pdbk ) ) {
2990                                         $colours[$pdbk] = 1;
2991                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
2992                                         $colours[$pdbk] = 0;
2993                                 } else {
2994                                         # Not in the link cache, add it to the query
2995                                         if ( !isset( $current ) ) {
2996                                                 $current = $val;
2997                                                 $tables = $page;
2998                                                 $join = '';
2999                                                 $query =  "SELECT page_id, page_namespace, page_title";
3000                                                 if ( $threshold > 0 ) {
3001                                                         $textTable = $dbr->tableName( 'text' );
3002                                                         $query .= ', LENGTH(old_text) AS page_len, page_is_redirect';
3003                                                         $tables .= ", $textTable";
3004                                                         $join = 'page_latest=old_id AND';
3005                                                 }
3006                                                 $query .= " FROM $tables WHERE $join (page_namespace=$val AND page_title IN(";
3007                                         } elseif ( $current != $val ) {
3008                                                 $current = $val;
3009                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
3010                                         } else {
3011                                                 $query .= ', ';
3012                                         }
3013
3014                                         $query .= $dbr->addQuotes( $wgLinkHolders['dbkeys'][$key] );
3015                                 }
3016                         }
3017                         if ( $query ) {
3018                                 $query .= '))';
3019                                 if ( $options & RLH_FOR_UPDATE ) {
3020                                         $query .= ' FOR UPDATE';
3021                                 }
3022
3023                                 $res = $dbr->query( $query, $fname );
3024
3025                                 # Fetch data and form into an associative array
3026                                 # non-existent = broken
3027                                 # 1 = known
3028                                 # 2 = stub
3029                                 while ( $s = $dbr->fetchObject($res) ) {
3030                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
3031                                         $pdbk = $title->getPrefixedDBkey();
3032                                         $wgLinkCache->addGoodLink( $s->page_id, $pdbk );
3033
3034                                         if ( $threshold >  0 ) {
3035                                                 $size = $s->page_len;
3036                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $length < $threshold ) {
3037                                                         $colours[$pdbk] = 1;
3038                                                 } else {
3039                                                         $colours[$pdbk] = 2;
3040                                                 }
3041                                         } else {
3042                                                 $colours[$pdbk] = 1;
3043                                         }
3044                                 }
3045                         }
3046                         wfProfileOut( $fname.'-check' );
3047
3048                         # Construct search and replace arrays
3049                         wfProfileIn( $fname.'-construct' );
3050                         $outputReplace = array();
3051                         foreach ( $wgLinkHolders['namespaces'] as $key => $ns ) {
3052                                 $pdbk = $pdbks[$key];
3053                                 $searchkey = '<!--LINK '.$key.'-->';
3054                                 $title = $wgLinkHolders['titles'][$key];
3055                                 if ( empty( $colours[$pdbk] ) ) {
3056                                         $wgLinkCache->addBadLink( $pdbk );
3057                                         $colours[$pdbk] = 0;
3058                                         $outputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3059                                                                         $wgLinkHolders['texts'][$key],
3060                                                                         $wgLinkHolders['queries'][$key] );
3061                                 } elseif ( $colours[$pdbk] == 1 ) {
3062                                         $outputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3063                                                                         $wgLinkHolders['texts'][$key],
3064                                                                         $wgLinkHolders['queries'][$key] );
3065                                 } elseif ( $colours[$pdbk] == 2 ) {
3066                                         $outputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3067                                                                         $wgLinkHolders['texts'][$key],
3068                                                                         $wgLinkHolders['queries'][$key] );
3069                                 }
3070                         }
3071                         wfProfileOut( $fname.'-construct' );
3072
3073                         # Do the thing
3074                         wfProfileIn( $fname.'-replace' );
3075
3076                         $text = preg_replace_callback(
3077                                 '/(<!--LINK .*?-->)/',
3078                                 "outputReplaceMatches",
3079                                 $text);
3080                         wfProfileOut( $fname.'-replace' );
3081                 }
3082
3083                 if ( !empty( $wgInterwikiLinkHolders ) ) {
3084                         wfProfileIn( $fname.'-interwiki' );
3085                         $outputReplace = $wgInterwikiLinkHolders;
3086                         $text = preg_replace_callback(
3087                                 '/<!--IWLINK (.*?)-->/',
3088                                 "outputReplaceMatches",
3089                                 $text );
3090                         wfProfileOut( $fname.'-interwiki' );
3091                 }
3092
3093                 wfProfileOut( $fname );
3094                 return $colours;
3095         }
3096
3097         /**
3098          * Renders an image gallery from a text with one line per image.
3099          * text labels may be given by using |-style alternative text. E.g.
3100          *   Image:one.jpg|The number "1"
3101          *   Image:tree.jpg|A tree
3102          * given as text will return the HTML of a gallery with two images,
3103          * labeled 'The number "1"' and
3104          * 'A tree'.
3105          */
3106         function renderImageGallery( $text ) {
3107                 global $wgLinkCache;
3108                 $ig = new ImageGallery();
3109                 $ig->setShowBytes( false );
3110                 $ig->setShowFilename( false );
3111                 $lines = explode( "\n", $text );
3112
3113                 foreach ( $lines as $line ) {
3114                         # match lines like these:
3115                         # Image:someimage.jpg|This is some image
3116                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3117                         # Skip empty lines
3118                         if ( count( $matches ) == 0 ) {
3119                                 continue;
3120                         }
3121                         $nt = Title::newFromURL( $matches[1] );
3122                         if( is_null( $nt ) ) {
3123                                 # Bogus title. Ignore these so we don't bomb out later.
3124                                 continue;
3125                         }
3126                         if ( isset( $matches[3] ) ) {
3127                                 $label = $matches[3];
3128                         } else {
3129                                 $label = '';
3130                         }
3131
3132                         # FIXME: Use the full wiki parser and add its links
3133                         # to the page's links.
3134                         $html = $this->mOptions->mSkin->formatComment( $label );
3135
3136                         $ig->add( Image::newFromTitle( $nt ), $html );
3137                         $wgLinkCache->addImageLinkObj( $nt );
3138                 }
3139                 return $ig->toHTML();
3140         }
3141 }
3142
3143 /**
3144  * @todo document
3145  * @package MediaWiki
3146  */
3147 class ParserOutput
3148 {
3149         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3150         var $mCacheTime; # Used in ParserCache
3151         var $mVersion;   # Compatibility check
3152
3153         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3154                 $containsOldMagic = false )
3155         {
3156                 $this->mText = $text;
3157                 $this->mLanguageLinks = $languageLinks;
3158                 $this->mCategoryLinks = $categoryLinks;
3159                 $this->mContainsOldMagic = $containsOldMagic;
3160                 $this->mCacheTime = '';
3161                 $this->mVersion = MW_PARSER_VERSION;
3162         }
3163
3164         function getText()                   { return $this->mText; }
3165         function getLanguageLinks()          { return $this->mLanguageLinks; }
3166         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
3167         function getCacheTime()              { return $this->mCacheTime; }
3168         function containsOldMagic()          { return $this->mContainsOldMagic; }
3169         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3170         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3171         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3172         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3173         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3174         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3175
3176         function merge( $other ) {
3177                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3178                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3179                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3180         }
3181
3182         /**
3183          * Return true if this cached output object predates the global or
3184          * per-article cache invalidation timestamps, or if it comes from
3185          * an incompatible older version.
3186          *
3187          * @param string $touched the affected article's last touched timestamp
3188          * @return bool
3189          * @access public
3190          */
3191         function expired( $touched ) {
3192                 global $wgCacheEpoch;
3193                 return $this->getCacheTime() <= $touched ||
3194                        $this->getCacheTime() <= $wgCacheEpoch ||
3195                        !isset( $this->mVersion ) ||
3196                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3197         }
3198 }
3199
3200 /**
3201  * Set options of the Parser
3202  * @todo document
3203  * @package MediaWiki
3204  */
3205 class ParserOptions
3206 {
3207         # All variables are private
3208         var $mUseTeX;                    # Use texvc to expand <math> tags
3209         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
3210         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3211         var $mAllowExternalImages;       # Allow external images inline
3212         var $mSkin;                      # Reference to the preferred skin
3213         var $mDateFormat;                # Date format index
3214         var $mEditSection;               # Create "edit section" links
3215         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
3216         var $mNumberHeadings;            # Automatically number headings
3217         var $mShowToc;                   # Show table of contents
3218
3219         function getUseTeX()                        { return $this->mUseTeX; }
3220         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3221         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3222         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3223         function getSkin()                          { return $this->mSkin; }
3224         function getDateFormat()                    { return $this->mDateFormat; }
3225         function getEditSection()                   { return $this->mEditSection; }
3226         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
3227         function getNumberHeadings()                { return $this->mNumberHeadings; }
3228         function getShowToc()                       { return $this->mShowToc; }
3229
3230         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3231         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3232         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3233         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3234         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3235         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3236         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
3237         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3238         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
3239
3240         function setSkin( &$x ) { $this->mSkin =& $x; }
3241
3242         # Get parser options
3243         /* static */ function newFromUser( &$user ) {
3244                 $popts = new ParserOptions;
3245                 $popts->initialiseFromUser( $user );
3246                 return $popts;
3247         }
3248
3249         # Get user options
3250         function initialiseFromUser( &$userInput ) {
3251                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
3252                 $fname = 'ParserOptions::initialiseFromUser';
3253                 wfProfileIn( $fname );
3254                 if ( !$userInput ) {
3255                         $user = new User;
3256                         $user->setLoaded( true );
3257                 } else {
3258                         $user =& $userInput;
3259                 }
3260
3261                 $this->mUseTeX = $wgUseTeX;
3262                 $this->mUseDynamicDates = $wgUseDynamicDates;
3263                 $this->mInterwikiMagic = $wgInterwikiMagic;
3264                 $this->mAllowExternalImages = $wgAllowExternalImages;
3265                 wfProfileIn( $fname.'-skin' );
3266                 $this->mSkin =& $user->getSkin();
3267                 wfProfileOut( $fname.'-skin' );
3268                 $this->mDateFormat = $user->getOption( 'date' );
3269                 $this->mEditSection = $user->getOption( 'editsection' );
3270                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
3271                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3272                 $this->mShowToc = $user->getOption( 'showtoc' );
3273                 wfProfileOut( $fname );
3274         }
3275
3276
3277 }
3278
3279 /**
3280  * Callback function used by Parser::replaceLinkHolders()
3281  * to substitute link placeholders.
3282  */
3283 function &outputReplaceMatches( $matches ) {
3284         global $outputReplace;
3285         return $outputReplace[$matches[1]];
3286 }
3287
3288 /**
3289  * Return the total number of articles
3290  */
3291 function wfNumberOfArticles() {
3292         global $wgNumberOfArticles;
3293
3294         wfLoadSiteStats();
3295         return $wgNumberOfArticles;
3296 }
3297
3298 /**
3299  * Get various statistics from the database
3300  * @private
3301  */
3302 function wfLoadSiteStats() {
3303         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3304         $fname = 'wfLoadSiteStats';
3305
3306         if ( -1 != $wgNumberOfArticles ) return;
3307         $dbr =& wfGetDB( DB_SLAVE );
3308         $s = $dbr->selectRow( 'site_stats',
3309                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3310                 array( 'ss_row_id' => 1 ), $fname
3311         );
3312
3313         if ( $s === false ) {
3314                 return;
3315         } else {
3316                 $wgTotalViews = $s->ss_total_views;
3317                 $wgTotalEdits = $s->ss_total_edits;
3318                 $wgNumberOfArticles = $s->ss_good_articles;
3319         }
3320 }
3321
3322 function wfEscapeHTMLTagsOnly( $in ) {
3323         return str_replace(
3324                 array( '"', '>', '<' ),
3325                 array( '&quot;', '&gt;', '&lt;' ),
3326                 $in );
3327 }
3328
3329 ?>