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