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', '1.6.0' );
  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
  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,      // ParserOptions object
 105                 $mTitle,        // Title context, used for self-link rendering and similar things
 106                 $mOutputType,   // Output type, one of the OT_xxx constants
 107             $mTemplates,        // cache of already loaded templates, avoids
 108                                 // multiple SQL queries for the same string
 109             $mTemplatePath,     // stores an unsorted hash of all the templates already loaded
 110                                 // in this path. Used for loop detection.
 111                 $mIWTransData = array(),
 112                 $mRevisionId;   // ID to display in {{REVISIONID}} tags
 113
 114         /**#@-*/
 115
 116         /**
 117          * Constructor
 118          *
 119          * @access public
 120          */
 121         function Parser() {
 122                 $this->mTemplates = array();
 123                 $this->mTemplatePath = array();
 124                 $this->mTagHooks = array();
 125                 $this->clearState();
 126         }
 127
 128         /**
 129          * Clear Parser state
 130          *
 131          * @access private
 132          */
 133         function clearState() {
 134                 $this->mOutput = new ParserOutput;
 135                 $this->mAutonumber = 0;
 136                 $this->mLastSection = '';
 137                 $this->mDTopen = false;
 138                 $this->mVariables = false;
 139                 $this->mIncludeCount = array();
 140                 $this->mStripState = array();
 141                 $this->mArgStack = array();
 142                 $this->mInPre = false;
 143                 $this->mInterwikiLinkHolders = array(
 144                         'texts' => array(),
 145                         'titles' => array()
 146                 );
 147                 $this->mLinkHolders = array(
 148                         'namespaces' => array(),
 149                         'dbkeys' => array(),
 150                         'queries' => array(),
 151                         'texts' => array(),
 152                         'titles' => array()
 153                 );
 154                 $this->mRevisionId = null;
 155                 $this->mUniqPrefix = 'UNIQ' . Parser::getRandomString();
 156
 157                 wfRunHooks( 'ParserClearState', array( &$this ) );
 158         }
 159
 160         /**
 161          * Convert wikitext to HTML
 162          * Do not call this function recursively.
 163          *
 164          * @access private
 165          * @param string $text Text we want to parse
 166          * @param Title &$title A title object
 167          * @param array $options
 168          * @param boolean $linestart
 169          * @param boolean $clearState
 170          * @param int $revid number to pass in {{REVISIONID}}
 171          * @return ParserOutput a ParserOutput
 172          */
 173         function parse( $text, &$title, $options, $linestart = true, $clearState = true, $revid = null ) {
 174                 /**
 175                  * First pass--just handle <nowiki> sections, pass the rest off
 176                  * to internalParse() which does all the real work.
 177                  */
 178
 179                 global $wgUseTidy, $wgContLang;
 180                 $fname = 'Parser::parse';
 181                 wfProfileIn( $fname );
 182
 183                 if ( $clearState ) {
 184                         $this->clearState();
 185                 }
 186
 187                 $this->mOptions = $options;
 188                 $this->mTitle =& $title;
 189                 $this->mRevisionId = $revid;
 190                 $this->mOutputType = OT_HTML;
 191
 192                 $this->mStripState = NULL;
 193
 194                 //$text = $this->strip( $text, $this->mStripState );
 195                 // VOODOO MAGIC FIX! Sometimes the above segfaults in PHP5.
 196                 $x =& $this->mStripState;
 197
 198                 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$x ) );
 199                 $text = $this->strip( $text, $x );
 200                 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$x ) );
 201
 202                 # Hook to suspend the parser in this state
 203                 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$x ) ) ) {
 204                         wfProfileOut( $fname );
 205                         return $text ;
 206                 }
 207
 208                 $text = $this->internalParse( $text );
 209
 210                 $text = $this->unstrip( $text, $this->mStripState );
 211
 212                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 213                 $fixtags = array(
 214                         # french spaces, last one Guillemet-left
 215                         # only if there is something before the space
 216                         '/(.) (?=\\?|:|;|!|\\302\\273)/' => '\\1&nbsp;\\2',
 217                         # french spaces, Guillemet-right
 218                         '/(\\302\\253) /' => '\\1&nbsp;',
 219                         '/<center *>(.*)<\\/center *>/i' => '<div class="center">\\1</div>',
 220                 );
 221                 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 222
 223                 # only once and last
 224                 $text = $this->doBlockLevels( $text, $linestart );
 225
 226                 $this->replaceLinkHolders( $text );
 227
 228                 # the position of the convert() call should not be changed. it
 229                 # assumes that the links are all replaces and the only thing left
 230                 # is the <nowiki> mark.
 231                 $text = $wgContLang->convert($text);
 232                 $this->mOutput->setTitleText($wgContLang->getParsedTitle());
 233
 234                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 235
 236                 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) );
 237
 238                 $text = Sanitizer::normalizeCharReferences( $text );
 239
 240                 if ($wgUseTidy) {
 241                         $text = Parser::tidy($text);
 242                 }
 243
 244                 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) );
 245
 246                 $this->mOutput->setText( $text );
 247                 wfProfileOut( $fname );
 248
 249                 return $this->mOutput;
 250         }
 251
 252         /**
 253          * Get a random string
 254          *
 255          * @access private
 256          * @static
 257          */
 258         function getRandomString() {
 259                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 260         }
 261
 262         /**
 263          * Replaces all occurrences of <$tag>content</$tag> in the text
 264          * with a random marker and returns the new text. the output parameter
 265          * $content will be an associative array filled with data on the form
 266          * $unique_marker => content.
 267          *
 268          * If $content is already set, the additional entries will be appended
 269          * If $tag is set to STRIP_COMMENTS, the function will extract
 270          * <!-- HTML comments -->
 271          *
 272          * @access private
 273          * @static
 274          */
 275         function extractTagsAndParams($tag, $text, &$content, &$tags, &$params, $uniq_prefix = ''){
 276                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 277                 if ( !$content ) {
 278                         $content = array( );
 279                 }
 280                 $n = 1;
 281                 $stripped = '';
 282
 283                 if ( !$tags ) {
 284                         $tags = array( );
 285                 }
 286
 287                 if ( !$params ) {
 288                         $params = array( );
 289                 }
 290
 291                 if( $tag == STRIP_COMMENTS ) {
 292                         $start = '/<!--()()/';
 293                         $end   = '/-->/';
 294                 } else {
 295                         $start = "/<$tag(\\s+[^\\/>]*|\\s*)(\\/?)>/i";
 296                         $end   = "/<\\/$tag\\s*>/i";
 297                 }
 298
 299                 while ( '' != $text ) {
 300                         $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
 301                         $stripped .= $p[0];
 302                         if( count( $p ) < 4 ) {
 303                                 break;
 304                         }
 305                         $attributes = $p[1];
 306                         $empty      = $p[2];
 307                         $inside     = $p[3];
 308
 309                         $marker = $rnd . sprintf('%08X', $n++);
 310                         $stripped .= $marker;
 311
 312                         $tags[$marker] = "<$tag$attributes$empty>";
 313                         $params[$marker] = Sanitizer::decodeTagAttributes( $attributes );
 314
 315                         if ( $empty === '/' ) {
 316                                 // Empty element tag, <tag />
 317                                 $content[$marker] = null;
 318                                 $text = $inside;
 319                         } else {
 320                                 $q = preg_split( $end, $inside, 2 );
 321                                 $content[$marker] = $q[0];
 322                                 if( count( $q ) < 2 ) {
 323                                         # No end tag -- let it run out to the end of the text.
 324                                         break;
 325                                 } else {
 326                                         $text = $q[1];
 327                                 }
 328                         }
 329                 }
 330                 return $stripped;
 331         }
 332
 333         /**
 334          * Wrapper function for extractTagsAndParams
 335          * for cases where $tags and $params isn't needed
 336          * i.e. where tags will never have params, like <nowiki>
 337          *
 338          * @access private
 339          * @static
 340          */
 341         function extractTags( $tag, $text, &$content, $uniq_prefix = '' ) {
 342                 $dummy_tags = array();
 343                 $dummy_params = array();
 344
 345                 return Parser::extractTagsAndParams( $tag, $text, $content,
 346                         $dummy_tags, $dummy_params, $uniq_prefix );
 347         }
 348
 349         /**
 350          * Strips and renders nowiki, pre, math, hiero
 351          * If $render is set, performs necessary rendering operations on plugins
 352          * Returns the text, and fills an array with data needed in unstrip()
 353          * If the $state is already a valid strip state, it adds to the state
 354          *
 355          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 356          *  will be stripped in addition to other tags. This is important
 357          *  for section editing, where these comments cause confusion when
 358          *  counting the sections in the wikisource
 359          *
 360          * @access private
 361          */
 362         function strip( $text, &$state, $stripcomments = false ) {
 363                 $render = ($this->mOutputType == OT_HTML);
 364                 $html_content = array();
 365                 $nowiki_content = array();
 366                 $math_content = array();
 367                 $pre_content = array();
 368                 $comment_content = array();
 369                 $ext_content = array();
 370                 $ext_tags = array();
 371                 $ext_params = array();
 372                 $gallery_content = array();
 373
 374                 # Replace any instances of the placeholders
 375                 $uniq_prefix = $this->mUniqPrefix;
 376                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 377
 378                 # html
 379                 global $wgRawHtml;
 380                 if( $wgRawHtml ) {
 381                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 382                         foreach( $html_content as $marker => $content ) {
 383                                 if ($render ) {
 384                                         # Raw and unchecked for validity.
 385                                         $html_content[$marker] = $content;
 386                                 } else {
 387                                         $html_content[$marker] = '<html>'.$content.'</html>';
 388                                 }
 389                         }
 390                 }
 391
 392                 # nowiki
 393                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 394                 foreach( $nowiki_content as $marker => $content ) {
 395                         if( $render ){
 396                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 397                         } else {
 398                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 399                         }
 400                 }
 401
 402                 # math
 403                 if( $this->mOptions->getUseTeX() ) {
 404                         $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 405                         foreach( $math_content as $marker => $content ){
 406                                 if( $render ) {
 407                                         $math_content[$marker] = renderMath( $content );
 408                                 } else {
 409                                         $math_content[$marker] = '<math>'.$content.'</math>';
 410                                 }
 411                         }
 412                 }
 413
 414                 # pre
 415                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 416                 foreach( $pre_content as $marker => $content ){
 417                         if( $render ){
 418                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 419                         } else {
 420                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 421                         }
 422                 }
 423
 424                 # gallery
 425                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 426                 foreach( $gallery_content as $marker => $content ) {
 427                         require_once( 'ImageGallery.php' );
 428                         if ( $render ) {
 429                                 $gallery_content[$marker] = $this->renderImageGallery( $content );
 430                         } else {
 431                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 432                         }
 433                 }
 434
 435                 # Comments
 436                 if($stripcomments) {
 437                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 438                         foreach( $comment_content as $marker => $content ){
 439                                 $comment_content[$marker] = '<!--'.$content.'-->';
 440                         }
 441                 }
 442
 443                 # Extensions
 444                 foreach ( $this->mTagHooks as $tag => $callback ) {
 445                         $ext_content[$tag] = array();
 446                         $text = Parser::extractTagsAndParams( $tag, $text, $ext_content[$tag],
 447                                 $ext_tags[$tag], $ext_params[$tag], $uniq_prefix );
 448                         foreach( $ext_content[$tag] as $marker => $content ) {
 449                                 $full_tag = $ext_tags[$tag][$marker];
 450                                 $params = $ext_params[$tag][$marker];
 451                                 if ( $render )
 452                                         $ext_content[$tag][$marker] = call_user_func_array( $callback, array( $content, $params, &$this ) );
 453                                 else {
 454                                         if ( is_null( $content ) ) {
 455                                                 // Empty element tag
 456                                                 $ext_content[$tag][$marker] = $full_tag;
 457                                         } else {
 458                                                 $ext_content[$tag][$marker] = "$full_tag$content</$tag>";
 459                                         }
 460                                 }
 461                         }
 462                 }
 463
 464                 # Merge state with the pre-existing state, if there is one
 465                 if ( $state ) {
 466                         $state['html'] = $state['html'] + $html_content;
 467                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 468                         $state['math'] = $state['math'] + $math_content;
 469                         $state['pre'] = $state['pre'] + $pre_content;
 470                         $state['comment'] = $state['comment'] + $comment_content;
 471                         $state['gallery'] = $state['gallery'] + $gallery_content;
 472
 473                         foreach( $ext_content as $tag => $array ) {
 474                                 if ( array_key_exists( $tag, $state ) ) {
 475                                         $state[$tag] = $state[$tag] + $array;
 476                                 }
 477                         }
 478                 } else {
 479                         $state = array(
 480                           'html' => $html_content,
 481                           'nowiki' => $nowiki_content,
 482                           'math' => $math_content,
 483                           'pre' => $pre_content,
 484                           'comment' => $comment_content,
 485                           'gallery' => $gallery_content,
 486                         ) + $ext_content;
 487                 }
 488                 return $text;
 489         }
 490
 491         /**
 492          * restores pre, math, and hiero removed by strip()
 493          *
 494          * always call unstripNoWiki() after this one
 495          * @access private
 496          */
 497         function unstrip( $text, &$state ) {
 498                 if ( !is_array( $state ) ) {
 499                         return $text;
 500                 }
 501
 502                 # Must expand in reverse order, otherwise nested tags will be corrupted
 503                 foreach( array_reverse( $state, true ) as $tag => $contentDict ) {
 504                         if( $tag != 'nowiki' && $tag != 'html' ) {
 505                                 foreach( array_reverse( $contentDict, true ) as $uniq => $content ) {
 506                                         $text = str_replace( $uniq, $content, $text );
 507                                 }
 508                         }
 509                 }
 510
 511                 return $text;
 512         }
 513
 514         /**
 515          * always call this after unstrip() to preserve the order
 516          *
 517          * @access private
 518          */
 519         function unstripNoWiki( $text, &$state ) {
 520                 if ( !is_array( $state ) ) {
 521                         return $text;
 522                 }
 523
 524                 # Must expand in reverse order, otherwise nested tags will be corrupted
 525                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 526                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 527                 }
 528
 529                 global $wgRawHtml;
 530                 if ($wgRawHtml) {
 531                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 532                                 $text = str_replace( key( $state['html'] ), $content, $text );
 533                         }
 534                 }
 535
 536                 return $text;
 537         }
 538
 539         /**
 540          * Add an item to the strip state
 541          * Returns the unique tag which must be inserted into the stripped text
 542          * The tag will be replaced with the original text in unstrip()
 543          *
 544          * @access private
 545          */
 546         function insertStripItem( $text, &$state ) {
 547                 $rnd = $this->mUniqPrefix . '-item' . Parser::getRandomString();
 548                 if ( !$state ) {
 549                         $state = array(
 550                           'html' => array(),
 551                           'nowiki' => array(),
 552                           'math' => array(),
 553                           'pre' => array(),
 554                           'comment' => array(),
 555                           'gallery' => array(),
 556                         );
 557                 }
 558                 $state['item'][$rnd] = $text;
 559                 return $rnd;
 560         }
 561
 562         /**
 563          * Interface with html tidy, used if $wgUseTidy = true.
 564          * If tidy isn't able to correct the markup, the original will be
 565          * returned in all its glory with a warning comment appended.
 566          *
 567          * Either the external tidy program or the in-process tidy extension
 568          * will be used depending on availability. Override the default
 569          * $wgTidyInternal setting to disable the internal if it's not working.
 570          *
 571          * @param string $text Hideous HTML input
 572          * @return string Corrected HTML output
 573          * @access public
 574          * @static
 575          */
 576         function tidy( $text ) {
 577                 global $wgTidyInternal;
 578                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 579 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 580 '<head><title>test</title></head><body>'.$text.'</body></html>';
 581                 if( $wgTidyInternal ) {
 582                         $correctedtext = Parser::internalTidy( $wrappedtext );
 583                 } else {
 584                         $correctedtext = Parser::externalTidy( $wrappedtext );
 585                 }
 586                 if( is_null( $correctedtext ) ) {
 587                         wfDebug( "Tidy error detected!\n" );
 588                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 589                 }
 590                 return $correctedtext;
 591         }
 592
 593         /**
 594          * Spawn an external HTML tidy process and get corrected markup back from it.
 595          *
 596          * @access private
 597          * @static
 598          */
 599         function externalTidy( $text ) {
 600                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 601                 $fname = 'Parser::externalTidy';
 602                 wfProfileIn( $fname );
 603
 604                 $cleansource = '';
 605                 $opts = ' -utf8';
 606
 607                 $descriptorspec = array(
 608                         0 => array('pipe', 'r'),
 609                         1 => array('pipe', 'w'),
 610                         2 => array('file', '/dev/null', 'a')
 611                 );
 612                 $pipes = array();
 613                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 614                 if (is_resource($process)) {
 615                         fwrite($pipes[0], $text);
 616                         fclose($pipes[0]);
 617                         while (!feof($pipes[1])) {
 618                                 $cleansource .= fgets($pipes[1], 1024);
 619                         }
 620                         fclose($pipes[1]);
 621                         proc_close($process);
 622                 }
 623
 624                 wfProfileOut( $fname );
 625
 626                 if( $cleansource == '' && $text != '') {
 627                         // Some kind of error happened, so we couldn't get the corrected text.
 628                         // Just give up; we'll use the source text and append a warning.
 629                         return null;
 630                 } else {
 631                         return $cleansource;
 632                 }
 633         }
 634
 635         /**
 636          * Use the HTML tidy PECL extension to use the tidy library in-process,
 637          * saving the overhead of spawning a new process. Currently written to
 638          * the PHP 4.3.x version of the extension, may not work on PHP 5.
 639          *
 640          * 'pear install tidy' should be able to compile the extension module.
 641          *
 642          * @access private
 643          * @static
 644          */
 645         function internalTidy( $text ) {
 646                 global $wgTidyConf;
 647                 $fname = 'Parser::internalTidy';
 648                 wfProfileIn( $fname );
 649
 650                 tidy_load_config( $wgTidyConf );
 651                 tidy_set_encoding( 'utf8' );
 652                 tidy_parse_string( $text );
 653                 tidy_clean_repair();
 654                 if( tidy_get_status() == 2 ) {
 655                         // 2 is magic number for fatal error
 656                         // http://www.php.net/manual/en/function.tidy-get-status.php
 657                         $cleansource = null;
 658                 } else {
 659                         $cleansource = tidy_get_output();
 660                 }
 661                 wfProfileOut( $fname );
 662                 return $cleansource;
 663         }
 664
 665         /**
 666          * parse the wiki syntax used to render tables
 667          *
 668          * @access private
 669          */
 670         function doTableStuff ( $t ) {
 671                 $fname = 'Parser::doTableStuff';
 672                 wfProfileIn( $fname );
 673
 674                 $t = explode ( "\n" , $t ) ;
 675                 $td = array () ; # Is currently a td tag open?
 676                 $ltd = array () ; # Was it TD or TH?
 677                 $tr = array () ; # Is currently a tr tag open?
 678                 $ltr = array () ; # tr attributes
 679                 $indent_level = 0; # indent level of the table
 680                 foreach ( $t AS $k => $x )
 681                 {
 682                         $x = trim ( $x ) ;
 683                         $fc = substr ( $x , 0 , 1 ) ;
 684                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 685                                 $indent_level = strlen( $matches[1] );
 686
 687                                 $attributes = $this->unstripForHTML( $matches[2] );
 688
 689                                 $t[$k] = str_repeat( '<dl><dd>', $indent_level ) .
 690                                         '<table' . Sanitizer::fixTagAttributes ( $attributes, 'table' ) . '>' ;
 691                                 array_push ( $td , false ) ;
 692                                 array_push ( $ltd , '' ) ;
 693                                 array_push ( $tr , false ) ;
 694                                 array_push ( $ltr , '' ) ;
 695                         }
 696                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 697                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 698                                 $z = "</table>" . substr ( $x , 2);
 699                                 $l = array_pop ( $ltd ) ;
 700                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 701                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 702                                 array_pop ( $ltr ) ;
 703                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 704                         }
 705                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 706                                 $x = substr ( $x , 1 ) ;
 707                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 708                                 $z = '' ;
 709                                 $l = array_pop ( $ltd ) ;
 710                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 711                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 712                                 array_pop ( $ltr ) ;
 713                                 $t[$k] = $z ;
 714                                 array_push ( $tr , false ) ;
 715                                 array_push ( $td , false ) ;
 716                                 array_push ( $ltd , '' ) ;
 717                                 $attributes = $this->unstripForHTML( $x );
 718                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $attributes, 'tr' ) ) ;
 719                         }
 720                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 721                                 # $x is a table row
 722                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 723                                         $fc = '+' ;
 724                                         $x = substr ( $x , 1 ) ;
 725                                 }
 726                                 $after = substr ( $x , 1 ) ;
 727                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 728                                 $after = explode ( '||' , $after ) ;
 729                                 $t[$k] = '' ;
 730
 731                                 # Loop through each table cell
 732                                 foreach ( $after AS $theline )
 733                                 {
 734                                         $z = '' ;
 735                                         if ( $fc != '+' )
 736                                         {
 737                                                 $tra = array_pop ( $ltr ) ;
 738                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 739                                                 array_push ( $tr , true ) ;
 740                                                 array_push ( $ltr , '' ) ;
 741                                         }
 742
 743                                         $l = array_pop ( $ltd ) ;
 744                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 745                                         if ( $fc == '|' ) $l = 'td' ;
 746                                         else if ( $fc == '!' ) $l = 'th' ;
 747                                         else if ( $fc == '+' ) $l = 'caption' ;
 748                                         else $l = '' ;
 749                                         array_push ( $ltd , $l ) ;
 750
 751                                         # Cell parameters
 752                                         $y = explode ( '|' , $theline , 2 ) ;
 753                                         # Note that a '|' inside an invalid link should not
 754                                         # be mistaken as delimiting cell parameters
 755                                         if ( strpos( $y[0], '[[' ) !== false ) {
 756                                                 $y = array ($theline);
 757                                         }
 758                                         if ( count ( $y ) == 1 )
 759                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 760                                         else {
 761                                                 $attributes = $this->unstripForHTML( $y[0] );
 762                                                 $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($attributes, $l).">{$y[1]}" ;
 763                                         }
 764                                         $t[$k] .= $y ;
 765                                         array_push ( $td , true ) ;
 766                                 }
 767                         }
 768                 }
 769
 770                 # Closing open td, tr && table
 771                 while ( count ( $td ) > 0 )
 772                 {
 773                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 774                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 775                         $t[] = '</table>' ;
 776                 }
 777
 778                 $t = implode ( "\n" , $t ) ;
 779                 wfProfileOut( $fname );
 780                 return $t ;
 781         }
 782
 783         /**
 784          * Helper function for parse() that transforms wiki markup into
 785          * HTML. Only called for $mOutputType == OT_HTML.
 786          *
 787          * @access private
 788          */
 789         function internalParse( $text ) {
 790                 global $wgContLang;
 791                 $args = array();
 792                 $isMain = true;
 793                 $fname = 'Parser::internalParse';
 794                 wfProfileIn( $fname );
 795
 796                 # Remove <noinclude> tags and <includeonly> sections
 797                 $text = strtr( $text, array( '<onlyinclude>' => '' , '</onlyinclude>' => '' ) );
 798                 $text = strtr( $text, array( '<noinclude>' => '', '</noinclude>' => '') );
 799                 $text = preg_replace( '/<includeonly>.*?<\/includeonly>/s', '', $text );
 800
 801                 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ) );
 802                 $text = $this->replaceVariables( $text, $args );
 803
 804                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 805
 806                 $text = $this->doHeadings( $text );
 807                 if($this->mOptions->getUseDynamicDates()) {
 808                         $df =& DateFormatter::getInstance();
 809                         $text = $df->reformat( $this->mOptions->getDateFormat(), $text );
 810                 }
 811                 $text = $this->doAllQuotes( $text );
 812                 $text = $this->replaceInternalLinks( $text );
 813                 $text = $this->replaceExternalLinks( $text );
 814
 815                 # replaceInternalLinks may sometimes leave behind
 816                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 817                 $text = str_replace($this->mUniqPrefix."NOPARSE", "", $text);
 818
 819                 $text = $this->doMagicLinks( $text );
 820                 $text = $this->doTableStuff( $text );
 821                 $text = $this->formatHeadings( $text, $isMain );
 822
 823                 $regex = '/<!--IW_TRANSCLUDE (\d+)-->/';
 824                 $text = preg_replace_callback($regex, array(&$this, 'scarySubstitution'), $text);
 825
 826                 wfProfileOut( $fname );
 827                 return $text;
 828         }
 829
 830         function scarySubstitution($matches) {
 831 #               return "[[".$matches[0]."]]";
 832                 return $this->mIWTransData[(int)$matches[0]];
 833         }
 834
 835         /**
 836          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 837          * magic external links.
 838          *
 839          * @access private
 840          */
 841         function &doMagicLinks( &$text ) {
 842                 $text = $this->magicISBN( $text );
 843                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 844                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 845                 return $text;
 846         }
 847
 848         /**
 849          * Parse headers and return html
 850          *
 851          * @access private
 852          */
 853         function doHeadings( $text ) {
 854                 $fname = 'Parser::doHeadings';
 855                 wfProfileIn( $fname );
 856                 for ( $i = 6; $i >= 1; --$i ) {
 857                         $h = str_repeat( '=', $i );
 858                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 859                           "<h{$i}>\\1</h{$i}>\\2", $text );
 860                 }
 861                 wfProfileOut( $fname );
 862                 return $text;
 863         }
 864
 865         /**
 866          * Replace single quotes with HTML markup
 867          * @access private
 868          * @return string the altered text
 869          */
 870         function doAllQuotes( $text ) {
 871                 $fname = 'Parser::doAllQuotes';
 872                 wfProfileIn( $fname );
 873                 $outtext = '';
 874                 $lines = explode( "\n", $text );
 875                 foreach ( $lines as $line ) {
 876                         $outtext .= $this->doQuotes ( $line ) . "\n";
 877                 }
 878                 $outtext = substr($outtext, 0,-1);
 879                 wfProfileOut( $fname );
 880                 return $outtext;
 881         }
 882
 883         /**
 884          * Helper function for doAllQuotes()
 885          * @access private
 886          */
 887         function doQuotes( $text ) {
 888                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 889                 if ( count( $arr ) == 1 )
 890                         return $text;
 891                 else
 892                 {
 893                         # First, do some preliminary work. This may shift some apostrophes from
 894                         # being mark-up to being text. It also counts the number of occurrences
 895                         # of bold and italics mark-ups.
 896                         $i = 0;
 897                         $numbold = 0;
 898                         $numitalics = 0;
 899                         foreach ( $arr as $r )
 900                         {
 901                                 if ( ( $i % 2 ) == 1 )
 902                                 {
 903                                         # If there are ever four apostrophes, assume the first is supposed to
 904                                         # be text, and the remaining three constitute mark-up for bold text.
 905                                         if ( strlen( $arr[$i] ) == 4 )
 906                                         {
 907                                                 $arr[$i-1] .= "'";
 908                                                 $arr[$i] = "'''";
 909                                         }
 910                                         # If there are more than 5 apostrophes in a row, assume they're all
 911                                         # text except for the last 5.
 912                                         else if ( strlen( $arr[$i] ) > 5 )
 913                                         {
 914                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 915                                                 $arr[$i] = "'''''";
 916                                         }
 917                                         # Count the number of occurrences of bold and italics mark-ups.
 918                                         # We are not counting sequences of five apostrophes.
 919                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 920                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 921                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 922                                 }
 923                                 $i++;
 924                         }
 925
 926                         # If there is an odd number of both bold and italics, it is likely
 927                         # that one of the bold ones was meant to be an apostrophe followed
 928                         # by italics. Which one we cannot know for certain, but it is more
 929                         # likely to be one that has a single-letter word before it.
 930                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 931                         {
 932                                 $i = 0;
 933                                 $firstsingleletterword = -1;
 934                                 $firstmultiletterword = -1;
 935                                 $firstspace = -1;
 936                                 foreach ( $arr as $r )
 937                                 {
 938                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
 939                                         {
 940                                                 $x1 = substr ($arr[$i-1], -1);
 941                                                 $x2 = substr ($arr[$i-1], -2, 1);
 942                                                 if ($x1 == ' ') {
 943                                                         if ($firstspace == -1) $firstspace = $i;
 944                                                 } else if ($x2 == ' ') {
 945                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 946                                                 } else {
 947                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 948                                                 }
 949                                         }
 950                                         $i++;
 951                                 }
 952
 953                                 # If there is a single-letter word, use it!
 954                                 if ($firstsingleletterword > -1)
 955                                 {
 956                                         $arr [ $firstsingleletterword ] = "''";
 957                                         $arr [ $firstsingleletterword-1 ] .= "'";
 958                                 }
 959                                 # If not, but there's a multi-letter word, use that one.
 960                                 else if ($firstmultiletterword > -1)
 961                                 {
 962                                         $arr [ $firstmultiletterword ] = "''";
 963                                         $arr [ $firstmultiletterword-1 ] .= "'";
 964                                 }
 965                                 # ... otherwise use the first one that has neither.
 966                                 # (notice that it is possible for all three to be -1 if, for example,
 967                                 # there is only one pentuple-apostrophe in the line)
 968                                 else if ($firstspace > -1)
 969                                 {
 970                                         $arr [ $firstspace ] = "''";
 971                                         $arr [ $firstspace-1 ] .= "'";
 972                                 }
 973                         }
 974
 975                         # Now let's actually convert our apostrophic mush to HTML!
 976                         $output = '';
 977                         $buffer = '';
 978                         $state = '';
 979                         $i = 0;
 980                         foreach ($arr as $r)
 981                         {
 982                                 if (($i % 2) == 0)
 983                                 {
 984                                         if ($state == 'both')
 985                                                 $buffer .= $r;
 986                                         else
 987                                                 $output .= $r;
 988                                 }
 989                                 else
 990                                 {
 991                                         if (strlen ($r) == 2)
 992                                         {
 993                                                 if ($state == 'i')
 994                                                 { $output .= '</i>'; $state = ''; }
 995                                                 else if ($state == 'bi')
 996                                                 { $output .= '</i>'; $state = 'b'; }
 997                                                 else if ($state == 'ib')
 998                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 999                                                 else if ($state == 'both')
1000                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
1001                                                 else # $state can be 'b' or ''
1002                                                 { $output .= '<i>'; $state .= 'i'; }
1003                                         }
1004                                         else if (strlen ($r) == 3)
1005                                         {
1006                                                 if ($state == 'b')
1007                                                 { $output .= '</b>'; $state = ''; }
1008                                                 else if ($state == 'bi')
1009                                                 { $output .= '</i></b><i>'; $state = 'i'; }
1010                                                 else if ($state == 'ib')
1011                                                 { $output .= '</b>'; $state = 'i'; }
1012                                                 else if ($state == 'both')
1013                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
1014                                                 else # $state can be 'i' or ''
1015                                                 { $output .= '<b>'; $state .= 'b'; }
1016                                         }
1017                                         else if (strlen ($r) == 5)
1018                                         {
1019                                                 if ($state == 'b')
1020                                                 { $output .= '</b><i>'; $state = 'i'; }
1021                                                 else if ($state == 'i')
1022                                                 { $output .= '</i><b>'; $state = 'b'; }
1023                                                 else if ($state == 'bi')
1024                                                 { $output .= '</i></b>'; $state = ''; }
1025                                                 else if ($state == 'ib')
1026                                                 { $output .= '</b></i>'; $state = ''; }
1027                                                 else if ($state == 'both')
1028                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
1029                                                 else # ($state == '')
1030                                                 { $buffer = ''; $state = 'both'; }
1031                                         }
1032                                 }
1033                                 $i++;
1034                         }
1035                         # Now close all remaining tags.  Notice that the order is important.
1036                         if ($state == 'b' || $state == 'ib')
1037                                 $output .= '</b>';
1038                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
1039                                 $output .= '</i>';
1040                         if ($state == 'bi')
1041                                 $output .= '</b>';
1042                         if ($state == 'both')
1043                                 $output .= '<b><i>'.$buffer.'</i></b>';
1044                         return $output;
1045                 }
1046         }
1047
1048         /**
1049          * Replace external links
1050          *
1051          * Note: this is all very hackish and the order of execution matters a lot.
1052          * Make sure to run maintenance/parserTests.php if you change this code.
1053          *
1054          * @access private
1055          */
1056         function replaceExternalLinks( $text ) {
1057                 global $wgContLang;
1058                 $fname = 'Parser::replaceExternalLinks';
1059                 wfProfileIn( $fname );
1060
1061                 $sk =& $this->mOptions->getSkin();
1062
1063                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1064
1065                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
1066
1067                 $i = 0;
1068                 while ( $i<count( $bits ) ) {
1069                         $url = $bits[$i++];
1070                         $protocol = $bits[$i++];
1071                         $text = $bits[$i++];
1072                         $trail = $bits[$i++];
1073
1074                         # The characters '<' and '>' (which were escaped by
1075                         # removeHTMLtags()) should not be included in
1076                         # URLs, per RFC 2396.
1077                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1078                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
1079                                 $url = substr($url, 0, $m2[0][1]);
1080                         }
1081
1082                         # If the link text is an image URL, replace it with an <img> tag
1083                         # This happened by accident in the original parser, but some people used it extensively
1084                         $img = $this->maybeMakeExternalImage( $text );
1085                         if ( $img !== false ) {
1086                                 $text = $img;
1087                         }
1088
1089                         $dtrail = '';
1090
1091                         # Set linktype for CSS - if URL==text, link is essentially free
1092                         $linktype = ($text == $url) ? 'free' : 'text';
1093
1094                         # No link text, e.g. [http://domain.tld/some.link]
1095                         if ( $text == '' ) {
1096                                 # Autonumber if allowed
1097                                 if ( strpos( HTTP_PROTOCOLS, str_replace('/','\/', $protocol) ) !== false ) {
1098                                         $text = '[' . ++$this->mAutonumber . ']';
1099                                         $linktype = 'autonumber';
1100                                 } else {
1101                                         # Otherwise just use the URL
1102                                         $text = htmlspecialchars( $url );
1103                                         $linktype = 'free';
1104                                 }
1105                         } else {
1106                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1107                                 # Check for trail
1108                                 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1109                         }
1110
1111                         $text = $wgContLang->markNoConversion($text);
1112
1113                         # Replace &amp; from obsolete syntax with &.
1114                         # All HTML entities will be escaped by makeExternalLink()
1115                         # or maybeMakeExternalImage()
1116                         $url = str_replace( '&amp;', '&', $url );
1117
1118                         # Process the trail (i.e. everything after this link up until start of the next link),
1119                         # replacing any non-bracketed links
1120                         $trail = $this->replaceFreeExternalLinks( $trail );
1121
1122
1123                         # Use the encoded URL
1124                         # This means that users can paste URLs directly into the text
1125                         # Funny characters like &ouml; aren't valid in URLs anyway
1126                         # This was changed in August 2004
1127                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
1128                 }
1129
1130                 wfProfileOut( $fname );
1131                 return $s;
1132         }
1133
1134         /**
1135          * Replace anything that looks like a URL with a link
1136          * @access private
1137          */
1138         function replaceFreeExternalLinks( $text ) {
1139                 global $wgContLang;
1140                 $fname = 'Parser::replaceFreeExternalLinks';
1141                 wfProfileIn( $fname );
1142
1143                 $bits = preg_split( '/(\b(?:' . wfUrlProtocols() . '))/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1144                 $s = array_shift( $bits );
1145                 $i = 0;
1146
1147                 $sk =& $this->mOptions->getSkin();
1148
1149                 while ( $i < count( $bits ) ){
1150                         $protocol = $bits[$i++];
1151                         $remainder = $bits[$i++];
1152
1153                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1154                                 # Found some characters after the protocol that look promising
1155                                 $url = $protocol . $m[1];
1156                                 $trail = $m[2];
1157
1158                                 # The characters '<' and '>' (which were escaped by
1159                                 # removeHTMLtags()) should not be included in
1160                                 # URLs, per RFC 2396.
1161                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1162                                         $trail = substr($url, $m2[0][1]) . $trail;
1163                                         $url = substr($url, 0, $m2[0][1]);
1164                                 }
1165
1166                                 # Move trailing punctuation to $trail
1167                                 $sep = ',;\.:!?';
1168                                 # If there is no left bracket, then consider right brackets fair game too
1169                                 if ( strpos( $url, '(' ) === false ) {
1170                                         $sep .= ')';
1171                                 }
1172
1173                                 $numSepChars = strspn( strrev( $url ), $sep );
1174                                 if ( $numSepChars ) {
1175                                         $trail = substr( $url, -$numSepChars ) . $trail;
1176                                         $url = substr( $url, 0, -$numSepChars );
1177                                 }
1178
1179                                 # Replace &amp; from obsolete syntax with &.
1180                                 # All HTML entities will be escaped by makeExternalLink()
1181                                 # or maybeMakeExternalImage()
1182                                 $url = str_replace( '&amp;', '&', $url );
1183
1184                                 # Is this an external image?
1185                                 $text = $this->maybeMakeExternalImage( $url );
1186                                 if ( $text === false ) {
1187                                         # Not an image, make a link
1188                                         $text = $sk->makeExternalLink( $url, $wgContLang->markNoConversion($url), true, 'free' );
1189                                 }
1190                                 $s .= $text . $trail;
1191                         } else {
1192                                 $s .= $protocol . $remainder;
1193                         }
1194                 }
1195                 wfProfileOut( $fname );
1196                 return $s;
1197         }
1198
1199         /**
1200          * make an image if it's allowed, either through the global
1201          * option or through the exception
1202          * @access private
1203          */
1204         function maybeMakeExternalImage( $url ) {
1205                 $sk =& $this->mOptions->getSkin();
1206                 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
1207                 $imagesexception = !empty($imagesfrom);
1208                 $text = false;
1209                 if ( $this->mOptions->getAllowExternalImages()
1210                      || ( $imagesexception && strpos( $url, $imagesfrom ) === 0 ) ) {
1211                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1212                                 # Image found
1213                                 $text = $sk->makeExternalImage( htmlspecialchars( $url ) );
1214                         }
1215                 }
1216                 return $text;
1217         }
1218
1219         /**
1220          * Process [[ ]] wikilinks
1221          *
1222          * @access private
1223          */
1224         function replaceInternalLinks( $s ) {
1225                 global $wgContLang;
1226                 static $fname = 'Parser::replaceInternalLinks' ;
1227
1228                 wfProfileIn( $fname );
1229
1230                 wfProfileIn( $fname.'-setup' );
1231                 static $tc = FALSE;
1232                 # the % is needed to support urlencoded titles as well
1233                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1234
1235                 $sk =& $this->mOptions->getSkin();
1236
1237                 #split the entire text string on occurences of [[
1238                 $a = explode( '[[', ' ' . $s );
1239                 #get the first element (all text up to first [[), and remove the space we added
1240                 $s = array_shift( $a );
1241                 $s = substr( $s, 1 );
1242
1243                 # Match a link having the form [[namespace:link|alternate]]trail
1244                 static $e1 = FALSE;
1245                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1246                 # Match cases where there is no "]]", which might still be images
1247                 static $e1_img = FALSE;
1248                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1249                 # Match the end of a line for a word that's not followed by whitespace,
1250                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1251                 $e2 = wfMsgForContent( 'linkprefix' );
1252
1253                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1254
1255                 if( is_null( $this->mTitle ) ) {
1256                         wfDebugDieBacktrace( 'nooo' );
1257                 }
1258                 $nottalk = !$this->mTitle->isTalkPage();
1259
1260                 if ( $useLinkPrefixExtension ) {
1261                         if ( preg_match( $e2, $s, $m ) ) {
1262                                 $first_prefix = $m[2];
1263                         } else {
1264                                 $first_prefix = false;
1265                         }
1266                 } else {
1267                         $prefix = '';
1268                 }
1269
1270                 $selflink = $this->mTitle->getPrefixedText();
1271                 wfProfileOut( $fname.'-setup' );
1272
1273                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1274                 $useSubpages = $this->areSubpagesAllowed();
1275
1276                 # Loop for each link
1277                 for ($k = 0; isset( $a[$k] ); $k++) {
1278                         $line = $a[$k];
1279                         if ( $useLinkPrefixExtension ) {
1280                                 wfProfileIn( $fname.'-prefixhandling' );
1281                                 if ( preg_match( $e2, $s, $m ) ) {
1282                                         $prefix = $m[2];
1283                                         $s = $m[1];
1284                                 } else {
1285                                         $prefix='';
1286                                 }
1287                                 # first link
1288                                 if($first_prefix) {
1289                                         $prefix = $first_prefix;
1290                                         $first_prefix = false;
1291                                 }
1292                                 wfProfileOut( $fname.'-prefixhandling' );
1293                         }
1294
1295                         $might_be_img = false;
1296
1297                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1298                                 $text = $m[2];
1299                                 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1300                                 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1301                                 # the real problem is with the $e1 regex
1302                                 # See bug 1300.
1303                                 #
1304                                 # Still some problems for cases where the ] is meant to be outside punctuation,
1305                                 # and no image is in sight. See bug 2095.
1306                                 #
1307                                 if( $text !== '' && preg_match( "/^\](.*)/s", $m[3], $n ) ) {
1308                                         $text .= ']'; # so that replaceExternalLinks($text) works later
1309                                         $m[3] = $n[1];
1310                                 }
1311                                 # fix up urlencoded title texts
1312                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1313                                 $trail = $m[3];
1314                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1315                                 $might_be_img = true;
1316                                 $text = $m[2];
1317                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1318                                 $trail = "";
1319                         } else { # Invalid form; output directly
1320                                 $s .= $prefix . '[[' . $line ;
1321                                 continue;
1322                         }
1323
1324                         # Don't allow internal links to pages containing
1325                         # PROTO: where PROTO is a valid URL protocol; these
1326                         # should be external links.
1327                         if (preg_match('/^(\b(?:' . wfUrlProtocols() . '))/', $m[1])) {
1328                                 $s .= $prefix . '[[' . $line ;
1329                                 continue;
1330                         }
1331
1332                         # Make subpage if necessary
1333                         if( $useSubpages ) {
1334                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1335                         } else {
1336                                 $link = $m[1];
1337                         }
1338
1339                         $noforce = (substr($m[1], 0, 1) != ':');
1340                         if (!$noforce) {
1341                                 # Strip off leading ':'
1342                                 $link = substr($link, 1);
1343                         }
1344
1345                         $nt = Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1346                         if( !$nt ) {
1347                                 $s .= $prefix . '[[' . $line;
1348                                 continue;
1349                         }
1350
1351                         #check other language variants of the link
1352                         #if the article does not exist
1353                         if( $checkVariantLink
1354                             && $nt->getArticleID() == 0 ) {
1355                                 $wgContLang->findVariantLink($link, $nt);
1356                         }
1357
1358                         $ns = $nt->getNamespace();
1359                         $iw = $nt->getInterWiki();
1360
1361                         if ($might_be_img) { # if this is actually an invalid link
1362                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1363                                         $found = false;
1364                                         while (isset ($a[$k+1]) ) {
1365                                                 #look at the next 'line' to see if we can close it there
1366                                                 $spliced = array_splice( $a, $k + 1, 1 );
1367                                                 $next_line = array_shift( $spliced );
1368                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1369                                                 # the first ]] closes the inner link, the second the image
1370                                                         $found = true;
1371                                                         $text .= '[[' . $m[1];
1372                                                         $trail = $m[2];
1373                                                         break;
1374                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1375                                                         #if there's exactly one ]] that's fine, we'll keep looking
1376                                                         $text .= '[[' . $m[0];
1377                                                 } else {
1378                                                         #if $next_line is invalid too, we need look no further
1379                                                         $text .= '[[' . $next_line;
1380                                                         break;
1381                                                 }
1382                                         }
1383                                         if ( !$found ) {
1384                                                 # we couldn't find the end of this imageLink, so output it raw
1385                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1386                                                 $text = $this->replaceInternalLinks($text);
1387                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1388                                                 # note: no $trail, because without an end, there *is* no trail
1389                                                 continue;
1390                                         }
1391                                 } else { #it's not an image, so output it raw
1392                                         $s .= $prefix . '[[' . $link . '|' . $text;
1393                                         # note: no $trail, because without an end, there *is* no trail
1394                                         continue;
1395                                 }
1396                         }
1397
1398                         $wasblank = ( '' == $text );
1399                         if( $wasblank ) $text = $link;
1400
1401
1402                         # Link not escaped by : , create the various objects
1403                         if( $noforce ) {
1404
1405                                 # Interwikis
1406                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1407                                         $this->mOutput->addLanguageLink( $nt->getFullText() );
1408                                         $s = rtrim($s . "\n");
1409                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1410                                         continue;
1411                                 }
1412
1413                                 if ( $ns == NS_IMAGE ) {
1414                                         wfProfileIn( "$fname-image" );
1415                                         if ( !wfIsBadImage( $nt->getDBkey() ) ) {
1416                                                 # recursively parse links inside the image caption
1417                                                 # actually, this will parse them in any other parameters, too,
1418                                                 # but it might be hard to fix that, and it doesn't matter ATM
1419                                                 $text = $this->replaceExternalLinks($text);
1420                                                 $text = $this->replaceInternalLinks($text);
1421
1422                                                 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1423                                                 $s .= $prefix . $this->armorLinks( $this->makeImage( $nt, $text ) ) . $trail;
1424                                                 $this->mOutput->addImage( $nt->getDBkey() );
1425
1426                                                 wfProfileOut( "$fname-image" );
1427                                                 continue;
1428                                         }
1429                                         wfProfileOut( "$fname-image" );
1430
1431                                 }
1432
1433                                 if ( $ns == NS_CATEGORY ) {
1434                                         wfProfileIn( "$fname-category" );
1435                                         $s = rtrim($s . "\n"); # bug 87
1436
1437                                         if ( $wasblank ) {
1438                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1439                                                         $sortkey = $this->mTitle->getText();
1440                                                 } else {
1441                                                         $sortkey = $this->mTitle->getPrefixedText();
1442                                                 }
1443                                         } else {
1444                                                 $sortkey = $text;
1445                                         }
1446                                         $sortkey = Sanitizer::decodeCharReferences( $sortkey );
1447                                         $sortkey = $wgContLang->convertCategoryKey( $sortkey );
1448                                         $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
1449
1450                                         /**
1451                                          * Strip the whitespace Category links produce, see bug 87
1452                                          * @todo We might want to use trim($tmp, "\n") here.
1453                                          */
1454                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1455
1456                                         wfProfileOut( "$fname-category" );
1457                                         continue;
1458                                 }
1459                         }
1460
1461                         if( ( $nt->getPrefixedText() === $selflink ) &&
1462                             ( $nt->getFragment() === '' ) ) {
1463                                 # Self-links are handled specially; generally de-link and change to bold.
1464                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1465                                 continue;
1466                         }
1467
1468                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1469                         if( $ns == NS_MEDIA ) {
1470                                 $link = $sk->makeMediaLinkObj( $nt, $text );
1471                                 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
1472                                 $s .= $prefix . $this->armorLinks( $link ) . $trail;
1473                                 $this->mOutput->addImage( $nt->getDBkey() );
1474                                 continue;
1475                         } elseif( $ns == NS_SPECIAL ) {
1476                                 $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix );
1477                                 continue;
1478                         } elseif( $ns == NS_IMAGE ) {
1479                                 $img = Image::newFromTitle( $nt );
1480                                 if( $img->exists() ) {
1481                                         // Force a blue link if the file exists; may be a remote
1482                                         // upload on the shared repository, and we want to see its
1483                                         // auto-generated page.
1484                                         $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix );
1485                                         continue;
1486                                 }
1487                         }
1488                         $s .= $this->makeLinkHolder( $nt, $text, '', $trail, $prefix );
1489                 }
1490                 wfProfileOut( $fname );
1491                 return $s;
1492         }
1493
1494         /**
1495          * Make a link placeholder. The text returned can be later resolved to a real link with
1496          * replaceLinkHolders(). This is done for two reasons: firstly to avoid further
1497          * parsing of interwiki links, and secondly to allow all extistence checks and
1498          * article length checks (for stub links) to be bundled into a single query.
1499          *
1500          */
1501         function makeLinkHolder( &$nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1502                 if ( ! is_object($nt) ) {
1503                         # Fail gracefully
1504                         $retVal = "<!-- ERROR -->{$prefix}{$text}{$trail}";
1505                 } else {
1506                         # Separate the link trail from the rest of the link
1507                         list( $inside, $trail ) = Linker::splitTrail( $trail );
1508
1509                         if ( $nt->isExternal() ) {
1510                                 $nr = array_push( $this->mInterwikiLinkHolders['texts'], $prefix.$text.$inside );
1511                                 $this->mInterwikiLinkHolders['titles'][] = $nt;
1512                                 $retVal = '<!--IWLINK '. ($nr-1) ."-->{$trail}";
1513                         } else {
1514                                 $nr = array_push( $this->mLinkHolders['namespaces'], $nt->getNamespace() );
1515                                 $this->mLinkHolders['dbkeys'][] = $nt->getDBkey();
1516                                 $this->mLinkHolders['queries'][] = $query;
1517                                 $this->mLinkHolders['texts'][] = $prefix.$text.$inside;
1518                                 $this->mLinkHolders['titles'][] = $nt;
1519
1520                                 $retVal = '<!--LINK '. ($nr-1) ."-->{$trail}";
1521                         }
1522                 }
1523                 return $retVal;
1524         }
1525
1526         /**
1527          * Render a forced-blue link inline; protect against double expansion of
1528          * URLs if we're in a mode that prepends full URL prefixes to internal links.
1529          * Since this little disaster has to split off the trail text to avoid
1530          * breaking URLs in the following text without breaking trails on the
1531          * wiki links, it's been made into a horrible function.
1532          *
1533          * @param Title $nt
1534          * @param string $text
1535          * @param string $query
1536          * @param string $trail
1537          * @param string $prefix
1538          * @return string HTML-wikitext mix oh yuck
1539          */
1540         function makeKnownLinkHolder( $nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1541                 list( $inside, $trail ) = Linker::splitTrail( $trail );
1542                 $sk =& $this->mOptions->getSkin();
1543                 $link = $sk->makeKnownLinkObj( $nt, $text, $query, $inside, $prefix );
1544                 return $this->armorLinks( $link ) . $trail;
1545         }
1546
1547         /**
1548          * Insert a NOPARSE hacky thing into any inline links in a chunk that's
1549          * going to go through further parsing steps before inline URL expansion.
1550          *
1551          * In particular this is important when using action=render, which causes
1552          * full URLs to be included.
1553          *
1554          * Oh man I hate our multi-layer parser!
1555          *
1556          * @param string more-or-less HTML
1557          * @return string less-or-more HTML with NOPARSE bits
1558          */
1559         function armorLinks( $text ) {
1560                 return preg_replace( "/\b(" . wfUrlProtocols() . ')/',
1561                         "{$this->mUniqPrefix}NOPARSE$1", $text );
1562         }
1563
1564         /**
1565          * Return true if subpage links should be expanded on this page.
1566          * @return bool
1567          */
1568         function areSubpagesAllowed() {
1569                 # Some namespaces don't allow subpages
1570                 global $wgNamespacesWithSubpages;
1571                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1572         }
1573
1574         /**
1575          * Handle link to subpage if necessary
1576          * @param string $target the source of the link
1577          * @param string &$text the link text, modified as necessary
1578          * @return string the full name of the link
1579          * @access private
1580          */
1581         function maybeDoSubpageLink($target, &$text) {
1582                 # Valid link forms:
1583                 # Foobar -- normal
1584                 # :Foobar -- override special treatment of prefix (images, language links)
1585                 # /Foobar -- convert to CurrentPage/Foobar
1586                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1587                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1588                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1589
1590                 $fname = 'Parser::maybeDoSubpageLink';
1591                 wfProfileIn( $fname );
1592                 $ret = $target; # default return value is no change
1593
1594                 # Some namespaces don't allow subpages,
1595                 # so only perform processing if subpages are allowed
1596                 if( $this->areSubpagesAllowed() ) {
1597                         # Look at the first character
1598                         if( $target != '' && $target{0} == '/' ) {
1599                                 # / at end means we don't want the slash to be shown
1600                                 if( substr( $target, -1, 1 ) == '/' ) {
1601                                         $target = substr( $target, 1, -1 );
1602                                         $noslash = $target;
1603                                 } else {
1604                                         $noslash = substr( $target, 1 );
1605                                 }
1606
1607                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1608                                 if( '' === $text ) {
1609                                         $text = $target;
1610                                 } # this might be changed for ugliness reasons
1611                         } else {
1612                                 # check for .. subpage backlinks
1613                                 $dotdotcount = 0;
1614                                 $nodotdot = $target;
1615                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1616                                         ++$dotdotcount;
1617                                         $nodotdot = substr( $nodotdot, 3 );
1618                                 }
1619                                 if($dotdotcount > 0) {
1620                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1621                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1622                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1623                                                 # / at the end means don't show full path
1624                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1625                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1626                                                         if( '' === $text ) {
1627                                                                 $text = $nodotdot;
1628                                                         }
1629                                                 }
1630                                                 $nodotdot = trim( $nodotdot );
1631                                                 if( $nodotdot != '' ) {
1632                                                         $ret .= '/' . $nodotdot;
1633                                                 }
1634                                         }
1635                                 }
1636                         }
1637                 }
1638
1639                 wfProfileOut( $fname );
1640                 return $ret;
1641         }
1642
1643         /**#@+
1644          * Used by doBlockLevels()
1645          * @access private
1646          */
1647         /* private */ function closeParagraph() {
1648                 $result = '';
1649                 if ( '' != $this->mLastSection ) {
1650                         $result = '</' . $this->mLastSection  . ">\n";
1651                 }
1652                 $this->mInPre = false;
1653                 $this->mLastSection = '';
1654                 return $result;
1655         }
1656         # getCommon() returns the length of the longest common substring
1657         # of both arguments, starting at the beginning of both.
1658         #
1659         /* private */ function getCommon( $st1, $st2 ) {
1660                 $fl = strlen( $st1 );
1661                 $shorter = strlen( $st2 );
1662                 if ( $fl < $shorter ) { $shorter = $fl; }
1663
1664                 for ( $i = 0; $i < $shorter; ++$i ) {
1665                         if ( $st1{$i} != $st2{$i} ) { break; }
1666                 }
1667                 return $i;
1668         }
1669         # These next three functions open, continue, and close the list
1670         # element appropriate to the prefix character passed into them.
1671         #
1672         /* private */ function openList( $char ) {
1673                 $result = $this->closeParagraph();
1674
1675                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1676                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1677                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1678                 else if ( ';' == $char ) {
1679                         $result .= '<dl><dt>';
1680                         $this->mDTopen = true;
1681                 }
1682                 else { $result = '<!-- ERR 1 -->'; }
1683
1684                 return $result;
1685         }
1686
1687         /* private */ function nextItem( $char ) {
1688                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1689                 else if ( ':' == $char || ';' == $char ) {
1690                         $close = '</dd>';
1691                         if ( $this->mDTopen ) { $close = '</dt>'; }
1692                         if ( ';' == $char ) {
1693                                 $this->mDTopen = true;
1694                                 return $close . '<dt>';
1695                         } else {
1696                                 $this->mDTopen = false;
1697                                 return $close . '<dd>';
1698                         }
1699                 }
1700                 return '<!-- ERR 2 -->';
1701         }
1702
1703         /* private */ function closeList( $char ) {
1704                 if ( '*' == $char ) { $text = '</li></ul>'; }
1705                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1706                 else if ( ':' == $char ) {
1707                         if ( $this->mDTopen ) {
1708                                 $this->mDTopen = false;
1709                                 $text = '</dt></dl>';
1710                         } else {
1711                                 $text = '</dd></dl>';
1712                         }
1713                 }
1714                 else {  return '<!-- ERR 3 -->'; }
1715                 return $text."\n";
1716         }
1717         /**#@-*/
1718
1719         /**
1720          * Make lists from lines starting with ':', '*', '#', etc.
1721          *
1722          * @access private
1723          * @return string the lists rendered as HTML
1724          */
1725         function doBlockLevels( $text, $linestart ) {
1726                 $fname = 'Parser::doBlockLevels';
1727                 wfProfileIn( $fname );
1728
1729                 # Parsing through the text line by line.  The main thing
1730                 # happening here is handling of block-level elements p, pre,
1731                 # and making lists from lines starting with * # : etc.
1732                 #
1733                 $textLines = explode( "\n", $text );
1734
1735                 $lastPrefix = $output = '';
1736                 $this->mDTopen = $inBlockElem = false;
1737                 $prefixLength = 0;
1738                 $paragraphStack = false;
1739
1740                 if ( !$linestart ) {
1741                         $output .= array_shift( $textLines );
1742                 }
1743                 foreach ( $textLines as $oLine ) {
1744                         $lastPrefixLength = strlen( $lastPrefix );
1745                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1746                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1747                         if ( !$this->mInPre ) {
1748                                 # Multiple prefixes may abut each other for nested lists.
1749                                 $prefixLength = strspn( $oLine, '*#:;' );
1750                                 $pref = substr( $oLine, 0, $prefixLength );
1751
1752                                 # eh?
1753                                 $pref2 = str_replace( ';', ':', $pref );
1754                                 $t = substr( $oLine, $prefixLength );
1755                                 $this->mInPre = !empty($preOpenMatch);
1756                         } else {
1757                                 # Don't interpret any other prefixes in preformatted text
1758                                 $prefixLength = 0;
1759                                 $pref = $pref2 = '';
1760                                 $t = $oLine;
1761                         }
1762
1763                         # List generation
1764                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1765                                 # Same as the last item, so no need to deal with nesting or opening stuff
1766                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1767                                 $paragraphStack = false;
1768
1769                                 if ( substr( $pref, -1 ) == ';') {
1770                                         # The one nasty exception: definition lists work like this:
1771                                         # ; title : definition text
1772                                         # So we check for : in the remainder text to split up the
1773                                         # title and definition, without b0rking links.
1774                                         $term = $t2 = '';
1775                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1776                                                 $t = $t2;
1777                                                 $output .= $term . $this->nextItem( ':' );
1778                                         }
1779                                 }
1780                         } elseif( $prefixLength || $lastPrefixLength ) {
1781                                 # Either open or close a level...
1782                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1783                                 $paragraphStack = false;
1784
1785                                 while( $commonPrefixLength < $lastPrefixLength ) {
1786                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1787                                         --$lastPrefixLength;
1788                                 }
1789                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1790                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1791                                 }
1792                                 while ( $prefixLength > $commonPrefixLength ) {
1793                                         $char = substr( $pref, $commonPrefixLength, 1 );
1794                                         $output .= $this->openList( $char );
1795
1796                                         if ( ';' == $char ) {
1797                                                 # FIXME: This is dupe of code above
1798                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1799                                                         $t = $t2;
1800                                                         $output .= $term . $this->nextItem( ':' );
1801                                                 }
1802                                         }
1803                                         ++$commonPrefixLength;
1804                                 }
1805                                 $lastPrefix = $pref2;
1806                         }
1807                         if( 0 == $prefixLength ) {
1808                                 wfProfileIn( "$fname-paragraph" );
1809                                 # No prefix (not in list)--go to paragraph mode
1810                                 // XXX: use a stack for nestable elements like span, table and div
1811                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1812                                 $closematch = preg_match(
1813                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1814                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1815                                 if ( $openmatch or $closematch ) {
1816                                         $paragraphStack = false;
1817                                         $output .= $this->closeParagraph();
1818                                         if ( $preOpenMatch and !$preCloseMatch ) {
1819                                                 $this->mInPre = true;
1820                                         }
1821                                         if ( $closematch ) {
1822                                                 $inBlockElem = false;
1823                                         } else {
1824                                                 $inBlockElem = true;
1825                                         }
1826                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1827                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1828                                                 // pre
1829                                                 if ($this->mLastSection != 'pre') {
1830                                                         $paragraphStack = false;
1831                                                         $output .= $this->closeParagraph().'<pre>';
1832                                                         $this->mLastSection = 'pre';
1833                                                 }
1834                                                 $t = substr( $t, 1 );
1835                                         } else {
1836                                                 // paragraph
1837                                                 if ( '' == trim($t) ) {
1838                                                         if ( $paragraphStack ) {
1839                                                                 $output .= $paragraphStack.'<br />';
1840                                                                 $paragraphStack = false;
1841                                                                 $this->mLastSection = 'p';
1842                                                         } else {
1843                                                                 if ($this->mLastSection != 'p' ) {
1844                                                                         $output .= $this->closeParagraph();
1845                                                                         $this->mLastSection = '';
1846                                                                         $paragraphStack = '<p>';
1847                                                                 } else {
1848                                                                         $paragraphStack = '</p><p>';
1849                                                                 }
1850                                                         }
1851                                                 } else {
1852                                                         if ( $paragraphStack ) {
1853                                                                 $output .= $paragraphStack;
1854                                                                 $paragraphStack = false;
1855                                                                 $this->mLastSection = 'p';
1856                                                         } else if ($this->mLastSection != 'p') {
1857                                                                 $output .= $this->closeParagraph().'<p>';
1858                                                                 $this->mLastSection = 'p';
1859                                                         }
1860                                                 }
1861                                         }
1862                                 }
1863                                 wfProfileOut( "$fname-paragraph" );
1864                         }
1865                         // somewhere above we forget to get out of pre block (bug 785)
1866                         if($preCloseMatch && $this->mInPre) {
1867                                 $this->mInPre = false;
1868                         }
1869                         if ($paragraphStack === false) {
1870                                 $output .= $t."\n";
1871                         }
1872                 }
1873                 while ( $prefixLength ) {
1874                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1875                         --$prefixLength;
1876                 }
1877                 if ( '' != $this->mLastSection ) {
1878                         $output .= '</' . $this->mLastSection . '>';
1879                         $this->mLastSection = '';
1880                 }
1881
1882                 wfProfileOut( $fname );
1883                 return $output;
1884         }
1885
1886         /**
1887          * Split up a string on ':', ignoring any occurences inside
1888          * <a>..</a> or <span>...</span>
1889          * @param string $str the string to split
1890          * @param string &$before set to everything before the ':'
1891          * @param string &$after set to everything after the ':'
1892          * return string the position of the ':', or false if none found
1893          */
1894         function findColonNoLinks($str, &$before, &$after) {
1895                 # I wonder if we should make this count all tags, not just <a>
1896                 # and <span>. That would prevent us from matching a ':' that
1897                 # comes in the middle of italics other such formatting....
1898                 # -- Wil
1899                 $fname = 'Parser::findColonNoLinks';
1900                 wfProfileIn( $fname );
1901                 $pos = 0;
1902                 do {
1903                         $colon = strpos($str, ':', $pos);
1904
1905                         if ($colon !== false) {
1906                                 $before = substr($str, 0, $colon);
1907                                 $after = substr($str, $colon + 1);
1908
1909                                 # Skip any ':' within <a> or <span> pairs
1910                                 $a = substr_count($before, '<a');
1911                                 $s = substr_count($before, '<span');
1912                                 $ca = substr_count($before, '</a>');
1913                                 $cs = substr_count($before, '</span>');
1914
1915                                 if ($a <= $ca and $s <= $cs) {
1916                                         # Tags are balanced before ':'; ok
1917                                         break;
1918                                 }
1919                                 $pos = $colon + 1;
1920                         }
1921                 } while ($colon !== false);
1922                 wfProfileOut( $fname );
1923                 return $colon;
1924         }
1925
1926         /**
1927          * Return value of a magic variable (like PAGENAME)
1928          *
1929          * @access private
1930          */
1931         function getVariableValue( $index ) {
1932                 global $wgContLang, $wgSitename, $wgServer, $wgServerName, $wgScriptPath;
1933
1934                 /**
1935                  * Some of these require message or data lookups and can be
1936                  * expensive to check many times.
1937                  */
1938                 static $varCache = array();
1939                 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$varCache ) ) )
1940                         if ( isset( $varCache[$index] ) )
1941                                 return $varCache[$index];
1942
1943                 $ts = time();
1944                 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) );
1945
1946                 switch ( $index ) {
1947                         case MAG_CURRENTMONTH:
1948                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm', $ts ) );
1949                         case MAG_CURRENTMONTHNAME:
1950                                 return $varCache[$index] = $wgContLang->getMonthName( date( 'n', $ts ) );
1951                         case MAG_CURRENTMONTHNAMEGEN:
1952                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date( 'n', $ts ) );
1953                         case MAG_CURRENTMONTHABBREV:
1954                                 return $varCache[$index] = $wgContLang->getMonthAbbreviation( date( 'n', $ts ) );
1955                         case MAG_CURRENTDAY:
1956                                 return $varCache[$index] = $wgContLang->formatNum( date( 'j', $ts ) );
1957                         case MAG_CURRENTDAY2:
1958                                 return $varCache[$index] = $wgContLang->formatNum( date( 'd', $ts ) );
1959                         case MAG_PAGENAME:
1960                                 return $this->mTitle->getText();
1961                         case MAG_PAGENAMEE:
1962                                 return $this->mTitle->getPartialURL();
1963                         case MAG_FULLPAGENAME:
1964                                 return $this->mTitle->getPrefixedText();
1965                         case MAG_FULLPAGENAMEE:
1966                                 return $this->mTitle->getPrefixedURL();
1967                         case MAG_REVISIONID:
1968                                 return $this->mRevisionId;
1969                         case MAG_NAMESPACE:
1970                                 return $wgContLang->getNsText( $this->mTitle->getNamespace() );
1971                         case MAG_NAMESPACEE:
1972                                 return wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
1973                         case MAG_CURRENTDAYNAME:
1974                                 return $varCache[$index] = $wgContLang->getWeekdayName( date( 'w', $ts ) + 1 );
1975                         case MAG_CURRENTYEAR:
1976                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y', $ts ), true );
1977                         case MAG_CURRENTTIME:
1978                                 return $varCache[$index] = $wgContLang->time( wfTimestamp( TS_MW, $ts ), false, false );
1979                         case MAG_CURRENTWEEK:
1980                                 return $varCache[$index] = $wgContLang->formatNum( date( 'W', $ts ) );
1981                         case MAG_CURRENTDOW:
1982                                 return $varCache[$index] = $wgContLang->formatNum( date( 'w', $ts ) );
1983                         case MAG_NUMBEROFARTICLES:
1984                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
1985                         case MAG_NUMBEROFFILES:
1986                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfFiles() );
1987                         case MAG_SITENAME:
1988                                 return $wgSitename;
1989                         case MAG_SERVER:
1990                                 return $wgServer;
1991                         case MAG_SERVERNAME:
1992                                 return $wgServerName;
1993                         case MAG_SCRIPTPATH:
1994                                 return $wgScriptPath;
1995                         default:
1996                                 $ret = null;
1997                                 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$varCache, &$index, &$ret ) ) )
1998                                         return $ret;
1999                                 else
2000                                         return null;
2001                 }
2002         }
2003
2004         /**
2005          * initialise the magic variables (like CURRENTMONTHNAME)
2006          *
2007          * @access private
2008          */
2009         function initialiseVariables() {
2010                 $fname = 'Parser::initialiseVariables';
2011                 wfProfileIn( $fname );
2012                 global $wgVariableIDs;
2013                 $this->mVariables = array();
2014                 foreach ( $wgVariableIDs as $id ) {
2015                         $mw =& MagicWord::get( $id );
2016                         $mw->addToArray( $this->mVariables, $id );
2017                 }
2018                 wfProfileOut( $fname );
2019         }
2020
2021         /**
2022          * parse any parentheses in format ((title|part|part))
2023          * and call callbacks to get a replacement text for any found piece
2024          *
2025          * @param string $text The text to parse
2026          * @param array $callbacks rules in form:
2027          *     '{' => array(                            # opening parentheses
2028          *                                      'end' => '}',   # closing parentheses
2029          *                                      'cb' => array(2 => callback,    # replacement callback to call if {{..}} is found
2030          *                                                                4 => callback         # replacement callback to call if {{{{..}}}} is found
2031          *                                                                )
2032          *                                      )
2033          * @access private
2034          */
2035         function replace_callback ($text, $callbacks) {
2036                 $openingBraceStack = array();   # this array will hold a stack of parentheses which are not closed yet
2037                 $lastOpeningBrace = -1;                 # last not closed parentheses
2038
2039                 for ($i = 0; $i < strlen($text); $i++) {
2040                         # check for any opening brace
2041                         $rule = null;
2042                         $nextPos = -1;
2043                         foreach ($callbacks as $key => $value) {
2044                                 $pos = strpos ($text, $key, $i);
2045                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)) {
2046                                         $rule = $value;
2047                                         $nextPos = $pos;
2048                                 }
2049                         }
2050
2051                         if ($lastOpeningBrace >= 0) {
2052                                 $pos = strpos ($text, $openingBraceStack[$lastOpeningBrace]['braceEnd'], $i);
2053
2054                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)){
2055                                         $rule = null;
2056                                         $nextPos = $pos;
2057                                 }
2058
2059                                 $pos = strpos ($text, '|', $i);
2060
2061                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)){
2062                                         $rule = null;
2063                                         $nextPos = $pos;
2064                                 }
2065                         }
2066
2067                         if ($nextPos == -1)
2068                                 break;
2069
2070                         $i = $nextPos;
2071
2072                         # found openning brace, lets add it to parentheses stack
2073                         if (null != $rule) {
2074                                 $piece = array('brace' => $text[$i],
2075                                                            'braceEnd' => $rule['end'],
2076                                                            'count' => 1,
2077                                                            'title' => '',
2078                                                            'parts' => null);
2079
2080                                 # count openning brace characters
2081                                 while ($i+1 < strlen($text) && $text[$i+1] == $piece['brace']) {
2082                                         $piece['count']++;
2083                                         $i++;
2084                                 }
2085
2086                                 $piece['startAt'] = $i+1;
2087                                 $piece['partStart'] = $i+1;
2088
2089                                 # we need to add to stack only if openning brace count is enough for any given rule
2090                                 foreach ($rule['cb'] as $cnt => $fn) {
2091                                         if ($piece['count'] >= $cnt) {
2092                                                 $lastOpeningBrace ++;
2093                                                 $openingBraceStack[$lastOpeningBrace] = $piece;
2094                                                 break;
2095                                         }
2096                                 }
2097
2098                                 continue;
2099                         }
2100                         else if ($lastOpeningBrace >= 0) {
2101                                 # first check if it is a closing brace
2102                                 if ($openingBraceStack[$lastOpeningBrace]['braceEnd'] == $text[$i]) {
2103                                         # lets check if it is enough characters for closing brace
2104                                         $count = 1;
2105                                         while ($i+$count < strlen($text) && $text[$i+$count] == $text[$i])
2106                                                 $count++;
2107
2108                                         # if there are more closing parentheses than opening ones, we parse less
2109                                         if ($openingBraceStack[$lastOpeningBrace]['count'] < $count)
2110                                                 $count = $openingBraceStack[$lastOpeningBrace]['count'];
2111
2112                                         # check for maximum matching characters (if there are 5 closing characters, we will probably need only 3 - depending on the rules)
2113                                         $matchingCount = 0;
2114                                         $matchingCallback = null;
2115                                         foreach ($callbacks[$openingBraceStack[$lastOpeningBrace]['brace']]['cb'] as $cnt => $fn) {
2116                                                 if ($count >= $cnt && $matchingCount < $cnt) {
2117                                                         $matchingCount = $cnt;
2118                                                         $matchingCallback = $fn;
2119                                                 }
2120                                         }
2121
2122                                         if ($matchingCount == 0) {
2123                                                 $i += $count - 1;
2124                                                 continue;
2125                                         }
2126
2127                                         # lets set a title or last part (if '|' was found)
2128                                         if (null === $openingBraceStack[$lastOpeningBrace]['parts'])
2129                                                 $openingBraceStack[$lastOpeningBrace]['title'] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2130                                         else
2131                                                 $openingBraceStack[$lastOpeningBrace]['parts'][] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2132
2133                                         $pieceStart = $openingBraceStack[$lastOpeningBrace]['startAt'] - $matchingCount;
2134                                         $pieceEnd = $i + $matchingCount;
2135
2136                                         if( is_callable( $matchingCallback ) ) {
2137                                                 $cbArgs = array (
2138                                                                                  'text' => substr($text, $pieceStart, $pieceEnd - $pieceStart),
2139                                                                                  'title' => trim($openingBraceStack[$lastOpeningBrace]['title']),
2140                                                                                  'parts' => $openingBraceStack[$lastOpeningBrace]['parts'],
2141                                                                                  'lineStart' => (($pieceStart > 0) && ($text[$pieceStart-1] == '\n')),
2142                                                                                  );
2143                                                 # finally we can call a user callback and replace piece of text
2144                                                 $replaceWith = call_user_func( $matchingCallback, $cbArgs );
2145                                                 $text = substr($text, 0, $pieceStart) . $replaceWith . substr($text, $pieceEnd);
2146                                                 $i = $pieceStart + strlen($replaceWith) - 1;
2147                                         }
2148                                         else {
2149                                                 # null value for callback means that parentheses should be parsed, but not replaced
2150                                                 $i += $matchingCount - 1;
2151                                         }
2152
2153                                         # reset last openning parentheses, but keep it in case there are unused characters
2154                                         $piece = array('brace' => $openingBraceStack[$lastOpeningBrace]['brace'],
2155                                                                    'braceEnd' => $openingBraceStack[$lastOpeningBrace]['braceEnd'],
2156                                                                    'count' => $openingBraceStack[$lastOpeningBrace]['count'],
2157                                                                    'title' => '',
2158                                                                    'parts' => null,
2159                                                                    'startAt' => $openingBraceStack[$lastOpeningBrace]['startAt']);
2160                                         $openingBraceStack[$lastOpeningBrace--] = null;
2161
2162                                         if ($matchingCount < $piece['count']) {
2163                                                 $piece['count'] -= $matchingCount;
2164                                                 $piece['startAt'] -= $matchingCount;
2165                                                 $piece['partStart'] = $piece['startAt'];
2166                                                 # do we still qualify for any callback with remaining count?
2167                                                 foreach ($callbacks[$piece['brace']]['cb'] as $cnt => $fn) {
2168                                                         if ($piece['count'] >= $cnt) {
2169                                                                 $lastOpeningBrace ++;
2170                                                                 $openingBraceStack[$lastOpeningBrace] = $piece;
2171                                                                 break;
2172                                                         }
2173                                                 }
2174                                         }
2175                                         continue;
2176                                 }
2177
2178                                 # lets set a title if it is a first separator, or next part otherwise
2179                                 if ($text[$i] == '|') {
2180                                         if (null === $openingBraceStack[$lastOpeningBrace]['parts']) {
2181                                                 $openingBraceStack[$lastOpeningBrace]['title'] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2182                                                 $openingBraceStack[$lastOpeningBrace]['parts'] = array();
2183                                         }
2184                                         else
2185                                                 $openingBraceStack[$lastOpeningBrace]['parts'][] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2186
2187                                         $openingBraceStack[$lastOpeningBrace]['partStart'] = $i + 1;
2188                                 }
2189                         }
2190                 }
2191
2192                 return $text;
2193         }
2194
2195         /**
2196          * Replace magic variables, templates, and template arguments
2197          * with the appropriate text. Templates are substituted recursively,
2198          * taking care to avoid infinite loops.
2199          *
2200          * Note that the substitution depends on value of $mOutputType:
2201          *  OT_WIKI: only {{subst:}} templates
2202          *  OT_MSG: only magic variables
2203          *  OT_HTML: all templates and magic variables
2204          *
2205          * @param string $tex The text to transform
2206          * @param array $args Key-value pairs representing template parameters to substitute
2207          * @access private
2208          */
2209         function replaceVariables( $text, $args = array() ) {
2210                 # Prevent too big inclusions
2211                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
2212                         return $text;
2213                 }
2214
2215                 $fname = 'Parser::replaceVariables';
2216                 wfProfileIn( $fname );
2217
2218                 # This function is called recursively. To keep track of arguments we need a stack:
2219                 array_push( $this->mArgStack, $args );
2220
2221                 $braceCallbacks = array();
2222                 $braceCallbacks[2] = array( &$this, 'braceSubstitution' );
2223                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
2224                         $braceCallbacks[3] = array( &$this, 'argSubstitution' );
2225                 }
2226                 $callbacks = array();
2227                 $callbacks['{'] = array('end' => '}', 'cb' => $braceCallbacks);
2228                 $callbacks['['] = array('end' => ']', 'cb' => array(2=>null));
2229                 $text = $this->replace_callback ($text, $callbacks);
2230
2231                 array_pop( $this->mArgStack );
2232
2233                 wfProfileOut( $fname );
2234                 return $text;
2235         }
2236
2237         /**
2238          * Replace magic variables
2239          * @access private
2240          */
2241         function variableSubstitution( $matches ) {
2242                 $fname = 'Parser::variableSubstitution';
2243                 $varname = $matches[1];
2244                 wfProfileIn( $fname );
2245                 if ( !$this->mVariables ) {
2246                         $this->initialiseVariables();
2247                 }
2248                 $skip = false;
2249                 if ( $this->mOutputType == OT_WIKI ) {
2250                         # Do only magic variables prefixed by SUBST
2251                         $mwSubst =& MagicWord::get( MAG_SUBST );
2252                         if (!$mwSubst->matchStartAndRemove( $varname ))
2253                                 $skip = true;
2254                         # Note that if we don't substitute the variable below,
2255                         # we don't remove the {{subst:}} magic word, in case
2256                         # it is a template rather than a magic variable.
2257                 }
2258                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
2259                         $id = $this->mVariables[$varname];
2260                         $text = $this->getVariableValue( $id );
2261                         $this->mOutput->mContainsOldMagic = true;
2262                 } else {
2263                         $text = $matches[0];
2264                 }
2265                 wfProfileOut( $fname );
2266                 return $text;
2267         }
2268
2269         # Split template arguments
2270         function getTemplateArgs( $argsString ) {
2271                 if ( $argsString === '' ) {
2272                         return array();
2273                 }
2274
2275                 $args = explode( '|', substr( $argsString, 1 ) );
2276
2277                 # If any of the arguments contains a '[[' but no ']]', it needs to be
2278                 # merged with the next arg because the '|' character between belongs
2279                 # to the link syntax and not the template parameter syntax.
2280                 $argc = count($args);
2281
2282                 for ( $i = 0; $i < $argc-1; $i++ ) {
2283                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
2284                                 $args[$i] .= '|'.$args[$i+1];
2285                                 array_splice($args, $i+1, 1);
2286                                 $i--;
2287                                 $argc--;
2288                         }
2289                 }
2290
2291                 return $args;
2292         }
2293
2294         /**
2295          * Return the text of a template, after recursively
2296          * replacing any variables or templates within the template.
2297          *
2298          * @param array $piece The parts of the template
2299          *  $piece['text']: matched text
2300          *  $piece['title']: the title, i.e. the part before the |
2301          *  $piece['parts']: the parameter array
2302          * @return string the text of the template
2303          * @access private
2304          */
2305         function braceSubstitution( $piece ) {
2306                 global $wgContLang;
2307                 $fname = 'Parser::braceSubstitution';
2308                 wfProfileIn( $fname );
2309
2310                 $found = false;
2311                 $nowiki = false;
2312                 $noparse = false;
2313
2314                 $title = NULL;
2315
2316                 $linestart = '';
2317
2318                 # $part1 is the bit before the first |, and must contain only title characters
2319                 # $args is a list of arguments, starting from index 0, not including $part1
2320
2321                 $part1 = $piece['title'];
2322                 # If the third subpattern matched anything, it will start with |
2323
2324                 if (null == $piece['parts']) {
2325                         $replaceWith = $this->variableSubstitution (array ($piece['text'], $piece['title']));
2326                         if ($replaceWith != $piece['text']) {
2327                                 $text = $replaceWith;
2328                                 $found = true;
2329                                 $noparse = true;
2330                         }
2331                 }
2332
2333                 $args = (null == $piece['parts']) ? array() : $piece['parts'];
2334                 $argc = count( $args );
2335
2336                 # SUBST
2337                 if ( !$found ) {
2338                         $mwSubst =& MagicWord::get( MAG_SUBST );
2339                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
2340                                 # One of two possibilities is true:
2341                                 # 1) Found SUBST but not in the PST phase
2342                                 # 2) Didn't find SUBST and in the PST phase
2343                                 # In either case, return without further processing
2344                                 $text = $piece['text'];
2345                                 $found = true;
2346                                 $noparse = true;
2347                         }
2348                 }
2349
2350                 # MSG, MSGNW and INT
2351                 if ( !$found ) {
2352                         # Check for MSGNW:
2353                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
2354                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
2355                                 $nowiki = true;
2356                         } else {
2357                                 # Remove obsolete MSG:
2358                                 $mwMsg =& MagicWord::get( MAG_MSG );
2359                                 $mwMsg->matchStartAndRemove( $part1 );
2360                         }
2361
2362                         # Check if it is an internal message
2363                         $mwInt =& MagicWord::get( MAG_INT );
2364                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
2365                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
2366                                         $text = $linestart . wfMsgReal( $part1, $args, true );
2367                                         $found = true;
2368                                 }
2369                         }
2370                 }
2371
2372                 # NS
2373                 if ( !$found ) {
2374                         # Check for NS: (namespace expansion)
2375                         $mwNs = MagicWord::get( MAG_NS );
2376                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
2377                                 if ( intval( $part1 ) ) {
2378                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
2379                                         $found = true;
2380                                 } else {
2381                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
2382                                         if ( !is_null( $index ) ) {
2383                                                 $text = $linestart . $wgContLang->getNsText( $index );
2384                                                 $found = true;
2385                                         }
2386                                 }
2387                         }
2388                 }
2389
2390                 # LCFIRST, UCFIRST, LC and UC
2391                 if ( !$found ) {
2392                         $lcfirst =& MagicWord::get( MAG_LCFIRST );
2393                         $ucfirst =& MagicWord::get( MAG_UCFIRST );
2394                         $lc =& MagicWord::get( MAG_LC );
2395                         $uc =& MagicWord::get( MAG_UC );
2396                         if ( $lcfirst->matchStartAndRemove( $part1 ) ) {
2397                                 $text = $linestart . $wgContLang->lcfirst( $part1 );
2398                                 $found = true;
2399                         } else if ( $ucfirst->matchStartAndRemove( $part1 ) ) {
2400                                 $text = $linestart . $wgContLang->ucfirst( $part1 );
2401                                 $found = true;
2402                         } else if ( $lc->matchStartAndRemove( $part1 ) ) {
2403                                 $text = $linestart . $wgContLang->lc( $part1 );
2404                                 $found = true;
2405                         } else if ( $uc->matchStartAndRemove( $part1 ) ) {
2406                                  $text = $linestart . $wgContLang->uc( $part1 );
2407                                  $found = true;
2408                         }
2409                 }
2410
2411                 # LOCALURL and FULLURL
2412                 if ( !$found ) {
2413                         $mwLocal =& MagicWord::get( MAG_LOCALURL );
2414                         $mwLocalE =& MagicWord::get( MAG_LOCALURLE );
2415                         $mwFull =& MagicWord::get( MAG_FULLURL );
2416                         $mwFullE =& MagicWord::get( MAG_FULLURLE );
2417
2418
2419                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2420                                 $func = 'getLocalURL';
2421                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2422                                 $func = 'escapeLocalURL';
2423                         } elseif ( $mwFull->matchStartAndRemove( $part1 ) ) {
2424                                 $func = 'getFullURL';
2425                         } elseif ( $mwFullE->matchStartAndRemove( $part1 ) ) {
2426                                 $func = 'escapeFullURL';
2427                         } else {
2428                                 $func = false;
2429                         }
2430
2431                         if ( $func !== false ) {
2432                                 $title = Title::newFromText( $part1 );
2433                                 if ( !is_null( $title ) ) {
2434                                         if ( $argc > 0 ) {
2435                                                 $text = $linestart . $title->$func( $args[0] );
2436                                         } else {
2437                                                 $text = $linestart . $title->$func();
2438                                         }
2439                                         $found = true;
2440                                 }
2441                         }
2442                 }
2443
2444                 # GRAMMAR
2445                 if ( !$found && $argc == 1 ) {
2446                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2447                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2448                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2449                                 $found = true;
2450                         }
2451                 }
2452
2453                 # PLURAL
2454                 if ( !$found && $argc >= 2 ) {
2455                         $mwPluralForm =& MagicWord::get( MAG_PLURAL );
2456                         if ( $mwPluralForm->matchStartAndRemove( $part1 ) ) {
2457                                 if ($argc==2) {$args[2]=$args[1];}
2458                                 $text = $linestart . $wgContLang->convertPlural( $part1, $args[0], $args[1], $args[2]);
2459                                 $found = true;
2460                         }
2461                 }
2462
2463                 # Template table test
2464
2465                 # Did we encounter this template already? If yes, it is in the cache
2466                 # and we need to check for loops.
2467                 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
2468                         $found = true;
2469
2470                         # Infinite loop test
2471                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2472                                 $noparse = true;
2473                                 $found = true;
2474                                 $text = $linestart .
2475                                         '{{' . $part1 . '}}' .
2476                                         '<!-- WARNING: template loop detected -->';
2477                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2478                         } else {
2479                                 # set $text to cached message.
2480                                 $text = $linestart . $this->mTemplates[$part1];
2481                         }
2482                 }
2483
2484                 # Load from database
2485                 $replaceHeadings = false;
2486                 $isHTML = false;
2487                 $lastPathLevel = $this->mTemplatePath;
2488                 if ( !$found ) {
2489                         $ns = NS_TEMPLATE;
2490                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage='' );
2491                         if ($subpage !== '') {
2492                                 $ns = $this->mTitle->getNamespace();
2493                         }
2494                         $title = Title::newFromText( $part1, $ns );
2495
2496                         if ($title) {
2497                                 $interwiki = Title::getInterwikiLink($title->getInterwiki());
2498                                 if ($interwiki != '' && $title->isTrans()) {
2499                                         return $this->scarytransclude($title, $interwiki);
2500                                 }
2501                         }
2502
2503                         if ( !is_null( $title ) && !$title->isExternal() ) {
2504                                 # Check for excessive inclusion
2505                                 $dbk = $title->getPrefixedDBkey();
2506                                 if ( $this->incrementIncludeCount( $dbk ) ) {
2507                                         if ( $title->getNamespace() == NS_SPECIAL && $this->mOptions->getAllowSpecialInclusion() ) {
2508                                                 # Capture special page output
2509                                                 $text = SpecialPage::capturePath( $title );
2510                                                 if ( is_string( $text ) ) {
2511                                                         $found = true;
2512                                                         $noparse = true;
2513                                                         $isHTML = true;
2514                                                         $this->disableCache();
2515                                                 }
2516                                         } else {
2517                                                 $articleContent = $this->fetchTemplate( $title );
2518                                                 if ( $articleContent !== false ) {
2519                                                         $found = true;
2520                                                         $text = $articleContent;
2521                                                         $replaceHeadings = true;
2522                                                 }
2523                                         }
2524                                 }
2525
2526                                 # If the title is valid but undisplayable, make a link to it
2527                                 if ( $this->mOutputType == OT_HTML && !$found ) {
2528                                         $text = '[['.$title->getPrefixedText().']]';
2529                                         $found = true;
2530                                 }
2531
2532                                 # Template cache array insertion
2533                                 if( $found ) {
2534                                         $this->mTemplates[$part1] = $text;
2535                                         $text = $linestart . $text;
2536                                 }
2537                         }
2538                 }
2539
2540                 # Recursive parsing, escaping and link table handling
2541                 # Only for HTML output
2542                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2543                         $text = wfEscapeWikiText( $text );
2544                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
2545                         # Clean up argument array
2546                         $assocArgs = array();
2547                         $index = 1;
2548                         foreach( $args as $arg ) {
2549                                 $eqpos = strpos( $arg, '=' );
2550                                 if ( $eqpos === false ) {
2551                                         $assocArgs[$index++] = $arg;
2552                                 } else {
2553                                         $name = trim( substr( $arg, 0, $eqpos ) );
2554                                         $value = trim( substr( $arg, $eqpos+1 ) );
2555                                         if ( $value === false ) {
2556                                                 $value = '';
2557                                         }
2558                                         if ( $name !== false ) {
2559                                                 $assocArgs[$name] = $value;
2560                                         }
2561                                 }
2562                         }
2563
2564                         # Add a new element to the templace recursion path
2565                         $this->mTemplatePath[$part1] = 1;
2566
2567                         # If there are any <onlyinclude> tags, only include them
2568                         if ( in_string( '<onlyinclude>', $text ) && in_string( '</onlyinclude>', $text ) ) {
2569                                 preg_match_all( '/<onlyinclude>(.*?)\n?<\/onlyinclude>/s', $text, $m );
2570                                 $text = '';
2571                                 foreach ($m[1] as $piece)
2572                                         $text .= $piece;
2573                         }
2574                         # Remove <noinclude> sections and <includeonly> tags
2575                         $text = preg_replace( '/<noinclude>.*?<\/noinclude>/s', '', $text );
2576                         $text = strtr( $text, array( '<includeonly>' => '' , '</includeonly>' => '' ) );
2577
2578                         if( $this->mOutputType == OT_HTML ) {
2579                                 # Strip <nowiki>, <pre>, etc.
2580                                 $text = $this->strip( $text, $this->mStripState );
2581                                 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'replaceVariables' ), $assocArgs );
2582                         }
2583                         $text = $this->replaceVariables( $text, $assocArgs );
2584
2585                         # If the template begins with a table or block-level
2586                         # element, it should be treated as beginning a new line.
2587                         if (!$piece['lineStart'] && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2588                                 $text = "\n" . $text;
2589                         }
2590                 }
2591                 # Prune lower levels off the recursion check path
2592                 $this->mTemplatePath = $lastPathLevel;
2593
2594                 if ( !$found ) {
2595                         wfProfileOut( $fname );
2596                         return $piece['text'];
2597                 } else {
2598                         if ( $isHTML ) {
2599                                 # Replace raw HTML by a placeholder
2600                                 # Add a blank line preceding, to prevent it from mucking up
2601                                 # immediately preceding headings
2602                                 $text = "\n\n" . $this->insertStripItem( $text, $this->mStripState );
2603                         } else {
2604                                 # replace ==section headers==
2605                                 # XXX this needs to go away once we have a better parser.
2606                                 if ( $this->mOutputType != OT_WIKI && $replaceHeadings ) {
2607                                         if( !is_null( $title ) )
2608                                                 $encodedname = base64_encode($title->getPrefixedDBkey());
2609                                         else
2610                                                 $encodedname = base64_encode("");
2611                                         $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2612                                                 PREG_SPLIT_DELIM_CAPTURE);
2613                                         $text = '';
2614                                         $nsec = 0;
2615                                         for( $i = 0; $i < count($m); $i += 2 ) {
2616                                                 $text .= $m[$i];
2617                                                 if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2618                                                 $hl = $m[$i + 1];
2619                                                 if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2620                                                         $text .= $hl;
2621                                                         continue;
2622                                                 }
2623                                                 preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2624                                                 $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2625                                                         . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2626
2627                                                 $nsec++;
2628                                         }
2629                                 }
2630                         }
2631                 }
2632
2633                 # Prune lower levels off the recursion check path
2634                 $this->mTemplatePath = $lastPathLevel;
2635
2636                 if ( !$found ) {
2637                         wfProfileOut( $fname );
2638                         return $piece['text'];
2639                 } else {
2640                         wfProfileOut( $fname );
2641                         return $text;
2642                 }
2643         }
2644
2645         /**
2646          * Fetch the unparsed text of a template and register a reference to it.
2647          */
2648         function fetchTemplate( $title ) {
2649                 $text = false;
2650                 // Loop to fetch the article, with up to 1 redirect
2651                 for ( $i = 0; $i < 2 && is_object( $title ); $i++ ) {
2652                         $rev = Revision::newFromTitle( $title );
2653                         $this->mOutput->addTemplate( $title, $title->getArticleID() );
2654                         if ( !$rev ) {
2655                                 break;
2656                         }
2657                         $text = $rev->getText();
2658                         if ( $text === false ) {
2659                                 break;
2660                         }
2661                         // Redirect?
2662                         $title = Title::newFromRedirect( $text );
2663                 }
2664                 return $text;
2665         }
2666
2667         /**
2668          * Translude an interwiki link.
2669          */
2670         function scarytransclude($title, $interwiki) {
2671                 global $wgEnableScaryTranscluding;
2672
2673                 if (!$wgEnableScaryTranscluding)
2674                         return wfMsg('scarytranscludedisabled');
2675
2676                 $articlename = "Template:" . $title->getDBkey();
2677                 $url = str_replace('$1', urlencode($articlename), $interwiki);
2678                 if (strlen($url) > 255)
2679                         return wfMsg('scarytranscludetoolong');
2680                 $text = $this->fetchScaryTemplateMaybeFromCache($url);
2681                 $this->mIWTransData[] = $text;
2682                 return "<!--IW_TRANSCLUDE ".(count($this->mIWTransData) - 1)."-->";
2683         }
2684
2685         function fetchScaryTemplateMaybeFromCache($url) {
2686                 $dbr =& wfGetDB(DB_SLAVE);
2687                 $obj = $dbr->selectRow('transcache', array('tc_time', 'tc_contents'),
2688                                 array('tc_url' => $url));
2689                 if ($obj) {
2690                         $time = $obj->tc_time;
2691                         $text = $obj->tc_contents;
2692                         if ($time && $time < (time() + (60*60))) {
2693                                 return $text;
2694                         }
2695                 }
2696
2697                 $text = wfGetHTTP($url . '?action=render');
2698                 if (!$text)
2699                         return wfMsg('scarytranscludefailed', $url);
2700
2701                 $dbw =& wfGetDB(DB_MASTER);
2702                 $dbw->replace('transcache', array(), array(
2703                         'tc_url' => $url,
2704                         'tc_time' => time(),
2705                         'tc_contents' => $text));
2706                 return $text;
2707         }
2708
2709
2710         /**
2711          * Triple brace replacement -- used for template arguments
2712          * @access private
2713          */
2714         function argSubstitution( $matches ) {
2715                 $arg = trim( $matches['title'] );
2716                 $text = $matches['text'];
2717                 $inputArgs = end( $this->mArgStack );
2718
2719                 if ( array_key_exists( $arg, $inputArgs ) ) {
2720                         $text = $inputArgs[$arg];
2721                 } else if ($this->mOutputType == OT_HTML && null != $matches['parts'] && count($matches['parts']) > 0) {
2722                         $text = $matches['parts'][0];
2723                 }
2724
2725                 return $text;
2726         }
2727
2728         /**
2729          * Returns true if the function is allowed to include this entity
2730          * @access private
2731          */
2732         function incrementIncludeCount( $dbk ) {
2733                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2734                         $this->mIncludeCount[$dbk] = 0;
2735                 }
2736                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2737                         return true;
2738                 } else {
2739                         return false;
2740                 }
2741         }
2742
2743         /**
2744          * This function accomplishes several tasks:
2745          * 1) Auto-number headings if that option is enabled
2746          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2747          * 3) Add a Table of contents on the top for users who have enabled the option
2748          * 4) Auto-anchor headings
2749          *
2750          * It loops through all headlines, collects the necessary data, then splits up the
2751          * string and re-inserts the newly formatted headlines.
2752          *
2753          * @param string $text
2754          * @param boolean $isMain
2755          * @access private
2756          */
2757         function formatHeadings( $text, $isMain=true ) {
2758                 global $wgMaxTocLevel, $wgContLang;
2759
2760                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2761                 $doShowToc = true;
2762                 $forceTocHere = false;
2763                 if( !$this->mTitle->userCanEdit() ) {
2764                         $showEditLink = 0;
2765                 } else {
2766                         $showEditLink = $this->mOptions->getEditSection();
2767                 }
2768
2769                 # Inhibit editsection links if requested in the page
2770                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2771                 if( $esw->matchAndRemove( $text ) ) {
2772                         $showEditLink = 0;
2773                 }
2774                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2775                 # do not add TOC
2776                 $mw =& MagicWord::get( MAG_NOTOC );
2777                 if( $mw->matchAndRemove( $text ) ) {
2778                         $doShowToc = false;
2779                 }
2780
2781                 # Get all headlines for numbering them and adding funky stuff like [edit]
2782                 # links - this is for later, but we need the number of headlines right now
2783                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2784
2785                 # if there are fewer than 4 headlines in the article, do not show TOC
2786                 if( $numMatches < 4 ) {
2787                         $doShowToc = false;
2788                 }
2789
2790                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2791                 # override above conditions and always show TOC at that place
2792
2793                 $mw =& MagicWord::get( MAG_TOC );
2794                 if($mw->match( $text ) ) {
2795                         $doShowToc = true;
2796                         $forceTocHere = true;
2797                 } else {
2798                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2799                         # override above conditions and always show TOC above first header
2800                         $mw =& MagicWord::get( MAG_FORCETOC );
2801                         if ($mw->matchAndRemove( $text ) ) {
2802                                 $doShowToc = true;
2803                         }
2804                 }
2805
2806                 # Never ever show TOC if no headers
2807                 if( $numMatches < 1 ) {
2808                         $doShowToc = false;
2809                 }
2810
2811                 # We need this to perform operations on the HTML
2812                 $sk =& $this->mOptions->getSkin();
2813
2814                 # headline counter
2815                 $headlineCount = 0;
2816                 $sectionCount = 0; # headlineCount excluding template sections
2817
2818                 # Ugh .. the TOC should have neat indentation levels which can be
2819                 # passed to the skin functions. These are determined here
2820                 $toc = '';
2821                 $full = '';
2822                 $head = array();
2823                 $sublevelCount = array();
2824                 $levelCount = array();
2825                 $toclevel = 0;
2826                 $level = 0;
2827                 $prevlevel = 0;
2828                 $toclevel = 0;
2829                 $prevtoclevel = 0;
2830
2831                 foreach( $matches[3] as $headline ) {
2832                         $istemplate = 0;
2833                         $templatetitle = '';
2834                         $templatesection = 0;
2835                         $numbering = '';
2836
2837                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2838                                 $istemplate = 1;
2839                                 $templatetitle = base64_decode($mat[1]);
2840                                 $templatesection = 1 + (int)base64_decode($mat[2]);
2841                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2842                         }
2843
2844                         if( $toclevel ) {
2845                                 $prevlevel = $level;
2846                                 $prevtoclevel = $toclevel;
2847                         }
2848                         $level = $matches[1][$headlineCount];
2849
2850                         if( $doNumberHeadings || $doShowToc ) {
2851
2852                                 if ( $level > $prevlevel ) {
2853                                         # Increase TOC level
2854                                         $toclevel++;
2855                                         $sublevelCount[$toclevel] = 0;
2856                                         $toc .= $sk->tocIndent();
2857                                 }
2858                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
2859                                         # Decrease TOC level, find level to jump to
2860
2861                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
2862                                                 # Can only go down to level 1
2863                                                 $toclevel = 1;
2864                                         } else {
2865                                                 for ($i = $toclevel; $i > 0; $i--) {
2866                                                         if ( $levelCount[$i] == $level ) {
2867                                                                 # Found last matching level
2868                                                                 $toclevel = $i;
2869                                                                 break;
2870                                                         }
2871                                                         elseif ( $levelCount[$i] < $level ) {
2872                                                                 # Found first matching level below current level
2873                                                                 $toclevel = $i + 1;
2874                                                                 break;
2875                                                         }
2876                                                 }
2877                                         }
2878
2879                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
2880                                 }
2881                                 else {
2882                                         # No change in level, end TOC line
2883                                         $toc .= $sk->tocLineEnd();
2884                                 }
2885
2886                                 $levelCount[$toclevel] = $level;
2887
2888                                 # count number of headlines for each level
2889                                 @$sublevelCount[$toclevel]++;
2890                                 $dot = 0;
2891                                 for( $i = 1; $i <= $toclevel; $i++ ) {
2892                                         if( !empty( $sublevelCount[$i] ) ) {
2893                                                 if( $dot ) {
2894                                                         $numbering .= '.';
2895                                                 }
2896                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2897                                                 $dot = 1;
2898                                         }
2899                                 }
2900                         }
2901
2902                         # The canonized header is a version of the header text safe to use for links
2903                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2904                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2905                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
2906
2907                         # Remove link placeholders by the link text.
2908                         #     <!--LINK number-->
2909                         # turns into
2910                         #     link text with suffix
2911                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2912                                                             "\$this->mLinkHolders['texts'][\$1]",
2913                                                             $canonized_headline );
2914                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
2915                                                             "\$this->mInterwikiLinkHolders['texts'][\$1]",
2916                                                             $canonized_headline );
2917
2918                         # strip out HTML
2919                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2920                         $tocline = trim( $canonized_headline );
2921                         $canonized_headline = Sanitizer::escapeId( $tocline );
2922                         $refers[$headlineCount] = $canonized_headline;
2923
2924                         # count how many in assoc. array so we can track dupes in anchors
2925                         @$refers[$canonized_headline]++;
2926                         $refcount[$headlineCount]=$refers[$canonized_headline];
2927
2928                         # Don't number the heading if it is the only one (looks silly)
2929                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
2930                                 # the two are different if the line contains a link
2931                                 $headline=$numbering . ' ' . $headline;
2932                         }
2933
2934                         # Create the anchor for linking from the TOC to the section
2935                         $anchor = $canonized_headline;
2936                         if($refcount[$headlineCount] > 1 ) {
2937                                 $anchor .= '_' . $refcount[$headlineCount];
2938                         }
2939                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2940                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
2941                         }
2942                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2943                                 if ( empty( $head[$headlineCount] ) ) {
2944                                         $head[$headlineCount] = '';
2945                                 }
2946                                 if( $istemplate )
2947                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2948                                 else
2949                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1);
2950                         }
2951
2952                         # give headline the correct <h#> tag
2953                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2954
2955                         $headlineCount++;
2956                         if( !$istemplate )
2957                                 $sectionCount++;
2958                 }
2959
2960                 if( $doShowToc ) {
2961                         $toc .= $sk->tocUnindent( $toclevel - 1 );
2962                         $toc = $sk->tocList( $toc );
2963                 }
2964
2965                 # split up and insert constructed headlines
2966
2967                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2968                 $i = 0;
2969
2970                 foreach( $blocks as $block ) {
2971                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2972                                 # This is the [edit] link that appears for the top block of text when
2973                                 # section editing is enabled
2974
2975                                 # Disabled because it broke block formatting
2976                                 # For example, a bullet point in the top line
2977                                 # $full .= $sk->editSectionLink(0);
2978                         }
2979                         $full .= $block;
2980                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2981                         # Top anchor now in skin
2982                                 $full = $full.$toc;
2983                         }
2984
2985                         if( !empty( $head[$i] ) ) {
2986                                 $full .= $head[$i];
2987                         }
2988                         $i++;
2989                 }
2990                 if($forceTocHere) {
2991                         $mw =& MagicWord::get( MAG_TOC );
2992                         return $mw->replace( $toc, $full );
2993                 } else {
2994                         return $full;
2995                 }
2996         }
2997
2998         /**
2999          * Return an HTML link for the "ISBN 123456" text
3000          * @access private
3001          */
3002         function magicISBN( $text ) {
3003                 $fname = 'Parser::magicISBN';
3004                 wfProfileIn( $fname );
3005
3006                 $a = split( 'ISBN ', ' '.$text );
3007                 if ( count ( $a ) < 2 ) {
3008                         wfProfileOut( $fname );
3009                         return $text;
3010                 }
3011                 $text = substr( array_shift( $a ), 1);
3012                 $valid = '0123456789-Xx';
3013
3014                 foreach ( $a as $x ) {
3015                         $isbn = $blank = '' ;
3016                         while ( ' ' == $x{0} ) {
3017                                 $blank .= ' ';
3018                                 $x = substr( $x, 1 );
3019                         }
3020                         if ( $x == '' ) { # blank isbn
3021                                 $text .= "ISBN $blank";
3022                                 continue;
3023                         }
3024                         while ( strstr( $valid, $x{0} ) != false ) {
3025                                 $isbn .= $x{0};
3026                                 $x = substr( $x, 1 );
3027                         }
3028                         $num = str_replace( '-', '', $isbn );
3029                         $num = str_replace( ' ', '', $num );
3030                         $num = str_replace( 'x', 'X', $num );
3031
3032                         if ( '' == $num ) {
3033                                 $text .= "ISBN $blank$x";
3034                         } else {
3035                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
3036                                 $text .= '<a href="' .
3037                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
3038                                         "\" class=\"internal\">ISBN $isbn</a>";
3039                                 $text .= $x;
3040                         }
3041                 }
3042                 wfProfileOut( $fname );
3043                 return $text;
3044         }
3045
3046         /**
3047          * Return an HTML link for the "RFC 1234" text
3048          *
3049          * @access private
3050          * @param string $text     Text to be processed
3051          * @param string $keyword  Magic keyword to use (default RFC)
3052          * @param string $urlmsg   Interface message to use (default rfcurl)
3053          * @return string
3054          */
3055         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
3056
3057                 $valid = '0123456789';
3058                 $internal = false;
3059
3060                 $a = split( $keyword, ' '.$text );
3061                 if ( count ( $a ) < 2 ) {
3062                         return $text;
3063                 }
3064                 $text = substr( array_shift( $a ), 1);
3065
3066                 /* Check if keyword is preceed by [[.
3067                  * This test is made here cause of the array_shift above
3068                  * that prevent the test to be done in the foreach.
3069                  */
3070                 if ( substr( $text, -2 ) == '[[' ) {
3071                         $internal = true;
3072                 }
3073
3074                 foreach ( $a as $x ) {
3075                         /* token might be empty if we have RFC RFC 1234 */
3076                         if ( $x=='' ) {
3077                                 $text.=$keyword;
3078                                 continue;
3079                                 }
3080
3081                         $id = $blank = '' ;
3082
3083                         /** remove and save whitespaces in $blank */
3084                         while ( $x{0} == ' ' ) {
3085                                 $blank .= ' ';
3086                                 $x = substr( $x, 1 );
3087                         }
3088
3089                         /** remove and save the rfc number in $id */
3090                         while ( strstr( $valid, $x{0} ) != false ) {
3091                                 $id .= $x{0};
3092                                 $x = substr( $x, 1 );
3093                         }
3094
3095                         if ( $id == '' ) {
3096                                 /* call back stripped spaces*/
3097                                 $text .= $keyword.$blank.$x;
3098                         } elseif( $internal ) {
3099                                 /* normal link */
3100                                 $text .= $keyword.$id.$x;
3101                         } else {
3102                                 /* build the external link*/
3103                                 $url = wfMsg( $urlmsg, $id);
3104                                 $sk =& $this->mOptions->getSkin();
3105                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
3106                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
3107                         }
3108
3109                         /* Check if the next RFC keyword is preceed by [[ */
3110                         $internal = ( substr($x,-2) == '[[' );
3111                 }
3112                 return $text;
3113         }
3114
3115         /**
3116          * Transform wiki markup when saving a page by doing \r\n -> \n
3117          * conversion, substitting signatures, {{subst:}} templates, etc.
3118          *
3119          * @param string $text the text to transform
3120          * @param Title &$title the Title object for the current article
3121          * @param User &$user the User object describing the current user
3122          * @param ParserOptions $options parsing options
3123          * @param bool $clearState whether to clear the parser state first
3124          * @return string the altered wiki markup
3125          * @access public
3126          */
3127         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
3128                 $this->mOptions = $options;
3129                 $this->mTitle =& $title;
3130                 $this->mOutputType = OT_WIKI;
3131
3132                 if ( $clearState ) {
3133                         $this->clearState();
3134                 }
3135
3136                 $stripState = false;
3137                 $pairs = array(
3138                         "\r\n" => "\n",
3139                 );
3140                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
3141                 $text = $this->strip( $text, $stripState, true );
3142                 $text = $this->pstPass2( $text, $user );
3143                 $text = $this->unstrip( $text, $stripState );
3144                 $text = $this->unstripNoWiki( $text, $stripState );
3145                 return $text;
3146         }
3147
3148         /**
3149          * Pre-save transform helper function
3150          * @access private
3151          */
3152         function pstPass2( $text, &$user ) {
3153                 global $wgContLang, $wgLocaltimezone;
3154
3155                 /* Note: This is the timestamp saved as hardcoded wikitext to
3156                  * the database, we use $wgContLang here in order to give
3157                  * everyone the same signiture and use the default one rather
3158                  * than the one selected in each users preferences.
3159                  */
3160                 if ( isset( $wgLocaltimezone ) ) {
3161                         $oldtz = getenv( 'TZ' );
3162                         putenv( 'TZ='.$wgLocaltimezone );
3163                 }
3164                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false, false) .
3165                   ' (' . date( 'T' ) . ')';
3166                 if ( isset( $wgLocaltimezone ) ) {
3167                         putenv( 'TZ='.$oldtz );
3168                 }
3169
3170                 # Variable replacement
3171                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
3172                 $text = $this->replaceVariables( $text );
3173
3174                 # Signatures
3175                 $sigText = $this->getUserSig( $user );
3176                 $text = strtr( $text, array(
3177                         '~~~~~' => $d,
3178                         '~~~~' => "$sigText $d",
3179                         '~~~' => $sigText
3180                 ) );
3181
3182                 # Context links: [[|name]] and [[name (context)|]]
3183                 #
3184                 global $wgLegalTitleChars;
3185                 $tc = "[$wgLegalTitleChars]";
3186                 $np = str_replace( array( '(', ')' ), array( '', '' ), $tc ); # No parens
3187
3188                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
3189                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
3190
3191                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
3192                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
3193                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
3194                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
3195                 $context = '';
3196                 $t = $this->mTitle->getText();
3197                 if ( preg_match( $conpat, $t, $m ) ) {
3198                         $context = $m[2];
3199                 }
3200                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
3201                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
3202                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
3203
3204                 if ( '' == $context ) {
3205                         $text = preg_replace( $p2, '[[\\1]]', $text );
3206                 } else {
3207                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
3208                 }
3209
3210                 # Trim trailing whitespace
3211                 # MAG_END (__END__) tag allows for trailing
3212                 # whitespace to be deliberately included
3213                 $text = rtrim( $text );
3214                 $mw =& MagicWord::get( MAG_END );
3215                 $mw->matchAndRemove( $text );
3216
3217                 return $text;
3218         }
3219
3220         /**
3221          * Fetch the user's signature text, if any, and normalize to
3222          * validated, ready-to-insert wikitext.
3223          *
3224          * @param User $user
3225          * @return string
3226          * @access private
3227          */
3228         function getUserSig( &$user ) {
3229                 global $wgContLang;
3230
3231                 $username = $user->getName();
3232                 $nickname = $user->getOption( 'nickname' );
3233                 $nickname = $nickname === '' ? $username : $nickname;
3234
3235                 if( $user->getBoolOption( 'fancysig' ) !== false ) {
3236                         # Sig. might contain markup; validate this
3237                         if( $this->validateSig( $nickname ) !== false ) {
3238                                 # Validated; clean up (if needed) and return it
3239                                 return( $this->cleanSig( $nickname ) );
3240                         } else {
3241                                 # Failed to validate; fall back to the default
3242                                 $nickname = $username;
3243                                 wfDebug( "Parser::getUserSig: $username has bad XML tags in signature.\n" );
3244                         }
3245                 }
3246
3247                 # If we're still here, make it a link to the user page
3248                 $userpage = $user->getUserPage();
3249                 return( '[[' . $userpage->getPrefixedText() . '|' . wfEscapeWikiText( $nickname ) . ']]' );
3250         }
3251
3252         /**
3253          * Check that the user's signature contains no bad XML
3254          *
3255          * @param string $text
3256          * @return mixed An expanded string, or false if invalid.
3257          */
3258         function validateSig( $text ) {
3259                 return( wfIsWellFormedXmlFragment( $text ) ? $text : false );
3260         }
3261
3262         /**
3263          * Clean up signature text
3264          *
3265          * 1) Force transclusions to be substituted
3266          * 2) Strip ~~~, ~~~~ and ~~~~~ out of signatures
3267          *
3268          * @static
3269          * @param string $text
3270          * @return string Text
3271          */
3272         function cleanSig( $text ) {
3273
3274                 $mw = MagicWord::get( MAG_SUBST );
3275                 $substre = $mw->getBaseRegex();
3276                 $subst = $mw->getSynonym( 0 );
3277                 $i = $mw->getRegexCase();
3278
3279                 $text = preg_replace(
3280                         "/
3281                                 \{\{
3282                                         (?!
3283                                                 (?:
3284                                                         $substre
3285                                                 )
3286                                         )
3287                         /x$i",
3288                         '{{' . $subst,
3289                         $text
3290                 );
3291
3292                 $text = preg_replace( '/~{3,5}/', '', $text );
3293                 $text = $this->replaceVariables( $text );
3294
3295                 return $text;
3296         }
3297
3298         /**
3299          * Set up some variables which are usually set up in parse()
3300          * so that an external function can call some class members with confidence
3301          * @access public
3302          */
3303         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
3304                 $this->mTitle =& $title;
3305                 $this->mOptions = $options;
3306                 $this->mOutputType = $outputType;
3307                 if ( $clearState ) {
3308                         $this->clearState();
3309                 }
3310         }
3311
3312         /**
3313          * Transform a MediaWiki message by replacing magic variables.
3314          *
3315          * @param string $text the text to transform
3316          * @param ParserOptions $options  options
3317          * @return string the text with variables substituted
3318          * @access public
3319          */
3320         function transformMsg( $text, $options ) {
3321                 global $wgTitle;
3322                 static $executing = false;
3323
3324                 # Guard against infinite recursion
3325                 if ( $executing ) {
3326                         return $text;
3327                 }
3328                 $executing = true;
3329
3330                 $this->mTitle = $wgTitle;
3331                 $this->mOptions = $options;
3332                 $this->mOutputType = OT_MSG;
3333                 $this->clearState();
3334                 $text = $this->replaceVariables( $text );
3335
3336                 $executing = false;
3337                 return $text;
3338         }
3339
3340         /**
3341          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
3342          * Callback will be called with the text within
3343          * Transform and return the text within
3344          *
3345          * @access public
3346          *
3347          * @param mixed $tag The tag to use, e.g. 'hook' for <hook>
3348          * @param mixed $callback The callback function (and object) to use for the tag
3349          *
3350          * @return The old value of the mTagHooks array associated with the hook
3351          */
3352         function setHook( $tag, $callback ) {
3353                 $oldVal = @$this->mTagHooks[$tag];
3354                 $this->mTagHooks[$tag] = $callback;
3355
3356                 return $oldVal;
3357         }
3358
3359         /**
3360          * Replace <!--LINK--> link placeholders with actual links, in the buffer
3361          * Placeholders created in Skin::makeLinkObj()
3362          * Returns an array of links found, indexed by PDBK:
3363          *  0 - broken
3364          *  1 - normal link
3365          *  2 - stub
3366          * $options is a bit field, RLH_FOR_UPDATE to select for update
3367          */
3368         function replaceLinkHolders( &$text, $options = 0 ) {
3369                 global $wgUser;
3370                 global $wgOutputReplace;
3371
3372                 $fname = 'Parser::replaceLinkHolders';
3373                 wfProfileIn( $fname );
3374
3375                 $pdbks = array();
3376                 $colours = array();
3377                 $sk =& $this->mOptions->getSkin();
3378                 $linkCache =& LinkCache::singleton();
3379
3380                 if ( !empty( $this->mLinkHolders['namespaces'] ) ) {
3381                         wfProfileIn( $fname.'-check' );
3382                         $dbr =& wfGetDB( DB_SLAVE );
3383                         $page = $dbr->tableName( 'page' );
3384                         $threshold = $wgUser->getOption('stubthreshold');
3385
3386                         # Sort by namespace
3387                         asort( $this->mLinkHolders['namespaces'] );
3388
3389                         # Generate query
3390                         $query = false;
3391                         foreach ( $this->mLinkHolders['namespaces'] as $key => $ns ) {
3392                                 # Make title object
3393                                 $title = $this->mLinkHolders['titles'][$key];
3394
3395                                 # Skip invalid entries.
3396                                 # Result will be ugly, but prevents crash.
3397                                 if ( is_null( $title ) ) {
3398                                         continue;
3399                                 }
3400                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
3401
3402                                 # Check if it's a static known link, e.g. interwiki
3403                                 if ( $title->isAlwaysKnown() ) {
3404                                         $colours[$pdbk] = 1;
3405                                 } elseif ( ( $id = $linkCache->getGoodLinkID( $pdbk ) ) != 0 ) {
3406                                         $colours[$pdbk] = 1;
3407                                         $this->mOutput->addLink( $title, $id );
3408                                 } elseif ( $linkCache->isBadLink( $pdbk ) ) {
3409                                         $colours[$pdbk] = 0;
3410                                 } else {
3411                                         # Not in the link cache, add it to the query
3412                                         if ( !isset( $current ) ) {
3413                                                 $current = $ns;
3414                                                 $query =  "SELECT page_id, page_namespace, page_title";
3415                                                 if ( $threshold > 0 ) {
3416                                                         $query .= ', page_len, page_is_redirect';
3417                                                 }
3418                                                 $query .= " FROM $page WHERE (page_namespace=$ns AND page_title IN(";
3419                                         } elseif ( $current != $ns ) {
3420                                                 $current = $ns;
3421                                                 $query .= ")) OR (page_namespace=$ns AND page_title IN(";
3422                                         } else {
3423                                                 $query .= ', ';
3424                                         }
3425
3426                                         $query .= $dbr->addQuotes( $this->mLinkHolders['dbkeys'][$key] );
3427                                 }
3428                         }
3429                         if ( $query ) {
3430                                 $query .= '))';
3431                                 if ( $options & RLH_FOR_UPDATE ) {
3432                                         $query .= ' FOR UPDATE';
3433                                 }
3434
3435                                 $res = $dbr->query( $query, $fname );
3436
3437                                 # Fetch data and form into an associative array
3438                                 # non-existent = broken
3439                                 # 1 = known
3440                                 # 2 = stub
3441                                 while ( $s = $dbr->fetchObject($res) ) {
3442                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
3443                                         $pdbk = $title->getPrefixedDBkey();
3444                                         $linkCache->addGoodLinkObj( $s->page_id, $title );
3445                                         $this->mOutput->addLink( $title, $s->page_id );
3446
3447                                         if ( $threshold >  0 ) {
3448                                                 $size = $s->page_len;
3449                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
3450                                                         $colours[$pdbk] = 1;
3451                                                 } else {
3452                                                         $colours[$pdbk] = 2;
3453                                                 }
3454                                         } else {
3455                                                 $colours[$pdbk] = 1;
3456                                         }
3457                                 }
3458                         }
3459                         wfProfileOut( $fname.'-check' );
3460
3461                         # Construct search and replace arrays
3462                         wfProfileIn( $fname.'-construct' );
3463                         $wgOutputReplace = array();
3464                         foreach ( $this->mLinkHolders['namespaces'] as $key => $ns ) {
3465                                 $pdbk = $pdbks[$key];
3466                                 $searchkey = "<!--LINK $key-->";
3467                                 $title = $this->mLinkHolders['titles'][$key];
3468                                 if ( empty( $colours[$pdbk] ) ) {
3469                                         $linkCache->addBadLinkObj( $title );
3470                                         $colours[$pdbk] = 0;
3471                                         $this->mOutput->addLink( $title, 0 );
3472                                         $wgOutputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3473                                                                         $this->mLinkHolders['texts'][$key],
3474                                                                         $this->mLinkHolders['queries'][$key] );
3475                                 } elseif ( $colours[$pdbk] == 1 ) {
3476                                         $wgOutputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3477                                                                         $this->mLinkHolders['texts'][$key],
3478                                                                         $this->mLinkHolders['queries'][$key] );
3479                                 } elseif ( $colours[$pdbk] == 2 ) {
3480                                         $wgOutputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3481                                                                         $this->mLinkHolders['texts'][$key],
3482                                                                         $this->mLinkHolders['queries'][$key] );
3483                                 }
3484                         }
3485                         wfProfileOut( $fname.'-construct' );
3486
3487                         # Do the thing
3488                         wfProfileIn( $fname.'-replace' );
3489
3490                         $text = preg_replace_callback(
3491                                 '/(<!--LINK .*?-->)/',
3492                                 "wfOutputReplaceMatches",
3493                                 $text);
3494
3495                         wfProfileOut( $fname.'-replace' );
3496                 }
3497
3498                 # Now process interwiki link holders
3499                 # This is quite a bit simpler than internal links
3500                 if ( !empty( $this->mInterwikiLinkHolders['texts'] ) ) {
3501                         wfProfileIn( $fname.'-interwiki' );
3502                         # Make interwiki link HTML
3503                         $wgOutputReplace = array();
3504                         foreach( $this->mInterwikiLinkHolders['texts'] as $key => $link ) {
3505                                 $title = $this->mInterwikiLinkHolders['titles'][$key];
3506                                 $wgOutputReplace[$key] = $sk->makeLinkObj( $title, $link );
3507                         }
3508
3509                         $text = preg_replace_callback(
3510                                 '/<!--IWLINK (.*?)-->/',
3511                                 "wfOutputReplaceMatches",
3512                                 $text );
3513                         wfProfileOut( $fname.'-interwiki' );
3514                 }
3515
3516                 wfProfileOut( $fname );
3517                 return $colours;
3518         }
3519
3520         /**
3521          * Replace <!--LINK--> link placeholders with plain text of links
3522          * (not HTML-formatted).
3523          * @param string $text
3524          * @return string
3525          */
3526         function replaceLinkHoldersText( $text ) {
3527                 global $wgUser;
3528                 global $wgOutputReplace;
3529
3530                 $fname = 'Parser::replaceLinkHoldersText';
3531                 wfProfileIn( $fname );
3532
3533                 $text = preg_replace_callback(
3534                         '/<!--(LINK|IWLINK) (.*?)-->/',
3535                         array( &$this, 'replaceLinkHoldersTextCallback' ),
3536                         $text );
3537
3538                 wfProfileOut( $fname );
3539                 return $text;
3540         }
3541
3542         /**
3543          * @param array $matches
3544          * @return string
3545          * @access private
3546          */
3547         function replaceLinkHoldersTextCallback( $matches ) {
3548                 $type = $matches[1];
3549                 $key  = $matches[2];
3550                 if( $type == 'LINK' ) {
3551                         if( isset( $this->mLinkHolders['texts'][$key] ) ) {
3552                                 return $this->mLinkHolders['texts'][$key];
3553                         }
3554                 } elseif( $type == 'IWLINK' ) {
3555                         if( isset( $this->mInterwikiLinkHolders['texts'][$key] ) ) {
3556                                 return $this->mInterwikiLinkHolders['texts'][$key];
3557                         }
3558                 }
3559                 return $matches[0];
3560         }
3561
3562         /**
3563          * Renders an image gallery from a text with one line per image.
3564          * text labels may be given by using |-style alternative text. E.g.
3565          *   Image:one.jpg|The number "1"
3566          *   Image:tree.jpg|A tree
3567          * given as text will return the HTML of a gallery with two images,
3568          * labeled 'The number "1"' and
3569          * 'A tree'.
3570          */
3571         function renderImageGallery( $text ) {
3572                 # Setup the parser
3573                 $parserOptions = new ParserOptions;
3574                 $localParser = new Parser();
3575
3576                 $ig = new ImageGallery();
3577                 $ig->setShowBytes( false );
3578                 $ig->setShowFilename( false );
3579                 $lines = explode( "\n", $text );
3580
3581                 foreach ( $lines as $line ) {
3582                         # match lines like these:
3583                         # Image:someimage.jpg|This is some image
3584                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3585                         # Skip empty lines
3586                         if ( count( $matches ) == 0 ) {
3587                                 continue;
3588                         }
3589                         $nt =& Title::newFromText( $matches[1] );
3590                         if( is_null( $nt ) ) {
3591                                 # Bogus title. Ignore these so we don't bomb out later.
3592                                 continue;
3593                         }
3594                         if ( isset( $matches[3] ) ) {
3595                                 $label = $matches[3];
3596                         } else {
3597                                 $label = '';
3598                         }
3599
3600                         $pout = $localParser->parse( $label , $this->mTitle, $parserOptions );
3601                         $html = $pout->getText();
3602
3603                         $ig->add( new Image( $nt ), $html );
3604                         $this->mOutput->addImage( $nt->getDBkey() );
3605                 }
3606                 return $ig->toHTML();
3607         }
3608
3609         /**
3610          * Parse image options text and use it to make an image
3611          */
3612         function makeImage( &$nt, $options ) {
3613                 global $wgContLang, $wgUseImageResize, $wgUser;
3614
3615                 $align = '';
3616
3617                 # Check if the options text is of the form "options|alt text"
3618                 # Options are:
3619                 #  * thumbnail          make a thumbnail with enlarge-icon and caption, alignment depends on lang
3620                 #  * left               no resizing, just left align. label is used for alt= only
3621                 #  * right              same, but right aligned
3622                 #  * none               same, but not aligned
3623                 #  * ___px              scale to ___ pixels width, no aligning. e.g. use in taxobox
3624                 #  * center             center the image
3625                 #  * framed             Keep original image size, no magnify-button.
3626
3627                 $part = explode( '|', $options);
3628
3629                 $mwThumb  =& MagicWord::get( MAG_IMG_THUMBNAIL );
3630                 $mwManualThumb =& MagicWord::get( MAG_IMG_MANUALTHUMB );
3631                 $mwLeft   =& MagicWord::get( MAG_IMG_LEFT );
3632                 $mwRight  =& MagicWord::get( MAG_IMG_RIGHT );
3633                 $mwNone   =& MagicWord::get( MAG_IMG_NONE );
3634                 $mwWidth  =& MagicWord::get( MAG_IMG_WIDTH );
3635                 $mwCenter =& MagicWord::get( MAG_IMG_CENTER );
3636                 $mwFramed =& MagicWord::get( MAG_IMG_FRAMED );
3637                 $caption = '';
3638
3639                 $width = $height = $framed = $thumb = false;
3640                 $manual_thumb = '' ;
3641
3642                 foreach( $part as $key => $val ) {
3643                         if ( $wgUseImageResize && ! is_null( $mwThumb->matchVariableStartToEnd($val) ) ) {
3644                                 $thumb=true;
3645                         } elseif ( ! is_null( $match = $mwManualThumb->matchVariableStartToEnd($val) ) ) {
3646                                 # use manually specified thumbnail
3647                                 $thumb=true;
3648                                 $manual_thumb = $match;
3649                         } elseif ( ! is_null( $mwRight->matchVariableStartToEnd($val) ) ) {
3650                                 # remember to set an alignment, don't render immediately
3651                                 $align = 'right';
3652                         } elseif ( ! is_null( $mwLeft->matchVariableStartToEnd($val) ) ) {
3653                                 # remember to set an alignment, don't render immediately
3654                                 $align = 'left';
3655                         } elseif ( ! is_null( $mwCenter->matchVariableStartToEnd($val) ) ) {
3656                                 # remember to set an alignment, don't render immediately
3657                                 $align = 'center';
3658                         } elseif ( ! is_null( $mwNone->matchVariableStartToEnd($val) ) ) {
3659                                 # remember to set an alignment, don't render immediately
3660                                 $align = 'none';
3661                         } elseif ( $wgUseImageResize && ! is_null( $match = $mwWidth->matchVariableStartToEnd($val) ) ) {
3662                                 wfDebug( "MAG_IMG_WIDTH match: $match\n" );
3663                                 # $match is the image width in pixels
3664                                 if ( preg_match( '/^([0-9]*)x([0-9]*)$/', $match, $m ) ) {
3665                                         $width = intval( $m[1] );
3666                                         $height = intval( $m[2] );
3667                                 } else {
3668                                         $width = intval($match);
3669                                 }
3670                         } elseif ( ! is_null( $mwFramed->matchVariableStartToEnd($val) ) ) {
3671                                 $framed=true;
3672                         } else {
3673                                 $caption = $val;
3674                         }
3675                 }
3676                 # Strip bad stuff out of the alt text
3677                 $alt = $this->replaceLinkHoldersText( $caption );
3678                 $alt = Sanitizer::stripAllTags( $alt );
3679
3680                 # Linker does the rest
3681                 $sk =& $this->mOptions->getSkin();
3682                 return $sk->makeImageLinkObj( $nt, $caption, $alt, $align, $width, $height, $framed, $thumb, $manual_thumb );
3683         }
3684
3685         /**
3686          * Set a flag in the output object indicating that the content is dynamic and
3687          * shouldn't be cached.
3688          */
3689         function disableCache() {
3690                 $this->mOutput->mCacheTime = -1;
3691         }
3692
3693         /**#@+
3694          * Callback from the Sanitizer for expanding items found in HTML attribute
3695          * values, so they can be safely tested and escaped.
3696          * @param string $text
3697          * @param array $args
3698          * @return string
3699          * @access private
3700          */
3701         function attributeStripCallback( &$text, $args ) {
3702                 $text = $this->replaceVariables( $text, $args );
3703                 $text = $this->unstripForHTML( $text );
3704                 return $text;
3705         }
3706
3707         function unstripForHTML( $text ) {
3708                 $text = $this->unstrip( $text, $this->mStripState );
3709                 $text = $this->unstripNoWiki( $text, $this->mStripState );
3710                 return $text;
3711         }
3712         /**#@-*/
3713
3714         /**#@+
3715          * Accessor/mutator
3716          */
3717         function Title( $x = NULL ) { return wfSetVar( $this->mTitle, $x ); }
3718         function Options( $x = NULL ) { return wfSetVar( $this->mOptions, $x ); }
3719         function OutputType( $x = NULL ) { return wfSetVar( $this->mOutputType, $x ); }
3720         /**#@-*/
3721
3722         /**#@+
3723          * Accessor
3724          */
3725         function getTags() { return array_keys( $this->mTagHooks ); }
3726         /**#@-*/
3727 }
3728
3729 /**
3730  * @todo document
3731  * @package MediaWiki
3732  */
3733 class ParserOutput
3734 {
3735         var $mText,             # The output text
3736                 $mLanguageLinks,    # List of the full text of language links, in the order they appear
3737                 $mCategories,       # Map of category names to sort keys
3738                 $mContainsOldMagic, # Boolean variable indicating if the input contained variables like {{CURRENTDAY}}
3739                 $mCacheTime,        # Timestamp on this article, or -1 for uncacheable. Used in ParserCache.
3740                 $mVersion,          # Compatibility check
3741                 $mTitleText,        # title text of the chosen language variant
3742                 $mLinks,            # 2-D map of NS/DBK to ID for the links in the document. ID=zero for broken.
3743                 $mTemplates,        # 2-D map of NS/DBK to ID for the template references. ID=zero for broken.
3744                 $mImages;           # DB keys of the images used, in the array key only
3745
3746         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3747                 $containsOldMagic = false, $titletext = '' )
3748         {
3749                 $this->mText = $text;
3750                 $this->mLanguageLinks = $languageLinks;
3751                 $this->mCategories = $categoryLinks;
3752                 $this->mContainsOldMagic = $containsOldMagic;
3753                 $this->mCacheTime = '';
3754                 $this->mVersion = MW_PARSER_VERSION;
3755                 $this->mTitleText = $titletext;
3756                 $this->mLinks = array();
3757                 $this->mTemplates = array();
3758                 $this->mImages = array();
3759         }
3760
3761         function getText()                   { return $this->mText; }
3762         function getLanguageLinks()          { return $this->mLanguageLinks; }
3763         function getCategoryLinks()          { return array_keys( $this->mCategories ); }
3764         function &getCategories()            { return $this->mCategories; }
3765         function getCacheTime()              { return $this->mCacheTime; }
3766         function getTitleText()              { return $this->mTitleText; }
3767         function &getLinks()                 { return $this->mLinks; }
3768         function &getTemplates()             { return $this->mTemplates; }
3769         function &getImages()                { return $this->mImages; }
3770
3771         function containsOldMagic()          { return $this->mContainsOldMagic; }
3772         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3773         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3774         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategories, $cl ); }
3775         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3776         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3777         function setTitleText( $t )          { return wfSetVar ($this->mTitleText, $t); }
3778
3779         function addCategory( $c, $sort )    { $this->mCategories[$c] = $sort; }
3780         function addImage( $name )           { $this->mImages[$name] = 1; }
3781         function addLanguageLink( $t )       { $this->mLanguageLinks[] = $t; }
3782
3783         function addLink( $title, $id ) {
3784                 $ns = $title->getNamespace();
3785                 $dbk = $title->getDBkey();
3786                 if ( !isset( $this->mLinks[$ns] ) ) {
3787                         $this->mLinks[$ns] = array();
3788                 }
3789                 $this->mLinks[$ns][$dbk] = $id;
3790         }
3791
3792         function addTemplate( $title, $id ) {
3793                 $ns = $title->getNamespace();
3794                 $dbk = $title->getDBkey();
3795                 if ( !isset( $this->mTemplates[$ns] ) ) {
3796                         $this->mTemplates[$ns] = array();
3797                 }
3798                 $this->mTemplates[$ns][$dbk] = $id;
3799         }
3800
3801         /**
3802          * @deprecated
3803          */
3804         /*
3805         function merge( $other ) {
3806                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3807                 $this->mCategories = array_merge( $this->mCategories, $this->mLanguageLinks );
3808                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3809         }*/
3810
3811         /**
3812          * Return true if this cached output object predates the global or
3813          * per-article cache invalidation timestamps, or if it comes from
3814          * an incompatible older version.
3815          *
3816          * @param string $touched the affected article's last touched timestamp
3817          * @return bool
3818          * @access public
3819          */
3820         function expired( $touched ) {
3821                 global $wgCacheEpoch;
3822                 return $this->getCacheTime() == -1 || // parser says it's uncacheable
3823                        $this->getCacheTime() < $touched ||
3824                        $this->getCacheTime() <= $wgCacheEpoch ||
3825                        !isset( $this->mVersion ) ||
3826                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
3827         }
3828 }
3829
3830 /**
3831  * Set options of the Parser
3832  * @todo document
3833  * @package MediaWiki
3834  */
3835 class ParserOptions
3836 {
3837         # All variables are private
3838         var $mUseTeX;                    # Use texvc to expand <math> tags
3839         var $mUseDynamicDates;           # Use DateFormatter to format dates
3840         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
3841         var $mAllowExternalImages;       # Allow external images inline
3842         var $mAllowExternalImagesFrom;   # If not, any exception?
3843         var $mSkin;                      # Reference to the preferred skin
3844         var $mDateFormat;                # Date format index
3845         var $mEditSection;               # Create "edit section" links
3846         var $mNumberHeadings;            # Automatically number headings
3847         var $mAllowSpecialInclusion;     # Allow inclusion of special pages
3848
3849         function getUseTeX()                        { return $this->mUseTeX; }
3850         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
3851         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
3852         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
3853         function getAllowExternalImagesFrom()       { return $this->mAllowExternalImagesFrom; }
3854         function &getSkin()                         { return $this->mSkin; }
3855         function getDateFormat()                    { return $this->mDateFormat; }
3856         function getEditSection()                   { return $this->mEditSection; }
3857         function getNumberHeadings()                { return $this->mNumberHeadings; }
3858         function getAllowSpecialInclusion()         { return $this->mAllowSpecialInclusion; }
3859
3860
3861         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
3862         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
3863         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
3864         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
3865         function setAllowExternalImagesFrom( $x )   { return wfSetVar( $this->mAllowExternalImagesFrom, $x ); }
3866         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
3867         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
3868         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
3869         function setAllowSpecialInclusion( $x )     { return wfSetVar( $this->mAllowSpecialInclusion, $x ); }
3870
3871         function setSkin( &$x ) { $this->mSkin =& $x; }
3872
3873         function ParserOptions() {
3874                 global $wgUser;
3875                 $this->initialiseFromUser( $wgUser );
3876         }
3877
3878         /**
3879          * Get parser options
3880          * @static
3881          */
3882         function newFromUser( &$user ) {
3883                 $popts = new ParserOptions;
3884                 $popts->initialiseFromUser( $user );
3885                 return $popts;
3886         }
3887
3888         /** Get user options */
3889         function initialiseFromUser( &$userInput ) {
3890                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages,
3891                        $wgAllowExternalImagesFrom, $wgAllowSpecialInclusion;
3892                 $fname = 'ParserOptions::initialiseFromUser';
3893                 wfProfileIn( $fname );
3894                 if ( !$userInput ) {
3895                         $user = new User;
3896                         $user->setLoaded( true );
3897                 } else {
3898                         $user =& $userInput;
3899                 }
3900
3901                 $this->mUseTeX = $wgUseTeX;
3902                 $this->mUseDynamicDates = $wgUseDynamicDates;
3903                 $this->mInterwikiMagic = $wgInterwikiMagic;
3904                 $this->mAllowExternalImages = $wgAllowExternalImages;
3905                 $this->mAllowExternalImagesFrom = $wgAllowExternalImagesFrom;
3906                 wfProfileIn( $fname.'-skin' );
3907                 $this->mSkin =& $user->getSkin();
3908                 wfProfileOut( $fname.'-skin' );
3909                 $this->mDateFormat = $user->getOption( 'date' );
3910                 $this->mEditSection = true;
3911                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
3912                 $this->mAllowSpecialInclusion = $wgAllowSpecialInclusion;
3913                 wfProfileOut( $fname );
3914         }
3915 }
3916
3917 /**
3918  * Callback function used by Parser::replaceLinkHolders()
3919  * to substitute link placeholders.
3920  */
3921 function &wfOutputReplaceMatches( $matches ) {
3922         global $wgOutputReplace;
3923         return $wgOutputReplace[$matches[1]];
3924 }
3925
3926 /**
3927  * Return the total number of articles
3928  */
3929 function wfNumberOfArticles() {
3930         global $wgNumberOfArticles;
3931
3932         wfLoadSiteStats();
3933         return $wgNumberOfArticles;
3934 }
3935
3936 /**
3937  * Return the number of files
3938  */
3939 function wfNumberOfFiles() {
3940         $fname = 'Parser::wfNumberOfFiles';
3941
3942         wfProfileIn( $fname );
3943         $dbr =& wfGetDB( DB_SLAVE );
3944         $res = $dbr->selectField('image', 'COUNT(*)', array(), $fname );
3945         wfProfileOut( $fname );
3946
3947         return $res;
3948 }
3949
3950 /**
3951  * Get various statistics from the database
3952  * @access private
3953  */
3954 function wfLoadSiteStats() {
3955         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
3956         $fname = 'wfLoadSiteStats';
3957
3958         if ( -1 != $wgNumberOfArticles ) return;
3959         $dbr =& wfGetDB( DB_SLAVE );
3960         $s = $dbr->selectRow( 'site_stats',
3961                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
3962                 array( 'ss_row_id' => 1 ), $fname
3963         );
3964
3965         if ( $s === false ) {
3966                 return;
3967         } else {
3968                 $wgTotalViews = $s->ss_total_views;
3969                 $wgTotalEdits = $s->ss_total_edits;
3970                 $wgNumberOfArticles = $s->ss_good_articles;
3971         }
3972 }
3973
3974 /**
3975  * Escape html tags
3976  * Basically replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
3977  *
3978  * @param string $in Text that might contain HTML tags
3979  * @return string Escaped string
3980  */
3981 function wfEscapeHTMLTagsOnly( $in ) {
3982         return str_replace(
3983                 array( '"', '>', '<' ),
3984                 array( '&quot;', '&gt;', '&lt;' ),
3985                 $in );
3986 }
3987
3988 ?>