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