includes/Parser.php

   1 <?php
   2 /**
   3  * File for Parser and related classes
   4  *
   5  * @package MediaWiki
   6  * @subpackage Parser
   7  */
   8
   9 /** */
  10 require_once( 'Sanitizer.php' );
  11 require_once( 'HttpFunctions.php' );
  12
  13 /**
  14  * Update this version number when the ParserOutput format
  15  * changes in an incompatible way, so the parser cache
  16  * can automatically discard old data.
  17  */
  18 define( 'MW_PARSER_VERSION', '1.5.0' );
  19
  20 /**
  21  * Variable substitution O(N^2) attack
  22  *
  23  * Without countermeasures, it would be possible to attack the parser by saving
  24  * a page filled with a large number of inclusions of large pages. The size of
  25  * the generated page would be proportional to the square of the input size.
  26  * Hence, we limit the number of inclusions of any given page, thus bringing any
  27  * attack back to O(N).
  28  */
  29
  30 define( 'MAX_INCLUDE_REPEAT', 100 );
  31 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  32
  33 define( 'RLH_FOR_UPDATE', 1 );
  34
  35 # Allowed values for $mOutputType
  36 define( 'OT_HTML', 1 );
  37 define( 'OT_WIKI', 2 );
  38 define( 'OT_MSG' , 3 );
  39
  40 # string parameter for extractTags which will cause it
  41 # to strip HTML comments in addition to regular
  42 # <XML>-style tags. This should not be anything we
  43 # may want to use in wikisyntax
  44 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  45
  46 # prefix for escaping, used in two functions at least
  47 define( 'UNIQ_PREFIX', 'NaodW29');
  48
  49 # Constants needed for external link processing
  50 define( 'HTTP_PROTOCOLS', 'http:\/\/|https:\/\/' );
  51 # Everything except bracket, space, or control characters
  52 define( 'EXT_LINK_URL_CLASS', '[^]<>"\\x00-\\x20\\x7F]' );
  53 # Including space
  54 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  55 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  56 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  57 define( 'EXT_LINK_BRACKETED',  '/\[(\b(' . wfUrlProtocols() . ')'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  58 define( 'EXT_IMAGE_REGEX',
  59         '/^('.HTTP_PROTOCOLS.')'.  # Protocol
  60         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  61         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  62 );
  63
  64 /**
  65  * PHP Parser
  66  *
  67  * Processes wiki markup
  68  *
  69  * <pre>
  70  * There are three main entry points into the Parser class:
  71  * parse()
  72  *   produces HTML output
  73  * preSaveTransform().
  74  *   produces altered wiki markup.
  75  * transformMsg()
  76  *   performs brace substitution on MediaWiki messages
  77  *
  78  * Globals used:
  79  *    objects:   $wgLang, $wgLinkCache
  80  *
  81  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  82  *
  83  * settings:
  84  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  85  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  86  *  $wgLocaltimezone, $wgAllowSpecialInclusion*
  87  *
  88  *  * only within ParserOptions
  89  * </pre>
  90  *
  91  * @package MediaWiki
  92  */
  93 class Parser
  94 {
  95         /**#@+
  96          * @access private
  97          */
  98         # Persistent:
  99         var $mTagHooks;
 100         var $mTagHookObjects;
 101
 102         # Cleared with clearState():
 103         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
 104         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
 105         var $mInterwikiLinkHolders, $mLinkHolders;
 106
 107         # Temporary:
 108         var $mOptions, $mTitle, $mOutputType,
 109             $mTemplates,        // cache of already loaded templates, avoids
 110                                 // multiple SQL queries for the same string
 111             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 112                                 // in this path. Used for loop detection.
 113
 114         var $mIWTransData = array();
 115
 116         /**#@-*/
 117
 118         /**
 119          * Constructor
 120          *
 121          * @access public
 122          */
 123         function Parser() {
 124                 $this->mTemplates = array();
 125                 $this->mTemplatePath = array();
 126                 $this->mTagHooks = array();
 127                 $this->mTagHookObjects = array();
 128                 $this->clearState();
 129         }
 130
 131         /**
 132          * Clear Parser state
 133          *
 134          * @access private
 135          */
 136         function clearState() {
 137                 $this->mOutput = new ParserOutput;
 138                 $this->mAutonumber = 0;
 139                 $this->mLastSection = '';
 140                 $this->mDTopen = false;
 141                 $this->mVariables = false;
 142                 $this->mIncludeCount = array();
 143                 $this->mStripState = array();
 144                 $this->mArgStack = array();
 145                 $this->mInPre = false;
 146                 $this->mInterwikiLinkHolders = array(
 147                         'texts' => array(),
 148                         'titles' => array()
 149                 );
 150                 $this->mLinkHolders = array(
 151                         'namespaces' => array(),
 152                         'dbkeys' => array(),
 153                         'queries' => array(),
 154                         'texts' => array(),
 155                         'titles' => array()
 156                 );
 157         }
 158
 159         /**
 160          * First pass--just handle <nowiki> sections, pass the rest off
 161          * to internalParse() which does all the real work.
 162          *
 163          * @access private
 164          * @param string $text Text we want to parse
 165          * @param Title &$title A title object
 166          * @param array $options
 167          * @param boolean $linestart
 168          * @param boolean $clearState
 169          * @return ParserOutput a ParserOutput
 170          */
 171         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 172                 global $wgUseTidy, $wgContLang;
 173                 $fname = 'Parser::parse';
 174                 wfProfileIn( $fname );
 175
 176                 if ( $clearState ) {
 177                         $this->clearState();
 178                 }
 179
 180                 $this->mOptions = $options;
 181                 $this->mTitle =& $title;
 182                 $this->mOutputType = OT_HTML;
 183
 184                 $this->mStripState = NULL;
 185
 186                 //$text = $this->strip( $text, $this->mStripState );
 187                 // VOODOO MAGIC FIX! Sometimes the above segfaults in PHP5.
 188                 $x =& $this->mStripState;
 189
 190                 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$x ) );
 191                 $text = $this->strip( $text, $x );
 192                 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$x ) );
 193
 194                 $text = $this->internalParse( $text );
 195
 196                 $text = $this->unstrip( $text, $this->mStripState );
 197
 198                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 199                 $fixtags = array(
 200                         # french spaces, last one Guillemet-left
 201                         # only if there is something before the space
 202                         '/(.) (?=\\?|:|;|!|\\302\\273)/' => '\\1&nbsp;\\2',
 203                         # french spaces, Guillemet-right
 204                         '/(\\302\\253) /' => '\\1&nbsp;',
 205                         '/<center *>(.*)<\\/center *>/i' => '<div class="center">\\1</div>',
 206                 );
 207                 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 208
 209                 # only once and last
 210                 $text = $this->doBlockLevels( $text, $linestart );
 211
 212                 $this->replaceLinkHolders( $text );
 213
 214                 # the position of the convert() call should not be changed. it
 215                 # assumes that the links are all replaces and the only thing left
 216                 # is the <nowiki> mark.
 217                 $text = $wgContLang->convert($text);
 218                 $this->mOutput->setTitleText($wgContLang->getParsedTitle());
 219
 220                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 221
 222                 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) );
 223
 224                 $text = Sanitizer::normalizeCharReferences( $text );
 225
 226                 if ($wgUseTidy) {
 227                         $text = Parser::tidy($text);
 228                 }
 229
 230                 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) );
 231
 232                 $this->mOutput->setText( $text );
 233                 wfProfileOut( $fname );
 234                 return $this->mOutput;
 235         }
 236
 237         /**
 238          * Get a random string
 239          *
 240          * @access private
 241          * @static
 242          */
 243         function getRandomString() {
 244                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 245         }
 246
 247         /**
 248          * Replaces all occurrences of <$tag>content</$tag> in the text
 249          * with a random marker and returns the new text. the output parameter
 250          * $content will be an associative array filled with data on the form
 251          * $unique_marker => content.
 252          *
 253          * If $content is already set, the additional entries will be appended
 254          * If $tag is set to STRIP_COMMENTS, the function will extract
 255          * <!-- HTML comments -->
 256          *
 257          * @access private
 258          * @static
 259          */
 260         function extractTagsAndParams($tag, $text, &$content, &$tags, &$params, $uniq_prefix = ''){
 261                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 262                 if ( !$content ) {
 263                         $content = array( );
 264                 }
 265                 $n = 1;
 266                 $stripped = '';
 267
 268                 if ( !$tags ) {
 269                         $tags = array( );
 270                 }
 271
 272                 if ( !$params ) {
 273                         $params = array( );
 274                 }
 275
 276                 if( $tag == STRIP_COMMENTS ) {
 277                         $start = '/<!--()()/';
 278                         $end   = '/-->/';
 279                 } else {
 280                         $start = "/<$tag(\\s+[^\\/>]*|\\s*)(\\/?)>/i";
 281                         $end   = "/<\\/$tag\\s*>/i";
 282                 }
 283
 284                 while ( '' != $text ) {
 285                         $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
 286                         $stripped .= $p[0];
 287                         if( count( $p ) < 4 ) {
 288                                 break;
 289                         }
 290                         $attributes = $p[1];
 291                         $empty      = $p[2];
 292                         $inside     = $p[3];
 293
 294                         $marker = $rnd . sprintf('%08X', $n++);
 295                         $stripped .= $marker;
 296
 297                         $tags[$marker] = "<$tag$attributes$empty>";
 298                         $params[$marker] = Sanitizer::decodeTagAttributes( $attributes );
 299
 300                         if ( $empty === '/' ) {
 301                                 // Empty element tag, <tag />
 302                                 $content[$marker] = null;
 303                                 $text = $inside;
 304                         } else {
 305                                 $q = preg_split( $end, $inside, 2 );
 306                                 $content[$marker] = $q[0];
 307                                 if( count( $q ) < 2 ) {
 308                                         # No end tag -- let it run out to the end of the text.
 309                                         break;
 310                                 } else {
 311                                         $text = $q[1];
 312                                 }
 313                         }
 314                 }
 315                 return $stripped;
 316         }
 317
 318         /**
 319          * Wrapper function for extractTagsAndParams
 320          * for cases where $tags and $params isn't needed
 321          * i.e. where tags will never have params, like <nowiki>
 322          *
 323          * @access private
 324          * @static
 325          */
 326         function extractTags( $tag, $text, &$content, $uniq_prefix = '' ) {
 327                 $dummy_tags = array();
 328                 $dummy_params = array();
 329
 330                 return Parser::extractTagsAndParams( $tag, $text, $content,
 331                         $dummy_tags, $dummy_params, $uniq_prefix );
 332         }
 333
 334         /**
 335          * Strips and renders nowiki, pre, math, hiero
 336          * If $render is set, performs necessary rendering operations on plugins
 337          * Returns the text, and fills an array with data needed in unstrip()
 338          * If the $state is already a valid strip state, it adds to the state
 339          *
 340          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 341          *  will be stripped in addition to other tags. This is important
 342          *  for section editing, where these comments cause confusion when
 343          *  counting the sections in the wikisource
 344          *
 345          * @access private
 346          */
 347         function strip( $text, &$state, $stripcomments = false ) {
 348                 $render = ($this->mOutputType == OT_HTML);
 349                 $html_content = array();
 350                 $nowiki_content = array();
 351                 $math_content = array();
 352                 $pre_content = array();
 353                 $comment_content = array();
 354                 $ext_content = array();
 355                 $ext_tags = array();
 356                 $ext_params = array();
 357                 $gallery_content = array();
 358
 359                 # Replace any instances of the placeholders
 360                 $uniq_prefix = UNIQ_PREFIX;
 361                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 362
 363                 # html
 364                 global $wgRawHtml;
 365                 if( $wgRawHtml ) {
 366                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 367                         foreach( $html_content as $marker => $content ) {
 368                                 if ($render ) {
 369                                         # Raw and unchecked for validity.
 370                                         $html_content[$marker] = $content;
 371                                 } else {
 372                                         $html_content[$marker] = '<html>'.$content.'</html>';
 373                                 }
 374                         }
 375                 }
 376
 377                 # nowiki
 378                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 379                 foreach( $nowiki_content as $marker => $content ) {
 380                         if( $render ){
 381                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 382                         } else {
 383                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 384                         }
 385                 }
 386
 387                 # math
 388                 if( $this->mOptions->getUseTeX() ) {
 389                         $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 390                         foreach( $math_content as $marker => $content ){
 391                                 if( $render ) {
 392                                         $math_content[$marker] = renderMath( $content );
 393                                 } else {
 394                                         $math_content[$marker] = '<math>'.$content.'</math>';
 395                                 }
 396                         }
 397                 }
 398
 399                 # pre
 400                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 401                 foreach( $pre_content as $marker => $content ){
 402                         if( $render ){
 403                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 404                         } else {
 405                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 406                         }
 407                 }
 408
 409                 # gallery
 410                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 411                 foreach( $gallery_content as $marker => $content ) {
 412                         require_once( 'ImageGallery.php' );
 413                         if ( $render ) {
 414                                 $gallery_content[$marker] = Parser::renderImageGallery( $content );
 415                         } else {
 416                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 417                         }
 418                 }
 419
 420                 # Comments
 421                 if($stripcomments) {
 422                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 423                         foreach( $comment_content as $marker => $content ){
 424                                 $comment_content[$marker] = '<!--'.$content.'-->';
 425                         }
 426                 }
 427
 428                 # Extensions
 429                 foreach ( $this->mTagHooks as $tag => $callback ) {
 430                         $ext_content[$tag] = array();
 431                         $text = Parser::extractTagsAndParams( $tag, $text, $ext_content[$tag],
 432                                 $ext_tags[$tag], $ext_params[$tag], $uniq_prefix );
 433                         foreach( $ext_content[$tag] as $marker => $content ) {
 434                                 $full_tag = $ext_tags[$tag][$marker];
 435                                 $params = $ext_params[$tag][$marker];
 436                                 if ( $render ) {
 437                                         if ( isset( $this->mTagHookObjects[$tag] ) )
 438                                                 $ext_content[$tag][$marker] = call_user_func_array(
 439                                                         array( &$this->mTagHookObjects[$tag], $callback ),
 440                                                         array( $content, $params, $this )
 441                                                 );
 442                                         else
 443                                                 $ext_content[$tag][$marker] = call_user_func_array(
 444                                                         $callback,
 445                                                         array( $content, $params, $this )
 446                                                 );
 447                                 } else {
 448                                         if ( is_null( $content ) ) {
 449                                                 // Empty element tag
 450                                                 $ext_content[$tag][$marker] = $full_tag;
 451                                         } else {
 452                                                 $ext_content[$tag][$marker] = "$full_tag$content</$tag>";
 453                                         }
 454                                 }
 455                         }
 456                 }
 457
 458                 # Merge state with the pre-existing state, if there is one
 459                 if ( $state ) {
 460                         $state['html'] = $state['html'] + $html_content;
 461                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 462                         $state['math'] = $state['math'] + $math_content;
 463                         $state['pre'] = $state['pre'] + $pre_content;
 464                         $state['comment'] = $state['comment'] + $comment_content;
 465                         $state['gallery'] = $state['gallery'] + $gallery_content;
 466
 467                         foreach( $ext_content as $tag => $array ) {
 468                                 if ( array_key_exists( $tag, $state ) ) {
 469                                         $state[$tag] = $state[$tag] + $array;
 470                                 }
 471                         }
 472                 } else {
 473                         $state = array(
 474                           'html' => $html_content,
 475                           'nowiki' => $nowiki_content,
 476                           'math' => $math_content,
 477                           'pre' => $pre_content,
 478                           'comment' => $comment_content,
 479                           'gallery' => $gallery_content,
 480                         ) + $ext_content;
 481                 }
 482                 return $text;
 483         }
 484
 485         /**
 486          * restores pre, math, and hiero removed by strip()
 487          *
 488          * always call unstripNoWiki() after this one
 489          * @access private
 490          */
 491         function unstrip( $text, &$state ) {
 492                 if ( !is_array( $state ) ) {
 493                         return $text;
 494                 }
 495
 496                 # Must expand in reverse order, otherwise nested tags will be corrupted
 497                 foreach( array_reverse( $state, true ) as $tag => $contentDict ) {
 498                         if( $tag != 'nowiki' && $tag != 'html' ) {
 499                                 foreach( array_reverse( $contentDict, true ) as $uniq => $content ) {
 500                                         $text = str_replace( $uniq, $content, $text );
 501                                 }
 502                         }
 503                 }
 504
 505                 return $text;
 506         }
 507
 508         /**
 509          * always call this after unstrip() to preserve the order
 510          *
 511          * @access private
 512          */
 513         function unstripNoWiki( $text, &$state ) {
 514                 if ( !is_array( $state ) ) {
 515                         return $text;
 516                 }
 517
 518                 # Must expand in reverse order, otherwise nested tags will be corrupted
 519                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 520                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 521                 }
 522
 523                 global $wgRawHtml;
 524                 if ($wgRawHtml) {
 525                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 526                                 $text = str_replace( key( $state['html'] ), $content, $text );
 527                         }
 528                 }
 529
 530                 return $text;
 531         }
 532
 533         /**
 534          * Add an item to the strip state
 535          * Returns the unique tag which must be inserted into the stripped text
 536          * The tag will be replaced with the original text in unstrip()
 537          *
 538          * @access private
 539          */
 540         function insertStripItem( $text, &$state ) {
 541                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 542                 if ( !$state ) {
 543                         $state = array(
 544                           'html' => array(),
 545                           'nowiki' => array(),
 546                           'math' => array(),
 547                           'pre' => array(),
 548                           'comment' => array(),
 549                           'gallery' => array(),
 550                         );
 551                 }
 552                 $state['item'][$rnd] = $text;
 553                 return $rnd;
 554         }
 555
 556         /**
 557          * Interface with html tidy, used if $wgUseTidy = true.
 558          * If tidy isn't able to correct the markup, the original will be
 559          * returned in all its glory with a warning comment appended.
 560          *
 561          * Either the external tidy program or the in-process tidy extension
 562          * will be used depending on availability. Override the default
 563          * $wgTidyInternal setting to disable the internal if it's not working.
 564          *
 565          * @param string $text Hideous HTML input
 566          * @return string Corrected HTML output
 567          * @access public
 568          * @static
 569          */
 570         function tidy( $text ) {
 571                 global $wgTidyInternal;
 572                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 573 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 574 '<head><title>test</title></head><body>'.$text.'</body></html>';
 575                 if( $wgTidyInternal ) {
 576                         $correctedtext = Parser::internalTidy( $wrappedtext );
 577                 } else {
 578                         $correctedtext = Parser::externalTidy( $wrappedtext );
 579                 }
 580                 if( is_null( $correctedtext ) ) {
 581                         wfDebug( "Tidy error detected!\n" );
 582                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 583                 }
 584                 return $correctedtext;
 585         }
 586
 587         /**
 588          * Spawn an external HTML tidy process and get corrected markup back from it.
 589          *
 590          * @access private
 591          * @static
 592          */
 593         function externalTidy( $text ) {
 594                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 595                 $fname = 'Parser::externalTidy';
 596                 wfProfileIn( $fname );
 597
 598                 $cleansource = '';
 599                 $opts = ' -utf8';
 600
 601                 $descriptorspec = array(
 602                         0 => array('pipe', 'r'),
 603                         1 => array('pipe', 'w'),
 604                         2 => array('file', '/dev/null', 'a')
 605                 );
 606                 $pipes = array();
 607                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 608                 if (is_resource($process)) {
 609                         fwrite($pipes[0], $text);
 610                         fclose($pipes[0]);
 611                         while (!feof($pipes[1])) {
 612                                 $cleansource .= fgets($pipes[1], 1024);
 613                         }
 614                         fclose($pipes[1]);
 615                         proc_close($process);
 616                 }
 617
 618                 wfProfileOut( $fname );
 619
 620                 if( $cleansource == '' && $text != '') {
 621                         // Some kind of error happened, so we couldn't get the corrected text.
 622                         // Just give up; we'll use the source text and append a warning.
 623                         return null;
 624                 } else {
 625                         return $cleansource;
 626                 }
 627         }
 628
 629         /**
 630          * Use the HTML tidy PECL extension to use the tidy library in-process,
 631          * saving the overhead of spawning a new process. Currently written to
 632          * the PHP 4.3.x version of the extension, may not work on PHP 5.
 633          *
 634          * 'pear install tidy' should be able to compile the extension module.
 635          *
 636          * @access private
 637          * @static
 638          */
 639         function internalTidy( $text ) {
 640                 global $wgTidyConf;
 641                 $fname = 'Parser::internalTidy';
 642                 wfProfileIn( $fname );
 643
 644                 tidy_load_config( $wgTidyConf );
 645                 tidy_set_encoding( 'utf8' );
 646                 tidy_parse_string( $text );
 647                 tidy_clean_repair();
 648                 if( tidy_get_status() == 2 ) {
 649                         // 2 is magic number for fatal error
 650                         // http://www.php.net/manual/en/function.tidy-get-status.php
 651                         $cleansource = null;
 652                 } else {
 653                         $cleansource = tidy_get_output();
 654                 }
 655                 wfProfileOut( $fname );
 656                 return $cleansource;
 657         }
 658
 659         /**
 660          * parse the wiki syntax used to render tables
 661          *
 662          * @access private
 663          */
 664         function doTableStuff ( $t ) {
 665                 $fname = 'Parser::doTableStuff';
 666                 wfProfileIn( $fname );
 667
 668                 $t = explode ( "\n" , $t ) ;
 669                 $td = array () ; # Is currently a td tag open?
 670                 $ltd = array () ; # Was it TD or TH?
 671                 $tr = array () ; # Is currently a tr tag open?
 672                 $ltr = array () ; # tr attributes
 673                 $indent_level = 0; # indent level of the table
 674                 foreach ( $t AS $k => $x )
 675                 {
 676                         $x = trim ( $x ) ;
 677                         $fc = substr ( $x , 0 , 1 ) ;
 678                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 679                                 $indent_level = strlen( $matches[1] );
 680
 681                                 $attributes = $this->unstripForHTML( $matches[2] );
 682
 683                                 $t[$k] = str_repeat( '<dl><dd>', $indent_level ) .
 684                                         '<table' . Sanitizer::fixTagAttributes ( $attributes, 'table' ) . '>' ;
 685                                 array_push ( $td , false ) ;
 686                                 array_push ( $ltd , '' ) ;
 687                                 array_push ( $tr , false ) ;
 688                                 array_push ( $ltr , '' ) ;
 689                         }
 690                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 691                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 692                                 $z = "</table>" . substr ( $x , 2);
 693                                 $l = array_pop ( $ltd ) ;
 694                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 695                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 696                                 array_pop ( $ltr ) ;
 697                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 698                         }
 699                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 700                                 $x = substr ( $x , 1 ) ;
 701                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 702                                 $z = '' ;
 703                                 $l = array_pop ( $ltd ) ;
 704                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 705                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 706                                 array_pop ( $ltr ) ;
 707                                 $t[$k] = $z ;
 708                                 array_push ( $tr , false ) ;
 709                                 array_push ( $td , false ) ;
 710                                 array_push ( $ltd , '' ) ;
 711                                 $attributes = $this->unstripForHTML( $x );
 712                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $attributes, 'tr' ) ) ;
 713                         }
 714                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 715                                 # $x is a table row
 716                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 717                                         $fc = '+' ;
 718                                         $x = substr ( $x , 1 ) ;
 719                                 }
 720                                 $after = substr ( $x , 1 ) ;
 721                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 722                                 $after = explode ( '||' , $after ) ;
 723                                 $t[$k] = '' ;
 724
 725                                 # Loop through each table cell
 726                                 foreach ( $after AS $theline )
 727                                 {
 728                                         $z = '' ;
 729                                         if ( $fc != '+' )
 730                                         {
 731                                                 $tra = array_pop ( $ltr ) ;
 732                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 733                                                 array_push ( $tr , true ) ;
 734                                                 array_push ( $ltr , '' ) ;
 735                                         }
 736
 737                                         $l = array_pop ( $ltd ) ;
 738                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 739                                         if ( $fc == '|' ) $l = 'td' ;
 740                                         else if ( $fc == '!' ) $l = 'th' ;
 741                                         else if ( $fc == '+' ) $l = 'caption' ;
 742                                         else $l = '' ;
 743                                         array_push ( $ltd , $l ) ;
 744
 745                                         # Cell parameters
 746                                         $y = explode ( '|' , $theline , 2 ) ;
 747                                         # Note that a '|' inside an invalid link should not
 748                                         # be mistaken as delimiting cell parameters
 749                                         if ( strpos( $y[0], '[[' ) !== false ) {
 750                                                 $y = array ($theline);
 751                                         }
 752                                         if ( count ( $y ) == 1 )
 753                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 754                                         else {
 755                                                 $attributes = $this->unstripForHTML( $y[0] );
 756                                                 $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($attributes, $l).">{$y[1]}" ;
 757                                         }
 758                                         $t[$k] .= $y ;
 759                                         array_push ( $td , true ) ;
 760                                 }
 761                         }
 762                 }
 763
 764                 # Closing open td, tr && table
 765                 while ( count ( $td ) > 0 )
 766                 {
 767                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 768                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 769                         $t[] = '</table>' ;
 770                 }
 771
 772                 $t = implode ( "\n" , $t ) ;
 773                 wfProfileOut( $fname );
 774                 return $t ;
 775         }
 776
 777         /**
 778          * Helper function for parse() that transforms wiki markup into
 779          * HTML. Only called for $mOutputType == OT_HTML.
 780          *
 781          * @access private
 782          */
 783         function internalParse( $text ) {
 784                 global $wgContLang;
 785                 $args = array();
 786                 $isMain = true;
 787                 $fname = 'Parser::internalParse';
 788                 wfProfileIn( $fname );
 789
 790                 # Remove <noinclude> tags and <includeonly> sections
 791                 $text = strtr( $text, array( '<onlyinclude>' => '' , '</onlyinclude>' => '' ) );
 792                 $text = strtr( $text, array( '<noinclude>' => '', '</noinclude>' => '') );
 793                 $text = preg_replace( '/<includeonly>.*?<\/includeonly>/s', '', $text );
 794
 795                 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ) );
 796                 $text = $this->replaceVariables( $text, $args );
 797
 798                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 799
 800                 $text = $this->doHeadings( $text );
 801                 if($this->mOptions->getUseDynamicDates()) {
 802                         $df =& DateFormatter::getInstance();
 803                         $text = $df->reformat( $this->mOptions->getDateFormat(), $text );
 804                 }
 805                 $text = $this->doAllQuotes( $text );
 806                 $text = $this->replaceInternalLinks( $text );
 807                 $text = $this->replaceExternalLinks( $text );
 808
 809                 # replaceInternalLinks may sometimes leave behind
 810                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 811                 $text = str_replace(UNIQ_PREFIX."NOPARSE", "", $text);
 812
 813                 $text = $this->doMagicLinks( $text );
 814                 $text = $this->doTableStuff( $text );
 815                 $text = $this->formatHeadings( $text, $isMain );
 816
 817                 $regex = '/<!--IW_TRANSCLUDE (\d+)-->/';
 818                 $text = preg_replace_callback($regex, array(&$this, 'scarySubstitution'), $text);
 819
 820                 wfProfileOut( $fname );
 821                 return $text;
 822         }
 823
 824         function scarySubstitution($matches) {
 825 #               return "[[".$matches[0]."]]";
 826                 return $this->mIWTransData[(int)$matches[0]];
 827         }
 828
 829         /**
 830          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 831          * magic external links.
 832          *
 833          * @access private
 834          */
 835         function &doMagicLinks( &$text ) {
 836                 $text = $this->magicISBN( $text );
 837                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 838                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 839                 return $text;
 840         }
 841
 842         /**
 843          * Parse ^^ tokens and return html
 844          *
 845          * @access private
 846          */
 847         function doExponent( $text ) {
 848                 $fname = 'Parser::doExponent';
 849                 wfProfileIn( $fname );
 850                 $text = preg_replace('/\^\^(.*?)\^\^/','<small><sup>\\1</sup></small>', $text);
 851                 wfProfileOut( $fname );
 852                 return $text;
 853         }
 854
 855         /**
 856          * Parse headers and return html
 857          *
 858          * @access private
 859          */
 860         function doHeadings( $text ) {
 861                 $fname = 'Parser::doHeadings';
 862                 wfProfileIn( $fname );
 863                 for ( $i = 6; $i >= 1; --$i ) {
 864                         $h = substr( '======', 0, $i );
 865                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 866                           "<h{$i}>\\1</h{$i}>\\2", $text );
 867                 }
 868                 wfProfileOut( $fname );
 869                 return $text;
 870         }
 871
 872         /**
 873          * Replace single quotes with HTML markup
 874          * @access private
 875          * @return string the altered text
 876          */
 877         function doAllQuotes( $text ) {
 878                 $fname = 'Parser::doAllQuotes';
 879                 wfProfileIn( $fname );
 880                 $outtext = '';
 881                 $lines = explode( "\n", $text );
 882                 foreach ( $lines as $line ) {
 883                         $outtext .= $this->doQuotes ( $line ) . "\n";
 884                 }
 885                 $outtext = substr($outtext, 0,-1);
 886                 wfProfileOut( $fname );
 887                 return $outtext;
 888         }
 889
 890         /**
 891          * Helper function for doAllQuotes()
 892          * @access private
 893          */
 894         function doQuotes( $text ) {
 895                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 896                 if ( count( $arr ) == 1 )
 897                         return $text;
 898                 else
 899                 {
 900                         # First, do some preliminary work. This may shift some apostrophes from
 901                         # being mark-up to being text. It also counts the number of occurrences
 902                         # of bold and italics mark-ups.
 903                         $i = 0;
 904                         $numbold = 0;
 905                         $numitalics = 0;
 906                         foreach ( $arr as $r )
 907                         {
 908                                 if ( ( $i % 2 ) == 1 )
 909                                 {
 910                                         # If there are ever four apostrophes, assume the first is supposed to
 911                                         # be text, and the remaining three constitute mark-up for bold text.
 912                                         if ( strlen( $arr[$i] ) == 4 )
 913                                         {
 914                                                 $arr[$i-1] .= "'";
 915                                                 $arr[$i] = "'''";
 916                                         }
 917                                         # If there are more than 5 apostrophes in a row, assume they're all
 918                                         # text except for the last 5.
 919                                         else if ( strlen( $arr[$i] ) > 5 )
 920                                         {
 921                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 922                                                 $arr[$i] = "'''''";
 923                                         }
 924                                         # Count the number of occurrences of bold and italics mark-ups.
 925                                         # We are not counting sequences of five apostrophes.
 926                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 927                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 928                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 929                                 }
 930                                 $i++;
 931                         }
 932
 933                         # If there is an odd number of both bold and italics, it is likely
 934                         # that one of the bold ones was meant to be an apostrophe followed
 935                         # by italics. Which one we cannot know for certain, but it is more
 936                         # likely to be one that has a single-letter word before it.
 937                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 938                         {
 939                                 $i = 0;
 940                                 $firstsingleletterword = -1;
 941                                 $firstmultiletterword = -1;
 942                                 $firstspace = -1;
 943                                 foreach ( $arr as $r )
 944                                 {
 945                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 946                                         {
 947                                                 $x1 = substr ($arr[$i-1], -1);
 948                                                 $x2 = substr ($arr[$i-1], -2, 1);
 949                                                 if ($x1 == ' ') {
 950                                                         if ($firstspace == -1) $firstspace = $i;
 951                                                 } else if ($x2 == ' ') {
 952                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 953                                                 } else {
 954                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 955                                                 }
 956                                         }
 957                                         $i++;
 958                                 }
 959
 960                                 # If there is a single-letter word, use it!
 961                                 if ($firstsingleletterword > -1)
 962                                 {
 963                                         $arr [ $firstsingleletterword ] = "''";
 964                                         $arr [ $firstsingleletterword-1 ] .= "'";
 965                                 }
 966                                 # If not, but there's a multi-letter word, use that one.
 967                                 else if ($firstmultiletterword > -1)
 968                                 {
 969                                         $arr [ $firstmultiletterword ] = "''";
 970                                         $arr [ $firstmultiletterword-1 ] .= "'";
 971                                 }
 972                                 # ... otherwise use the first one that has neither.
 973                                 # (notice that it is possible for all three to be -1 if, for example,
 974                                 # there is only one pentuple-apostrophe in the line)
 975                                 else if ($firstspace > -1)
 976                                 {
 977                                         $arr [ $firstspace ] = "''";
 978                                         $arr [ $firstspace-1 ] .= "'";
 979                                 }
 980                         }
 981
 982                         # Now let's actually convert our apostrophic mush to HTML!
 983                         $output = '';
 984                         $buffer = '';
 985                         $state = '';
 986                         $i = 0;
 987                         foreach ($arr as $r)
 988                         {
 989                                 if (($i % 2) == 0)
 990                                 {
 991                                         if ($state == 'both')
 992                                                 $buffer .= $r;
 993                                         else
 994                                                 $output .= $r;
 995                                 }
 996                                 else
 997                                 {
 998                                         if (strlen ($r) == 2)
 999                                         {
1000                                                 if ($state == 'i')
1001                                                 { $output .= '</i>'; $state = ''; }
1002                                                 else if ($state == 'bi')
1003                                                 { $output .= '</i>'; $state = 'b'; }
1004                                                 else if ($state == 'ib')
1005                                                 { $output .= '</b></i><b>'; $state = 'b'; }
1006                                                 else if ($state == 'both')
1007                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
1008                                                 else # $state can be 'b' or ''
1009                                                 { $output .= '<i>'; $state .= 'i'; }
1010                                         }
1011                                         else if (strlen ($r) == 3)
1012                                         {
1013                                                 if ($state == 'b')
1014                                                 { $output .= '</b>'; $state = ''; }
1015                                                 else if ($state == 'bi')
1016                                                 { $output .= '</i></b><i>'; $state = 'i'; }
1017                                                 else if ($state == 'ib')
1018                                                 { $output .= '</b>'; $state = 'i'; }
1019                                                 else if ($state == 'both')
1020                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
1021                                                 else # $state can be 'i' or ''
1022                                                 { $output .= '<b>'; $state .= 'b'; }
1023                                         }
1024                                         else if (strlen ($r) == 5)
1025                                         {
1026                                                 if ($state == 'b')
1027                                                 { $output .= '</b><i>'; $state = 'i'; }
1028                                                 else if ($state == 'i')
1029                                                 { $output .= '</i><b>'; $state = 'b'; }
1030                                                 else if ($state == 'bi')
1031                                                 { $output .= '</i></b>'; $state = ''; }
1032                                                 else if ($state == 'ib')
1033                                                 { $output .= '</b></i>'; $state = ''; }
1034                                                 else if ($state == 'both')
1035                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
1036                                                 else # ($state == '')
1037                                                 { $buffer = ''; $state = 'both'; }
1038                                         }
1039                                 }
1040                                 $i++;
1041                         }
1042                         # Now close all remaining tags.  Notice that the order is important.
1043                         if ($state == 'b' || $state == 'ib')
1044                                 $output .= '</b>';
1045                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
1046                                 $output .= '</i>';
1047                         if ($state == 'bi')
1048                                 $output .= '</b>';
1049                         if ($state == 'both')
1050                                 $output .= '<b><i>'.$buffer.'</i></b>';
1051                         return $output;
1052                 }
1053         }
1054
1055         /**
1056          * Replace external links
1057          *
1058          * Note: this is all very hackish and the order of execution matters a lot.
1059          * Make sure to run maintenance/parserTests.php if you change this code.
1060          *
1061          * @access private
1062          */
1063         function replaceExternalLinks( $text ) {
1064                 global $wgContLang;
1065                 $fname = 'Parser::replaceExternalLinks';
1066                 wfProfileIn( $fname );
1067
1068                 $sk =& $this->mOptions->getSkin();
1069
1070                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1071
1072                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
1073
1074                 $i = 0;
1075                 while ( $i<count( $bits ) ) {
1076                         $url = $bits[$i++];
1077                         $protocol = $bits[$i++];
1078                         $text = $bits[$i++];
1079                         $trail = $bits[$i++];
1080
1081                         # The characters '<' and '>' (which were escaped by
1082                         # removeHTMLtags()) should not be included in
1083                         # URLs, per RFC 2396.
1084                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1085                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
1086                                 $url = substr($url, 0, $m2[0][1]);
1087                         }
1088
1089                         # If the link text is an image URL, replace it with an <img> tag
1090                         # This happened by accident in the original parser, but some people used it extensively
1091                         $img = $this->maybeMakeExternalImage( $text );
1092                         if ( $img !== false ) {
1093                                 $text = $img;
1094                         }
1095
1096                         $dtrail = '';
1097
1098                         # Set linktype for CSS - if URL==text, link is essentially free
1099                         $linktype = ($text == $url) ? 'free' : 'text';
1100
1101                         # No link text, e.g. [http://domain.tld/some.link]
1102                         if ( $text == '' ) {
1103                                 # Autonumber if allowed
1104                                 if ( strpos( HTTP_PROTOCOLS, str_replace('/','\/', $protocol) ) !== false ) {
1105                                         $text = '[' . ++$this->mAutonumber . ']';
1106                                         $linktype = 'autonumber';
1107                                 } else {
1108                                         # Otherwise just use the URL
1109                                         $text = htmlspecialchars( $url );
1110                                         $linktype = 'free';
1111                                 }
1112                         } else {
1113                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1114                                 # Check for trail
1115                                 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1116                         }
1117
1118                         $text = $wgContLang->markNoConversion($text);
1119
1120                         # Replace &amp; from obsolete syntax with &.
1121                         # All HTML entities will be escaped by makeExternalLink()
1122                         # or maybeMakeExternalImage()
1123                         $url = str_replace( '&amp;', '&', $url );
1124
1125                         # Process the trail (i.e. everything after this link up until start of the next link),
1126                         # replacing any non-bracketed links
1127                         $trail = $this->replaceFreeExternalLinks( $trail );
1128
1129
1130                         # Use the encoded URL
1131                         # This means that users can paste URLs directly into the text
1132                         # Funny characters like &ouml; aren't valid in URLs anyway
1133                         # This was changed in August 2004
1134                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
1135                 }
1136
1137                 wfProfileOut( $fname );
1138                 return $s;
1139         }
1140
1141         /**
1142          * Replace anything that looks like a URL with a link
1143          * @access private
1144          */
1145         function replaceFreeExternalLinks( $text ) {
1146                 global $wgContLang;
1147                 $fname = 'Parser::replaceFreeExternalLinks';
1148                 wfProfileIn( $fname );
1149
1150                 $bits = preg_split( '/(\b(?:' . wfUrlProtocols() . '))/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1151                 $s = array_shift( $bits );
1152                 $i = 0;
1153
1154                 $sk =& $this->mOptions->getSkin();
1155
1156                 while ( $i < count( $bits ) ){
1157                         $protocol = $bits[$i++];
1158                         $remainder = $bits[$i++];
1159
1160                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1161                                 # Found some characters after the protocol that look promising
1162                                 $url = $protocol . $m[1];
1163                                 $trail = $m[2];
1164
1165                                 # The characters '<' and '>' (which were escaped by
1166                                 # removeHTMLtags()) should not be included in
1167                                 # URLs, per RFC 2396.
1168                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1169                                         $trail = substr($url, $m2[0][1]) . $trail;
1170                                         $url = substr($url, 0, $m2[0][1]);
1171                                 }
1172
1173                                 # Move trailing punctuation to $trail
1174                                 $sep = ',;\.:!?';
1175                                 # If there is no left bracket, then consider right brackets fair game too
1176                                 if ( strpos( $url, '(' ) === false ) {
1177                                         $sep .= ')';
1178                                 }
1179
1180                                 $numSepChars = strspn( strrev( $url ), $sep );
1181                                 if ( $numSepChars ) {
1182                                         $trail = substr( $url, -$numSepChars ) . $trail;
1183                                         $url = substr( $url, 0, -$numSepChars );
1184                                 }
1185
1186                                 # Replace &amp; from obsolete syntax with &.
1187                                 # All HTML entities will be escaped by makeExternalLink()
1188                                 # or maybeMakeExternalImage()
1189                                 $url = str_replace( '&amp;', '&', $url );
1190
1191                                 # Is this an external image?
1192                                 $text = $this->maybeMakeExternalImage( $url );
1193                                 if ( $text === false ) {
1194                                         # Not an image, make a link
1195                                         $text = $sk->makeExternalLink( $url, $wgContLang->markNoConversion($url), true, 'free' );
1196                                 }
1197                                 $s .= $text . $trail;
1198                         } else {
1199                                 $s .= $protocol . $remainder;
1200                         }
1201                 }
1202                 wfProfileOut( $fname );
1203                 return $s;
1204         }
1205
1206         /**
1207          * make an image if it's allowed, either through the global
1208          * option or through the exception
1209          * @access private
1210          */
1211         function maybeMakeExternalImage( $url ) {
1212                 $sk =& $this->mOptions->getSkin();
1213                 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
1214                 $imagesexception = !empty($imagesfrom);
1215                 $text = false;
1216                 if ( $this->mOptions->getAllowExternalImages()
1217                      || ( $imagesexception && strpos( $url, $imagesfrom ) === 0 ) ) {
1218                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1219                                 # Image found
1220                                 $text = $sk->makeExternalImage( htmlspecialchars( $url ) );
1221                         }
1222                 }
1223                 return $text;
1224         }
1225
1226         /**
1227          * Process [[ ]] wikilinks
1228          *
1229          * @access private
1230          */
1231         function replaceInternalLinks( $s ) {
1232                 global $wgContLang, $wgLinkCache;
1233                 static $fname = 'Parser::replaceInternalLinks' ;
1234
1235                 wfProfileIn( $fname );
1236
1237                 wfProfileIn( $fname.'-setup' );
1238                 static $tc = FALSE;
1239                 # the % is needed to support urlencoded titles as well
1240                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1241
1242                 $sk =& $this->mOptions->getSkin();
1243
1244                 #split the entire text string on occurences of [[
1245                 $a = explode( '[[', ' ' . $s );
1246                 #get the first element (all text up to first [[), and remove the space we added
1247                 $s = array_shift( $a );
1248                 $s = substr( $s, 1 );
1249
1250                 # Match a link having the form [[namespace:link|alternate]]trail
1251                 static $e1 = FALSE;
1252                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1253                 # Match cases where there is no "]]", which might still be images
1254                 static $e1_img = FALSE;
1255                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1256                 # Match the end of a line for a word that's not followed by whitespace,
1257                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1258                 $e2 = wfMsgForContent( 'linkprefix' );
1259
1260                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1261
1262                 if( is_null( $this->mTitle ) ) {
1263                         wfDebugDieBacktrace( 'nooo' );
1264                 }
1265                 $nottalk = !$this->mTitle->isTalkPage();
1266
1267                 if ( $useLinkPrefixExtension ) {
1268                         if ( preg_match( $e2, $s, $m ) ) {
1269                                 $first_prefix = $m[2];
1270                         } else {
1271                                 $first_prefix = false;
1272                         }
1273                 } else {
1274                         $prefix = '';
1275                 }
1276
1277                 $selflink = $this->mTitle->getPrefixedText();
1278                 wfProfileOut( $fname.'-setup' );
1279
1280                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1281                 $useSubpages = $this->areSubpagesAllowed();
1282
1283                 # Loop for each link
1284                 for ($k = 0; isset( $a[$k] ); $k++) {
1285                         $line = $a[$k];
1286                         if ( $useLinkPrefixExtension ) {
1287                                 wfProfileIn( $fname.'-prefixhandling' );
1288                                 if ( preg_match( $e2, $s, $m ) ) {
1289                                         $prefix = $m[2];
1290                                         $s = $m[1];
1291                                 } else {
1292                                         $prefix='';
1293                                 }
1294                                 # first link
1295                                 if($first_prefix) {
1296                                         $prefix = $first_prefix;
1297                                         $first_prefix = false;
1298                                 }
1299                                 wfProfileOut( $fname.'-prefixhandling' );
1300                         }
1301
1302                         $might_be_img = false;
1303
1304                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1305                                 $text = $m[2];
1306                                 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1307                                 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1308                                 # the real problem is with the $e1 regex
1309                                 # See bug 1300.
1310                                 #
1311                                 # Still some problems for cases where the ] is meant to be outside punctuation,
1312                                 # and no image is in sight. See bug 2095.
1313                                 #
1314                                 if( $text !== '' && preg_match( "/^\](.*)/s", $m[3], $n ) ) {
1315                                         $text .= ']'; # so that replaceExternalLinks($text) works later
1316                                         $m[3] = $n[1];
1317                                 }
1318                                 # fix up urlencoded title texts
1319                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1320                                 $trail = $m[3];
1321                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1322                                 $might_be_img = true;
1323                                 $text = $m[2];
1324                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1325                                 $trail = "";
1326                         } else { # Invalid form; output directly
1327                                 $s .= $prefix . '[[' . $line ;
1328                                 continue;
1329                         }
1330
1331                         # Don't allow internal links to pages containing
1332                         # PROTO: where PROTO is a valid URL protocol; these
1333                         # should be external links.
1334                         if (preg_match('/^(\b(?:' . wfUrlProtocols() . '))/', $m[1])) {
1335                                 $s .= $prefix . '[[' . $line ;
1336                                 continue;
1337                         }
1338
1339                         # Make subpage if necessary
1340                         if( $useSubpages ) {
1341                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1342                         } else {
1343                                 $link = $m[1];
1344                         }
1345
1346                         $noforce = (substr($m[1], 0, 1) != ':');
1347                         if (!$noforce) {
1348                                 # Strip off leading ':'
1349                                 $link = substr($link, 1);
1350                         }
1351
1352                         $nt = Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1353                         if( !$nt ) {
1354                                 $s .= $prefix . '[[' . $line;
1355                                 continue;
1356                         }
1357
1358                         #check other language variants of the link
1359                         #if the article does not exist
1360                         if( $checkVariantLink
1361                             && $nt->getArticleID() == 0 ) {
1362                                 $wgContLang->findVariantLink($link, $nt);
1363                         }
1364
1365                         $ns = $nt->getNamespace();
1366                         $iw = $nt->getInterWiki();
1367
1368                         if ($might_be_img) { # if this is actually an invalid link
1369                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1370                                         $found = false;
1371                                         while (isset ($a[$k+1]) ) {
1372                                                 #look at the next 'line' to see if we can close it there
1373                                                 $spliced = array_splice( $a, $k + 1, 1 );
1374                                                 $next_line = array_shift( $spliced );
1375                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1376                                                 # the first ]] closes the inner link, the second the image
1377                                                         $found = true;
1378                                                         $text .= '[[' . $m[1];
1379                                                         $trail = $m[2];
1380                                                         break;
1381                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1382                                                         #if there's exactly one ]] that's fine, we'll keep looking
1383                                                         $text .= '[[' . $m[0];
1384                                                 } else {
1385                                                         #if $next_line is invalid too, we need look no further
1386                                                         $text .= '[[' . $next_line;
1387                                                         break;
1388                                                 }
1389                                         }
1390                                         if ( !$found ) {
1391                                                 # we couldn't find the end of this imageLink, so output it raw
1392                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1393                                                 $text = $this->replaceInternalLinks($text);
1394                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1395                                                 # note: no $trail, because without an end, there *is* no trail
1396                                                 continue;
1397                                         }
1398                                 } else { #it's not an image, so output it raw
1399                                         $s .= $prefix . '[[' . $link . '|' . $text;
1400                                         # note: no $trail, because without an end, there *is* no trail
1401                                         continue;
1402                                 }
1403                         }
1404
1405                         $wasblank = ( '' == $text );
1406                         if( $wasblank ) $text = $link;
1407
1408
1409                         # Link not escaped by : , create the various objects
1410                         if( $noforce ) {
1411
1412                                 # Interwikis
1413                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1414                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1415                                         $s = rtrim($s . "\n");
1416                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1417                                         continue;
1418                                 }
1419
1420                                 if ( $ns == NS_IMAGE ) {
1421                                         wfProfileIn( "$fname-image" );
1422                                         if ( !wfIsBadImage( $nt->getDBkey() ) ) {
1423                                                 # recursively parse links inside the image caption
1424                                                 # actually, this will parse them in any other parameters, too,
1425                                                 # but it might be hard to fix that, and it doesn't matter ATM
1426                                                 $text = $this->replaceExternalLinks($text);
1427                                                 $text = $this->replaceInternalLinks($text);
1428
1429                                                 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1430                                                 $s .= $prefix . preg_replace( "/\b(" . wfUrlProtocols() . ')/', UNIQ_PREFIX."NOPARSE$1", $this->makeImage( $nt, $text) ) . $trail;
1431                                                 $wgLinkCache->addImageLinkObj( $nt );
1432
1433                                                 wfProfileOut( "$fname-image" );
1434                                                 continue;
1435                                         }
1436                                         wfProfileOut( "$fname-image" );
1437
1438                                 }
1439
1440                                 if ( $ns == NS_CATEGORY ) {
1441                                         wfProfileIn( "$fname-category" );
1442                                         $t = $wgContLang->convertHtml( $nt->getText() );
1443                                         $s = rtrim($s . "\n"); # bug 87
1444
1445                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1446                                         $t = $sk->makeLinkObj( $nt, $t, '', '' , $prefix );
1447                                         $wgLinkCache->resume();
1448
1449                                         if ( $wasblank ) {
1450                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1451                                                         $sortkey = $this->mTitle->getText();
1452                                                 } else {
1453                                                         $sortkey = $this->mTitle->getPrefixedText();
1454                                                 }
1455                                         } else {
1456                                                 $sortkey = $text;
1457                                         }
1458                                         $sortkey = $wgContLang->convertCategoryKey( $sortkey );
1459                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1460                                         $this->mOutput->addCategoryLink( $t );
1461
1462                                         /**
1463                                          * Strip the whitespace Category links produce, see bug 87
1464                                          * @todo We might want to use trim($tmp, "\n") here.
1465                                          */
1466                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1467
1468                                         wfProfileOut( "$fname-category" );
1469                                         continue;
1470                                 }
1471                         }
1472
1473                         if( ( $nt->getPrefixedText() === $selflink ) &&
1474                             ( $nt->getFragment() === '' ) ) {
1475                                 # Self-links are handled specially; generally de-link and change to bold.
1476                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1477                                 continue;
1478                         }
1479
1480                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1481                         if( $ns == NS_MEDIA ) {
1482                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text, true ) . $trail;
1483                                 $wgLinkCache->addImageLinkObj( $nt );
1484                                 continue;
1485                         } elseif( $ns == NS_SPECIAL ) {
1486                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1487                                 continue;
1488                         }
1489                         $s .= $this->makeLinkHolder( $nt, $text, '', $trail, $prefix );
1490                 }
1491                 wfProfileOut( $fname );
1492                 return $s;
1493         }
1494
1495         /**
1496          * Make a link placeholder. The text returned can be later resolved to a real link with
1497          * replaceLinkHolders(). This is done for two reasons: firstly to avoid further
1498          * parsing of interwiki links, and secondly to allow all extistence checks and
1499          * article length checks (for stub links) to be bundled into a single query.
1500          *
1501          */
1502         function makeLinkHolder( &$nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1503                 if ( ! is_object($nt) ) {
1504                         # Fail gracefully
1505                         $retVal = "<!-- ERROR -->{$prefix}{$text}{$trail}";
1506                 } else {
1507                         # Separate the link trail from the rest of the link
1508                         list( $inside, $trail ) = Linker::splitTrail( $trail );
1509
1510                         if ( $nt->isExternal() ) {
1511                                 $nr = array_push( $this->mInterwikiLinkHolders['texts'], $prefix.$text.$inside );
1512                                 $this->mInterwikiLinkHolders['titles'][] = $nt;
1513                                 $retVal = '<!--IWLINK '. ($nr-1) ."-->{$trail}";
1514                         } else {
1515                                 $nr = array_push( $this->mLinkHolders['namespaces'], $nt->getNamespace() );
1516                                 $this->mLinkHolders['dbkeys'][] = $nt->getDBkey();
1517                                 $this->mLinkHolders['queries'][] = $query;
1518                                 $this->mLinkHolders['texts'][] = $prefix.$text.$inside;
1519                                 $this->mLinkHolders['titles'][] = $nt;
1520
1521                                 $retVal = '<!--LINK '. ($nr-1) ."-->{$trail}";
1522                         }
1523                 }
1524                 return $retVal;
1525         }
1526
1527         /**
1528          * Return true if subpage links should be expanded on this page.
1529          * @return bool
1530          */
1531         function areSubpagesAllowed() {
1532                 # Some namespaces don't allow subpages
1533                 global $wgNamespacesWithSubpages;
1534                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1535         }
1536
1537         /**
1538          * Handle link to subpage if necessary
1539          * @param string $target the source of the link
1540          * @param string &$text the link text, modified as necessary
1541          * @return string the full name of the link
1542          * @access private
1543          */
1544         function maybeDoSubpageLink($target, &$text) {
1545                 # Valid link forms:
1546                 # Foobar -- normal
1547                 # :Foobar -- override special treatment of prefix (images, language links)
1548                 # /Foobar -- convert to CurrentPage/Foobar
1549                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1550                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1551                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1552
1553                 $fname = 'Parser::maybeDoSubpageLink';
1554                 wfProfileIn( $fname );
1555                 $ret = $target; # default return value is no change
1556
1557                 # Some namespaces don't allow subpages,
1558                 # so only perform processing if subpages are allowed
1559                 if( $this->areSubpagesAllowed() ) {
1560                         # Look at the first character
1561                         if( $target != '' && $target{0} == '/' ) {
1562                                 # / at end means we don't want the slash to be shown
1563                                 if( substr( $target, -1, 1 ) == '/' ) {
1564                                         $target = substr( $target, 1, -1 );
1565                                         $noslash = $target;
1566                                 } else {
1567                                         $noslash = substr( $target, 1 );
1568                                 }
1569
1570                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1571                                 if( '' === $text ) {
1572                                         $text = $target;
1573                                 } # this might be changed for ugliness reasons
1574                         } else {
1575                                 # check for .. subpage backlinks
1576                                 $dotdotcount = 0;
1577                                 $nodotdot = $target;
1578                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1579                                         ++$dotdotcount;
1580                                         $nodotdot = substr( $nodotdot, 3 );
1581                                 }
1582                                 if($dotdotcount > 0) {
1583                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1584                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1585                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1586                                                 # / at the end means don't show full path
1587                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1588                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1589                                                         if( '' === $text ) {
1590                                                                 $text = $nodotdot;
1591                                                         }
1592                                                 }
1593                                                 $nodotdot = trim( $nodotdot );
1594                                                 if( $nodotdot != '' ) {
1595                                                         $ret .= '/' . $nodotdot;
1596                                                 }
1597                                         }
1598                                 }
1599                         }
1600                 }
1601
1602                 wfProfileOut( $fname );
1603                 return $ret;
1604         }
1605
1606         /**#@+
1607          * Used by doBlockLevels()
1608          * @access private
1609          */
1610         /* private */ function closeParagraph() {
1611                 $result = '';
1612                 if ( '' != $this->mLastSection ) {
1613                         $result = '</' . $this->mLastSection  . ">\n";
1614                 }
1615                 $this->mInPre = false;
1616                 $this->mLastSection = '';
1617                 return $result;
1618         }
1619         # getCommon() returns the length of the longest common substring
1620         # of both arguments, starting at the beginning of both.
1621         #
1622         /* private */ function getCommon( $st1, $st2 ) {
1623                 $fl = strlen( $st1 );
1624                 $shorter = strlen( $st2 );
1625                 if ( $fl < $shorter ) { $shorter = $fl; }
1626
1627                 for ( $i = 0; $i < $shorter; ++$i ) {
1628                         if ( $st1{$i} != $st2{$i} ) { break; }
1629                 }
1630                 return $i;
1631         }
1632         # These next three functions open, continue, and close the list
1633         # element appropriate to the prefix character passed into them.
1634         #
1635         /* private */ function openList( $char ) {
1636                 $result = $this->closeParagraph();
1637
1638                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1639                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1640                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1641                 else if ( ';' == $char ) {
1642                         $result .= '<dl><dt>';
1643                         $this->mDTopen = true;
1644                 }
1645                 else { $result = '<!-- ERR 1 -->'; }
1646
1647                 return $result;
1648         }
1649
1650         /* private */ function nextItem( $char ) {
1651                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1652                 else if ( ':' == $char || ';' == $char ) {
1653                         $close = '</dd>';
1654                         if ( $this->mDTopen ) { $close = '</dt>'; }
1655                         if ( ';' == $char ) {
1656                                 $this->mDTopen = true;
1657                                 return $close . '<dt>';
1658                         } else {
1659                                 $this->mDTopen = false;
1660                                 return $close . '<dd>';
1661                         }
1662                 }
1663                 return '<!-- ERR 2 -->';
1664         }
1665
1666         /* private */ function closeList( $char ) {
1667                 if ( '*' == $char ) { $text = '</li></ul>'; }
1668                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1669                 else if ( ':' == $char ) {
1670                         if ( $this->mDTopen ) {
1671                                 $this->mDTopen = false;
1672                                 $text = '</dt></dl>';
1673                         } else {
1674                                 $text = '</dd></dl>';
1675                         }
1676                 }
1677                 else {  return '<!-- ERR 3 -->'; }
1678                 return $text."\n";
1679         }
1680         /**#@-*/
1681
1682         /**
1683          * Make lists from lines starting with ':', '*', '#', etc.
1684          *
1685          * @access private
1686          * @return string the lists rendered as HTML
1687          */
1688         function doBlockLevels( $text, $linestart ) {
1689                 $fname = 'Parser::doBlockLevels';
1690                 wfProfileIn( $fname );
1691
1692                 # Parsing through the text line by line.  The main thing
1693                 # happening here is handling of block-level elements p, pre,
1694                 # and making lists from lines starting with * # : etc.
1695                 #
1696                 $textLines = explode( "\n", $text );
1697
1698                 $lastPrefix = $output = '';
1699                 $this->mDTopen = $inBlockElem = false;
1700                 $prefixLength = 0;
1701                 $paragraphStack = false;
1702
1703                 if ( !$linestart ) {
1704                         $output .= array_shift( $textLines );
1705                 }
1706                 foreach ( $textLines as $oLine ) {
1707                         $lastPrefixLength = strlen( $lastPrefix );
1708                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1709                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1710                         if ( !$this->mInPre ) {
1711                                 # Multiple prefixes may abut each other for nested lists.
1712                                 $prefixLength = strspn( $oLine, '*#:;' );
1713                                 $pref = substr( $oLine, 0, $prefixLength );
1714
1715                                 # eh?
1716                                 $pref2 = str_replace( ';', ':', $pref );
1717                                 $t = substr( $oLine, $prefixLength );
1718                                 $this->mInPre = !empty($preOpenMatch);
1719                         } else {
1720                                 # Don't interpret any other prefixes in preformatted text
1721                                 $prefixLength = 0;
1722                                 $pref = $pref2 = '';
1723                                 $t = $oLine;
1724                         }
1725
1726                         # List generation
1727                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1728                                 # Same as the last item, so no need to deal with nesting or opening stuff
1729                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1730                                 $paragraphStack = false;
1731
1732                                 if ( substr( $pref, -1 ) == ';') {
1733                                         # The one nasty exception: definition lists work like this:
1734                                         # ; title : definition text
1735                                         # So we check for : in the remainder text to split up the
1736                                         # title and definition, without b0rking links.
1737                                         $term = $t2 = '';
1738                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1739                                                 $t = $t2;
1740                                                 $output .= $term . $this->nextItem( ':' );
1741                                         }
1742                                 }
1743                         } elseif( $prefixLength || $lastPrefixLength ) {
1744                                 # Either open or close a level...
1745                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1746                                 $paragraphStack = false;
1747
1748                                 while( $commonPrefixLength < $lastPrefixLength ) {
1749                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1750                                         --$lastPrefixLength;
1751                                 }
1752                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1753                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1754                                 }
1755                                 while ( $prefixLength > $commonPrefixLength ) {
1756                                         $char = substr( $pref, $commonPrefixLength, 1 );
1757                                         $output .= $this->openList( $char );
1758
1759                                         if ( ';' == $char ) {
1760                                                 # FIXME: This is dupe of code above
1761                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1762                                                         $t = $t2;
1763                                                         $output .= $term . $this->nextItem( ':' );
1764                                                 }
1765                                         }
1766                                         ++$commonPrefixLength;
1767                                 }
1768                                 $lastPrefix = $pref2;
1769                         }
1770                         if( 0 == $prefixLength ) {
1771                                 wfProfileIn( "$fname-paragraph" );
1772                                 # No prefix (not in list)--go to paragraph mode
1773                                 $uniq_prefix = UNIQ_PREFIX;
1774                                 // XXX: use a stack for nestable elements like span, table and div
1775                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1776                                 $closematch = preg_match(
1777                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1778                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1779                                 if ( $openmatch or $closematch ) {
1780                                         $paragraphStack = false;
1781                                         $output .= $this->closeParagraph();
1782                                         if ( $preOpenMatch and !$preCloseMatch ) {
1783                                                 $this->mInPre = true;
1784                                         }
1785                                         if ( $closematch ) {
1786                                                 $inBlockElem = false;
1787                                         } else {
1788                                                 $inBlockElem = true;
1789                                         }
1790                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1791                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1792                                                 // pre
1793                                                 if ($this->mLastSection != 'pre') {
1794                                                         $paragraphStack = false;
1795                                                         $output .= $this->closeParagraph().'<pre>';
1796                                                         $this->mLastSection = 'pre';
1797                                                 }
1798                                                 $t = substr( $t, 1 );
1799                                         } else {
1800                                                 // paragraph
1801                                                 if ( '' == trim($t) ) {
1802                                                         if ( $paragraphStack ) {
1803                                                                 $output .= $paragraphStack.'<br />';
1804                                                                 $paragraphStack = false;
1805                                                                 $this->mLastSection = 'p';
1806                                                         } else {
1807                                                                 if ($this->mLastSection != 'p' ) {
1808                                                                         $output .= $this->closeParagraph();
1809                                                                         $this->mLastSection = '';
1810                                                                         $paragraphStack = '<p>';
1811                                                                 } else {
1812                                                                         $paragraphStack = '</p><p>';
1813                                                                 }
1814                                                         }
1815                                                 } else {
1816                                                         if ( $paragraphStack ) {
1817                                                                 $output .= $paragraphStack;
1818                                                                 $paragraphStack = false;
1819                                                                 $this->mLastSection = 'p';
1820                                                         } else if ($this->mLastSection != 'p') {
1821                                                                 $output .= $this->closeParagraph().'<p>';
1822                                                                 $this->mLastSection = 'p';
1823                                                         }
1824                                                 }
1825                                         }
1826                                 }
1827                                 wfProfileOut( "$fname-paragraph" );
1828                         }
1829                         // somewhere above we forget to get out of pre block (bug 785)
1830                         if($preCloseMatch && $this->mInPre) {
1831                                 $this->mInPre = false;
1832                         }
1833                         if ($paragraphStack === false) {
1834                                 $output .= $t."\n";
1835                         }
1836                 }
1837                 while ( $prefixLength ) {
1838                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1839                         --$prefixLength;
1840                 }
1841                 if ( '' != $this->mLastSection ) {
1842                         $output .= '</' . $this->mLastSection . '>';
1843                         $this->mLastSection = '';
1844                 }
1845
1846                 wfProfileOut( $fname );
1847                 return $output;
1848         }
1849
1850         /**
1851          * Split up a string on ':', ignoring any occurences inside
1852          * <a>..</a> or <span>...</span>
1853          * @param string $str the string to split
1854          * @param string &$before set to everything before the ':'
1855          * @param string &$after set to everything after the ':'
1856          * return string the position of the ':', or false if none found
1857          */
1858         function findColonNoLinks($str, &$before, &$after) {
1859                 # I wonder if we should make this count all tags, not just <a>
1860                 # and <span>. That would prevent us from matching a ':' that
1861                 # comes in the middle of italics other such formatting....
1862                 # -- Wil
1863                 $fname = 'Parser::findColonNoLinks';
1864                 wfProfileIn( $fname );
1865                 $pos = 0;
1866                 do {
1867                         $colon = strpos($str, ':', $pos);
1868
1869                         if ($colon !== false) {
1870                                 $before = substr($str, 0, $colon);
1871                                 $after = substr($str, $colon + 1);
1872
1873                                 # Skip any ':' within <a> or <span> pairs
1874                                 $a = substr_count($before, '<a');
1875                                 $s = substr_count($before, '<span');
1876                                 $ca = substr_count($before, '</a>');
1877                                 $cs = substr_count($before, '</span>');
1878
1879                                 if ($a <= $ca and $s <= $cs) {
1880                                         # Tags are balanced before ':'; ok
1881                                         break;
1882                                 }
1883                                 $pos = $colon + 1;
1884                         }
1885                 } while ($colon !== false);
1886                 wfProfileOut( $fname );
1887                 return $colon;
1888         }
1889
1890         /**
1891          * Return value of a magic variable (like PAGENAME)
1892          *
1893          * @access private
1894          */
1895         function getVariableValue( $index ) {
1896                 global $wgContLang, $wgSitename, $wgServer, $wgServerName, $wgArticle, $wgScriptPath;
1897
1898                 /**
1899                  * Some of these require message or data lookups and can be
1900                  * expensive to check many times.
1901                  */
1902                 static $varCache = array();
1903                 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$varCache ) ) )
1904                         if ( isset( $varCache[$index] ) )
1905                                 return $varCache[$index];
1906
1907                 $ts = time();
1908                 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) );
1909                 $revid = $wgArticle->getRevIdFetched();
1910                 wfRunHooks( 'ParserGetVariableValueRevid', array( &$this, &$revid ) );
1911
1912                 switch ( $index ) {
1913                         case MAG_CURRENTMONTH:
1914                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm', $ts ) );
1915                         case MAG_CURRENTMONTHNAME:
1916                                 return $varCache[$index] = $wgContLang->getMonthName( date( 'n', $ts ) );
1917                         case MAG_CURRENTMONTHNAMEGEN:
1918                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date( 'n', $ts ) );
1919                         case MAG_CURRENTMONTHABBREV:
1920                                 return $varCache[$index] = $wgContLang->getMonthAbbreviation( date( 'n', $ts ) );
1921                         case MAG_CURRENTDAY:
1922                                 return $varCache[$index] = $wgContLang->formatNum( date( 'j', $ts ) );
1923                         case MAG_PAGENAME:
1924                                 return $this->mTitle->getText();
1925                         case MAG_PAGENAMEE:
1926                                 return $this->mTitle->getPartialURL();
1927                         case MAG_FULLPAGENAME:
1928                                 return $this->mTitle->getPrefixedText();
1929                         case MAG_FULLPAGENAMEE:
1930                                 return wfUrlencode( $this->mTitle->getPrefixedText() );
1931                         case MAG_REVISIONID:
1932                                 return $revid;
1933                         case MAG_NAMESPACE:
1934                                 return $wgContLang->getNsText( $this->mTitle->getNamespace() );
1935                         case MAG_NAMESPACEE:
1936                                 return wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
1937                         case MAG_CURRENTDAYNAME:
1938                                 return $varCache[$index] = $wgContLang->getWeekdayName( date( 'w', $ts ) + 1 );
1939                         case MAG_CURRENTYEAR:
1940                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y', $ts ), true );
1941                         case MAG_CURRENTTIME:
1942                                 return $varCache[$index] = $wgContLang->time( wfTimestamp( TS_MW, $ts ), false, false );
1943                         case MAG_CURRENTWEEK:
1944                                 return $varCache[$index] = $wgContLang->formatNum( date( 'W', $ts ) );
1945                         case MAG_CURRENTDOW:
1946                                 return $varCache[$index] = $wgContLang->formatNum( date( 'w', $ts ) );
1947                         case MAG_NUMBEROFARTICLES:
1948                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1949                         case MAG_NUMBEROFFILES:
1950                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfFiles() );
1951                         case MAG_SITENAME:
1952                                 return $wgSitename;
1953                         case MAG_SERVER:
1954                                 return $wgServer;
1955                         case MAG_SERVERNAME:
1956                                 return $wgServerName;
1957                         case MAG_SCRIPTPATH:
1958                                 return $wgScriptPath;
1959                         default:
1960                                 $ret = null;
1961                                 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$varCache, &$index, &$ret ) ) )
1962                                         return $ret;
1963                                 else
1964                                         return null;
1965                 }
1966         }
1967
1968         /**
1969          * initialise the magic variables (like CURRENTMONTHNAME)
1970          *
1971          * @access private
1972          */
1973         function initialiseVariables() {
1974                 $fname = 'Parser::initialiseVariables';
1975                 wfProfileIn( $fname );
1976                 global $wgVariableIDs;
1977                 $this->mVariables = array();
1978                 foreach ( $wgVariableIDs as $id ) {
1979                         $mw =& MagicWord::get( $id );
1980                         $mw->addToArray( $this->mVariables, $id );
1981                 }
1982                 wfProfileOut( $fname );
1983         }
1984
1985         /**
1986          * parse any parentheses in format ((title|part|part))
1987          * and call callbacks to get a replacement text for any found piece
1988          *
1989          * @param string $text The text to parse
1990          * @param array $callbacks rules in form:
1991          *     '{' => array(                            # opening parentheses
1992          *                                      'end' => '}',   # closing parentheses
1993          *                                      'cb' => array(2 => callback,    # replacement callback to call if {{..}} is found
1994          *                                                                4 => callback         # replacement callback to call if {{{{..}}}} is found
1995          *                                                                )
1996          *                                      )
1997          * @access private
1998          */
1999         function replace_callback ($text, $callbacks) {
2000                 $openingBraceStack = array();   # this array will hold a stack of parentheses which are not closed yet
2001                 $lastOpeningBrace = -1;                 # last not closed parentheses
2002
2003                 for ($i = 0; $i < strlen($text); $i++) {
2004                         # check for any opening brace
2005                         $rule = null;
2006                         $nextPos = -1;
2007                         foreach ($callbacks as $key => $value) {
2008                                 $pos = strpos ($text, $key, $i);
2009                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)) {
2010                                         $rule = $value;
2011                                         $nextPos = $pos;
2012                                 }
2013                         }
2014
2015                         if ($lastOpeningBrace >= 0) {
2016                                 $pos = strpos ($text, $openingBraceStack[$lastOpeningBrace]['braceEnd'], $i);
2017
2018                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)){
2019                                         $rule = null;
2020                                         $nextPos = $pos;
2021                                 }
2022
2023                                 $pos = strpos ($text, '|', $i);
2024
2025                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)){
2026                                         $rule = null;
2027                                         $nextPos = $pos;
2028                                 }
2029                         }
2030
2031                         if ($nextPos == -1)
2032                                 break;
2033
2034                         $i = $nextPos;
2035
2036                         # found openning brace, lets add it to parentheses stack
2037                         if (null != $rule) {
2038                                 $piece = array('brace' => $text[$i],
2039                                                            'braceEnd' => $rule['end'],
2040                                                            'count' => 1,
2041                                                            'title' => '',
2042                                                            'parts' => null);
2043
2044                                 # count openning brace characters
2045                                 while ($i+1 < strlen($text) && $text[$i+1] == $piece['brace']) {
2046                                         $piece['count']++;
2047                                         $i++;
2048                                 }
2049
2050                                 $piece['startAt'] = $i+1;
2051                                 $piece['partStart'] = $i+1;
2052
2053                                 # we need to add to stack only if openning brace count is enough for any given rule
2054                                 foreach ($rule['cb'] as $cnt => $fn) {
2055                                         if ($piece['count'] >= $cnt) {
2056                                                 $lastOpeningBrace ++;
2057                                                 $openingBraceStack[$lastOpeningBrace] = $piece;
2058                                                 break;
2059                                         }
2060                                 }
2061
2062                                 continue;
2063                         }
2064                         else if ($lastOpeningBrace >= 0) {
2065                                 # first check if it is a closing brace
2066                                 if ($openingBraceStack[$lastOpeningBrace]['braceEnd'] == $text[$i]) {
2067                                         # lets check if it is enough characters for closing brace
2068                                         $count = 1;
2069                                         while ($i+$count < strlen($text) && $text[$i+$count] == $text[$i])
2070                                                 $count++;
2071
2072                                         # if there are more closing parentheses than opening ones, we parse less
2073                                         if ($openingBraceStack[$lastOpeningBrace]['count'] < $count)
2074                                                 $count = $openingBraceStack[$lastOpeningBrace]['count'];
2075
2076                                         # check for maximum matching characters (if there are 5 closing characters, we will probably need only 3 - depending on the rules)
2077                                         $matchingCount = 0;
2078                                         $matchingCallback = null;
2079                                         foreach ($callbacks[$openingBraceStack[$lastOpeningBrace]['brace']]['cb'] as $cnt => $fn) {
2080                                                 if ($count >= $cnt && $matchingCount < $cnt) {
2081                                                         $matchingCount = $cnt;
2082                                                         $matchingCallback = $fn;
2083                                                 }
2084                                         }
2085
2086                                         if ($matchingCount == 0) {
2087                                                 $i += $count - 1;
2088                                                 continue;
2089                                         }
2090
2091                                         # lets set a title or last part (if '|' was found)
2092                                         if (null === $openingBraceStack[$lastOpeningBrace]['parts'])
2093                                                 $openingBraceStack[$lastOpeningBrace]['title'] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2094                                         else
2095                                                 $openingBraceStack[$lastOpeningBrace]['parts'][] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2096
2097                                         $pieceStart = $openingBraceStack[$lastOpeningBrace]['startAt'] - $matchingCount;
2098                                         $pieceEnd = $i + $matchingCount;
2099
2100                                         if( is_callable( $matchingCallback ) ) {
2101                                                 $cbArgs = array (
2102                                                                                  'text' => substr($text, $pieceStart, $pieceEnd - $pieceStart),
2103                                                                                  'title' => trim($openingBraceStack[$lastOpeningBrace]['title']),
2104                                                                                  'parts' => $openingBraceStack[$lastOpeningBrace]['parts'],
2105                                                                                  'lineStart' => (($pieceStart > 0) && ($text[$pieceStart-1] == '\n')),
2106                                                                                  );
2107                                                 # finally we can call a user callback and replace piece of text
2108                                                 $replaceWith = call_user_func( $matchingCallback, $cbArgs );
2109                                                 $text = substr($text, 0, $pieceStart) . $replaceWith . substr($text, $pieceEnd);
2110                                                 $i = $pieceStart + strlen($replaceWith) - 1;
2111                                         }
2112                                         else {
2113                                                 # null value for callback means that parentheses should be parsed, but not replaced
2114                                                 $i += $matchingCount - 1;
2115                                         }
2116
2117                                         # reset last openning parentheses, but keep it in case there are unused characters
2118                                         $piece = array('brace' => $openingBraceStack[$lastOpeningBrace]['brace'],
2119                                                                    'braceEnd' => $openingBraceStack[$lastOpeningBrace]['braceEnd'],
2120                                                                    'count' => $openingBraceStack[$lastOpeningBrace]['count'],
2121                                                                    'title' => '',
2122                                                                    'parts' => null,
2123                                                                    'startAt' => $openingBraceStack[$lastOpeningBrace]['startAt']);
2124                                         $openingBraceStack[$lastOpeningBrace--] = null;
2125
2126                                         if ($matchingCount < $piece['count']) {
2127                                                 $piece['count'] -= $matchingCount;
2128                                                 $piece['startAt'] -= $matchingCount;
2129                                                 $piece['partStart'] = $piece['startAt'];
2130                                                 # do we still qualify for any callback with remaining count?
2131                                                 foreach ($callbacks[$piece['brace']]['cb'] as $cnt => $fn) {
2132                                                         if ($piece['count'] >= $cnt) {
2133                                                                 $lastOpeningBrace ++;
2134                                                                 $openingBraceStack[$lastOpeningBrace] = $piece;
2135                                                                 break;
2136                                                         }
2137                                                 }
2138                                         }
2139                                         continue;
2140                                 }
2141
2142                                 # lets set a title if it is a first separator, or next part otherwise
2143                                 if ($text[$i] == '|') {
2144                                         if (null === $openingBraceStack[$lastOpeningBrace]['parts']) {
2145                                                 $openingBraceStack[$lastOpeningBrace]['title'] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2146                                                 $openingBraceStack[$lastOpeningBrace]['parts'] = array();
2147                                         }
2148                                         else
2149                                                 $openingBraceStack[$lastOpeningBrace]['parts'][] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2150
2151                                         $openingBraceStack[$lastOpeningBrace]['partStart'] = $i + 1;
2152                                 }
2153                         }
2154                 }
2155
2156                 return $text;
2157         }
2158
2159         /**
2160          * Replace magic variables, templates, and template arguments
2161          * with the appropriate text. Templates are substituted recursively,
2162          * taking care to avoid infinite loops.
2163          *
2164          * Note that the substitution depends on value of $mOutputType:
2165          *  OT_WIKI: only {{subst:}} templates
2166          *  OT_MSG: only magic variables
2167          *  OT_HTML: all templates and magic variables
2168          *
2169          * @param string $tex The text to transform
2170          * @param array $args Key-value pairs representing template parameters to substitute
2171          * @access private
2172          */
2173         function replaceVariables( $text, $args = array() ) {
2174                 # Prevent too big inclusions
2175                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
2176                         return $text;
2177                 }
2178
2179                 $fname = 'Parser::replaceVariables';
2180                 wfProfileIn( $fname );
2181
2182                 # This function is called recursively. To keep track of arguments we need a stack:
2183                 array_push( $this->mArgStack, $args );
2184
2185                 $braceCallbacks = array();
2186                 $braceCallbacks[2] = array( &$this, 'braceSubstitution' );
2187                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
2188                         $braceCallbacks[3] = array( &$this, 'argSubstitution' );
2189                 }
2190                 $callbacks = array();
2191                 $callbacks['{'] = array('end' => '}', 'cb' => $braceCallbacks);
2192                 $callbacks['['] = array('end' => ']', 'cb' => array(2=>null));
2193                 $text = $this->replace_callback ($text, $callbacks);
2194
2195                 array_pop( $this->mArgStack );
2196
2197                 wfProfileOut( $fname );
2198                 return $text;
2199         }
2200
2201         /**
2202          * Replace magic variables
2203          * @access private
2204          */
2205         function variableSubstitution( $matches ) {
2206                 $fname = 'parser::variableSubstitution';
2207                 $varname = $matches[1];
2208                 wfProfileIn( $fname );
2209                 if ( !$this->mVariables ) {
2210                         $this->initialiseVariables();
2211                 }
2212                 $skip = false;
2213                 if ( $this->mOutputType == OT_WIKI ) {
2214                         # Do only magic variables prefixed by SUBST
2215                         $mwSubst =& MagicWord::get( MAG_SUBST );
2216                         if (!$mwSubst->matchStartAndRemove( $varname ))
2217                                 $skip = true;
2218                         # Note that if we don't substitute the variable below,
2219                         # we don't remove the {{subst:}} magic word, in case
2220                         # it is a template rather than a magic variable.
2221                 }
2222                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
2223                         $id = $this->mVariables[$varname];
2224                         $text = $this->getVariableValue( $id );
2225                         $this->mOutput->mContainsOldMagic = true;
2226                 } else {
2227                         $text = $matches[0];
2228                 }
2229                 wfProfileOut( $fname );
2230                 return $text;
2231         }
2232
2233         # Split template arguments
2234         function getTemplateArgs( $argsString ) {
2235                 if ( $argsString === '' ) {
2236                         return array();
2237                 }
2238
2239                 $args = explode( '|', substr( $argsString, 1 ) );
2240
2241                 # If any of the arguments contains a '[[' but no ']]', it needs to be
2242                 # merged with the next arg because the '|' character between belongs
2243                 # to the link syntax and not the template parameter syntax.
2244                 $argc = count($args);
2245
2246                 for ( $i = 0; $i < $argc-1; $i++ ) {
2247                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
2248                                 $args[$i] .= '|'.$args[$i+1];
2249                                 array_splice($args, $i+1, 1);
2250                                 $i--;
2251                                 $argc--;
2252                         }
2253                 }
2254
2255                 return $args;
2256         }
2257
2258         /**
2259          * Return the text of a template, after recursively
2260          * replacing any variables or templates within the template.
2261          *
2262          * @param array $piece The parts of the template
2263          *  $piece['text']: matched text
2264          *  $piece['title']: the title, i.e. the part before the |
2265          *  $piece['parts']: the parameter array
2266          * @return string the text of the template
2267          * @access private
2268          */
2269         function braceSubstitution( $piece ) {
2270                 global $wgLinkCache, $wgContLang;
2271                 $fname = 'Parser::braceSubstitution';
2272                 wfProfileIn( $fname );
2273
2274                 $found = false;
2275                 $nowiki = false;
2276                 $noparse = false;
2277
2278                 $title = NULL;
2279
2280                 $linestart = '';
2281
2282                 # $part1 is the bit before the first |, and must contain only title characters
2283                 # $args is a list of arguments, starting from index 0, not including $part1
2284
2285                 $part1 = $piece['title'];
2286                 # If the third subpattern matched anything, it will start with |
2287
2288                 if (null == $piece['parts']) {
2289                         $replaceWith = $this->variableSubstitution (array ($piece['text'], $piece['title']));
2290                         if ($replaceWith != $piece['text']) {
2291                                 $text = $replaceWith;
2292                                 $found = true;
2293                                 $noparse = true;
2294                         }
2295                 }
2296
2297                 $args = (null == $piece['parts']) ? array() : $piece['parts'];
2298                 $argc = count( $args );
2299
2300                 # SUBST
2301                 if ( !$found ) {
2302                         $mwSubst =& MagicWord::get( MAG_SUBST );
2303                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
2304                                 # One of two possibilities is true:
2305                                 # 1) Found SUBST but not in the PST phase
2306                                 # 2) Didn't find SUBST and in the PST phase
2307                                 # In either case, return without further processing
2308                                 $text = $piece['text'];
2309                                 $found = true;
2310                                 $noparse = true;
2311                         }
2312                 }
2313
2314                 # MSG, MSGNW and INT
2315                 if ( !$found ) {
2316                         # Check for MSGNW:
2317                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
2318                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
2319                                 $nowiki = true;
2320                         } else {
2321                                 # Remove obsolete MSG:
2322                                 $mwMsg =& MagicWord::get( MAG_MSG );
2323                                 $mwMsg->matchStartAndRemove( $part1 );
2324                         }
2325
2326                         # Check if it is an internal message
2327                         $mwInt =& MagicWord::get( MAG_INT );
2328                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
2329                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
2330                                         $text = $linestart . wfMsgReal( $part1, $args, true );
2331                                         $found = true;
2332                                 }
2333                         }
2334                 }
2335
2336                 # NS
2337                 if ( !$found ) {
2338                         # Check for NS: (namespace expansion)
2339                         $mwNs = MagicWord::get( MAG_NS );
2340                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
2341                                 if ( intval( $part1 ) ) {
2342                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
2343                                         $found = true;
2344                                 } else {
2345                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
2346                                         if ( !is_null( $index ) ) {
2347                                                 $text = $linestart . $wgContLang->getNsText( $index );
2348                                                 $found = true;
2349                                         }
2350                                 }
2351                         }
2352                 }
2353
2354                 # LCFIRST, UCFIRST, LC and UC
2355                 if ( !$found ) {
2356                         $lcfirst =& MagicWord::get( MAG_LCFIRST );
2357                         $ucfirst =& MagicWord::get( MAG_UCFIRST );
2358                         $lc =& MagicWord::get( MAG_LC );
2359                         $uc =& MagicWord::get( MAG_UC );
2360                         if ( $lcfirst->matchStartAndRemove( $part1 ) ) {
2361                                 $text = $linestart . $wgContLang->lcfirst( $part1 );
2362                                 $found = true;
2363                         } else if ( $ucfirst->matchStartAndRemove( $part1 ) ) {
2364                                 $text = $linestart . $wgContLang->ucfirst( $part1 );
2365                                 $found = true;
2366                         } else if ( $lc->matchStartAndRemove( $part1 ) ) {
2367                                 $text = $linestart . $wgContLang->lc( $part1 );
2368                                 $found = true;
2369                         } else if ( $uc->matchStartAndRemove( $part1 ) ) {
2370                                  $text = $linestart . $wgContLang->uc( $part1 );
2371                                  $found = true;
2372                         }
2373                 }
2374
2375                 # LOCALURL and FULLURL
2376                 if ( !$found ) {
2377                         $mwLocal =& MagicWord::get( MAG_LOCALURL );
2378                         $mwLocalE =& MagicWord::get( MAG_LOCALURLE );
2379                         $mwFull =& MagicWord::get( MAG_FULLURL );
2380                         $mwFullE =& MagicWord::get( MAG_FULLURLE );
2381
2382
2383                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2384                                 $func = 'getLocalURL';
2385                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2386                                 $func = 'escapeLocalURL';
2387                         } elseif ( $mwFull->matchStartAndRemove( $part1 ) ) {
2388                                 $func = 'getFullURL';
2389                         } elseif ( $mwFullE->matchStartAndRemove( $part1 ) ) {
2390                                 $func = 'escapeFullURL';
2391                         } else {
2392                                 $func = false;
2393                         }
2394
2395                         if ( $func !== false ) {
2396                                 $title = Title::newFromText( $part1 );
2397                                 if ( !is_null( $title ) ) {
2398                                         if ( $argc > 0 ) {
2399                                                 $text = $linestart . $title->$func( $args[0] );
2400                                         } else {
2401                                                 $text = $linestart . $title->$func();
2402                                         }
2403                                         $found = true;
2404                                 }
2405                         }
2406                 }
2407
2408                 # GRAMMAR
2409                 if ( !$found && $argc == 1 ) {
2410                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2411                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2412                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2413                                 $found = true;
2414                         }
2415                 }
2416
2417                 # PLURAL
2418                 if ( !$found && $argc >= 2 ) {
2419                         $mwPluralForm =& MagicWord::get( MAG_PLURAL );
2420                         if ( $mwPluralForm->matchStartAndRemove( $part1 ) ) {
2421                                 if ($argc==2) {$args[2]=$args[1];}
2422                                 $text = $linestart . $wgContLang->convertPlural( $part1, $args[0], $args[1], $args[2]);
2423                                 $found = true;
2424                         }
2425                 }
2426
2427                 # Template table test
2428
2429                 # Did we encounter this template already? If yes, it is in the cache
2430                 # and we need to check for loops.
2431                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2432                         $found = true;
2433
2434                         # Infinite loop test
2435                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2436                                 $noparse = true;
2437                                 $found = true;
2438                                 $text = $linestart .
2439                                         "\{\{$part1}}" .
2440                                         '<!-- WARNING: template loop detected -->';
2441                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2442                         } else {
2443                                 # set $text to cached message.
2444                                 $text = $linestart . $this->mTemplates[$part1];
2445                         }
2446                 }
2447
2448                 # Load from database
2449                 $replaceHeadings = false;
2450                 $isHTML = false;
2451                 $lastPathLevel = $this->mTemplatePath;
2452                 if ( !$found ) {
2453                         $ns = NS_TEMPLATE;
2454                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2455                         if ($subpage !== '') {
2456                                 $ns = $this->mTitle->getNamespace();
2457                         }
2458                         $title = Title::newFromText( $part1, $ns );
2459
2460                         if ($title) {
2461                                 $interwiki = Title::getInterwikiLink($title->getInterwiki());
2462                                 if ($interwiki != '' && $title->isTrans()) {
2463                                         return $this->scarytransclude($title, $interwiki);
2464                                 }
2465                         }
2466
2467                         if ( !is_null( $title ) && !$title->isExternal() ) {
2468                                 # Check for excessive inclusion
2469                                 $dbk = $title->getPrefixedDBkey();
2470                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2471                                         if ( $title->getNamespace() == NS_SPECIAL && $this->mOptions->getAllowSpecialInclusion() ) {
2472                                                 # Capture special page output
2473                                                 $text = SpecialPage::capturePath( $title );
2474                                                 if ( is_string( $text ) ) {
2475                                                         $found = true;
2476                                                         $noparse = true;
2477                                                         $isHTML = true;
2478                                                         $this->disableCache();
2479                                                 }
2480                                         } else {
2481                                                 $article = new Article( $title );
2482                                                 $articleContent = $article->fetchContent(0, false);
2483                                                 if ( $articleContent !== false ) {
2484                                                         $found = true;
2485                                                         $text = $articleContent;
2486                                                         $replaceHeadings = true;
2487                                                 }
2488                                         }
2489                                 }
2490
2491                                 # If the title is valid but undisplayable, make a link to it
2492                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2493                                         $text = '[['.$title->getPrefixedText().']]';
2494                                         $found = true;
2495                                 }
2496
2497                                 # Template cache array insertion
2498                                 if( $found ) {
2499                                         $this->mTemplates[$part1] = $text;
2500                                         $text = $linestart . $text;
2501                                 }
2502                         }
2503                 }
2504
2505                 # Recursive parsing, escaping and link table handling
2506                 # Only for HTML output
2507                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2508                         $text = wfEscapeWikiText( $text );
2509                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2510                         # Clean up argument array
2511                         $assocArgs = array();
2512                         $index = 1;
2513                         foreach( $args as $arg ) {
2514                                 $eqpos = strpos( $arg, '=' );
2515                                 if ( $eqpos === false ) {
2516                                         $assocArgs[$index++] = $arg;
2517                                 } else {
2518                                         $name = trim( substr( $arg, 0, $eqpos ) );
2519                                         $value = trim( substr( $arg, $eqpos+1 ) );
2520                                         if ( $value === false ) {
2521                                                 $value = '';
2522                                         }
2523                                         if ( $name !== false ) {
2524                                                 $assocArgs[$name] = $value;
2525                                         }
2526                                 }
2527                         }
2528
2529                         # Add a new element to the templace recursion path
2530                         $this->mTemplatePath[$part1] = 1;
2531
2532                         # If there are any <onlyinclude> tags, only include them
2533                         if ( in_string( '<onlyinclude>', $text ) && in_string( '</onlyinclude>', $text ) ) {
2534                                 preg_match_all( '/<onlyinclude>(.*?)<\/onlyinclude>/s', $text, $m );
2535                                 $text = '';
2536                                 foreach ($m[1] as $piece)
2537                                         $text .= $this->trimOnlyinclude( $piece );
2538                         }
2539                         # Remove <noinclude> sections and <includeonly> tags
2540                         $text = preg_replace( '/<noinclude>.*?<\/noinclude>/s', '', $text );
2541                         $text = strtr( $text, array( '<includeonly>' => '' , '</includeonly>' => '' ) );
2542
2543                         if( $this->mOutputType == OT_HTML ) {
2544                                 # Strip <nowiki>, <pre>, etc.
2545                                 $text = $this->strip( $text, $this->mStripState );
2546                                 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'replaceVariables' ), $assocArgs );
2547                         }
2548                         $text = $this->replaceVariables( $text, $assocArgs );
2549
2550                         # Resume the link cache and register the inclusion as a link
2551                         if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
2552                                 $wgLinkCache->addLinkObj( $title );
2553                         }
2554
2555                         # If the template begins with a table or block-level
2556                         # element, it should be treated as beginning a new line.
2557                         if (!$piece['lineStart'] && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2558                                 $text = "\n" . $text;
2559                         }
2560                 }
2561                 # Prune lower levels off the recursion check path
2562                 $this->mTemplatePath = $lastPathLevel;
2563
2564                 if ( !$found ) {
2565                         wfProfileOut( $fname );
2566                         return $piece['text'];
2567                 } else {
2568                         if ( $isHTML ) {
2569                                 # Replace raw HTML by a placeholder
2570                                 # Add a blank line preceding, to prevent it from mucking up
2571                                 # immediately preceding headings
2572                                 $text = "\n\n" . $this->insertStripItem( $text, $this->mStripState );
2573                         } else {
2574                                 # replace ==section headers==
2575                                 # XXX this needs to go away once we have a better parser.
2576                                 if ( $this->mOutputType != OT_WIKI && $replaceHeadings ) {
2577                                         if( !is_null( $title ) )
2578                                                 $encodedname = base64_encode($title->getPrefixedDBkey());
2579                                         else
2580                                                 $encodedname = base64_encode("");
2581                                         $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2582                                                 PREG_SPLIT_DELIM_CAPTURE);
2583                                         $text = '';
2584                                         $nsec = 0;
2585                                         for( $i = 0; $i < count($m); $i += 2 ) {
2586                                                 $text .= $m[$i];
2587                                                 if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2588                                                 $hl = $m[$i + 1];
2589                                                 if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2590                                                         $text .= $hl;
2591                                                         continue;
2592                                                 }
2593                                                 preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2594                                                 $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2595                                                         . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2596
2597                                                 $nsec++;
2598                                         }
2599                                 }
2600                         }
2601                 }
2602
2603                 # Prune lower levels off the recursion check path
2604                 $this->mTemplatePath = $lastPathLevel;
2605
2606                 if ( !$found ) {
2607                         wfProfileOut( $fname );
2608                         return $piece['text'];
2609                 } else {
2610                         wfProfileOut( $fname );
2611                         return $text;
2612                 }
2613         }
2614
2615         /**
2616          * Trim the first and last newlines of a string, this is not equivalent
2617          * to trim( $str, "\n" ) which would trim them all.
2618          *
2619          * @param string $str The string to trim
2620          * @return string
2621          */
2622         function trimOnlyinclude( $str ) {
2623                 $str = preg_replace( "/^\n/", '', $str );
2624                 $str = preg_replace( "/\n$/", '', $str );
2625                 return $str;
2626         }
2627
2628         /**
2629          * Translude an interwiki link.
2630          */
2631         function scarytransclude($title, $interwiki) {
2632                 global $wgEnableScaryTranscluding;
2633
2634                 if (!$wgEnableScaryTranscluding)
2635                         return wfMsg('scarytranscludedisabled');
2636
2637                 $articlename = "Template:" . $title->getDBkey();
2638                 $url = str_replace('$1', urlencode($articlename), $interwiki);
2639                 if (strlen($url) > 255)
2640                         return wfMsg('scarytranscludetoolong');
2641                 $text = $this->fetchScaryTemplateMaybeFromCache($url);
2642                 $this->mIWTransData[] = $text;
2643                 return "<!--IW_TRANSCLUDE ".(count($this->mIWTransData) - 1)."-->";
2644         }
2645
2646         function fetchScaryTemplateMaybeFromCache($url) {
2647                 $dbr =& wfGetDB(DB_SLAVE);
2648                 $obj = $dbr->selectRow('transcache', array('tc_time', 'tc_contents'),
2649                                 array('tc_url' => $url));
2650                 if ($obj) {
2651                         $time = $obj->tc_time;
2652                         $text = $obj->tc_contents;
2653                         if ($time && $time < (time() + (60*60))) {
2654                                 return $text;
2655                         }
2656                 }
2657
2658                 $text = wfGetHTTP($url . '?action=render');
2659                 if (!$text)
2660                         return wfMsg('scarytranscludefailed', $url);
2661
2662                 $dbw =& wfGetDB(DB_MASTER);
2663                 $dbw->replace('transcache', array(), array(
2664                         'tc_url' => $url,
2665                         'tc_time' => time(),
2666                         'tc_contents' => $text));
2667                 return $text;
2668         }
2669
2670
2671         /**
2672          * Triple brace replacement -- used for template arguments
2673          * @access private
2674          */
2675         function argSubstitution( $matches ) {
2676                 $arg = trim( $matches['title'] );
2677                 $text = $matches['text'];
2678                 $inputArgs = end( $this->mArgStack );
2679
2680                 if ( array_key_exists( $arg, $inputArgs ) ) {
2681                         $text = $inputArgs[$arg];
2682                 } else if ($this->mOutputType == OT_HTML && null != $matches['parts'] && count($matches['parts']) > 0) {
2683                         $text = $matches['parts'][0];
2684                 }
2685
2686                 return $text;
2687         }
2688
2689         /**
2690          * Returns true if the function is allowed to include this entity
2691          * @access private
2692          */
2693         function incrementIncludeCount( $dbk ) {
2694                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2695                         $this->mIncludeCount[$dbk] = 0;
2696                 }
2697                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2698                         return true;
2699                 } else {
2700                         return false;
2701                 }
2702         }
2703
2704         /**
2705          * This function accomplishes several tasks:
2706          * 1) Auto-number headings if that option is enabled
2707          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2708          * 3) Add a Table of contents on the top for users who have enabled the option
2709          * 4) Auto-anchor headings
2710          *
2711          * It loops through all headlines, collects the necessary data, then splits up the
2712          * string and re-inserts the newly formatted headlines.
2713          *
2714          * @param string $text
2715          * @param boolean $isMain
2716          * @access private
2717          */
2718         function formatHeadings( $text, $isMain=true ) {
2719                 global $wgMaxTocLevel, $wgContLang, $wgLinkHolders, $wgInterwikiLinkHolders;
2720
2721                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2722                 $doShowToc = true;
2723                 $forceTocHere = false;
2724                 if( !$this->mTitle->userCanEdit() ) {
2725                         $showEditLink = 0;
2726                 } else {
2727                         $showEditLink = $this->mOptions->getEditSection();
2728                 }
2729
2730                 # Inhibit editsection links if requested in the page
2731                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2732                 if( $esw->matchAndRemove( $text ) ) {
2733                         $showEditLink = 0;
2734                 }
2735                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2736                 # do not add TOC
2737                 $mw =& MagicWord::get( MAG_NOTOC );
2738                 if( $mw->matchAndRemove( $text ) ) {
2739                         $doShowToc = false;
2740                 }
2741
2742                 # Get all headlines for numbering them and adding funky stuff like [edit]
2743                 # links - this is for later, but we need the number of headlines right now
2744                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2745
2746                 # if there are fewer than 4 headlines in the article, do not show TOC
2747                 if( $numMatches < 4 ) {
2748                         $doShowToc = false;
2749                 }
2750
2751                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2752                 # override above conditions and always show TOC at that place
2753
2754                 $mw =& MagicWord::get( MAG_TOC );
2755                 if($mw->match( $text ) ) {
2756                         $doShowToc = true;
2757                         $forceTocHere = true;
2758                 } else {
2759                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2760                         # override above conditions and always show TOC above first header
2761                         $mw =& MagicWord::get( MAG_FORCETOC );
2762                         if ($mw->matchAndRemove( $text ) ) {
2763                                 $doShowToc = true;
2764                         }
2765                 }
2766
2767                 # Never ever show TOC if no headers
2768                 if( $numMatches < 1 ) {
2769                         $doShowToc = false;
2770                 }
2771
2772                 # We need this to perform operations on the HTML
2773                 $sk =& $this->mOptions->getSkin();
2774
2775                 # headline counter
2776                 $headlineCount = 0;
2777                 $sectionCount = 0; # headlineCount excluding template sections
2778
2779                 # Ugh .. the TOC should have neat indentation levels which can be
2780                 # passed to the skin functions. These are determined here
2781                 $toc = '';
2782                 $full = '';
2783                 $head = array();
2784                 $sublevelCount = array();
2785                 $levelCount = array();
2786                 $toclevel = 0;
2787                 $level = 0;
2788                 $prevlevel = 0;
2789                 $toclevel = 0;
2790                 $prevtoclevel = 0;
2791
2792                 foreach( $matches[3] as $headline ) {
2793                         $istemplate = 0;
2794                         $templatetitle = '';
2795                         $templatesection = 0;
2796                         $numbering = '';
2797
2798                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2799                                 $istemplate = 1;
2800                                 $templatetitle = base64_decode($mat[1]);
2801                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2802                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2803                         }
2804
2805                         if( $toclevel ) {
2806                                 $prevlevel = $level;
2807                                 $prevtoclevel = $toclevel;
2808                         }
2809                         $level = $matches[1][$headlineCount];
2810
2811                         if( $doNumberHeadings || $doShowToc ) {
2812
2813                                 if ( $level > $prevlevel ) {
2814                                         # Increase TOC level
2815                                         $toclevel++;
2816                                         $sublevelCount[$toclevel] = 0;
2817                                         $toc .= $sk->tocIndent();
2818                                 }
2819                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2820                                         # Decrease TOC level, find level to jump to
2821
2822                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2823                                                 # Can only go down to level 1
2824                                                 $toclevel = 1;
2825                                         } else {
2826                                                 for ($i = $toclevel; $i > 0; $i--) {
2827                                                         if ( $levelCount[$i] == $level ) {
2828                                                                 # Found last matching level
2829                                                                 $toclevel = $i;
2830                                                                 break;
2831                                                         }
2832                                                         elseif ( $levelCount[$i] < $level ) {
2833                                                                 # Found first matching level below current level
2834                                                                 $toclevel = $i + 1;
2835                                                                 break;
2836                                                         }
2837                                                 }
2838                                         }
2839
2840                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2841                                 }
2842                                 else {
2843                                         # No change in level, end TOC line
2844                                         $toc .= $sk->tocLineEnd();
2845                                 }
2846
2847                                 $levelCount[$toclevel] = $level;
2848
2849                                 # count number of headlines for each level
2850                                 @$sublevelCount[$toclevel]++;
2851                                 $dot = 0;
2852                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2853                                         if( !empty( $sublevelCount[$i] ) ) {
2854                                                 if( $dot ) {
2855                                                         $numbering .= '.';
2856                                                 }
2857                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2858                                                 $dot = 1;
2859                                         }
2860                                 }
2861                         }
2862
2863                         # The canonized header is a version of the header text safe to use for links
2864                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2865                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2866                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2867
2868                         # Remove link placeholders by the link text.
2869                         #     <!--LINK number-->
2870                         # turns into
2871                         #     link text with suffix
2872                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2873                                                             "\$this->mLinkHolders['texts'][\$1]",
2874                                                             $canonized_headline );
2875                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2876                                                             "\$this->mInterwikiLinkHolders['texts'][\$1]",
2877                                                             $canonized_headline );
2878
2879                         # strip out HTML
2880                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2881                         $tocline = trim( $canonized_headline );
2882                         $canonized_headline = urlencode( Sanitizer::decodeCharReferences( str_replace(' ', '_', $tocline) ) );
2883                         $replacearray = array(
2884                                 '%3A' => ':',
2885                                 '%' => '.'
2886                         );
2887                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2888                         $refers[$headlineCount] = $canonized_headline;
2889
2890                         # count how many in assoc. array so we can track dupes in anchors
2891                         @$refers[$canonized_headline]++;
2892                         $refcount[$headlineCount]=$refers[$canonized_headline];
2893
2894                         # Don't number the heading if it is the only one (looks silly)
2895                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2896                                 # the two are different if the line contains a link
2897                                 $headline=$numbering . ' ' . $headline;
2898                         }
2899
2900                         # Create the anchor for linking from the TOC to the section
2901                         $anchor = $canonized_headline;
2902                         if($refcount[$headlineCount] > 1 ) {
2903                                 $anchor .= '_' . $refcount[$headlineCount];
2904                         }
2905                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2906                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2907                         }
2908                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2909                                 if ( empty( $head[$headlineCount] ) ) {
2910                                         $head[$headlineCount] = '';
2911                                 }
2912                                 if( $istemplate )
2913                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2914                                 else
2915                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2916                         }
2917
2918                         # give headline the correct <h#> tag
2919                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2920
2921                         $headlineCount++;
2922                         if( !$istemplate )
2923                                 $sectionCount++;
2924                 }
2925
2926                 if( $doShowToc ) {
2927                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2928                         $toc = $sk->tocList( $toc );
2929                 }
2930
2931                 # split up and insert constructed headlines
2932
2933                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2934                 $i = 0;
2935
2936                 foreach( $blocks as $block ) {
2937                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2938                                 # This is the [edit] link that appears for the top block of text when
2939                                 # section editing is enabled
2940
2941                                 # Disabled because it broke block formatting
2942                                 # For example, a bullet point in the top line
2943                                 # $full .= $sk->editSectionLink(0);
2944                         }
2945                         $full .= $block;
2946                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2947                         # Top anchor now in skin
2948                                 $full = $full.$toc;
2949                         }
2950
2951                         if( !empty( $head[$i] ) ) {
2952                                 $full .= $head[$i];
2953                         }
2954                         $i++;
2955                 }
2956                 if($forceTocHere) {
2957                         $mw =& MagicWord::get( MAG_TOC );
2958                         return $mw->replace( $toc, $full );
2959                 } else {
2960                         return $full;
2961                 }
2962         }
2963
2964         /**
2965          * Return an HTML link for the "ISBN 123456" text
2966          * @access private
2967          */
2968         function magicISBN( $text ) {
2969                 $fname = 'Parser::magicISBN';
2970                 wfProfileIn( $fname );
2971
2972                 $a = split( 'ISBN ', ' '.$text );
2973                 if ( count ( $a ) < 2 ) {
2974                         wfProfileOut( $fname );
2975                         return $text;
2976                 }
2977                 $text = substr( array_shift( $a ), 1);
2978                 $valid = '0123456789-Xx';
2979
2980                 foreach ( $a as $x ) {
2981                         $isbn = $blank = '' ;
2982                         while ( ' ' == $x{0} ) {
2983                                 $blank .= ' ';
2984                                 $x = substr( $x, 1 );
2985                         }
2986                         if ( $x == '' ) { # blank isbn
2987                                 $text .= "ISBN $blank";
2988                                 continue;
2989                         }
2990                         while ( strstr( $valid, $x{0} ) != false ) {
2991                                 $isbn .= $x{0};
2992                                 $x = substr( $x, 1 );
2993                         }
2994                         $num = str_replace( '-', '', $isbn );
2995                         $num = str_replace( ' ', '', $num );
2996                         $num = str_replace( 'x', 'X', $num );
2997
2998                         if ( '' == $num ) {
2999                                 $text .= "ISBN $blank$x";
3000                         } else {
3001                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
3002                                 $text .= '<a href="' .
3003                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
3004                                         "\" class=\"internal\">ISBN $isbn</a>";
3005                                 $text .= $x;
3006                         }
3007                 }
3008                 wfProfileOut( $fname );
3009                 return $text;
3010         }
3011
3012         /**
3013          * Return an HTML link for the "RFC 1234" text
3014          *
3015          * @access private
3016          * @param string $text     Text to be processed
3017          * @param string $keyword  Magic keyword to use (default RFC)
3018          * @param string $urlmsg   Interface message to use (default rfcurl)
3019          * @return string
3020          */
3021         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
3022
3023                 $valid = '0123456789';
3024                 $internal = false;
3025
3026                 $a = split( $keyword, ' '.$text );
3027                 if ( count ( $a ) < 2 ) {
3028                         return $text;
3029                 }
3030                 $text = substr( array_shift( $a ), 1);
3031
3032                 /* Check if keyword is preceed by [[.
3033                  * This test is made here cause of the array_shift above
3034                  * that prevent the test to be done in the foreach.
3035                  */
3036                 if ( substr( $text, -2 ) == '[[' ) {
3037                         $internal = true;
3038                 }
3039
3040                 foreach ( $a as $x ) {
3041                         /* token might be empty if we have RFC RFC 1234 */
3042                         if ( $x=='' ) {
3043                                 $text.=$keyword;
3044                                 continue;
3045                                 }
3046
3047                         $id = $blank = '' ;
3048
3049                         /** remove and save whitespaces in $blank */
3050                         while ( $x{0} == ' ' ) {
3051                                 $blank .= ' ';
3052                                 $x = substr( $x, 1 );
3053                         }
3054
3055                         /** remove and save the rfc number in $id */
3056                         while ( strstr( $valid, $x{0} ) != false ) {
3057                                 $id .= $x{0};
3058                                 $x = substr( $x, 1 );
3059                         }
3060
3061                         if ( $id == '' ) {
3062                                 /* call back stripped spaces*/
3063                                 $text .= $keyword.$blank.$x;
3064                         } elseif( $internal ) {
3065                                 /* normal link */
3066                                 $text .= $keyword.$id.$x;
3067                         } else {
3068                                 /* build the external link*/
3069                                 $url = wfMsg( $urlmsg, $id);
3070                                 $sk =& $this->mOptions->getSkin();
3071                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
3072                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
3073                         }
3074
3075                         /* Check if the next RFC keyword is preceed by [[ */
3076                         $internal = ( substr($x,-2) == '[[' );
3077                 }
3078                 return $text;
3079         }
3080
3081         /**
3082          * Transform wiki markup when saving a page by doing \r\n -> \n
3083          * conversion, substitting signatures, {{subst:}} templates, etc.
3084          *
3085          * @param string $text the text to transform
3086          * @param Title &$title the Title object for the current article
3087          * @param User &$user the User object describing the current user
3088          * @param ParserOptions $options parsing options
3089          * @param bool $clearState whether to clear the parser state first
3090          * @return string the altered wiki markup
3091          * @access public
3092          */
3093         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
3094                 $this->mOptions = $options;
3095                 $this->mTitle =& $title;
3096                 $this->mOutputType = OT_WIKI;
3097
3098                 if ( $clearState ) {
3099                         $this->clearState();
3100                 }
3101
3102                 $stripState = false;
3103                 $pairs = array(
3104                         "\r\n" => "\n",
3105                 );
3106                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
3107                 $text = $this->strip( $text, $stripState, true );
3108                 $text = $this->pstPass2( $text, $user );
3109                 $text = $this->unstrip( $text, $stripState );
3110                 $text = $this->unstripNoWiki( $text, $stripState );
3111                 return $text;
3112         }
3113
3114         /**
3115          * Pre-save transform helper function
3116          * @access private
3117          */
3118         function pstPass2( $text, &$user ) {
3119                 global $wgContLang, $wgLocaltimezone;
3120
3121                 # Variable replacement
3122                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
3123                 $text = $this->replaceVariables( $text );
3124
3125                 # Signatures
3126                 #
3127                 $sigText = $this->getUserSig( $user );
3128
3129                 /* Note: This is the timestamp saved as hardcoded wikitext to
3130                  * the database, we use $wgContLang here in order to give
3131                  * everyone the same signiture and use the default one rather
3132                  * than the one selected in each users preferences.
3133                  */
3134                 if ( isset( $wgLocaltimezone ) ) {
3135                         $oldtz = getenv( 'TZ' );
3136                         putenv( 'TZ='.$wgLocaltimezone );
3137                 }
3138                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false, false) .
3139                   ' (' . date( 'T' ) . ')';
3140                 if ( isset( $wgLocaltimezone ) ) {
3141                         putenv( 'TZ='.$oldtz );
3142                 }
3143
3144                 $text = preg_replace( '/~~~~~/', $d, $text );
3145                 $text = preg_replace( '/~~~~/', "$sigText $d", $text );
3146                 $text = preg_replace( '/~~~/', $sigText, $text );
3147
3148                 # Context links: [[|name]] and [[name (context)|]]
3149                 #
3150                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
3151                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
3152                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
3153                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
3154
3155                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
3156                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
3157                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
3158                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
3159                 $context = '';
3160                 $t = $this->mTitle->getText();
3161                 if ( preg_match( $conpat, $t, $m ) ) {
3162                         $context = $m[2];
3163                 }
3164                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
3165                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
3166                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
3167
3168                 if ( '' == $context ) {
3169                         $text = preg_replace( $p2, '[[\\1]]', $text );
3170                 } else {
3171                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
3172                 }
3173
3174                 # Trim trailing whitespace
3175                 # MAG_END (__END__) tag allows for trailing
3176                 # whitespace to be deliberately included
3177                 $text = rtrim( $text );
3178                 $mw =& MagicWord::get( MAG_END );
3179                 $mw->matchAndRemove( $text );
3180
3181                 return $text;
3182         }
3183
3184         /**
3185          * Fetch the user's signature text, if any, and normalize to
3186          * validated, ready-to-insert wikitext.
3187          *
3188          * @param User $user
3189          * @return string
3190          * @access private
3191          */
3192         function getUserSig( &$user ) {
3193                 $name = $user->getName();
3194                 $nick = trim( $user->getOption( 'nickname' ) );
3195                 if ( '' == $nick ) {
3196                         $nick = $name;
3197                 }
3198
3199                 if( $user->getOption( 'fancysig' ) ) {
3200                         // A wikitext signature.
3201                         $valid = $this->validateSig( $nick );
3202                         if( $valid === false ) {
3203                                 // Fall back to default sig
3204                                 $nick = $name;
3205                                 wfDebug( "Parser::getUserSig: $name has bad XML tags in signature.\n" );
3206                         } else {
3207                                 return $nick;
3208                         }
3209                 }
3210
3211                 // Plain text linking to the user's homepage
3212                 global $wgContLang;
3213                 $page = $user->getUserPage();
3214                 return '[[' .
3215                         $page->getPrefixedText() .
3216                         "|" .
3217                         wfEscapeWikIText( $nick ) .
3218                         "]]";
3219         }
3220
3221         /**
3222          * We want to enforce two rules on wikitext sigs here:
3223          * 1) Expand any templates at save time (forced subst:)
3224          * 2) Check for unbalanced XML tags, and reject if so.
3225          *
3226          * @param string $text
3227          * @return mixed An expanded string, or false if invalid.
3228          *
3229          * @todo Run brace substitutions
3230          * @todo ?? Check for unbalanced '' and ''' quotes, etc
3231          */
3232         function validateSig( $text ) {
3233                 if( wfIsWellFormedXmlFragment( $text ) ) {
3234                         return $text;
3235                 } else {
3236                         return false;
3237                 }
3238         }
3239
3240         /**
3241          * Set up some variables which are usually set up in parse()
3242          * so that an external function can call some class members with confidence
3243          * @access public
3244          */
3245         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
3246                 $this->mTitle =& $title;
3247                 $this->mOptions = $options;
3248                 $this->mOutputType = $outputType;
3249                 if ( $clearState ) {
3250                         $this->clearState();
3251                 }
3252         }
3253
3254         /**
3255          * Transform a MediaWiki message by replacing magic variables.
3256          *
3257          * @param string $text the text to transform
3258          * @param ParserOptions $options  options
3259          * @return string the text with variables substituted
3260          * @access public
3261          */
3262         function transformMsg( $text, $options ) {
3263                 global $wgTitle;
3264                 static $executing = false;
3265
3266                 # Guard against infinite recursion
3267                 if ( $executing ) {
3268                         return $text;
3269                 }
3270                 $executing = true;
3271
3272                 $this->mTitle = $wgTitle;
3273                 $this->mOptions = $options;
3274                 $this->mOutputType = OT_MSG;
3275                 $this->clearState();
3276                 $text = $this->replaceVariables( $text );
3277
3278                 $executing = false;
3279                 return $text;
3280         }
3281
3282         /**
3283          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
3284          * Callback will be called with the text within
3285          * Transform and return the text within
3286          *
3287          * @access public
3288          *
3289          * @param mixed $tag The tag to use, e.g. 'hook' for <hook>
3290          * @param mixed $callback The callback function to use for the tag
3291          * @param mixed $object The object which contains the callback function, optional
3292          *
3293          * @return The old value of the mTagHooks array associated with the hook
3294          */
3295         function setHook( $tag, $callback, $object = null ) {
3296                 $oldVal = @$this->mTagHooks[$tag];
3297                 $this->mTagHooks[$tag] = $callback;
3298
3299                 if ( isset( $object ) )
3300                         $this->mTagHookObjects[$tag] =& $object;
3301                 else
3302                         $this->mTagHookObjects[$tag] = null;
3303
3304                 return $oldVal;
3305         }
3306
3307         /**
3308          * Replace <!--LINK--> link placeholders with actual links, in the buffer
3309          * Placeholders created in Skin::makeLinkObj()
3310          * Returns an array of links found, indexed by PDBK:
3311          *  0 - broken
3312          *  1 - normal link
3313          *  2 - stub
3314          * $options is a bit field, RLH_FOR_UPDATE to select for update
3315          */
3316         function replaceLinkHolders( &$text, $options = 0 ) {
3317                 global $wgUser, $wgLinkCache;
3318                 global $wgOutputReplace;
3319
3320                 $fname = 'Parser::replaceLinkHolders';
3321                 wfProfileIn( $fname );
3322
3323                 $pdbks = array();
3324                 $colours = array();
3325                 $sk = $this->mOptions->getSkin();
3326
3327                 if ( !empty( $this->mLinkHolders['namespaces'] ) ) {
3328                         wfProfileIn( $fname.'-check' );
3329                         $dbr =& wfGetDB( DB_SLAVE );
3330                         $page = $dbr->tableName( 'page' );
3331                         $threshold = $wgUser->getOption('stubthreshold');
3332
3333                         # Sort by namespace
3334                         asort( $this->mLinkHolders['namespaces'] );
3335
3336                         # Generate query
3337                         $query = false;
3338                         foreach ( $this->mLinkHolders['namespaces'] as $key => $val ) {
3339                                 # Make title object
3340                                 $title = $this->mLinkHolders['titles'][$key];
3341
3342                                 # Skip invalid entries.
3343                                 # Result will be ugly, but prevents crash.
3344                                 if ( is_null( $title ) ) {
3345                                         continue;
3346                                 }
3347                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
3348
3349                                 # Check if it's in the link cache already
3350                                 if ( $title->isAlwaysKnown() || $wgLinkCache->getGoodLinkID( $pdbk ) ) {
3351                                         $colours[$pdbk] = 1;
3352                                 } elseif ( $wgLinkCache->isBadLink( $pdbk ) ) {
3353                                         $colours[$pdbk] = 0;
3354                                 } else {
3355                                         # Not in the link cache, add it to the query
3356                                         if ( !isset( $current ) ) {
3357                                                 $current = $val;
3358                                                 $query =  "SELECT page_id, page_namespace, page_title";
3359                                                 if ( $threshold > 0 ) {
3360                                                         $query .= ', page_len, page_is_redirect';
3361                                                 }
3362                                                 $query .= " FROM $page WHERE (page_namespace=$val AND page_title IN(";
3363                                         } elseif ( $current != $val ) {
3364                                                 $current = $val;
3365                                                 $query .= ")) OR (page_namespace=$val AND page_title IN(";
3366                                         } else {
3367                                                 $query .= ', ';
3368                                         }
3369
3370                                         $query .= $dbr->addQuotes( $this->mLinkHolders['dbkeys'][$key] );
3371                                 }
3372                         }
3373                         if ( $query ) {
3374                                 $query .= '))';
3375                                 if ( $options & RLH_FOR_UPDATE ) {
3376                                         $query .= ' FOR UPDATE';
3377                                 }
3378
3379                                 $res = $dbr->query( $query, $fname );
3380
3381                                 # Fetch data and form into an associative array
3382                                 # non-existent = broken
3383                                 # 1 = known
3384                                 # 2 = stub
3385                                 while ( $s = $dbr->fetchObject($res) ) {
3386                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
3387                                         $pdbk = $title->getPrefixedDBkey();
3388                                         $wgLinkCache->addGoodLinkObj( $s->page_id, $title );
3389
3390                                         if ( $threshold >  0 ) {
3391                                                 $size = $s->page_len;
3392                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
3393                                                         $colours[$pdbk] = 1;
3394                                                 } else {
3395                                                         $colours[$pdbk] = 2;
3396                                                 }
3397                                         } else {
3398                                                 $colours[$pdbk] = 1;
3399                                         }
3400                                 }
3401                         }
3402                         wfProfileOut( $fname.'-check' );
3403
3404                         # Construct search and replace arrays
3405                         wfProfileIn( $fname.'-construct' );
3406                         $wgOutputReplace = array();
3407                         foreach ( $this->mLinkHolders['namespaces'] as $key => $ns ) {
3408                                 $pdbk = $pdbks[$key];
3409                                 $searchkey = "<!--LINK $key-->";
3410                                 $title = $this->mLinkHolders['titles'][$key];
3411                                 if ( empty( $colours[$pdbk] ) ) {
3412                                         $wgLinkCache->addBadLinkObj( $title );
3413                                         $colours[$pdbk] = 0;
3414                                         $wgOutputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3415                                                                         $this->mLinkHolders['texts'][$key],
3416                                                                         $this->mLinkHolders['queries'][$key] );
3417                                 } elseif ( $colours[$pdbk] == 1 ) {
3418                                         $wgOutputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3419                                                                         $this->mLinkHolders['texts'][$key],
3420                                                                         $this->mLinkHolders['queries'][$key] );
3421                                 } elseif ( $colours[$pdbk] == 2 ) {
3422                                         $wgOutputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3423                                                                         $this->mLinkHolders['texts'][$key],
3424                                                                         $this->mLinkHolders['queries'][$key] );
3425                                 }
3426                         }
3427                         wfProfileOut( $fname.'-construct' );
3428
3429                         # Do the thing
3430                         wfProfileIn( $fname.'-replace' );
3431
3432                         $text = preg_replace_callback(
3433                                 '/(<!--LINK .*?-->)/',
3434                                 "wfOutputReplaceMatches",
3435                                 $text);
3436
3437                         wfProfileOut( $fname.'-replace' );
3438                 }
3439
3440                 # Now process interwiki link holders
3441                 # This is quite a bit simpler than internal links
3442                 if ( !empty( $this->mInterwikiLinkHolders['texts'] ) ) {
3443                         wfProfileIn( $fname.'-interwiki' );
3444                         # Make interwiki link HTML
3445                         $wgOutputReplace = array();
3446                         foreach( $this->mInterwikiLinkHolders['texts'] as $key => $link ) {
3447                                 $title = $this->mInterwikiLinkHolders['titles'][$key];
3448                                 $wgOutputReplace[$key] = $sk->makeLinkObj( $title, $link );
3449                         }
3450
3451                         $text = preg_replace_callback(
3452                                 '/<!--IWLINK (.*?)-->/',
3453                                 "wfOutputReplaceMatches",
3454                                 $text );
3455                         wfProfileOut( $fname.'-interwiki' );
3456                 }
3457
3458                 wfProfileOut( $fname );
3459                 return $colours;
3460         }
3461
3462         /**
3463          * Replace <!--LINK--> link placeholders with plain text of links
3464          * (not HTML-formatted).
3465          * @param string $text
3466          * @return string
3467          */
3468         function replaceLinkHoldersText( $text ) {
3469                 global $wgUser, $wgLinkCache;
3470                 global $wgOutputReplace;
3471
3472                 $fname = 'Parser::replaceLinkHoldersText';
3473                 wfProfileIn( $fname );
3474
3475                 $text = preg_replace_callback(
3476                         '/<!--(LINK|IWLINK) (.*?)-->/',
3477                         array( &$this, 'replaceLinkHoldersTextCallback' ),
3478                         $text );
3479
3480                 wfProfileOut( $fname );
3481                 return $text;
3482         }
3483
3484         /**
3485          * @param array $matches
3486          * @return string
3487          * @access private
3488          */
3489         function replaceLinkHoldersTextCallback( $matches ) {
3490                 $type = $matches[1];
3491                 $key  = $matches[2];
3492                 if( $type == 'LINK' ) {
3493                         if( isset( $this->mLinkHolders['texts'][$key] ) ) {
3494                                 return $this->mLinkHolders['texts'][$key];
3495                         }
3496                 } elseif( $type == 'IWLINK' ) {
3497                         if( isset( $this->mInterwikiLinkHolders['texts'][$key] ) ) {
3498                                 return $this->mInterwikiLinkHolders['texts'][$key];
3499                         }
3500                 }
3501                 return $matches[0];
3502         }
3503
3504         /**
3505          * Renders an image gallery from a text with one line per image.
3506          * text labels may be given by using |-style alternative text. E.g.
3507          *   Image:one.jpg|The number "1"
3508          *   Image:tree.jpg|A tree
3509          * given as text will return the HTML of a gallery with two images,
3510          * labeled 'The number "1"' and
3511          * 'A tree'.
3512          *
3513          * @static
3514          */
3515         function renderImageGallery( $text ) {
3516                 # Setup the parser
3517                 global $wgUser, $wgTitle;
3518                 $parserOptions = ParserOptions::newFromUser( $wgUser );
3519                 $localParser = new Parser();
3520
3521                 global $wgLinkCache;
3522                 $ig = new ImageGallery();
3523                 $ig->setShowBytes( false );
3524                 $ig->setShowFilename( false );
3525                 $lines = explode( "\n", $text );
3526
3527                 foreach ( $lines as $line ) {
3528                         # match lines like these:
3529                         # Image:someimage.jpg|This is some image
3530                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3531                         # Skip empty lines
3532                         if ( count( $matches ) == 0 ) {
3533                                 continue;
3534                         }
3535                         $nt = Title::newFromURL( $matches[1] );
3536                         if( is_null( $nt ) ) {
3537                                 # Bogus title. Ignore these so we don't bomb out later.
3538                                 continue;
3539                         }
3540                         if ( isset( $matches[3] ) ) {
3541                                 $label = $matches[3];
3542                         } else {
3543                                 $label = '';
3544                         }
3545
3546                         $html = $localParser->parse( $label , $wgTitle, $parserOptions );
3547                         $html = $html->mText;
3548
3549                         $ig->add( new Image( $nt ), $html );
3550                         $wgLinkCache->addImageLinkObj( $nt );
3551                 }
3552                 return $ig->toHTML();
3553         }
3554
3555         /**
3556          * Parse image options text and use it to make an image
3557          */
3558         function makeImage( &$nt, $options ) {
3559                 global $wgContLang, $wgUseImageResize;
3560                 global $wgUser, $wgThumbLimits;
3561
3562                 $align = '';
3563
3564                 # Check if the options text is of the form "options|alt text"
3565                 # Options are:
3566                 #  * thumbnail          make a thumbnail with enlarge-icon and caption, alignment depends on lang
3567                 #  * left               no resizing, just left align. label is used for alt= only
3568                 #  * right              same, but right aligned
3569                 #  * none               same, but not aligned
3570                 #  * ___px              scale to ___ pixels width, no aligning. e.g. use in taxobox
3571                 #  * center             center the image
3572                 #  * framed             Keep original image size, no magnify-button.
3573
3574                 $part = explode( '|', $options);
3575
3576                 $mwThumb  =& MagicWord::get( MAG_IMG_THUMBNAIL );
3577                 $mwManualThumb =& MagicWord::get( MAG_IMG_MANUALTHUMB );
3578                 $mwLeft   =& MagicWord::get( MAG_IMG_LEFT );
3579                 $mwRight  =& MagicWord::get( MAG_IMG_RIGHT );
3580                 $mwNone   =& MagicWord::get( MAG_IMG_NONE );
3581                 $mwWidth  =& MagicWord::get( MAG_IMG_WIDTH );
3582                 $mwCenter =& MagicWord::get( MAG_IMG_CENTER );
3583                 $mwFramed =& MagicWord::get( MAG_IMG_FRAMED );
3584                 $caption = '';
3585
3586                 $width = $height = $framed = $thumb = false;
3587                 $manual_thumb = '' ;
3588
3589                 foreach( $part as $key => $val ) {
3590                         if ( $wgUseImageResize && ! is_null( $mwThumb->matchVariableStartToEnd($val) ) ) {
3591                                 $thumb=true;
3592                         } elseif ( ! is_null( $match = $mwManualThumb->matchVariableStartToEnd($val) ) ) {
3593                                 # use manually specified thumbnail
3594                                 $thumb=true;
3595                                 $manual_thumb = $match;
3596                         } elseif ( ! is_null( $mwRight->matchVariableStartToEnd($val) ) ) {
3597                                 # remember to set an alignment, don't render immediately
3598                                 $align = 'right';
3599                         } elseif ( ! is_null( $mwLeft->matchVariableStartToEnd($val) ) ) {
3600                                 # remember to set an alignment, don't render immediately
3601                                 $align = 'left';
3602                         } elseif ( ! is_null( $mwCenter->matchVariableStartToEnd($val) ) ) {
3603                                 # remember to set an alignment, don't render immediately
3604                                 $align = 'center';
3605                         } elseif ( ! is_null( $mwNone->matchVariableStartToEnd($val) ) ) {
3606                                 # remember to set an alignment, don't render immediately
3607                                 $align = 'none';
3608                         } elseif ( $wgUseImageResize && ! is_null( $match = $mwWidth->matchVariableStartToEnd($val) ) ) {
3609                                 wfDebug( "MAG_IMG_WIDTH match: $match\n" );
3610                                 # $match is the image width in pixels
3611                                 if ( preg_match( '/^([0-9]*)x([0-9]*)$/', $match, $m ) ) {
3612                                         $width = intval( $m[1] );
3613                                         $height = intval( $m[2] );
3614                                 } else {
3615                                         $width = intval($match);
3616                                 }
3617                         } elseif ( ! is_null( $mwFramed->matchVariableStartToEnd($val) ) ) {
3618                                 $framed=true;
3619                         } else {
3620                                 $caption = $val;
3621                         }
3622                 }
3623                 # Strip bad stuff out of the alt text
3624                 $alt = $this->replaceLinkHoldersText( $caption );
3625                 $alt = Sanitizer::stripAllTags( $alt );
3626
3627                 # Linker does the rest
3628                 $sk =& $this->mOptions->getSkin();
3629                 return $sk->makeImageLinkObj( $nt, $caption, $alt, $align, $width, $height, $framed, $thumb, $manual_thumb );
3630         }
3631
3632         /**
3633          * Set a flag in the output object indicating that the content is dynamic and
3634          * shouldn't be cached.
3635          */
3636         function disableCache() {
3637                 $this->mOutput->mCacheTime = -1;
3638         }
3639
3640         /**
3641          * Callback from the Sanitizer for expanding items found in HTML attribute
3642          * values, so they can be safely tested and escaped.
3643          * @param string $text
3644          * @param array $args
3645          * @return string
3646          * @access private
3647          */
3648         function attributeStripCallback( &$text, $args ) {
3649                 $text = $this->replaceVariables( $text, $args );
3650                 $text = $this->unstripForHTML( $text );
3651                 return $text;
3652         }
3653
3654         function unstripForHTML( $text ) {
3655                 $text = $this->unstrip( $text, $this->mStripState );
3656                 $text = $this->unstripNoWiki( $text, $this->mStripState );
3657                 return $text;
3658         }
3659 }
3660
3661 /**
3662  * @todo document
3663  * @package MediaWiki
3664  */
3665 class ParserOutput
3666 {
3667         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
3668         var $mCacheTime; # Timestamp on this article, or -1 for uncacheable. Used in ParserCache.
3669         var $mVersion;   # Compatibility check
3670         var $mTitleText; # title text of the chosen language variant
3671
3672         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3673                 $containsOldMagic = false, $titletext = '' )
3674         {
3675                 $this->mText = $text;
3676                 $this->mLanguageLinks = $languageLinks;
3677                 $this->mCategoryLinks = $categoryLinks;
3678                 $this->mContainsOldMagic = $containsOldMagic;
3679                 $this->mCacheTime = '';
3680                 $this->mVersion = MW_PARSER_VERSION;
3681                 $this->mTitleText = $titletext;
3682         }
3683
3684         function getText()                   { return $this->mText; }
3685         function getLanguageLinks()          { return $this->mLanguageLinks; }
3686         function getCategoryLinks()          { return array_keys( $this->mCategoryLinks ); }
3687         function getCacheTime()              { return $this->mCacheTime; }
3688         function getTitleText()              { return $this->mTitleText; }
3689         function containsOldMagic()          { return $this->mContainsOldMagic; }
3690         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3691         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3692         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategoryLinks, $cl ); }
3693         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3694         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3695         function setTitleText( $t )          { return wfSetVar ($this->mTitleText, $t); }
3696
3697         function addCategoryLink( $c )       { $this->mCategoryLinks[$c] = 1; }
3698
3699         function merge( $other ) {
3700                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3701                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
3702                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3703         }
3704
3705         /**
3706          * Return true if this cached output object predates the global or
3707          * per-article cache invalidation timestamps, or if it comes from
3708          * an incompatible older version.
3709          *
3710          * @param string $touched the affected article's last touched timestamp
3711          * @return bool
3712          * @access public
3713          */
3714         function expired( $touched ) {
3715                 global $wgCacheEpoch;
3716                 return $this->getCacheTime() == -1 || // parser says it's uncacheable
3717                        $this->getCacheTime() < $touched ||
3718                        $this->getCacheTime() <= $wgCacheEpoch ||
3719                        !isset( $this->mVersion ) ||
3720                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3721         }
3722 }
3723
3724 /**
3725  * Set options of the Parser
3726  * @todo document
3727  * @package MediaWiki
3728  */
3729 class ParserOptions
3730 {
3731         # All variables are private
3732         var $mUseTeX;                    # Use texvc to expand <math> tags
3733         var $mUseDynamicDates;           # Use DateFormatter to format dates
3734         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3735         var $mAllowExternalImages;       # Allow external images inline
3736         var $mAllowExternalImagesFrom;   # If not, any exception?
3737         var $mSkin;                      # Reference to the preferred skin
3738         var $mDateFormat;                # Date format index
3739         var $mEditSection;               # Create "edit section" links
3740         var $mNumberHeadings;            # Automatically number headings
3741         var $mAllowSpecialInclusion;     # Allow inclusion of special pages
3742
3743         function getUseTeX()                        { return $this->mUseTeX; }
3744         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3745         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3746         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3747         function getAllowExternalImagesFrom()       { return $this->mAllowExternalImagesFrom; }
3748         function &getSkin()                         { return $this->mSkin; }
3749         function getDateFormat()                    { return $this->mDateFormat; }
3750         function getEditSection()                   { return $this->mEditSection; }
3751         function getNumberHeadings()                { return $this->mNumberHeadings; }
3752         function getAllowSpecialInclusion()         { return $this->mAllowSpecialInclusion; }
3753
3754
3755         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3756         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3757         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3758         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3759         function setAllowExternalImagesFrom( $x )   { return wfSetVar( $this->mAllowExternalImagesFrom, $x ); }
3760         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3761         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3762         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3763         function setAllowSpecialInclusion( $x )     { return wfSetVar( $this->mAllowSpecialInclusion, $x ); }
3764
3765         function setSkin( &$x ) { $this->mSkin =& $x; }
3766
3767         function ParserOptions() {
3768                 global $wgUser;
3769                 $this->initialiseFromUser( $wgUser );
3770         }
3771
3772         /**
3773          * Get parser options
3774          * @static
3775          */
3776         function newFromUser( &$user ) {
3777                 $popts = new ParserOptions;
3778                 $popts->initialiseFromUser( $user );
3779                 return $popts;
3780         }
3781
3782         /** Get user options */
3783         function initialiseFromUser( &$userInput ) {
3784                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages,
3785                        $wgAllowExternalImagesFrom, $wgAllowSpecialInclusion;
3786                 $fname = 'ParserOptions::initialiseFromUser';
3787                 wfProfileIn( $fname );
3788                 if ( !$userInput ) {
3789                         $user = new User;
3790                         $user->setLoaded( true );
3791                 } else {
3792                         $user =& $userInput;
3793                 }
3794
3795                 $this->mUseTeX = $wgUseTeX;
3796                 $this->mUseDynamicDates = $wgUseDynamicDates;
3797                 $this->mInterwikiMagic = $wgInterwikiMagic;
3798                 $this->mAllowExternalImages = $wgAllowExternalImages;
3799                 $this->mAllowExternalImagesFrom = $wgAllowExternalImagesFrom;
3800                 wfProfileIn( $fname.'-skin' );
3801                 $this->mSkin =& $user->getSkin();
3802                 wfProfileOut( $fname.'-skin' );
3803                 $this->mDateFormat = $user->getOption( 'date' );
3804                 $this->mEditSection = true;
3805                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3806                 $this->mAllowSpecialInclusion = $wgAllowSpecialInclusion;
3807                 wfProfileOut( $fname );
3808         }
3809 }
3810
3811 /**
3812  * Callback function used by Parser::replaceLinkHolders()
3813  * to substitute link placeholders.
3814  */
3815 function &wfOutputReplaceMatches( $matches ) {
3816         global $wgOutputReplace;
3817         return $wgOutputReplace[$matches[1]];
3818 }
3819
3820 /**
3821  * Return the total number of articles
3822  */
3823 function wfNumberOfArticles() {
3824         global $wgNumberOfArticles;
3825
3826         wfLoadSiteStats();
3827         return $wgNumberOfArticles;
3828 }
3829
3830 /**
3831  * Return the number of files
3832  */
3833 function wfNumberOfFiles() {
3834         $fname = 'Parser::wfNumberOfFiles';
3835
3836         wfProfileIn( $fname );
3837         $dbr =& wfGetDB( DB_SLAVE );
3838         $res = $dbr->selectField('image', 'COUNT(*)', array(), $fname );
3839         wfProfileOut( $fname );
3840
3841         return $res;
3842 }
3843
3844 /**
3845  * Get various statistics from the database
3846  * @private
3847  */
3848 function wfLoadSiteStats() {
3849         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3850         $fname = 'wfLoadSiteStats';
3851
3852         if ( -1 != $wgNumberOfArticles ) return;
3853         $dbr =& wfGetDB( DB_SLAVE );
3854         $s = $dbr->selectRow( 'site_stats',
3855                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3856                 array( 'ss_row_id' => 1 ), $fname
3857         );
3858
3859         if ( $s === false ) {
3860                 return;
3861         } else {
3862                 $wgTotalViews = $s->ss_total_views;
3863                 $wgTotalEdits = $s->ss_total_edits;
3864                 $wgNumberOfArticles = $s->ss_good_articles;
3865         }
3866 }
3867
3868 /**
3869  * Escape html tags
3870  * Basicly replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3871  *
3872  * @param string $in Text that might contain HTML tags
3873  * @return string Escaped string
3874  */
3875 function wfEscapeHTMLTagsOnly( $in ) {
3876         return str_replace(
3877                 array( '"', '>', '<' ),
3878                 array( '&quot;', '&gt;', '&lt;' ),
3879                 $in );
3880 }
3881
3882 ?>