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