includes/Parser.php

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