includes/Parser.php

   1 <?php
   2 /**
   3  * File for Parser and related classes
   4  *
   5  * @package MediaWiki
   6  * @subpackage Parser
   7  */
   8
   9 /** */
  10 require_once( 'Sanitizer.php' );
  11 require_once( 'HttpFunctions.php' );
  12
  13 /**
  14  * Update this version number when the ParserOutput format
  15  * changes in an incompatible way, so the parser cache
  16  * can automatically discard old data.
  17  */
  18 define( 'MW_PARSER_VERSION', '1.6.0' );
  19
  20 /**
  21  * Variable substitution O(N^2) attack
  22  *
  23  * Without countermeasures, it would be possible to attack the parser by saving
  24  * a page filled with a large number of inclusions of large pages. The size of
  25  * the generated page would be proportional to the square of the input size.
  26  * Hence, we limit the number of inclusions of any given page, thus bringing any
  27  * attack back to O(N).
  28  */
  29
  30 define( 'MAX_INCLUDE_REPEAT', 100 );
  31 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  32
  33 define( 'RLH_FOR_UPDATE', 1 );
  34
  35 # Allowed values for $mOutputType
  36 define( 'OT_HTML', 1 );
  37 define( 'OT_WIKI', 2 );
  38 define( 'OT_MSG' , 3 );
  39
  40 # string parameter for extractTags which will cause it
  41 # to strip HTML comments in addition to regular
  42 # <XML>-style tags. This should not be anything we
  43 # may want to use in wikisyntax
  44 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  45
  46 # Constants needed for external link processing
  47 define( 'HTTP_PROTOCOLS', 'http:\/\/|https:\/\/' );
  48 # Everything except bracket, space, or control characters
  49 define( 'EXT_LINK_URL_CLASS', '[^][<>"\\x00-\\x20\\x7F]' );
  50 # Including space
  51 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  52 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  53 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  54 define( 'EXT_LINK_BRACKETED',  '/\[(\b(' . wfUrlProtocols() . ')'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  55 define( 'EXT_IMAGE_REGEX',
  56         '/^('.HTTP_PROTOCOLS.')'.  # Protocol
  57         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  58         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  59 );
  60
  61 /**
  62  * PHP Parser
  63  *
  64  * Processes wiki markup
  65  *
  66  * <pre>
  67  * There are three main entry points into the Parser class:
  68  * parse()
  69  *   produces HTML output
  70  * preSaveTransform().
  71  *   produces altered wiki markup.
  72  * transformMsg()
  73  *   performs brace substitution on MediaWiki messages
  74  *
  75  * Globals used:
  76  *    objects:   $wgLang, $wgContLang
  77  *
  78  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  79  *
  80  * settings:
  81  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  82  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  83  *  $wgLocaltimezone, $wgAllowSpecialInclusion*
  84  *
  85  *  * only within ParserOptions
  86  * </pre>
  87  *
  88  * @package MediaWiki
  89  */
  90 class Parser
  91 {
  92         /**#@+
  93          * @access private
  94          */
  95         # Persistent:
  96         var $mTagHooks;
  97
  98         # Cleared with clearState():
  99         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
 100         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
 101         var $mInterwikiLinkHolders, $mLinkHolders, $mUniqPrefix;
 102         var $mTemplates,        // cache of already loaded templates, avoids
 103                                 // multiple SQL queries for the same string
 104             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
 105                                 // in this path. Used for loop detection.
 106
 107         # Temporary
 108         # These are variables reset at least once per parse regardless of $clearState
 109         var $mOptions,      // ParserOptions object
 110                 $mTitle,        // Title context, used for self-link rendering and similar things
 111                 $mOutputType,   // Output type, one of the OT_xxx constants
 112                 $mRevisionId;   // ID to display in {{REVISIONID}} tags
 113
 114         /**#@-*/
 115
 116         /**
 117          * Constructor
 118          *
 119          * @access public
 120          */
 121         function Parser() {
 122                 $this->mTagHooks = array();
 123                 $this->clearState();
 124         }
 125
 126         /**
 127          * Clear Parser state
 128          *
 129          * @access private
 130          */
 131         function clearState() {
 132                 $this->mOutput = new ParserOutput;
 133                 $this->mAutonumber = 0;
 134                 $this->mLastSection = '';
 135                 $this->mDTopen = false;
 136                 $this->mVariables = false;
 137                 $this->mIncludeCount = array();
 138                 $this->mStripState = array();
 139                 $this->mArgStack = array();
 140                 $this->mInPre = false;
 141                 $this->mInterwikiLinkHolders = array(
 142                         'texts' => array(),
 143                         'titles' => array()
 144                 );
 145                 $this->mLinkHolders = array(
 146                         'namespaces' => array(),
 147                         'dbkeys' => array(),
 148                         'queries' => array(),
 149                         'texts' => array(),
 150                         'titles' => array()
 151                 );
 152                 $this->mRevisionId = null;
 153                 $this->mUniqPrefix = 'UNIQ' . Parser::getRandomString();
 154
 155                 # Clear these on every parse, bug 4549
 156                 $this->mTemplates = array();
 157                 $this->mTemplatePath = array();
 158
 159                 wfRunHooks( 'ParserClearState', array( &$this ) );
 160         }
 161
 162         /**
 163          * Accessor for mUniqPrefix.
 164          *
 165          * @access public
 166          */
 167         function UniqPrefix() {
 168                 return $this->mUniqPrefix;
 169         }
 170
 171         /**
 172          * Convert wikitext to HTML
 173          * Do not call this function recursively.
 174          *
 175          * @access private
 176          * @param string $text Text we want to parse
 177          * @param Title &$title A title object
 178          * @param array $options
 179          * @param boolean $linestart
 180          * @param boolean $clearState
 181          * @param int $revid number to pass in {{REVISIONID}}
 182          * @return ParserOutput a ParserOutput
 183          */
 184         function parse( $text, &$title, $options, $linestart = true, $clearState = true, $revid = null ) {
 185                 /**
 186                  * First pass--just handle <nowiki> sections, pass the rest off
 187                  * to internalParse() which does all the real work.
 188                  */
 189
 190                 global $wgUseTidy, $wgAlwaysUseTidy, $wgContLang;
 191                 $fname = 'Parser::parse';
 192                 wfProfileIn( $fname );
 193
 194                 if ( $clearState ) {
 195                         $this->clearState();
 196                 }
 197
 198                 $this->mOptions = $options;
 199                 $this->mTitle =& $title;
 200                 $this->mRevisionId = $revid;
 201                 $this->mOutputType = OT_HTML;
 202
 203                 $this->mStripState = NULL;
 204
 205                 //$text = $this->strip( $text, $this->mStripState );
 206                 // VOODOO MAGIC FIX! Sometimes the above segfaults in PHP5.
 207                 $x =& $this->mStripState;
 208
 209                 wfRunHooks( 'ParserBeforeStrip', array( &$this, &$text, &$x ) );
 210                 $text = $this->strip( $text, $x );
 211                 wfRunHooks( 'ParserAfterStrip', array( &$this, &$text, &$x ) );
 212
 213                 # Hook to suspend the parser in this state
 214                 if ( !wfRunHooks( 'ParserBeforeInternalParse', array( &$this, &$text, &$x ) ) ) {
 215                         wfProfileOut( $fname );
 216                         return $text ;
 217                 }
 218
 219                 $text = $this->internalParse( $text );
 220
 221                 $text = $this->unstrip( $text, $this->mStripState );
 222
 223                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 224                 $fixtags = array(
 225                         # french spaces, last one Guillemet-left
 226                         # only if there is something before the space
 227                         '/(.) (?=\\?|:|;|!|\\302\\273)/' => '\\1&nbsp;\\2',
 228                         # french spaces, Guillemet-right
 229                         '/(\\302\\253) /' => '\\1&nbsp;',
 230                         '/<center *>(.*)<\\/center *>/i' => '<div class="center">\\1</div>',
 231                 );
 232                 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 233
 234                 # only once and last
 235                 $text = $this->doBlockLevels( $text, $linestart );
 236
 237                 $this->replaceLinkHolders( $text );
 238
 239                 # the position of the parserConvert() call should not be changed. it
 240                 # assumes that the links are all replaced and the only thing left
 241                 # is the <nowiki> mark.
 242                 # Side-effects: this calls $this->mOutput->setTitleText()
 243                 $text = $wgContLang->parserConvert( $text, $this );
 244
 245                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 246
 247                 wfRunHooks( 'ParserBeforeTidy', array( &$this, &$text ) );
 248
 249                 $text = Sanitizer::normalizeCharReferences( $text );
 250
 251                 if (($wgUseTidy and $this->mOptions->mTidy) or $wgAlwaysUseTidy) {
 252                         $text = Parser::tidy($text);
 253                 } else {
 254                         # attempt to sanitize at least some nesting problems
 255                         # (bug #2702 and quite a few others)
 256                         $tidyregs = array(
 257                                 # ''Something [http://www.cool.com cool''] -->
 258                                 # <i>Something</i><a href="http://www.cool.com"..><i>cool></i></a>
 259                                 '/(<([bi])>)(<([bi])>)?([^<]*)(<\/?a[^<]*>)([^<]*)(<\/\\4>)?(<\/\\2>)/' =>
 260                                 '\\1\\3\\5\\8\\9\\6\\1\\3\\7\\8\\9',
 261                                 # fix up an anchor inside another anchor, only
 262                                 # at least for a single single nested link (bug 3695)
 263                                 '/(<a[^>]+>)([^<]*)(<a[^>]+>[^<]*)<\/a>(.*)<\/a>/' =>
 264                                 '\\1\\2</a>\\3</a>\\1\\4</a>',
 265                                 # fix div inside inline elements- doBlockLevels won't wrap a line which
 266                                 # contains a div, so fix it up here; replace
 267                                 # div with escaped text
 268                                 '/(<([aib]) [^>]+>)([^<]*)(<div([^>]*)>)(.*)(<\/div>)([^<]*)(<\/\\2>)/' =>
 269                                 '\\1\\3&lt;div\\5&gt;\\6&lt;/div&gt;\\8\\9',
 270                                 # remove empty italic or bold tag pairs, some
 271                                 # introduced by rules above
 272                                 '/<([bi])><\/\\1>/' => ''
 273                         );
 274
 275                         $text = preg_replace(
 276                                 array_keys( $tidyregs ),
 277                                 array_values( $tidyregs ),
 278                                 $text );
 279                 }
 280
 281                 wfRunHooks( 'ParserAfterTidy', array( &$this, &$text ) );
 282
 283                 $this->mOutput->setText( $text );
 284                 wfProfileOut( $fname );
 285
 286                 return $this->mOutput;
 287         }
 288
 289         /**
 290          * Get a random string
 291          *
 292          * @access private
 293          * @static
 294          */
 295         function getRandomString() {
 296                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 297         }
 298
 299         function &getTitle() { return $this->mTitle; }
 300         function getOptions() { return $this->mOptions; }
 301
 302         /**
 303          * Replaces all occurrences of <$tag>content</$tag> in the text
 304          * with a random marker and returns the new text. the output parameter
 305          * $content will be an associative array filled with data on the form
 306          * $unique_marker => content.
 307          *
 308          * If $content is already set, the additional entries will be appended
 309          * If $tag is set to STRIP_COMMENTS, the function will extract
 310          * <!-- HTML comments -->
 311          *
 312          * @access private
 313          * @static
 314          */
 315         function extractTagsAndParams($tag, $text, &$content, &$tags, &$params, $uniq_prefix = ''){
 316                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 317                 if ( !$content ) {
 318                         $content = array( );
 319                 }
 320                 $n = 1;
 321                 $stripped = '';
 322
 323                 if ( !$tags ) {
 324                         $tags = array( );
 325                 }
 326
 327                 if ( !$params ) {
 328                         $params = array( );
 329                 }
 330
 331                 if( $tag == STRIP_COMMENTS ) {
 332                         $start = '/<!--()()/';
 333                         $end   = '/-->/';
 334                 } else {
 335                         $start = "/<$tag(\\s+[^\\/>]*|\\s*)(\\/?)>/i";
 336                         $end   = "/<\\/$tag\\s*>/i";
 337                 }
 338
 339                 while ( '' != $text ) {
 340                         $p = preg_split( $start, $text, 2, PREG_SPLIT_DELIM_CAPTURE );
 341                         $stripped .= $p[0];
 342                         if( count( $p ) < 4 ) {
 343                                 break;
 344                         }
 345                         $attributes = $p[1];
 346                         $empty      = $p[2];
 347                         $inside     = $p[3];
 348
 349                         $marker = $rnd . sprintf('%08X', $n++);
 350                         $stripped .= $marker;
 351
 352                         $tags[$marker] = "<$tag$attributes$empty>";
 353                         $params[$marker] = Sanitizer::decodeTagAttributes( $attributes );
 354
 355                         if ( $empty === '/' ) {
 356                                 // Empty element tag, <tag />
 357                                 $content[$marker] = null;
 358                                 $text = $inside;
 359                         } else {
 360                                 $q = preg_split( $end, $inside, 2 );
 361                                 $content[$marker] = $q[0];
 362                                 if( count( $q ) < 2 ) {
 363                                         # No end tag -- let it run out to the end of the text.
 364                                         break;
 365                                 } else {
 366                                         $text = $q[1];
 367                                 }
 368                         }
 369                 }
 370                 return $stripped;
 371         }
 372
 373         /**
 374          * Wrapper function for extractTagsAndParams
 375          * for cases where $tags and $params isn't needed
 376          * i.e. where tags will never have params, like <nowiki>
 377          *
 378          * @access private
 379          * @static
 380          */
 381         function extractTags( $tag, $text, &$content, $uniq_prefix = '' ) {
 382                 $dummy_tags = array();
 383                 $dummy_params = array();
 384
 385                 return Parser::extractTagsAndParams( $tag, $text, $content,
 386                         $dummy_tags, $dummy_params, $uniq_prefix );
 387         }
 388
 389         /**
 390          * Strips and renders nowiki, pre, math, hiero
 391          * If $render is set, performs necessary rendering operations on plugins
 392          * Returns the text, and fills an array with data needed in unstrip()
 393          * If the $state is already a valid strip state, it adds to the state
 394          *
 395          * @param bool $stripcomments when set, HTML comments <!-- like this -->
 396          *  will be stripped in addition to other tags. This is important
 397          *  for section editing, where these comments cause confusion when
 398          *  counting the sections in the wikisource
 399          *
 400          * @access private
 401          */
 402         function strip( $text, &$state, $stripcomments = false ) {
 403                 $render = ($this->mOutputType == OT_HTML);
 404                 $html_content = array();
 405                 $nowiki_content = array();
 406                 $math_content = array();
 407                 $pre_content = array();
 408                 $comment_content = array();
 409                 $ext_content = array();
 410                 $ext_tags = array();
 411                 $ext_params = array();
 412                 $gallery_content = array();
 413
 414                 # Replace any instances of the placeholders
 415                 $uniq_prefix = $this->mUniqPrefix;
 416                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 417
 418                 # html
 419                 global $wgRawHtml;
 420                 if( $wgRawHtml ) {
 421                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 422                         foreach( $html_content as $marker => $content ) {
 423                                 if ($render ) {
 424                                         # Raw and unchecked for validity.
 425                                         $html_content[$marker] = $content;
 426                                 } else {
 427                                         $html_content[$marker] = '<html>'.$content.'</html>';
 428                                 }
 429                         }
 430                 }
 431
 432                 # nowiki
 433                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 434                 foreach( $nowiki_content as $marker => $content ) {
 435                         if( $render ){
 436                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 437                         } else {
 438                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 439                         }
 440                 }
 441
 442                 # math
 443                 if( $this->mOptions->getUseTeX() ) {
 444                         $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 445                         foreach( $math_content as $marker => $content ){
 446                                 if( $render ) {
 447                                         $math_content[$marker] = renderMath( $content );
 448                                 } else {
 449                                         $math_content[$marker] = '<math>'.$content.'</math>';
 450                                 }
 451                         }
 452                 }
 453
 454                 # pre
 455                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 456                 foreach( $pre_content as $marker => $content ){
 457                         if( $render ){
 458                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 459                         } else {
 460                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 461                         }
 462                 }
 463
 464                 # gallery
 465                 $text = Parser::extractTags('gallery', $text, $gallery_content, $uniq_prefix);
 466                 foreach( $gallery_content as $marker => $content ) {
 467                         require_once( 'ImageGallery.php' );
 468                         if ( $render ) {
 469                                 $gallery_content[$marker] = $this->renderImageGallery( $content );
 470                         } else {
 471                                 $gallery_content[$marker] = '<gallery>'.$content.'</gallery>';
 472                         }
 473                 }
 474
 475                 # Comments
 476                 $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 477                 foreach( $comment_content as $marker => $content ){
 478                         $comment_content[$marker] = '<!--'.$content.'-->';
 479                 }
 480
 481                 # Extensions
 482                 foreach ( $this->mTagHooks as $tag => $callback ) {
 483                         $ext_content[$tag] = array();
 484                         $text = Parser::extractTagsAndParams( $tag, $text, $ext_content[$tag],
 485                                 $ext_tags[$tag], $ext_params[$tag], $uniq_prefix );
 486                         foreach( $ext_content[$tag] as $marker => $content ) {
 487                                 $full_tag = $ext_tags[$tag][$marker];
 488                                 $params = $ext_params[$tag][$marker];
 489                                 if ( $render )
 490                                         $ext_content[$tag][$marker] = call_user_func_array( $callback, array( $content, $params, &$this ) );
 491                                 else {
 492                                         if ( is_null( $content ) ) {
 493                                                 // Empty element tag
 494                                                 $ext_content[$tag][$marker] = $full_tag;
 495                                         } else {
 496                                                 $ext_content[$tag][$marker] = "$full_tag$content</$tag>";
 497                                         }
 498                                 }
 499                         }
 500                 }
 501
 502                 # Unstrip comments unless explicitly told otherwise.
 503                 # (The comments are always stripped prior to this point, so as to
 504                 # not invoke any extension tags / parser hooks contained within
 505                 # a comment.)
 506                 if ( !$stripcomments ) {
 507                         $tempstate = array( 'comment' => $comment_content );
 508                         $text = $this->unstrip( $text, $tempstate );
 509                         $comment_content = array();
 510                 }
 511
 512                 # Merge state with the pre-existing state, if there is one
 513                 if ( $state ) {
 514                         $state['html'] = $state['html'] + $html_content;
 515                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 516                         $state['math'] = $state['math'] + $math_content;
 517                         $state['pre'] = $state['pre'] + $pre_content;
 518                         $state['gallery'] = $state['gallery'] + $gallery_content;
 519                         $state['comment'] = $state['comment'] + $comment_content;
 520
 521                         foreach( $ext_content as $tag => $array ) {
 522                                 if ( array_key_exists( $tag, $state ) ) {
 523                                         $state[$tag] = $state[$tag] + $array;
 524                                 }
 525                         }
 526                 } else {
 527                         $state = array(
 528                           'html' => $html_content,
 529                           'nowiki' => $nowiki_content,
 530                           'math' => $math_content,
 531                           'pre' => $pre_content,
 532                           'gallery' => $gallery_content,
 533                           'comment' => $comment_content,
 534                         ) + $ext_content;
 535                 }
 536                 return $text;
 537         }
 538
 539         /**
 540          * restores pre, math, and hiero removed by strip()
 541          *
 542          * always call unstripNoWiki() after this one
 543          * @access private
 544          */
 545         function unstrip( $text, &$state ) {
 546                 if ( !is_array( $state ) ) {
 547                         return $text;
 548                 }
 549
 550                 # Must expand in reverse order, otherwise nested tags will be corrupted
 551                 foreach( array_reverse( $state, true ) as $tag => $contentDict ) {
 552                         if( $tag != 'nowiki' && $tag != 'html' ) {
 553                                 foreach( array_reverse( $contentDict, true ) as $uniq => $content ) {
 554                                         $text = str_replace( $uniq, $content, $text );
 555                                 }
 556                         }
 557                 }
 558
 559                 return $text;
 560         }
 561
 562         /**
 563          * always call this after unstrip() to preserve the order
 564          *
 565          * @access private
 566          */
 567         function unstripNoWiki( $text, &$state ) {
 568                 if ( !is_array( $state ) ) {
 569                         return $text;
 570                 }
 571
 572                 # Must expand in reverse order, otherwise nested tags will be corrupted
 573                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 574                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 575                 }
 576
 577                 global $wgRawHtml;
 578                 if ($wgRawHtml) {
 579                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 580                                 $text = str_replace( key( $state['html'] ), $content, $text );
 581                         }
 582                 }
 583
 584                 return $text;
 585         }
 586
 587         /**
 588          * Add an item to the strip state
 589          * Returns the unique tag which must be inserted into the stripped text
 590          * The tag will be replaced with the original text in unstrip()
 591          *
 592          * @access private
 593          */
 594         function insertStripItem( $text, &$state ) {
 595                 $rnd = $this->mUniqPrefix . '-item' . Parser::getRandomString();
 596                 if ( !$state ) {
 597                         $state = array(
 598                           'html' => array(),
 599                           'nowiki' => array(),
 600                           'math' => array(),
 601                           'pre' => array(),
 602                           'comment' => array(),
 603                           'gallery' => array(),
 604                         );
 605                 }
 606                 $state['item'][$rnd] = $text;
 607                 return $rnd;
 608         }
 609
 610         /**
 611          * Interface with html tidy, used if $wgUseTidy = true.
 612          * If tidy isn't able to correct the markup, the original will be
 613          * returned in all its glory with a warning comment appended.
 614          *
 615          * Either the external tidy program or the in-process tidy extension
 616          * will be used depending on availability. Override the default
 617          * $wgTidyInternal setting to disable the internal if it's not working.
 618          *
 619          * @param string $text Hideous HTML input
 620          * @return string Corrected HTML output
 621          * @access public
 622          * @static
 623          */
 624         function tidy( $text ) {
 625                 global $wgTidyInternal;
 626                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 627 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 628 '<head><title>test</title></head><body>'.$text.'</body></html>';
 629                 if( $wgTidyInternal ) {
 630                         $correctedtext = Parser::internalTidy( $wrappedtext );
 631                 } else {
 632                         $correctedtext = Parser::externalTidy( $wrappedtext );
 633                 }
 634                 if( is_null( $correctedtext ) ) {
 635                         wfDebug( "Tidy error detected!\n" );
 636                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 637                 }
 638                 return $correctedtext;
 639         }
 640
 641         /**
 642          * Spawn an external HTML tidy process and get corrected markup back from it.
 643          *
 644          * @access private
 645          * @static
 646          */
 647         function externalTidy( $text ) {
 648                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 649                 $fname = 'Parser::externalTidy';
 650                 wfProfileIn( $fname );
 651
 652                 $cleansource = '';
 653                 $opts = ' -utf8';
 654
 655                 $descriptorspec = array(
 656                         0 => array('pipe', 'r'),
 657                         1 => array('pipe', 'w'),
 658                         2 => array('file', '/dev/null', 'a')
 659                 );
 660                 $pipes = array();
 661                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
 662                 if (is_resource($process)) {
 663                         // Theoretically, this style of communication could cause a deadlock
 664                         // here. If the stdout buffer fills up, then writes to stdin could
 665                         // block. This doesn't appear to happen with tidy, because tidy only
 666                         // writes to stdout after it's finished reading from stdin. Search
 667                         // for tidyParseStdin and tidySaveStdout in console/tidy.c
 668                         fwrite($pipes[0], $text);
 669                         fclose($pipes[0]);
 670                         while (!feof($pipes[1])) {
 671                                 $cleansource .= fgets($pipes[1], 1024);
 672                         }
 673                         fclose($pipes[1]);
 674                         proc_close($process);
 675                 }
 676
 677                 wfProfileOut( $fname );
 678
 679                 if( $cleansource == '' && $text != '') {
 680                         // Some kind of error happened, so we couldn't get the corrected text.
 681                         // Just give up; we'll use the source text and append a warning.
 682                         return null;
 683                 } else {
 684                         return $cleansource;
 685                 }
 686         }
 687
 688         /**
 689          * Use the HTML tidy PECL extension to use the tidy library in-process,
 690          * saving the overhead of spawning a new process. Currently written to
 691          * the PHP 4.3.x version of the extension, may not work on PHP 5.
 692          *
 693          * 'pear install tidy' should be able to compile the extension module.
 694          *
 695          * @access private
 696          * @static
 697          */
 698         function internalTidy( $text ) {
 699                 global $wgTidyConf;
 700                 $fname = 'Parser::internalTidy';
 701                 wfProfileIn( $fname );
 702
 703                 tidy_load_config( $wgTidyConf );
 704                 tidy_set_encoding( 'utf8' );
 705                 tidy_parse_string( $text );
 706                 tidy_clean_repair();
 707                 if( tidy_get_status() == 2 ) {
 708                         // 2 is magic number for fatal error
 709                         // http://www.php.net/manual/en/function.tidy-get-status.php
 710                         $cleansource = null;
 711                 } else {
 712                         $cleansource = tidy_get_output();
 713                 }
 714                 wfProfileOut( $fname );
 715                 return $cleansource;
 716         }
 717
 718         /**
 719          * parse the wiki syntax used to render tables
 720          *
 721          * @access private
 722          */
 723         function doTableStuff ( $t ) {
 724                 $fname = 'Parser::doTableStuff';
 725                 wfProfileIn( $fname );
 726
 727                 $t = explode ( "\n" , $t ) ;
 728                 $td = array () ; # Is currently a td tag open?
 729                 $ltd = array () ; # Was it TD or TH?
 730                 $tr = array () ; # Is currently a tr tag open?
 731                 $ltr = array () ; # tr attributes
 732                 $has_opened_tr = array(); # Did this table open a <tr> element?
 733                 $indent_level = 0; # indent level of the table
 734                 foreach ( $t AS $k => $x )
 735                 {
 736                         $x = trim ( $x ) ;
 737                         $fc = substr ( $x , 0 , 1 ) ;
 738                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 739                                 $indent_level = strlen( $matches[1] );
 740
 741                                 $attributes = $this->unstripForHTML( $matches[2] );
 742
 743                                 $t[$k] = str_repeat( '<dl><dd>', $indent_level ) .
 744                                         '<table' . Sanitizer::fixTagAttributes ( $attributes, 'table' ) . '>' ;
 745                                 array_push ( $td , false ) ;
 746                                 array_push ( $ltd , '' ) ;
 747                                 array_push ( $tr , false ) ;
 748                                 array_push ( $ltr , '' ) ;
 749                                 array_push ( $has_opened_tr, false );
 750                         }
 751                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 752                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 753                                 $z = "</table>" . substr ( $x , 2);
 754                                 $l = array_pop ( $ltd ) ;
 755                                 if ( !array_pop ( $has_opened_tr ) ) $z = "<tr><td></td></tr>" . $z ;
 756                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 757                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 758                                 array_pop ( $ltr ) ;
 759                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 760                         }
 761                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 762                                 $x = substr ( $x , 1 ) ;
 763                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 764                                 $z = '' ;
 765                                 $l = array_pop ( $ltd ) ;
 766                                 array_pop ( $has_opened_tr );
 767                                 array_push ( $has_opened_tr , true ) ;
 768                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 769                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 770                                 array_pop ( $ltr ) ;
 771                                 $t[$k] = $z ;
 772                                 array_push ( $tr , false ) ;
 773                                 array_push ( $td , false ) ;
 774                                 array_push ( $ltd , '' ) ;
 775                                 $attributes = $this->unstripForHTML( $x );
 776                                 array_push ( $ltr , Sanitizer::fixTagAttributes ( $attributes, 'tr' ) ) ;
 777                         }
 778                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 779                                 # $x is a table row
 780                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 781                                         $fc = '+' ;
 782                                         $x = substr ( $x , 1 ) ;
 783                                 }
 784                                 $after = substr ( $x , 1 ) ;
 785                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 786
 787                                 // Split up multiple cells on the same line.
 788                                 // FIXME: This can result in improper nesting of tags processed
 789                                 // by earlier parser steps, but should avoid splitting up eg
 790                                 // attribute values containing literal "||".
 791                                 $after = wfExplodeMarkup( '||', $after );
 792
 793                                 $t[$k] = '' ;
 794
 795                                 # Loop through each table cell
 796                                 foreach ( $after AS $theline )
 797                                 {
 798                                         $z = '' ;
 799                                         if ( $fc != '+' )
 800                                         {
 801                                                 $tra = array_pop ( $ltr ) ;
 802                                                 if ( !array_pop ( $tr ) ) $z = '<tr'.$tra.">\n" ;
 803                                                 array_push ( $tr , true ) ;
 804                                                 array_push ( $ltr , '' ) ;
 805                                                 array_pop ( $has_opened_tr );
 806                                                 array_push ( $has_opened_tr , true ) ;
 807                                         }
 808
 809                                         $l = array_pop ( $ltd ) ;
 810                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 811                                         if ( $fc == '|' ) $l = 'td' ;
 812                                         else if ( $fc == '!' ) $l = 'th' ;
 813                                         else if ( $fc == '+' ) $l = 'caption' ;
 814                                         else $l = '' ;
 815                                         array_push ( $ltd , $l ) ;
 816
 817                                         # Cell parameters
 818                                         $y = explode ( '|' , $theline , 2 ) ;
 819                                         # Note that a '|' inside an invalid link should not
 820                                         # be mistaken as delimiting cell parameters
 821                                         if ( strpos( $y[0], '[[' ) !== false ) {
 822                                                 $y = array ($theline);
 823                                         }
 824                                         if ( count ( $y ) == 1 )
 825                                                 $y = "{$z}<{$l}>{$y[0]}" ;
 826                                         else {
 827                                                 $attributes = $this->unstripForHTML( $y[0] );
 828                                                 $y = "{$z}<{$l}".Sanitizer::fixTagAttributes($attributes, $l).">{$y[1]}" ;
 829                                         }
 830                                         $t[$k] .= $y ;
 831                                         array_push ( $td , true ) ;
 832                                 }
 833                         }
 834                 }
 835
 836                 # Closing open td, tr && table
 837                 while ( count ( $td ) > 0 )
 838                 {
 839                         $l = array_pop ( $ltd ) ;
 840                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 841                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 842                         if ( !array_pop ( $has_opened_tr ) ) $t[] = "<tr><td></td></tr>" ;
 843                         $t[] = '</table>' ;
 844                 }
 845
 846                 $t = implode ( "\n" , $t ) ;
 847                 # special case: don't return empty table
 848                 if($t == "<table>\n<tr><td></td></tr>\n</table>")
 849                         $t = '';
 850                 wfProfileOut( $fname );
 851                 return $t ;
 852         }
 853
 854         /**
 855          * Helper function for parse() that transforms wiki markup into
 856          * HTML. Only called for $mOutputType == OT_HTML.
 857          *
 858          * @access private
 859          */
 860         function internalParse( $text ) {
 861                 $args = array();
 862                 $isMain = true;
 863                 $fname = 'Parser::internalParse';
 864                 wfProfileIn( $fname );
 865
 866                 # Remove <noinclude> tags and <includeonly> sections
 867                 $text = strtr( $text, array( '<onlyinclude>' => '' , '</onlyinclude>' => '' ) );
 868                 $text = strtr( $text, array( '<noinclude>' => '', '</noinclude>' => '') );
 869                 $text = preg_replace( '/<includeonly>.*?<\/includeonly>/s', '', $text );
 870
 871                 $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'attributeStripCallback' ) );
 872                 $text = $this->replaceVariables( $text, $args );
 873
 874                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 875
 876                 $text = $this->doHeadings( $text );
 877                 if($this->mOptions->getUseDynamicDates()) {
 878                         $df =& DateFormatter::getInstance();
 879                         $text = $df->reformat( $this->mOptions->getDateFormat(), $text );
 880                 }
 881                 $text = $this->doAllQuotes( $text );
 882                 $text = $this->replaceInternalLinks( $text );
 883                 $text = $this->replaceExternalLinks( $text );
 884
 885                 # replaceInternalLinks may sometimes leave behind
 886                 # absolute URLs, which have to be masked to hide them from replaceExternalLinks
 887                 $text = str_replace($this->mUniqPrefix."NOPARSE", "", $text);
 888
 889                 $text = $this->doMagicLinks( $text );
 890                 $text = $this->doTableStuff( $text );
 891                 $text = $this->formatHeadings( $text, $isMain );
 892
 893                 wfProfileOut( $fname );
 894                 return $text;
 895         }
 896
 897         /**
 898          * Replace special strings like "ISBN xxx" and "RFC xxx" with
 899          * magic external links.
 900          *
 901          * @access private
 902          */
 903         function &doMagicLinks( &$text ) {
 904                 $text = $this->magicISBN( $text );
 905                 $text = $this->magicRFC( $text, 'RFC ', 'rfcurl' );
 906                 $text = $this->magicRFC( $text, 'PMID ', 'pubmedurl' );
 907                 return $text;
 908         }
 909
 910         /**
 911          * Parse headers and return html
 912          *
 913          * @access private
 914          */
 915         function doHeadings( $text ) {
 916                 $fname = 'Parser::doHeadings';
 917                 wfProfileIn( $fname );
 918                 for ( $i = 6; $i >= 1; --$i ) {
 919                         $h = str_repeat( '=', $i );
 920                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 921                           "<h{$i}>\\1</h{$i}>\\2", $text );
 922                 }
 923                 wfProfileOut( $fname );
 924                 return $text;
 925         }
 926
 927         /**
 928          * Replace single quotes with HTML markup
 929          * @access private
 930          * @return string the altered text
 931          */
 932         function doAllQuotes( $text ) {
 933                 $fname = 'Parser::doAllQuotes';
 934                 wfProfileIn( $fname );
 935                 $outtext = '';
 936                 $lines = explode( "\n", $text );
 937                 foreach ( $lines as $line ) {
 938                         $outtext .= $this->doQuotes ( $line ) . "\n";
 939                 }
 940                 $outtext = substr($outtext, 0,-1);
 941                 wfProfileOut( $fname );
 942                 return $outtext;
 943         }
 944
 945         /**
 946          * Helper function for doAllQuotes()
 947          * @access private
 948          */
 949         function doQuotes( $text ) {
 950                 $arr = preg_split( "/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 951                 if ( count( $arr ) == 1 )
 952                         return $text;
 953                 else
 954                 {
 955                         # First, do some preliminary work. This may shift some apostrophes from
 956                         # being mark-up to being text. It also counts the number of occurrences
 957                         # of bold and italics mark-ups.
 958                         $i = 0;
 959                         $numbold = 0;
 960                         $numitalics = 0;
 961                         foreach ( $arr as $r )
 962                         {
 963                                 if ( ( $i % 2 ) == 1 )
 964                                 {
 965                                         # If there are ever four apostrophes, assume the first is supposed to
 966                                         # be text, and the remaining three constitute mark-up for bold text.
 967                                         if ( strlen( $arr[$i] ) == 4 )
 968                                         {
 969                                                 $arr[$i-1] .= "'";
 970                                                 $arr[$i] = "'''";
 971                                         }
 972                                         # If there are more than 5 apostrophes in a row, assume they're all
 973                                         # text except for the last 5.
 974                                         else if ( strlen( $arr[$i] ) > 5 )
 975                                         {
 976                                                 $arr[$i-1] .= str_repeat( "'", strlen( $arr[$i] ) - 5 );
 977                                                 $arr[$i] = "'''''";
 978                                         }
 979                                         # Count the number of occurrences of bold and italics mark-ups.
 980                                         # We are not counting sequences of five apostrophes.
 981                                         if ( strlen( $arr[$i] ) == 2 ) $numitalics++;  else
 982                                         if ( strlen( $arr[$i] ) == 3 ) $numbold++;     else
 983                                         if ( strlen( $arr[$i] ) == 5 ) { $numitalics++; $numbold++; }
 984                                 }
 985                                 $i++;
 986                         }
 987
 988                         # If there is an odd number of both bold and italics, it is likely
 989                         # that one of the bold ones was meant to be an apostrophe followed
 990                         # by italics. Which one we cannot know for certain, but it is more
 991                         # likely to be one that has a single-letter word before it.
 992                         if ( ( $numbold % 2 == 1 ) && ( $numitalics % 2 == 1 ) )
 993                         {
 994                                 $i = 0;
 995                                 $firstsingleletterword = -1;
 996                                 $firstmultiletterword = -1;
 997                                 $firstspace = -1;
 998                                 foreach ( $arr as $r )
 999                                 {
1000                                         if ( ( $i % 2 == 1 ) and ( strlen( $r ) == 3 ) )
1001                                         {
1002                                                 $x1 = substr ($arr[$i-1], -1);
1003                                                 $x2 = substr ($arr[$i-1], -2, 1);
1004                                                 if ($x1 == ' ') {
1005                                                         if ($firstspace == -1) $firstspace = $i;
1006                                                 } else if ($x2 == ' ') {
1007                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
1008                                                 } else {
1009                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
1010                                                 }
1011                                         }
1012                                         $i++;
1013                                 }
1014
1015                                 # If there is a single-letter word, use it!
1016                                 if ($firstsingleletterword > -1)
1017                                 {
1018                                         $arr [ $firstsingleletterword ] = "''";
1019                                         $arr [ $firstsingleletterword-1 ] .= "'";
1020                                 }
1021                                 # If not, but there's a multi-letter word, use that one.
1022                                 else if ($firstmultiletterword > -1)
1023                                 {
1024                                         $arr [ $firstmultiletterword ] = "''";
1025                                         $arr [ $firstmultiletterword-1 ] .= "'";
1026                                 }
1027                                 # ... otherwise use the first one that has neither.
1028                                 # (notice that it is possible for all three to be -1 if, for example,
1029                                 # there is only one pentuple-apostrophe in the line)
1030                                 else if ($firstspace > -1)
1031                                 {
1032                                         $arr [ $firstspace ] = "''";
1033                                         $arr [ $firstspace-1 ] .= "'";
1034                                 }
1035                         }
1036
1037                         # Now let's actually convert our apostrophic mush to HTML!
1038                         $output = '';
1039                         $buffer = '';
1040                         $state = '';
1041                         $i = 0;
1042                         foreach ($arr as $r)
1043                         {
1044                                 if (($i % 2) == 0)
1045                                 {
1046                                         if ($state == 'both')
1047                                                 $buffer .= $r;
1048                                         else
1049                                                 $output .= $r;
1050                                 }
1051                                 else
1052                                 {
1053                                         if (strlen ($r) == 2)
1054                                         {
1055                                                 if ($state == 'i')
1056                                                 { $output .= '</i>'; $state = ''; }
1057                                                 else if ($state == 'bi')
1058                                                 { $output .= '</i>'; $state = 'b'; }
1059                                                 else if ($state == 'ib')
1060                                                 { $output .= '</b></i><b>'; $state = 'b'; }
1061                                                 else if ($state == 'both')
1062                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
1063                                                 else # $state can be 'b' or ''
1064                                                 { $output .= '<i>'; $state .= 'i'; }
1065                                         }
1066                                         else if (strlen ($r) == 3)
1067                                         {
1068                                                 if ($state == 'b')
1069                                                 { $output .= '</b>'; $state = ''; }
1070                                                 else if ($state == 'bi')
1071                                                 { $output .= '</i></b><i>'; $state = 'i'; }
1072                                                 else if ($state == 'ib')
1073                                                 { $output .= '</b>'; $state = 'i'; }
1074                                                 else if ($state == 'both')
1075                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
1076                                                 else # $state can be 'i' or ''
1077                                                 { $output .= '<b>'; $state .= 'b'; }
1078                                         }
1079                                         else if (strlen ($r) == 5)
1080                                         {
1081                                                 if ($state == 'b')
1082                                                 { $output .= '</b><i>'; $state = 'i'; }
1083                                                 else if ($state == 'i')
1084                                                 { $output .= '</i><b>'; $state = 'b'; }
1085                                                 else if ($state == 'bi')
1086                                                 { $output .= '</i></b>'; $state = ''; }
1087                                                 else if ($state == 'ib')
1088                                                 { $output .= '</b></i>'; $state = ''; }
1089                                                 else if ($state == 'both')
1090                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
1091                                                 else # ($state == '')
1092                                                 { $buffer = ''; $state = 'both'; }
1093                                         }
1094                                 }
1095                                 $i++;
1096                         }
1097                         # Now close all remaining tags.  Notice that the order is important.
1098                         if ($state == 'b' || $state == 'ib')
1099                                 $output .= '</b>';
1100                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
1101                                 $output .= '</i>';
1102                         if ($state == 'bi')
1103                                 $output .= '</b>';
1104                         if ($state == 'both')
1105                                 $output .= '<b><i>'.$buffer.'</i></b>';
1106                         return $output;
1107                 }
1108         }
1109
1110         /**
1111          * Replace external links
1112          *
1113          * Note: this is all very hackish and the order of execution matters a lot.
1114          * Make sure to run maintenance/parserTests.php if you change this code.
1115          *
1116          * @access private
1117          */
1118         function replaceExternalLinks( $text ) {
1119                 global $wgContLang;
1120                 $fname = 'Parser::replaceExternalLinks';
1121                 wfProfileIn( $fname );
1122
1123                 $sk =& $this->mOptions->getSkin();
1124
1125                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1126
1127                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
1128
1129                 $i = 0;
1130                 while ( $i<count( $bits ) ) {
1131                         $url = $bits[$i++];
1132                         $protocol = $bits[$i++];
1133                         $text = $bits[$i++];
1134                         $trail = $bits[$i++];
1135
1136                         # The characters '<' and '>' (which were escaped by
1137                         # removeHTMLtags()) should not be included in
1138                         # URLs, per RFC 2396.
1139                         if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1140                                 $text = substr($url, $m2[0][1]) . ' ' . $text;
1141                                 $url = substr($url, 0, $m2[0][1]);
1142                         }
1143
1144                         # If the link text is an image URL, replace it with an <img> tag
1145                         # This happened by accident in the original parser, but some people used it extensively
1146                         $img = $this->maybeMakeExternalImage( $text );
1147                         if ( $img !== false ) {
1148                                 $text = $img;
1149                         }
1150
1151                         $dtrail = '';
1152
1153                         # Set linktype for CSS - if URL==text, link is essentially free
1154                         $linktype = ($text == $url) ? 'free' : 'text';
1155
1156                         # No link text, e.g. [http://domain.tld/some.link]
1157                         if ( $text == '' ) {
1158                                 # Autonumber if allowed
1159                                 if ( strpos( HTTP_PROTOCOLS, str_replace('/','\/', $protocol) ) !== false ) {
1160                                         $text = '[' . ++$this->mAutonumber . ']';
1161                                         $linktype = 'autonumber';
1162                                 } else {
1163                                         # Otherwise just use the URL
1164                                         $text = htmlspecialchars( $url );
1165                                         $linktype = 'free';
1166                                 }
1167                         } else {
1168                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1169                                 # Check for trail
1170                                 list( $dtrail, $trail ) = Linker::splitTrail( $trail );
1171                         }
1172
1173                         $text = $wgContLang->markNoConversion($text);
1174
1175                         # Replace &amp; from obsolete syntax with &.
1176                         # All HTML entities will be escaped by makeExternalLink()
1177                         $url = str_replace( '&amp;', '&', $url );
1178
1179                         # Process the trail (i.e. everything after this link up until start of the next link),
1180                         # replacing any non-bracketed links
1181                         $trail = $this->replaceFreeExternalLinks( $trail );
1182
1183                         # Use the encoded URL
1184                         # This means that users can paste URLs directly into the text
1185                         # Funny characters like &ouml; aren't valid in URLs anyway
1186                         # This was changed in August 2004
1187                         $s .= $sk->makeExternalLink( $url, $text, false, $linktype ) . $dtrail . $trail;
1188
1189                         # Register link in the output object.
1190                         # Replace unnecessary URL escape codes with the referenced character
1191                         # This prevents spammers from hiding links from the filters
1192                         $pasteurized = Parser::replaceUnusualEscapes( $url );
1193                         $this->mOutput->addExternalLink( $pasteurized );
1194                 }
1195
1196                 wfProfileOut( $fname );
1197                 return $s;
1198         }
1199
1200         /**
1201          * Replace anything that looks like a URL with a link
1202          * @access private
1203          */
1204         function replaceFreeExternalLinks( $text ) {
1205                 global $wgContLang;
1206                 $fname = 'Parser::replaceFreeExternalLinks';
1207                 wfProfileIn( $fname );
1208
1209                 $bits = preg_split( '/(\b(?:' . wfUrlProtocols() . '))/S', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1210                 $s = array_shift( $bits );
1211                 $i = 0;
1212
1213                 $sk =& $this->mOptions->getSkin();
1214
1215                 while ( $i < count( $bits ) ){
1216                         $protocol = $bits[$i++];
1217                         $remainder = $bits[$i++];
1218
1219                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1220                                 # Found some characters after the protocol that look promising
1221                                 $url = $protocol . $m[1];
1222                                 $trail = $m[2];
1223
1224                                 # special case: handle urls as url args:
1225                                 # http://www.example.com/foo?=http://www.example.com/bar
1226                                 if(strlen($trail) == 0 &&
1227                                         isset($bits[$i]) &&
1228                                         preg_match('/^'. wfUrlProtocols() . '$/S', $bits[$i]) &&
1229                                         preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $bits[$i + 1], $m ))
1230                                 {
1231                                         # add protocol, arg
1232                                         $url .= $bits[$i] . $bits[$i + 1]; # protocol, url as arg to previous link
1233                                         $i += 2;
1234                                         $trail = $m[2];
1235                                 }
1236
1237                                 # The characters '<' and '>' (which were escaped by
1238                                 # removeHTMLtags()) should not be included in
1239                                 # URLs, per RFC 2396.
1240                                 if (preg_match('/&(lt|gt);/', $url, $m2, PREG_OFFSET_CAPTURE)) {
1241                                         $trail = substr($url, $m2[0][1]) . $trail;
1242                                         $url = substr($url, 0, $m2[0][1]);
1243                                 }
1244
1245                                 # Move trailing punctuation to $trail
1246                                 $sep = ',;\.:!?';
1247                                 # If there is no left bracket, then consider right brackets fair game too
1248                                 if ( strpos( $url, '(' ) === false ) {
1249                                         $sep .= ')';
1250                                 }
1251
1252                                 $numSepChars = strspn( strrev( $url ), $sep );
1253                                 if ( $numSepChars ) {
1254                                         $trail = substr( $url, -$numSepChars ) . $trail;
1255                                         $url = substr( $url, 0, -$numSepChars );
1256                                 }
1257
1258                                 # Replace &amp; from obsolete syntax with &.
1259                                 # All HTML entities will be escaped by makeExternalLink()
1260                                 # or maybeMakeExternalImage()
1261                                 $url = str_replace( '&amp;', '&', $url );
1262
1263                                 # Is this an external image?
1264                                 $text = $this->maybeMakeExternalImage( $url );
1265                                 if ( $text === false ) {
1266                                         # Not an image, make a link
1267                                         $text = $sk->makeExternalLink( $url, $wgContLang->markNoConversion($url), true, 'free' );
1268                                         # Register it in the output object...
1269                                         # Replace unnecessary URL escape codes with their equivalent characters
1270                                         $pasteurized = Parser::replaceUnusualEscapes( $url );
1271                                         $this->mOutput->addExternalLink( $pasteurized );
1272                                 }
1273                                 $s .= $text . $trail;
1274                         } else {
1275                                 $s .= $protocol . $remainder;
1276                         }
1277                 }
1278                 wfProfileOut( $fname );
1279                 return $s;
1280         }
1281
1282         /**
1283          * Replace unusual URL escape codes with their equivalent characters
1284          * @param string
1285          * @return string
1286          * @static
1287          * @fixme This can merge genuinely required bits in the path or query string,
1288          *        breaking legit URLs. A proper fix would treat the various parts of
1289          *        the URL differently; as a workaround, just use the output for
1290          *        statistical records, not for actual linking/output.
1291          */
1292         function replaceUnusualEscapes( $url ) {
1293                 return preg_replace_callback( '/%[0-9A-Fa-f]{2}/',
1294                         array( 'Parser', 'replaceUnusualEscapesCallback' ), $url );
1295         }
1296
1297         /**
1298          * Callback function used in replaceUnusualEscapes().
1299          * Replaces unusual URL escape codes with their equivalent character
1300          * @static
1301          * @access private
1302          */
1303         function replaceUnusualEscapesCallback( $matches ) {
1304                 $char = urldecode( $matches[0] );
1305                 $ord = ord( $char );
1306                 // Is it an unsafe or HTTP reserved character according to RFC 1738?
1307                 if ( $ord > 32 && $ord < 127 && strpos( '<>"#{}|\^~[]`;/?', $char ) === false ) {
1308                         // No, shouldn't be escaped
1309                         return $char;
1310                 } else {
1311                         // Yes, leave it escaped
1312                         return $matches[0];
1313                 }
1314         }
1315
1316         /**
1317          * make an image if it's allowed, either through the global
1318          * option or through the exception
1319          * @access private
1320          */
1321         function maybeMakeExternalImage( $url ) {
1322                 $sk =& $this->mOptions->getSkin();
1323                 $imagesfrom = $this->mOptions->getAllowExternalImagesFrom();
1324                 $imagesexception = !empty($imagesfrom);
1325                 $text = false;
1326                 if ( $this->mOptions->getAllowExternalImages()
1327                      || ( $imagesexception && strpos( $url, $imagesfrom ) === 0 ) ) {
1328                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1329                                 # Image found
1330                                 $text = $sk->makeExternalImage( htmlspecialchars( $url ) );
1331                         }
1332                 }
1333                 return $text;
1334         }
1335
1336         /**
1337          * Process [[ ]] wikilinks
1338          *
1339          * @access private
1340          */
1341         function replaceInternalLinks( $s ) {
1342                 global $wgContLang;
1343                 static $fname = 'Parser::replaceInternalLinks' ;
1344
1345                 wfProfileIn( $fname );
1346
1347                 wfProfileIn( $fname.'-setup' );
1348                 static $tc = FALSE;
1349                 # the % is needed to support urlencoded titles as well
1350                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1351
1352                 $sk =& $this->mOptions->getSkin();
1353
1354                 #split the entire text string on occurences of [[
1355                 $a = explode( '[[', ' ' . $s );
1356                 #get the first element (all text up to first [[), and remove the space we added
1357                 $s = array_shift( $a );
1358                 $s = substr( $s, 1 );
1359
1360                 # Match a link having the form [[namespace:link|alternate]]trail
1361                 static $e1 = FALSE;
1362                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD"; }
1363                 # Match cases where there is no "]]", which might still be images
1364                 static $e1_img = FALSE;
1365                 if ( !$e1_img ) { $e1_img = "/^([{$tc}]+)\\|(.*)\$/sD"; }
1366                 # Match the end of a line for a word that's not followed by whitespace,
1367                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1368                 $e2 = wfMsgForContent( 'linkprefix' );
1369
1370                 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1371
1372                 if( is_null( $this->mTitle ) ) {
1373                         wfDebugDieBacktrace( 'nooo' );
1374                 }
1375                 $nottalk = !$this->mTitle->isTalkPage();
1376
1377                 if ( $useLinkPrefixExtension ) {
1378                         if ( preg_match( $e2, $s, $m ) ) {
1379                                 $first_prefix = $m[2];
1380                         } else {
1381                                 $first_prefix = false;
1382                         }
1383                 } else {
1384                         $prefix = '';
1385                 }
1386
1387                 $selflink = $this->mTitle->getPrefixedText();
1388                 wfProfileOut( $fname.'-setup' );
1389
1390                 $checkVariantLink = sizeof($wgContLang->getVariants())>1;
1391                 $useSubpages = $this->areSubpagesAllowed();
1392
1393                 # Loop for each link
1394                 for ($k = 0; isset( $a[$k] ); $k++) {
1395                         $line = $a[$k];
1396                         if ( $useLinkPrefixExtension ) {
1397                                 wfProfileIn( $fname.'-prefixhandling' );
1398                                 if ( preg_match( $e2, $s, $m ) ) {
1399                                         $prefix = $m[2];
1400                                         $s = $m[1];
1401                                 } else {
1402                                         $prefix='';
1403                                 }
1404                                 # first link
1405                                 if($first_prefix) {
1406                                         $prefix = $first_prefix;
1407                                         $first_prefix = false;
1408                                 }
1409                                 wfProfileOut( $fname.'-prefixhandling' );
1410                         }
1411
1412                         $might_be_img = false;
1413
1414                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1415                                 $text = $m[2];
1416                                 # If we get a ] at the beginning of $m[3] that means we have a link that's something like:
1417                                 # [[Image:Foo.jpg|[http://example.com desc]]] <- having three ] in a row fucks up,
1418                                 # the real problem is with the $e1 regex
1419                                 # See bug 1300.
1420                                 #
1421                                 # Still some problems for cases where the ] is meant to be outside punctuation,
1422                                 # and no image is in sight. See bug 2095.
1423                                 #
1424                                 if( $text !== '' &&
1425                                         preg_match( "/^\](.*)/s", $m[3], $n ) &&
1426                                         strpos($text, '[') !== false
1427                                 )
1428                                 {
1429                                         $text .= ']'; # so that replaceExternalLinks($text) works later
1430                                         $m[3] = $n[1];
1431                                 }
1432                                 # fix up urlencoded title texts
1433                                 if(preg_match('/%/', $m[1] ))
1434                                         # Should anchors '#' also be rejected?
1435                                         $m[1] = str_replace( array('<', '>'), array('&lt;', '&gt;'), urldecode($m[1]) );
1436                                 $trail = $m[3];
1437                         } elseif( preg_match($e1_img, $line, $m) ) { # Invalid, but might be an image with a link in its caption
1438                                 $might_be_img = true;
1439                                 $text = $m[2];
1440                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1441                                 $trail = "";
1442                         } else { # Invalid form; output directly
1443                                 $s .= $prefix . '[[' . $line ;
1444                                 continue;
1445                         }
1446
1447                         # Don't allow internal links to pages containing
1448                         # PROTO: where PROTO is a valid URL protocol; these
1449                         # should be external links.
1450                         if (preg_match('/^(\b(?:' . wfUrlProtocols() . '))/', $m[1])) {
1451                                 $s .= $prefix . '[[' . $line ;
1452                                 continue;
1453                         }
1454
1455                         # Make subpage if necessary
1456                         if( $useSubpages ) {
1457                                 $link = $this->maybeDoSubpageLink( $m[1], $text );
1458                         } else {
1459                                 $link = $m[1];
1460                         }
1461
1462                         $noforce = (substr($m[1], 0, 1) != ':');
1463                         if (!$noforce) {
1464                                 # Strip off leading ':'
1465                                 $link = substr($link, 1);
1466                         }
1467
1468                         $nt = Title::newFromText( $this->unstripNoWiki($link, $this->mStripState) );
1469                         if( !$nt ) {
1470                                 $s .= $prefix . '[[' . $line;
1471                                 continue;
1472                         }
1473
1474                         #check other language variants of the link
1475                         #if the article does not exist
1476                         if( $checkVariantLink
1477                             && $nt->getArticleID() == 0 ) {
1478                                 $wgContLang->findVariantLink($link, $nt);
1479                         }
1480
1481                         $ns = $nt->getNamespace();
1482                         $iw = $nt->getInterWiki();
1483
1484                         if ($might_be_img) { # if this is actually an invalid link
1485                                 if ($ns == NS_IMAGE && $noforce) { #but might be an image
1486                                         $found = false;
1487                                         while (isset ($a[$k+1]) ) {
1488                                                 #look at the next 'line' to see if we can close it there
1489                                                 $spliced = array_splice( $a, $k + 1, 1 );
1490                                                 $next_line = array_shift( $spliced );
1491                                                 if( preg_match("/^(.*?]].*?)]](.*)$/sD", $next_line, $m) ) {
1492                                                 # the first ]] closes the inner link, the second the image
1493                                                         $found = true;
1494                                                         $text .= '[[' . $m[1];
1495                                                         $trail = $m[2];
1496                                                         break;
1497                                                 } elseif( preg_match("/^.*?]].*$/sD", $next_line, $m) ) {
1498                                                         #if there's exactly one ]] that's fine, we'll keep looking
1499                                                         $text .= '[[' . $m[0];
1500                                                 } else {
1501                                                         #if $next_line is invalid too, we need look no further
1502                                                         $text .= '[[' . $next_line;
1503                                                         break;
1504                                                 }
1505                                         }
1506                                         if ( !$found ) {
1507                                                 # we couldn't find the end of this imageLink, so output it raw
1508                                                 #but don't ignore what might be perfectly normal links in the text we've examined
1509                                                 $text = $this->replaceInternalLinks($text);
1510                                                 $s .= $prefix . '[[' . $link . '|' . $text;
1511                                                 # note: no $trail, because without an end, there *is* no trail
1512                                                 continue;
1513                                         }
1514                                 } else { #it's not an image, so output it raw
1515                                         $s .= $prefix . '[[' . $link . '|' . $text;
1516                                         # note: no $trail, because without an end, there *is* no trail
1517                                         continue;
1518                                 }
1519                         }
1520
1521                         $wasblank = ( '' == $text );
1522                         if( $wasblank ) $text = $link;
1523
1524
1525                         # Link not escaped by : , create the various objects
1526                         if( $noforce ) {
1527
1528                                 # Interwikis
1529                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1530                                         $this->mOutput->addLanguageLink( $nt->getFullText() );
1531                                         $s = rtrim($s . "\n");
1532                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1533                                         continue;
1534                                 }
1535
1536                                 if ( $ns == NS_IMAGE ) {
1537                                         wfProfileIn( "$fname-image" );
1538                                         if ( !wfIsBadImage( $nt->getDBkey() ) ) {
1539                                                 # recursively parse links inside the image caption
1540                                                 # actually, this will parse them in any other parameters, too,
1541                                                 # but it might be hard to fix that, and it doesn't matter ATM
1542                                                 $text = $this->replaceExternalLinks($text);
1543                                                 $text = $this->replaceInternalLinks($text);
1544
1545                                                 # cloak any absolute URLs inside the image markup, so replaceExternalLinks() won't touch them
1546                                                 $s .= $prefix . $this->armorLinks( $this->makeImage( $nt, $text ) ) . $trail;
1547                                                 $this->mOutput->addImage( $nt->getDBkey() );
1548
1549                                                 wfProfileOut( "$fname-image" );
1550                                                 continue;
1551                                         }
1552                                         wfProfileOut( "$fname-image" );
1553
1554                                 }
1555
1556                                 if ( $ns == NS_CATEGORY ) {
1557                                         wfProfileIn( "$fname-category" );
1558                                         $s = rtrim($s . "\n"); # bug 87
1559
1560                                         if ( $wasblank ) {
1561                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1562                                                         $sortkey = $this->mTitle->getText();
1563                                                 } else {
1564                                                         $sortkey = $this->mTitle->getPrefixedText();
1565                                                 }
1566                                         } else {
1567                                                 $sortkey = $text;
1568                                         }
1569                                         $sortkey = Sanitizer::decodeCharReferences( $sortkey );
1570                                         $sortkey = $wgContLang->convertCategoryKey( $sortkey );
1571                                         $this->mOutput->addCategory( $nt->getDBkey(), $sortkey );
1572
1573                                         /**
1574                                          * Strip the whitespace Category links produce, see bug 87
1575                                          * @todo We might want to use trim($tmp, "\n") here.
1576                                          */
1577                                         $s .= trim($prefix . $trail, "\n") == '' ? '': $prefix . $trail;
1578
1579                                         wfProfileOut( "$fname-category" );
1580                                         continue;
1581                                 }
1582                         }
1583
1584                         if( ( $nt->getPrefixedText() === $selflink ) &&
1585                             ( $nt->getFragment() === '' ) ) {
1586                                 # Self-links are handled specially; generally de-link and change to bold.
1587                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1588                                 continue;
1589                         }
1590
1591                         # Special and Media are pseudo-namespaces; no pages actually exist in them
1592                         if( $ns == NS_MEDIA ) {
1593                                 $link = $sk->makeMediaLinkObj( $nt, $text );
1594                                 # Cloak with NOPARSE to avoid replacement in replaceExternalLinks
1595                                 $s .= $prefix . $this->armorLinks( $link ) . $trail;
1596                                 $this->mOutput->addImage( $nt->getDBkey() );
1597                                 continue;
1598                         } elseif( $ns == NS_SPECIAL ) {
1599                                 $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix );
1600                                 continue;
1601                         } elseif( $ns == NS_IMAGE ) {
1602                                 $img = Image::newFromTitle( $nt );
1603                                 if( $img->exists() ) {
1604                                         // Force a blue link if the file exists; may be a remote
1605                                         // upload on the shared repository, and we want to see its
1606                                         // auto-generated page.
1607                                         $s .= $this->makeKnownLinkHolder( $nt, $text, '', $trail, $prefix );
1608                                         continue;
1609                                 }
1610                         }
1611                         $s .= $this->makeLinkHolder( $nt, $text, '', $trail, $prefix );
1612                 }
1613                 wfProfileOut( $fname );
1614                 return $s;
1615         }
1616
1617         /**
1618          * Make a link placeholder. The text returned can be later resolved to a real link with
1619          * replaceLinkHolders(). This is done for two reasons: firstly to avoid further
1620          * parsing of interwiki links, and secondly to allow all extistence checks and
1621          * article length checks (for stub links) to be bundled into a single query.
1622          *
1623          */
1624         function makeLinkHolder( &$nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1625                 if ( ! is_object($nt) ) {
1626                         # Fail gracefully
1627                         $retVal = "<!-- ERROR -->{$prefix}{$text}{$trail}";
1628                 } else {
1629                         # Separate the link trail from the rest of the link
1630                         list( $inside, $trail ) = Linker::splitTrail( $trail );
1631
1632                         if ( $nt->isExternal() ) {
1633                                 $nr = array_push( $this->mInterwikiLinkHolders['texts'], $prefix.$text.$inside );
1634                                 $this->mInterwikiLinkHolders['titles'][] = $nt;
1635                                 $retVal = '<!--IWLINK '. ($nr-1) ."-->{$trail}";
1636                         } else {
1637                                 $nr = array_push( $this->mLinkHolders['namespaces'], $nt->getNamespace() );
1638                                 $this->mLinkHolders['dbkeys'][] = $nt->getDBkey();
1639                                 $this->mLinkHolders['queries'][] = $query;
1640                                 $this->mLinkHolders['texts'][] = $prefix.$text.$inside;
1641                                 $this->mLinkHolders['titles'][] = $nt;
1642
1643                                 $retVal = '<!--LINK '. ($nr-1) ."-->{$trail}";
1644                         }
1645                 }
1646                 return $retVal;
1647         }
1648
1649         /**
1650          * Render a forced-blue link inline; protect against double expansion of
1651          * URLs if we're in a mode that prepends full URL prefixes to internal links.
1652          * Since this little disaster has to split off the trail text to avoid
1653          * breaking URLs in the following text without breaking trails on the
1654          * wiki links, it's been made into a horrible function.
1655          *
1656          * @param Title $nt
1657          * @param string $text
1658          * @param string $query
1659          * @param string $trail
1660          * @param string $prefix
1661          * @return string HTML-wikitext mix oh yuck
1662          */
1663         function makeKnownLinkHolder( $nt, $text = '', $query = '', $trail = '', $prefix = '' ) {
1664                 list( $inside, $trail ) = Linker::splitTrail( $trail );
1665                 $sk =& $this->mOptions->getSkin();
1666                 $link = $sk->makeKnownLinkObj( $nt, $text, $query, $inside, $prefix );
1667                 return $this->armorLinks( $link ) . $trail;
1668         }
1669
1670         /**
1671          * Insert a NOPARSE hacky thing into any inline links in a chunk that's
1672          * going to go through further parsing steps before inline URL expansion.
1673          *
1674          * In particular this is important when using action=render, which causes
1675          * full URLs to be included.
1676          *
1677          * Oh man I hate our multi-layer parser!
1678          *
1679          * @param string more-or-less HTML
1680          * @return string less-or-more HTML with NOPARSE bits
1681          */
1682         function armorLinks( $text ) {
1683                 return preg_replace( "/\b(" . wfUrlProtocols() . ')/',
1684                         "{$this->mUniqPrefix}NOPARSE$1", $text );
1685         }
1686
1687         /**
1688          * Return true if subpage links should be expanded on this page.
1689          * @return bool
1690          */
1691         function areSubpagesAllowed() {
1692                 # Some namespaces don't allow subpages
1693                 global $wgNamespacesWithSubpages;
1694                 return !empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()]);
1695         }
1696
1697         /**
1698          * Handle link to subpage if necessary
1699          * @param string $target the source of the link
1700          * @param string &$text the link text, modified as necessary
1701          * @return string the full name of the link
1702          * @access private
1703          */
1704         function maybeDoSubpageLink($target, &$text) {
1705                 # Valid link forms:
1706                 # Foobar -- normal
1707                 # :Foobar -- override special treatment of prefix (images, language links)
1708                 # /Foobar -- convert to CurrentPage/Foobar
1709                 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1710                 # ../ -- convert to CurrentPage, from CurrentPage/CurrentSubPage
1711                 # ../Foobar -- convert to CurrentPage/Foobar, from CurrentPage/CurrentSubPage
1712
1713                 $fname = 'Parser::maybeDoSubpageLink';
1714                 wfProfileIn( $fname );
1715                 $ret = $target; # default return value is no change
1716
1717                 # Some namespaces don't allow subpages,
1718                 # so only perform processing if subpages are allowed
1719                 if( $this->areSubpagesAllowed() ) {
1720                         # Look at the first character
1721                         if( $target != '' && $target{0} == '/' ) {
1722                                 # / at end means we don't want the slash to be shown
1723                                 if( substr( $target, -1, 1 ) == '/' ) {
1724                                         $target = substr( $target, 1, -1 );
1725                                         $noslash = $target;
1726                                 } else {
1727                                         $noslash = substr( $target, 1 );
1728                                 }
1729
1730                                 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1731                                 if( '' === $text ) {
1732                                         $text = $target;
1733                                 } # this might be changed for ugliness reasons
1734                         } else {
1735                                 # check for .. subpage backlinks
1736                                 $dotdotcount = 0;
1737                                 $nodotdot = $target;
1738                                 while( strncmp( $nodotdot, "../", 3 ) == 0 ) {
1739                                         ++$dotdotcount;
1740                                         $nodotdot = substr( $nodotdot, 3 );
1741                                 }
1742                                 if($dotdotcount > 0) {
1743                                         $exploded = explode( '/', $this->mTitle->GetPrefixedText() );
1744                                         if( count( $exploded ) > $dotdotcount ) { # not allowed to go below top level page
1745                                                 $ret = implode( '/', array_slice( $exploded, 0, -$dotdotcount ) );
1746                                                 # / at the end means don't show full path
1747                                                 if( substr( $nodotdot, -1, 1 ) == '/' ) {
1748                                                         $nodotdot = substr( $nodotdot, 0, -1 );
1749                                                         if( '' === $text ) {
1750                                                                 $text = $nodotdot;
1751                                                         }
1752                                                 }
1753                                                 $nodotdot = trim( $nodotdot );
1754                                                 if( $nodotdot != '' ) {
1755                                                         $ret .= '/' . $nodotdot;
1756                                                 }
1757                                         }
1758                                 }
1759                         }
1760                 }
1761
1762                 wfProfileOut( $fname );
1763                 return $ret;
1764         }
1765
1766         /**#@+
1767          * Used by doBlockLevels()
1768          * @access private
1769          */
1770         /* private */ function closeParagraph() {
1771                 $result = '';
1772                 if ( '' != $this->mLastSection ) {
1773                         $result = '</' . $this->mLastSection  . ">\n";
1774                 }
1775                 $this->mInPre = false;
1776                 $this->mLastSection = '';
1777                 return $result;
1778         }
1779         # getCommon() returns the length of the longest common substring
1780         # of both arguments, starting at the beginning of both.
1781         #
1782         /* private */ function getCommon( $st1, $st2 ) {
1783                 $fl = strlen( $st1 );
1784                 $shorter = strlen( $st2 );
1785                 if ( $fl < $shorter ) { $shorter = $fl; }
1786
1787                 for ( $i = 0; $i < $shorter; ++$i ) {
1788                         if ( $st1{$i} != $st2{$i} ) { break; }
1789                 }
1790                 return $i;
1791         }
1792         # These next three functions open, continue, and close the list
1793         # element appropriate to the prefix character passed into them.
1794         #
1795         /* private */ function openList( $char ) {
1796                 $result = $this->closeParagraph();
1797
1798                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1799                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1800                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1801                 else if ( ';' == $char ) {
1802                         $result .= '<dl><dt>';
1803                         $this->mDTopen = true;
1804                 }
1805                 else { $result = '<!-- ERR 1 -->'; }
1806
1807                 return $result;
1808         }
1809
1810         /* private */ function nextItem( $char ) {
1811                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1812                 else if ( ':' == $char || ';' == $char ) {
1813                         $close = '</dd>';
1814                         if ( $this->mDTopen ) { $close = '</dt>'; }
1815                         if ( ';' == $char ) {
1816                                 $this->mDTopen = true;
1817                                 return $close . '<dt>';
1818                         } else {
1819                                 $this->mDTopen = false;
1820                                 return $close . '<dd>';
1821                         }
1822                 }
1823                 return '<!-- ERR 2 -->';
1824         }
1825
1826         /* private */ function closeList( $char ) {
1827                 if ( '*' == $char ) { $text = '</li></ul>'; }
1828                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1829                 else if ( ':' == $char ) {
1830                         if ( $this->mDTopen ) {
1831                                 $this->mDTopen = false;
1832                                 $text = '</dt></dl>';
1833                         } else {
1834                                 $text = '</dd></dl>';
1835                         }
1836                 }
1837                 else {  return '<!-- ERR 3 -->'; }
1838                 return $text."\n";
1839         }
1840         /**#@-*/
1841
1842         /**
1843          * Make lists from lines starting with ':', '*', '#', etc.
1844          *
1845          * @access private
1846          * @return string the lists rendered as HTML
1847          */
1848         function doBlockLevels( $text, $linestart ) {
1849                 $fname = 'Parser::doBlockLevels';
1850                 wfProfileIn( $fname );
1851
1852                 # Parsing through the text line by line.  The main thing
1853                 # happening here is handling of block-level elements p, pre,
1854                 # and making lists from lines starting with * # : etc.
1855                 #
1856                 $textLines = explode( "\n", $text );
1857
1858                 $lastPrefix = $output = '';
1859                 $this->mDTopen = $inBlockElem = false;
1860                 $prefixLength = 0;
1861                 $paragraphStack = false;
1862
1863                 if ( !$linestart ) {
1864                         $output .= array_shift( $textLines );
1865                 }
1866                 foreach ( $textLines as $oLine ) {
1867                         $lastPrefixLength = strlen( $lastPrefix );
1868                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1869                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1870                         if ( !$this->mInPre ) {
1871                                 # Multiple prefixes may abut each other for nested lists.
1872                                 $prefixLength = strspn( $oLine, '*#:;' );
1873                                 $pref = substr( $oLine, 0, $prefixLength );
1874
1875                                 # eh?
1876                                 $pref2 = str_replace( ';', ':', $pref );
1877                                 $t = substr( $oLine, $prefixLength );
1878                                 $this->mInPre = !empty($preOpenMatch);
1879                         } else {
1880                                 # Don't interpret any other prefixes in preformatted text
1881                                 $prefixLength = 0;
1882                                 $pref = $pref2 = '';
1883                                 $t = $oLine;
1884                         }
1885
1886                         # List generation
1887                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1888                                 # Same as the last item, so no need to deal with nesting or opening stuff
1889                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1890                                 $paragraphStack = false;
1891
1892                                 if ( substr( $pref, -1 ) == ';') {
1893                                         # The one nasty exception: definition lists work like this:
1894                                         # ; title : definition text
1895                                         # So we check for : in the remainder text to split up the
1896                                         # title and definition, without b0rking links.
1897                                         $term = $t2 = '';
1898                                         if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1899                                                 $t = $t2;
1900                                                 $output .= $term . $this->nextItem( ':' );
1901                                         }
1902                                 }
1903                         } elseif( $prefixLength || $lastPrefixLength ) {
1904                                 # Either open or close a level...
1905                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1906                                 $paragraphStack = false;
1907
1908                                 while( $commonPrefixLength < $lastPrefixLength ) {
1909                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1910                                         --$lastPrefixLength;
1911                                 }
1912                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1913                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1914                                 }
1915                                 while ( $prefixLength > $commonPrefixLength ) {
1916                                         $char = substr( $pref, $commonPrefixLength, 1 );
1917                                         $output .= $this->openList( $char );
1918
1919                                         if ( ';' == $char ) {
1920                                                 # FIXME: This is dupe of code above
1921                                                 if ($this->findColonNoLinks($t, $term, $t2) !== false) {
1922                                                         $t = $t2;
1923                                                         $output .= $term . $this->nextItem( ':' );
1924                                                 }
1925                                         }
1926                                         ++$commonPrefixLength;
1927                                 }
1928                                 $lastPrefix = $pref2;
1929                         }
1930                         if( 0 == $prefixLength ) {
1931                                 wfProfileIn( "$fname-paragraph" );
1932                                 # No prefix (not in list)--go to paragraph mode
1933                                 // XXX: use a stack for nestable elements like span, table and div
1934                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/iS', $t );
1935                                 $closematch = preg_match(
1936                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1937                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul)/iS', $t );
1938                                 if ( $openmatch or $closematch ) {
1939                                         $paragraphStack = false;
1940                                         $output .= $this->closeParagraph();
1941                                         if ( $preOpenMatch and !$preCloseMatch ) {
1942                                                 $this->mInPre = true;
1943                                         }
1944                                         if ( $closematch ) {
1945                                                 $inBlockElem = false;
1946                                         } else {
1947                                                 $inBlockElem = true;
1948                                         }
1949                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1950                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1951                                                 // pre
1952                                                 if ($this->mLastSection != 'pre') {
1953                                                         $paragraphStack = false;
1954                                                         $output .= $this->closeParagraph().'<pre>';
1955                                                         $this->mLastSection = 'pre';
1956                                                 }
1957                                                 $t = substr( $t, 1 );
1958                                         } else {
1959                                                 // paragraph
1960                                                 if ( '' == trim($t) ) {
1961                                                         if ( $paragraphStack ) {
1962                                                                 $output .= $paragraphStack.'<br />';
1963                                                                 $paragraphStack = false;
1964                                                                 $this->mLastSection = 'p';
1965                                                         } else {
1966                                                                 if ($this->mLastSection != 'p' ) {
1967                                                                         $output .= $this->closeParagraph();
1968                                                                         $this->mLastSection = '';
1969                                                                         $paragraphStack = '<p>';
1970                                                                 } else {
1971                                                                         $paragraphStack = '</p><p>';
1972                                                                 }
1973                                                         }
1974                                                 } else {
1975                                                         if ( $paragraphStack ) {
1976                                                                 $output .= $paragraphStack;
1977                                                                 $paragraphStack = false;
1978                                                                 $this->mLastSection = 'p';
1979                                                         } else if ($this->mLastSection != 'p') {
1980                                                                 $output .= $this->closeParagraph().'<p>';
1981                                                                 $this->mLastSection = 'p';
1982                                                         }
1983                                                 }
1984                                         }
1985                                 }
1986                                 wfProfileOut( "$fname-paragraph" );
1987                         }
1988                         // somewhere above we forget to get out of pre block (bug 785)
1989                         if($preCloseMatch && $this->mInPre) {
1990                                 $this->mInPre = false;
1991                         }
1992                         if ($paragraphStack === false) {
1993                                 $output .= $t."\n";
1994                         }
1995                 }
1996                 while ( $prefixLength ) {
1997                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1998                         --$prefixLength;
1999                 }
2000                 if ( '' != $this->mLastSection ) {
2001                         $output .= '</' . $this->mLastSection . '>';
2002                         $this->mLastSection = '';
2003                 }
2004
2005                 wfProfileOut( $fname );
2006                 return $output;
2007         }
2008
2009         /**
2010          * Split up a string on ':', ignoring any occurences inside
2011          * <a>..</a> or <span>...</span>
2012          * @param string $str the string to split
2013          * @param string &$before set to everything before the ':'
2014          * @param string &$after set to everything after the ':'
2015          * return string the position of the ':', or false if none found
2016          */
2017         function findColonNoLinks($str, &$before, &$after) {
2018                 # I wonder if we should make this count all tags, not just <a>
2019                 # and <span>. That would prevent us from matching a ':' that
2020                 # comes in the middle of italics other such formatting....
2021                 # -- Wil
2022                 $fname = 'Parser::findColonNoLinks';
2023                 wfProfileIn( $fname );
2024                 $pos = 0;
2025                 do {
2026                         $colon = strpos($str, ':', $pos);
2027
2028                         if ($colon !== false) {
2029                                 $before = substr($str, 0, $colon);
2030                                 $after = substr($str, $colon + 1);
2031
2032                                 # Skip any ':' within <a> or <span> pairs
2033                                 $a = substr_count($before, '<a');
2034                                 $s = substr_count($before, '<span');
2035                                 $ca = substr_count($before, '</a>');
2036                                 $cs = substr_count($before, '</span>');
2037
2038                                 if ($a <= $ca and $s <= $cs) {
2039                                         # Tags are balanced before ':'; ok
2040                                         break;
2041                                 }
2042                                 $pos = $colon + 1;
2043                         }
2044                 } while ($colon !== false);
2045                 wfProfileOut( $fname );
2046                 return $colon;
2047         }
2048
2049         /**
2050          * Return value of a magic variable (like PAGENAME)
2051          *
2052          * @access private
2053          */
2054         function getVariableValue( $index ) {
2055                 global $wgContLang, $wgSitename, $wgServer, $wgServerName, $wgScriptPath;
2056
2057                 /**
2058                  * Some of these require message or data lookups and can be
2059                  * expensive to check many times.
2060                  */
2061                 static $varCache = array();
2062                 if ( wfRunHooks( 'ParserGetVariableValueVarCache', array( &$this, &$varCache ) ) )
2063                         if ( isset( $varCache[$index] ) )
2064                                 return $varCache[$index];
2065
2066                 $ts = time();
2067                 wfRunHooks( 'ParserGetVariableValueTs', array( &$this, &$ts ) );
2068
2069                 switch ( $index ) {
2070                         case MAG_CURRENTMONTH:
2071                                 return $varCache[$index] = $wgContLang->formatNum( date( 'm', $ts ) );
2072                         case MAG_CURRENTMONTHNAME:
2073                                 return $varCache[$index] = $wgContLang->getMonthName( date( 'n', $ts ) );
2074                         case MAG_CURRENTMONTHNAMEGEN:
2075                                 return $varCache[$index] = $wgContLang->getMonthNameGen( date( 'n', $ts ) );
2076                         case MAG_CURRENTMONTHABBREV:
2077                                 return $varCache[$index] = $wgContLang->getMonthAbbreviation( date( 'n', $ts ) );
2078                         case MAG_CURRENTDAY:
2079                                 return $varCache[$index] = $wgContLang->formatNum( date( 'j', $ts ) );
2080                         case MAG_CURRENTDAY2:
2081                                 return $varCache[$index] = $wgContLang->formatNum( date( 'd', $ts ) );
2082                         case MAG_PAGENAME:
2083                                 return $this->mTitle->getText();
2084                         case MAG_PAGENAMEE:
2085                                 return $this->mTitle->getPartialURL();
2086                         case MAG_FULLPAGENAME:
2087                                 return $this->mTitle->getPrefixedText();
2088                         case MAG_FULLPAGENAMEE:
2089                                 return $this->mTitle->getPrefixedURL();
2090                         case MAG_SUBPAGENAME:
2091                                 return $this->mTitle->getSubpageText();
2092                         case MAG_SUBPAGENAMEE:
2093                                 return $this->mTitle->getSubpageUrlForm();
2094                         case MAG_REVISIONID:
2095                                 return $this->mRevisionId;
2096                         case MAG_NAMESPACE:
2097                                 return $wgContLang->getNsText( $this->mTitle->getNamespace() );
2098                         case MAG_NAMESPACEE:
2099                                 return wfUrlencode( $wgContLang->getNsText( $this->mTitle->getNamespace() ) );
2100                         case MAG_CURRENTDAYNAME:
2101                                 return $varCache[$index] = $wgContLang->getWeekdayName( date( 'w', $ts ) + 1 );
2102                         case MAG_CURRENTYEAR:
2103                                 return $varCache[$index] = $wgContLang->formatNum( date( 'Y', $ts ), true );
2104                         case MAG_CURRENTTIME:
2105                                 return $varCache[$index] = $wgContLang->time( wfTimestamp( TS_MW, $ts ), false, false );
2106                         case MAG_CURRENTWEEK:
2107                                 // @bug 4594 PHP5 has it zero padded, PHP4 does not, cast to
2108                                 // int to remove the padding
2109                                 return $varCache[$index] = $wgContLang->formatNum( (int)date( 'W', $ts ) );
2110                         case MAG_CURRENTDOW:
2111                                 return $varCache[$index] = $wgContLang->formatNum( date( 'w', $ts ) );
2112                         case MAG_NUMBEROFARTICLES:
2113                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfArticles() );
2114                         case MAG_NUMBEROFFILES:
2115                                 return $varCache[$index] = $wgContLang->formatNum( wfNumberOfFiles() );
2116                         case MAG_SITENAME:
2117                                 return $wgSitename;
2118                         case MAG_SERVER:
2119                                 return $wgServer;
2120                         case MAG_SERVERNAME:
2121                                 return $wgServerName;
2122                         case MAG_SCRIPTPATH:
2123                                 return $wgScriptPath;
2124                         default:
2125                                 $ret = null;
2126                                 if ( wfRunHooks( 'ParserGetVariableValueSwitch', array( &$this, &$varCache, &$index, &$ret ) ) )
2127                                         return $ret;
2128                                 else
2129                                         return null;
2130                 }
2131         }
2132
2133         /**
2134          * initialise the magic variables (like CURRENTMONTHNAME)
2135          *
2136          * @access private
2137          */
2138         function initialiseVariables() {
2139                 $fname = 'Parser::initialiseVariables';
2140                 wfProfileIn( $fname );
2141                 global $wgVariableIDs;
2142                 $this->mVariables = array();
2143                 foreach ( $wgVariableIDs as $id ) {
2144                         $mw =& MagicWord::get( $id );
2145                         $mw->addToArray( $this->mVariables, $id );
2146                 }
2147                 wfProfileOut( $fname );
2148         }
2149
2150         /**
2151          * parse any parentheses in format ((title|part|part))
2152          * and call callbacks to get a replacement text for any found piece
2153          *
2154          * @param string $text The text to parse
2155          * @param array $callbacks rules in form:
2156          *     '{' => array(                            # opening parentheses
2157          *                                      'end' => '}',   # closing parentheses
2158          *                                      'cb' => array(2 => callback,    # replacement callback to call if {{..}} is found
2159          *                                                                4 => callback         # replacement callback to call if {{{{..}}}} is found
2160          *                                                                )
2161          *                                      )
2162          * @access private
2163          */
2164         function replace_callback ($text, $callbacks) {
2165                 $openingBraceStack = array();   # this array will hold a stack of parentheses which are not closed yet
2166                 $lastOpeningBrace = -1;                 # last not closed parentheses
2167
2168                 for ($i = 0; $i < strlen($text); $i++) {
2169                         # check for any opening brace
2170                         $rule = null;
2171                         $nextPos = -1;
2172                         foreach ($callbacks as $key => $value) {
2173                                 $pos = strpos ($text, $key, $i);
2174                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)) {
2175                                         $rule = $value;
2176                                         $nextPos = $pos;
2177                                 }
2178                         }
2179
2180                         if ($lastOpeningBrace >= 0) {
2181                                 $pos = strpos ($text, $openingBraceStack[$lastOpeningBrace]['braceEnd'], $i);
2182
2183                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)){
2184                                         $rule = null;
2185                                         $nextPos = $pos;
2186                                 }
2187
2188                                 $pos = strpos ($text, '|', $i);
2189
2190                                 if (false !== $pos && (-1 == $nextPos || $pos < $nextPos)){
2191                                         $rule = null;
2192                                         $nextPos = $pos;
2193                                 }
2194                         }
2195
2196                         if ($nextPos == -1)
2197                                 break;
2198
2199                         $i = $nextPos;
2200
2201                         # found openning brace, lets add it to parentheses stack
2202                         if (null != $rule) {
2203                                 $piece = array('brace' => $text[$i],
2204                                                            'braceEnd' => $rule['end'],
2205                                                            'count' => 1,
2206                                                            'title' => '',
2207                                                            'parts' => null);
2208
2209                                 # count openning brace characters
2210                                 while ($i+1 < strlen($text) && $text[$i+1] == $piece['brace']) {
2211                                         $piece['count']++;
2212                                         $i++;
2213                                 }
2214
2215                                 $piece['startAt'] = $i+1;
2216                                 $piece['partStart'] = $i+1;
2217
2218                                 # we need to add to stack only if openning brace count is enough for any given rule
2219                                 foreach ($rule['cb'] as $cnt => $fn) {
2220                                         if ($piece['count'] >= $cnt) {
2221                                                 $lastOpeningBrace ++;
2222                                                 $openingBraceStack[$lastOpeningBrace] = $piece;
2223                                                 break;
2224                                         }
2225                                 }
2226
2227                                 continue;
2228                         }
2229                         else if ($lastOpeningBrace >= 0) {
2230                                 # first check if it is a closing brace
2231                                 if ($openingBraceStack[$lastOpeningBrace]['braceEnd'] == $text[$i]) {
2232                                         # lets check if it is enough characters for closing brace
2233                                         $count = 1;
2234                                         while ($i+$count < strlen($text) && $text[$i+$count] == $text[$i])
2235                                                 $count++;
2236
2237                                         # if there are more closing parentheses than opening ones, we parse less
2238                                         if ($openingBraceStack[$lastOpeningBrace]['count'] < $count)
2239                                                 $count = $openingBraceStack[$lastOpeningBrace]['count'];
2240
2241                                         # check for maximum matching characters (if there are 5 closing characters, we will probably need only 3 - depending on the rules)
2242                                         $matchingCount = 0;
2243                                         $matchingCallback = null;
2244                                         foreach ($callbacks[$openingBraceStack[$lastOpeningBrace]['brace']]['cb'] as $cnt => $fn) {
2245                                                 if ($count >= $cnt && $matchingCount < $cnt) {
2246                                                         $matchingCount = $cnt;
2247                                                         $matchingCallback = $fn;
2248                                                 }
2249                                         }
2250
2251                                         if ($matchingCount == 0) {
2252                                                 $i += $count - 1;
2253                                                 continue;
2254                                         }
2255
2256                                         # lets set a title or last part (if '|' was found)
2257                                         if (null === $openingBraceStack[$lastOpeningBrace]['parts'])
2258                                                 $openingBraceStack[$lastOpeningBrace]['title'] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2259                                         else
2260                                                 $openingBraceStack[$lastOpeningBrace]['parts'][] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2261
2262                                         $pieceStart = $openingBraceStack[$lastOpeningBrace]['startAt'] - $matchingCount;
2263                                         $pieceEnd = $i + $matchingCount;
2264
2265                                         if( is_callable( $matchingCallback ) ) {
2266                                                 $cbArgs = array (
2267                                                                                  'text' => substr($text, $pieceStart, $pieceEnd - $pieceStart),
2268                                                                                  'title' => trim($openingBraceStack[$lastOpeningBrace]['title']),
2269                                                                                  'parts' => $openingBraceStack[$lastOpeningBrace]['parts'],
2270                                                                                  'lineStart' => (($pieceStart > 0) && ($text[$pieceStart-1] == '\n')),
2271                                                                                  );
2272                                                 # finally we can call a user callback and replace piece of text
2273                                                 $replaceWith = call_user_func( $matchingCallback, $cbArgs );
2274                                                 $text = substr($text, 0, $pieceStart) . $replaceWith . substr($text, $pieceEnd);
2275                                                 $i = $pieceStart + strlen($replaceWith) - 1;
2276                                         }
2277                                         else {
2278                                                 # null value for callback means that parentheses should be parsed, but not replaced
2279                                                 $i += $matchingCount - 1;
2280                                         }
2281
2282                                         # reset last openning parentheses, but keep it in case there are unused characters
2283                                         $piece = array('brace' => $openingBraceStack[$lastOpeningBrace]['brace'],
2284                                                                    'braceEnd' => $openingBraceStack[$lastOpeningBrace]['braceEnd'],
2285                                                                    'count' => $openingBraceStack[$lastOpeningBrace]['count'],
2286                                                                    'title' => '',
2287                                                                    'parts' => null,
2288                                                                    'startAt' => $openingBraceStack[$lastOpeningBrace]['startAt']);
2289                                         $openingBraceStack[$lastOpeningBrace--] = null;
2290
2291                                         if ($matchingCount < $piece['count']) {
2292                                                 $piece['count'] -= $matchingCount;
2293                                                 $piece['startAt'] -= $matchingCount;
2294                                                 $piece['partStart'] = $piece['startAt'];
2295                                                 # do we still qualify for any callback with remaining count?
2296                                                 foreach ($callbacks[$piece['brace']]['cb'] as $cnt => $fn) {
2297                                                         if ($piece['count'] >= $cnt) {
2298                                                                 $lastOpeningBrace ++;
2299                                                                 $openingBraceStack[$lastOpeningBrace] = $piece;
2300                                                                 break;
2301                                                         }
2302                                                 }
2303                                         }
2304                                         continue;
2305                                 }
2306
2307                                 # lets set a title if it is a first separator, or next part otherwise
2308                                 if ($text[$i] == '|') {
2309                                         if (null === $openingBraceStack[$lastOpeningBrace]['parts']) {
2310                                                 $openingBraceStack[$lastOpeningBrace]['title'] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2311                                                 $openingBraceStack[$lastOpeningBrace]['parts'] = array();
2312                                         }
2313                                         else
2314                                                 $openingBraceStack[$lastOpeningBrace]['parts'][] = substr($text, $openingBraceStack[$lastOpeningBrace]['partStart'], $i - $openingBraceStack[$lastOpeningBrace]['partStart']);
2315
2316                                         $openingBraceStack[$lastOpeningBrace]['partStart'] = $i + 1;
2317                                 }
2318                         }
2319                 }
2320
2321                 return $text;
2322         }
2323
2324         /**
2325          * Replace magic variables, templates, and template arguments
2326          * with the appropriate text. Templates are substituted recursively,
2327          * taking care to avoid infinite loops.
2328          *
2329          * Note that the substitution depends on value of $mOutputType:
2330          *  OT_WIKI: only {{subst:}} templates
2331          *  OT_MSG: only magic variables
2332          *  OT_HTML: all templates and magic variables
2333          *
2334          * @param string $tex The text to transform
2335          * @param array $args Key-value pairs representing template parameters to substitute
2336          * @param bool $argsOnly Only do argument (triple-brace) expansion, not double-brace expansion
2337          * @access private
2338          */
2339         function replaceVariables( $text, $args = array(), $argsOnly = false ) {
2340                 # Prevent too big inclusions
2341                 if( strlen( $text ) > MAX_INCLUDE_SIZE ) {
2342                         return $text;
2343                 }
2344
2345                 $fname = 'Parser::replaceVariables';
2346                 wfProfileIn( $fname );
2347
2348                 # This function is called recursively. To keep track of arguments we need a stack:
2349                 array_push( $this->mArgStack, $args );
2350
2351                 $braceCallbacks = array();
2352                 if ( !$argsOnly ) {
2353                         $braceCallbacks[2] = array( &$this, 'braceSubstitution' );
2354                 }
2355                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
2356                         $braceCallbacks[3] = array( &$this, 'argSubstitution' );
2357                 }
2358                 $callbacks = array();
2359                 $callbacks['{'] = array('end' => '}', 'cb' => $braceCallbacks);
2360                 $callbacks['['] = array('end' => ']', 'cb' => array(2=>null));
2361                 $text = $this->replace_callback ($text, $callbacks);
2362
2363                 array_pop( $this->mArgStack );
2364
2365                 wfProfileOut( $fname );
2366                 return $text;
2367         }
2368
2369         /**
2370          * Replace magic variables
2371          * @access private
2372          */
2373         function variableSubstitution( $matches ) {
2374                 $fname = 'Parser::variableSubstitution';
2375                 $varname = $matches[1];
2376                 wfProfileIn( $fname );
2377                 if ( !$this->mVariables ) {
2378                         $this->initialiseVariables();
2379                 }
2380                 $skip = false;
2381                 if ( $this->mOutputType == OT_WIKI ) {
2382                         # Do only magic variables prefixed by SUBST
2383                         $mwSubst =& MagicWord::get( MAG_SUBST );
2384                         if (!$mwSubst->matchStartAndRemove( $varname ))
2385                                 $skip = true;
2386                         # Note that if we don't substitute the variable below,
2387                         # we don't remove the {{subst:}} magic word, in case
2388                         # it is a template rather than a magic variable.
2389                 }
2390                 if ( !$skip && array_key_exists( $varname, $this->mVariables ) ) {
2391                         $id = $this->mVariables[$varname];
2392                         $text = $this->getVariableValue( $id );
2393                         $this->mOutput->mContainsOldMagic = true;
2394                 } else {
2395                         $text = $matches[0];
2396                 }
2397                 wfProfileOut( $fname );
2398                 return $text;
2399         }
2400
2401         # Split template arguments
2402         function getTemplateArgs( $argsString ) {
2403                 if ( $argsString === '' ) {
2404                         return array();
2405                 }
2406
2407                 $args = explode( '|', substr( $argsString, 1 ) );
2408
2409                 # If any of the arguments contains a '[[' but no ']]', it needs to be
2410                 # merged with the next arg because the '|' character between belongs
2411                 # to the link syntax and not the template parameter syntax.
2412                 $argc = count($args);
2413
2414                 for ( $i = 0; $i < $argc-1; $i++ ) {
2415                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
2416                                 $args[$i] .= '|'.$args[$i+1];
2417                                 array_splice($args, $i+1, 1);
2418                                 $i--;
2419                                 $argc--;
2420                         }
2421                 }
2422
2423                 return $args;
2424         }
2425
2426         /**
2427          * Return the text of a template, after recursively
2428          * replacing any variables or templates within the template.
2429          *
2430          * @param array $piece The parts of the template
2431          *  $piece['text']: matched text
2432          *  $piece['title']: the title, i.e. the part before the |
2433          *  $piece['parts']: the parameter array
2434          * @return string the text of the template
2435          * @access private
2436          */
2437         function braceSubstitution( $piece ) {
2438                 global $wgContLang;
2439                 $fname = 'Parser::braceSubstitution';
2440                 wfProfileIn( $fname );
2441
2442                 # Flags
2443                 $found = false;             # $text has been filled
2444                 $nowiki = false;            # wiki markup in $text should be escaped
2445                 $noparse = false;           # Unsafe HTML tags should not be stripped, etc.
2446                 $noargs = false;            # Don't replace triple-brace arguments in $text
2447                 $replaceHeadings = false;   # Make the edit section links go to the template not the article
2448                 $isHTML = false;            # $text is HTML, armour it against wikitext transformation
2449                 $forceRawInterwiki = false; # Force interwiki transclusion to be done in raw mode not rendered
2450
2451                 # Title object, where $text came from
2452                 $title = NULL;
2453
2454                 $linestart = '';
2455
2456                 # $part1 is the bit before the first |, and must contain only title characters
2457                 # $args is a list of arguments, starting from index 0, not including $part1
2458
2459                 $part1 = $piece['title'];
2460                 # If the third subpattern matched anything, it will start with |
2461
2462                 if (null == $piece['parts']) {
2463                         $replaceWith = $this->variableSubstitution (array ($piece['text'], $piece['title']));
2464                         if ($replaceWith != $piece['text']) {
2465                                 $text = $replaceWith;
2466                                 $found = true;
2467                                 $noparse = true;
2468                                 $noargs = true;
2469                         }
2470                 }
2471
2472                 $args = (null == $piece['parts']) ? array() : $piece['parts'];
2473                 $argc = count( $args );
2474
2475                 # SUBST
2476                 if ( !$found ) {
2477                         $mwSubst =& MagicWord::get( MAG_SUBST );
2478                         if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
2479                                 # One of two possibilities is true:
2480                                 # 1) Found SUBST but not in the PST phase
2481                                 # 2) Didn't find SUBST and in the PST phase
2482                                 # In either case, return without further processing
2483                                 $text = $piece['text'];
2484                                 $found = true;
2485                                 $noparse = true;
2486                                 $noargs = true;
2487                         }
2488                 }
2489
2490                 # MSG, MSGNW, INT and RAW
2491                 if ( !$found ) {
2492                         # Check for MSGNW:
2493                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
2494                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
2495                                 $nowiki = true;
2496                         } else {
2497                                 # Remove obsolete MSG:
2498                                 $mwMsg =& MagicWord::get( MAG_MSG );
2499                                 $mwMsg->matchStartAndRemove( $part1 );
2500                         }
2501
2502                         # Check for RAW:
2503                         $mwRaw =& MagicWord::get( MAG_RAW );
2504                         if ( $mwRaw->matchStartAndRemove( $part1 ) ) {
2505                                 $forceRawInterwiki = true;
2506                         }
2507
2508                         # Check if it is an internal message
2509                         $mwInt =& MagicWord::get( MAG_INT );
2510                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
2511                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
2512                                         $text = $linestart . wfMsgReal( $part1, $args, true );
2513                                         $found = true;
2514                                 }
2515                         }
2516                 }
2517
2518                 # NS
2519                 if ( !$found ) {
2520                         # Check for NS: (namespace expansion)
2521                         $mwNs = MagicWord::get( MAG_NS );
2522                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
2523                                 if ( intval( $part1 ) || $part1 == "0" ) {
2524                                         $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
2525                                         $found = true;
2526                                 } else {
2527                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
2528                                         if ( !is_null( $index ) ) {
2529                                                 $text = $linestart . $wgContLang->getNsText( $index );
2530                                                 $found = true;
2531                                         }
2532                                 }
2533                         }
2534                 }
2535
2536                 # LCFIRST, UCFIRST, LC and UC
2537                 if ( !$found ) {
2538                         $lcfirst =& MagicWord::get( MAG_LCFIRST );
2539                         $ucfirst =& MagicWord::get( MAG_UCFIRST );
2540                         $lc =& MagicWord::get( MAG_LC );
2541                         $uc =& MagicWord::get( MAG_UC );
2542                         if ( $lcfirst->matchStartAndRemove( $part1 ) ) {
2543                                 $text = $linestart . $wgContLang->lcfirst( $part1 );
2544                                 $found = true;
2545                         } else if ( $ucfirst->matchStartAndRemove( $part1 ) ) {
2546                                 $text = $linestart . $wgContLang->ucfirst( $part1 );
2547                                 $found = true;
2548                         } else if ( $lc->matchStartAndRemove( $part1 ) ) {
2549                                 $text = $linestart . $wgContLang->lc( $part1 );
2550                                 $found = true;
2551                         } else if ( $uc->matchStartAndRemove( $part1 ) ) {
2552                                  $text = $linestart . $wgContLang->uc( $part1 );
2553                                  $found = true;
2554                         }
2555                 }
2556
2557                 # LOCALURL and FULLURL
2558                 if ( !$found ) {
2559                         $mwLocal =& MagicWord::get( MAG_LOCALURL );
2560                         $mwLocalE =& MagicWord::get( MAG_LOCALURLE );
2561                         $mwFull =& MagicWord::get( MAG_FULLURL );
2562                         $mwFullE =& MagicWord::get( MAG_FULLURLE );
2563
2564
2565                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
2566                                 $func = 'getLocalURL';
2567                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
2568                                 $func = 'escapeLocalURL';
2569                         } elseif ( $mwFull->matchStartAndRemove( $part1 ) ) {
2570                                 $func = 'getFullURL';
2571                         } elseif ( $mwFullE->matchStartAndRemove( $part1 ) ) {
2572                                 $func = 'escapeFullURL';
2573                         } else {
2574                                 $func = false;
2575                         }
2576
2577                         if ( $func !== false ) {
2578                                 $title = Title::newFromText( $part1 );
2579                                 if ( !is_null( $title ) ) {
2580                                         if ( $argc > 0 ) {
2581                                                 $text = $linestart . $title->$func( $args[0] );
2582                                         } else {
2583                                                 $text = $linestart . $title->$func();
2584                                         }
2585                                         $found = true;
2586                                 }
2587                         }
2588                 }
2589
2590                 # GRAMMAR
2591                 if ( !$found && $argc == 1 ) {
2592                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
2593                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
2594                                 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
2595                                 $found = true;
2596                         }
2597                 }
2598
2599                 # PLURAL
2600                 if ( !$found && $argc >= 2 ) {
2601                         $mwPluralForm =& MagicWord::get( MAG_PLURAL );
2602                         if ( $mwPluralForm->matchStartAndRemove( $part1 ) ) {
2603                                 if ($argc==2) {$args[2]=$args[1];}
2604                                 $text = $linestart . $wgContLang->convertPlural( $part1, $args[0], $args[1], $args[2]);
2605                                 $found = true;
2606                         }
2607                 }
2608
2609                 # Template table test
2610
2611                 # Did we encounter this template already? If yes, it is in the cache
2612                 # and we need to check for loops.
2613                 if ( !$found && isset( $this->mTemplates[$piece['title']] ) ) {
2614                         $found = true;
2615
2616                         # Infinite loop test
2617                         if ( isset( $this->mTemplatePath[$part1] ) ) {
2618                                 $noparse = true;
2619                                 $noargs = true;
2620                                 $found = true;
2621                                 $text = $linestart .
2622                                         '{{' . $part1 . '}}' .
2623                                         '<!-- WARNING: template loop detected -->';
2624                                 wfDebug( "$fname: template loop broken at '$part1'\n" );
2625                         } else {
2626                                 # set $text to cached message.
2627                                 $text = $linestart . $this->mTemplates[$piece['title']];
2628                         }
2629                 }
2630
2631                 # Load from database
2632                 $lastPathLevel = $this->mTemplatePath;
2633                 if ( !$found ) {
2634                         $ns = NS_TEMPLATE;
2635                         # declaring $subpage directly in the function call
2636                         # does not work correctly with references and breaks
2637                         # {{/subpage}}-style inclusions
2638                         $subpage = '';
2639                         $part1 = $this->maybeDoSubpageLink( $part1, $subpage );
2640                         if ($subpage !== '') {
2641                                 $ns = $this->mTitle->getNamespace();
2642                         }
2643                         $title = Title::newFromText( $part1, $ns );
2644
2645                         if ( !is_null( $title ) ) {
2646                                 if ( !$title->isExternal() ) {
2647                                         # Check for excessive inclusion
2648                                         $dbk = $title->getPrefixedDBkey();
2649                                         if ( $this->incrementIncludeCount( $dbk ) ) {
2650                                                 if ( $title->getNamespace() == NS_SPECIAL && $this->mOptions->getAllowSpecialInclusion() ) {
2651                                                         # Capture special page output
2652                                                         $text = SpecialPage::capturePath( $title );
2653                                                         if ( is_string( $text ) ) {
2654                                                                 $found = true;
2655                                                                 $noparse = true;
2656                                                                 $noargs = true;
2657                                                                 $isHTML = true;
2658                                                                 $this->disableCache();
2659                                                         }
2660                                                 } else {
2661                                                         $articleContent = $this->fetchTemplate( $title );
2662                                                         if ( $articleContent !== false ) {
2663                                                                 $found = true;
2664                                                                 $text = $articleContent;
2665                                                                 $replaceHeadings = true;
2666                                                         }
2667                                                 }
2668                                         }
2669
2670                                         # If the title is valid but undisplayable, make a link to it
2671                                         if ( $this->mOutputType == OT_HTML && !$found ) {
2672                                                 $text = '[['.$title->getPrefixedText().']]';
2673                                                 $found = true;
2674                                         }
2675                                 } elseif ( $title->isTrans() ) {
2676                                         // Interwiki transclusion
2677                                         if ( $this->mOutputType == OT_HTML && !$forceRawInterwiki ) {
2678                                                 $text = $this->interwikiTransclude( $title, 'render' );
2679                                                 $isHTML = true;
2680                                                 $noparse = true;
2681                                         } else {
2682                                                 $text = $this->interwikiTransclude( $title, 'raw' );
2683                                                 $replaceHeadings = true;
2684                                         }
2685                                         $found = true;
2686                                 }
2687
2688                                 # Template cache array insertion
2689                                 # Use the original $piece['title'] not the mangled $part1, so that
2690                                 # modifiers such as RAW: produce separate cache entries
2691                                 if( $found ) {
2692                                         $this->mTemplates[$piece['title']] = $text;
2693                                         $text = $linestart . $text;
2694                                 }
2695                         }
2696                 }
2697
2698                 # Recursive parsing, escaping and link table handling
2699                 # Only for HTML output
2700                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
2701                         $text = wfEscapeWikiText( $text );
2702                 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found ) {
2703                         if ( !$noargs ) {
2704                                 # Clean up argument array
2705                                 $assocArgs = array();
2706                                 $index = 1;
2707                                 foreach( $args as $arg ) {
2708                                         $eqpos = strpos( $arg, '=' );
2709                                         if ( $eqpos === false ) {
2710                                                 $assocArgs[$index++] = $arg;
2711                                         } else {
2712                                                 $name = trim( substr( $arg, 0, $eqpos ) );
2713                                                 $value = trim( substr( $arg, $eqpos+1 ) );
2714                                                 if ( $value === false ) {
2715                                                         $value = '';
2716                                                 }
2717                                                 if ( $name !== false ) {
2718                                                         $assocArgs[$name] = $value;
2719                                                 }
2720                                         }
2721                                 }
2722
2723                                 # Add a new element to the templace recursion path
2724                                 $this->mTemplatePath[$part1] = 1;
2725                         }
2726
2727                         if ( !$noparse ) {
2728                                 # If there are any <onlyinclude> tags, only include them
2729                                 if ( in_string( '<onlyinclude>', $text ) && in_string( '</onlyinclude>', $text ) ) {
2730                                         preg_match_all( '/<onlyinclude>(.*?)\n?<\/onlyinclude>/s', $text, $m );
2731                                         $text = '';
2732                                         foreach ($m[1] as $piece)
2733                                                 $text .= $piece;
2734                                 }
2735                                 # Remove <noinclude> sections and <includeonly> tags
2736                                 $text = preg_replace( '/<noinclude>.*?<\/noinclude>/s', '', $text );
2737                                 $text = strtr( $text, array( '<includeonly>' => '' , '</includeonly>' => '' ) );
2738
2739                                 if( $this->mOutputType == OT_HTML ) {
2740                                         # Strip <nowiki>, <pre>, etc.
2741                                         $text = $this->strip( $text, $this->mStripState );
2742                                         $text = Sanitizer::removeHTMLtags( $text, array( &$this, 'replaceVariables' ), $assocArgs );
2743                                 }
2744                                 $text = $this->replaceVariables( $text, $assocArgs );
2745
2746                                 # If the template begins with a table or block-level
2747                                 # element, it should be treated as beginning a new line.
2748                                 if (!$piece['lineStart'] && preg_match('/^({\\||:|;|#|\*)/', $text)) {
2749                                         $text = "\n" . $text;
2750                                 }
2751                         } elseif ( !$noargs ) {
2752                                 # $noparse and !$noargs
2753                                 # Just replace the arguments, not any double-brace items
2754                                 # This is used for rendered interwiki transclusion
2755                                 $text = $this->replaceVariables( $text, $assocArgs, true );
2756                         }
2757                 }
2758                 # Prune lower levels off the recursion check path
2759                 $this->mTemplatePath = $lastPathLevel;
2760
2761                 if ( !$found ) {
2762                         wfProfileOut( $fname );
2763                         return $piece['text'];
2764                 } else {
2765                         if ( $isHTML ) {
2766                                 # Replace raw HTML by a placeholder
2767                                 # Add a blank line preceding, to prevent it from mucking up
2768                                 # immediately preceding headings
2769                                 $text = "\n\n" . $this->insertStripItem( $text, $this->mStripState );
2770                         } else {
2771                                 # replace ==section headers==
2772                                 # XXX this needs to go away once we have a better parser.
2773                                 if ( $this->mOutputType != OT_WIKI && $replaceHeadings ) {
2774                                         if( !is_null( $title ) )
2775                                                 $encodedname = base64_encode($title->getPrefixedDBkey());
2776                                         else
2777                                                 $encodedname = base64_encode("");
2778                                         $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
2779                                                 PREG_SPLIT_DELIM_CAPTURE);
2780                                         $text = '';
2781                                         $nsec = 0;
2782                                         for( $i = 0; $i < count($m); $i += 2 ) {
2783                                                 $text .= $m[$i];
2784                                                 if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
2785                                                 $hl = $m[$i + 1];
2786                                                 if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
2787                                                         $text .= $hl;
2788                                                         continue;
2789                                                 }
2790                                                 preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
2791                                                 $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
2792                                                         . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
2793
2794                                                 $nsec++;
2795                                         }
2796                                 }
2797                         }
2798                 }
2799
2800                 # Prune lower levels off the recursion check path
2801                 $this->mTemplatePath = $lastPathLevel;
2802
2803                 if ( !$found ) {
2804                         wfProfileOut( $fname );
2805                         return $piece['text'];
2806                 } else {
2807                         wfProfileOut( $fname );
2808                         return $text;
2809                 }
2810         }
2811
2812         /**
2813          * Fetch the unparsed text of a template and register a reference to it.
2814          */
2815         function fetchTemplate( $title ) {
2816                 $text = false;
2817                 // Loop to fetch the article, with up to 1 redirect
2818                 for ( $i = 0; $i < 2 && is_object( $title ); $i++ ) {
2819                         $rev = Revision::newFromTitle( $title );
2820                         $this->mOutput->addTemplate( $title, $title->getArticleID() );
2821                         if ( !$rev ) {
2822                                 break;
2823                         }
2824                         $text = $rev->getText();
2825                         if ( $text === false ) {
2826                                 break;
2827                         }
2828                         // Redirect?
2829                         $title = Title::newFromRedirect( $text );
2830                 }
2831                 return $text;
2832         }
2833
2834         /**
2835          * Transclude an interwiki link.
2836          */
2837         function interwikiTransclude( $title, $action ) {
2838                 global $wgEnableScaryTranscluding, $wgCanonicalNamespaceNames;
2839
2840                 if (!$wgEnableScaryTranscluding)
2841                         return wfMsg('scarytranscludedisabled');
2842
2843                 // The namespace will actually only be 0 or 10, depending on whether there was a leading :
2844                 // But we'll handle it generally anyway
2845                 if ( $title->getNamespace() ) {
2846                         // Use the canonical namespace, which should work anywhere
2847                         $articleName = $wgCanonicalNamespaceNames[$title->getNamespace()] . ':' . $title->getDBkey();
2848                 } else {
2849                         $articleName = $title->getDBkey();
2850                 }
2851
2852                 $url = str_replace('$1', urlencode($articleName), Title::getInterwikiLink($title->getInterwiki()));
2853                 $url .= "?action=$action";
2854                 if (strlen($url) > 255)
2855                         return wfMsg('scarytranscludetoolong');
2856                 return $this->fetchScaryTemplateMaybeFromCache($url);
2857         }
2858
2859         function fetchScaryTemplateMaybeFromCache($url) {
2860                 global $wgTranscludeCacheExpiry;
2861                 $dbr =& wfGetDB(DB_SLAVE);
2862                 $obj = $dbr->selectRow('transcache', array('tc_time', 'tc_contents'),
2863                                 array('tc_url' => $url));
2864                 if ($obj) {
2865                         $time = $obj->tc_time;
2866                         $text = $obj->tc_contents;
2867                         if ($time && time() < $time + $wgTranscludeCacheExpiry ) {
2868                                 return $text;
2869                         }
2870                 }
2871
2872                 $text = wfGetHTTP($url);
2873                 if (!$text)
2874                         return wfMsg('scarytranscludefailed', $url);
2875
2876                 $dbw =& wfGetDB(DB_MASTER);
2877                 $dbw->replace('transcache', array('tc_url'), array(
2878                         'tc_url' => $url,
2879                         'tc_time' => time(),
2880                         'tc_contents' => $text));
2881                 return $text;
2882         }
2883
2884
2885         /**
2886          * Triple brace replacement -- used for template arguments
2887          * @access private
2888          */
2889         function argSubstitution( $matches ) {
2890                 $arg = trim( $matches['title'] );
2891                 $text = $matches['text'];
2892                 $inputArgs = end( $this->mArgStack );
2893
2894                 if ( array_key_exists( $arg, $inputArgs ) ) {
2895                         $text = $inputArgs[$arg];
2896                 } else if ($this->mOutputType == OT_HTML && null != $matches['parts'] && count($matches['parts']) > 0) {
2897                         $text = $matches['parts'][0];
2898                 }
2899
2900                 return $text;
2901         }
2902
2903         /**
2904          * Returns true if the function is allowed to include this entity
2905          * @access private
2906          */
2907         function incrementIncludeCount( $dbk ) {
2908                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
2909                         $this->mIncludeCount[$dbk] = 0;
2910                 }
2911                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
2912                         return true;
2913                 } else {
2914                         return false;
2915                 }
2916         }
2917
2918         /**
2919          * This function accomplishes several tasks:
2920          * 1) Auto-number headings if that option is enabled
2921          * 2) Add an [edit] link to sections for logged in users who have enabled the option
2922          * 3) Add a Table of contents on the top for users who have enabled the option
2923          * 4) Auto-anchor headings
2924          *
2925          * It loops through all headlines, collects the necessary data, then splits up the
2926          * string and re-inserts the newly formatted headlines.
2927          *
2928          * @param string $text
2929          * @param boolean $isMain
2930          * @access private
2931          */
2932         function formatHeadings( $text, $isMain=true ) {
2933                 global $wgMaxTocLevel, $wgContLang;
2934
2935                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2936                 $doShowToc = true;
2937                 $forceTocHere = false;
2938                 if( !$this->mTitle->userCanEdit() ) {
2939                         $showEditLink = 0;
2940                 } else {
2941                         $showEditLink = $this->mOptions->getEditSection();
2942                 }
2943
2944                 # Inhibit editsection links if requested in the page
2945                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2946                 if( $esw->matchAndRemove( $text ) ) {
2947                         $showEditLink = 0;
2948                 }
2949                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2950                 # do not add TOC
2951                 $mw =& MagicWord::get( MAG_NOTOC );
2952                 if( $mw->matchAndRemove( $text ) ) {
2953                         $doShowToc = false;
2954                 }
2955
2956                 # Get all headlines for numbering them and adding funky stuff like [edit]
2957                 # links - this is for later, but we need the number of headlines right now
2958                 $numMatches = preg_match_all( '/<H([1-6])(.*?'.'>)(.*?)<\/H[1-6] *>/i', $text, $matches );
2959
2960                 # if there are fewer than 4 headlines in the article, do not show TOC
2961                 if( $numMatches < 4 ) {
2962                         $doShowToc = false;
2963                 }
2964
2965                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2966                 # override above conditions and always show TOC at that place
2967
2968                 $mw =& MagicWord::get( MAG_TOC );
2969                 if($mw->match( $text ) ) {
2970                         $doShowToc = true;
2971                         $forceTocHere = true;
2972                 } else {
2973                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2974                         # override above conditions and always show TOC above first header
2975                         $mw =& MagicWord::get( MAG_FORCETOC );
2976                         if ($mw->matchAndRemove( $text ) ) {
2977                                 $doShowToc = true;
2978                         }
2979                 }
2980
2981                 # Never ever show TOC if no headers
2982                 if( $numMatches < 1 ) {
2983                         $doShowToc = false;
2984                 }
2985
2986                 # We need this to perform operations on the HTML
2987                 $sk =& $this->mOptions->getSkin();
2988
2989                 # headline counter
2990                 $headlineCount = 0;
2991                 $sectionCount = 0; # headlineCount excluding template sections
2992
2993                 # Ugh .. the TOC should have neat indentation levels which can be
2994                 # passed to the skin functions. These are determined here
2995                 $toc = '';
2996                 $full = '';
2997                 $head = array();
2998                 $sublevelCount = array();
2999                 $levelCount = array();
3000                 $toclevel = 0;
3001                 $level = 0;
3002                 $prevlevel = 0;
3003                 $toclevel = 0;
3004                 $prevtoclevel = 0;
3005
3006                 foreach( $matches[3] as $headline ) {
3007                         $istemplate = 0;
3008                         $templatetitle = '';
3009                         $templatesection = 0;
3010                         $numbering = '';
3011
3012                         if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
3013                                 $istemplate = 1;
3014                                 $templatetitle = base64_decode($mat[1]);
3015                                 $templatesection = 1 + (int)base64_decode($mat[2]);
3016                                 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
3017                         }
3018
3019                         if( $toclevel ) {
3020                                 $prevlevel = $level;
3021                                 $prevtoclevel = $toclevel;
3022                         }
3023                         $level = $matches[1][$headlineCount];
3024
3025                         if( $doNumberHeadings || $doShowToc ) {
3026
3027                                 if ( $level > $prevlevel ) {
3028                                         # Increase TOC level
3029                                         $toclevel++;
3030                                         $sublevelCount[$toclevel] = 0;
3031                                         $toc .= $sk->tocIndent();
3032                                 }
3033                                 elseif ( $level < $prevlevel && $toclevel > 1 ) {
3034                                         # Decrease TOC level, find level to jump to
3035
3036                                         if ( $toclevel == 2 && $level <= $levelCount[1] ) {
3037                                                 # Can only go down to level 1
3038                                                 $toclevel = 1;
3039                                         } else {
3040                                                 for ($i = $toclevel; $i > 0; $i--) {
3041                                                         if ( $levelCount[$i] == $level ) {
3042                                                                 # Found last matching level
3043                                                                 $toclevel = $i;
3044                                                                 break;
3045                                                         }
3046                                                         elseif ( $levelCount[$i] < $level ) {
3047                                                                 # Found first matching level below current level
3048                                                                 $toclevel = $i + 1;
3049                                                                 break;
3050                                                         }
3051                                                 }
3052                                         }
3053
3054                                         $toc .= $sk->tocUnindent( $prevtoclevel - $toclevel );
3055                                 }
3056                                 else {
3057                                         # No change in level, end TOC line
3058                                         $toc .= $sk->tocLineEnd();
3059                                 }
3060
3061                                 $levelCount[$toclevel] = $level;
3062
3063                                 # count number of headlines for each level
3064                                 @$sublevelCount[$toclevel]++;
3065                                 $dot = 0;
3066                                 for( $i = 1; $i <= $toclevel; $i++ ) {
3067                                         if( !empty( $sublevelCount[$i] ) ) {
3068                                                 if( $dot ) {
3069                                                         $numbering .= '.';
3070                                                 }
3071                                                 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
3072                                                 $dot = 1;
3073                                         }
3074                                 }
3075                         }
3076
3077                         # The canonized header is a version of the header text safe to use for links
3078                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
3079                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
3080                         $canonized_headline = $this->unstripNoWiki( $canonized_headline, $this->mStripState );
3081
3082                         # Remove link placeholders by the link text.
3083                         #     <!--LINK number-->
3084                         # turns into
3085                         #     link text with suffix
3086                         $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
3087                                                             "\$this->mLinkHolders['texts'][\$1]",
3088                                                             $canonized_headline );
3089                         $canonized_headline = preg_replace( '/<!--IWLINK ([0-9]*)-->/e',
3090                                                             "\$this->mInterwikiLinkHolders['texts'][\$1]",
3091                                                             $canonized_headline );
3092
3093                         # strip out HTML
3094                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
3095                         $tocline = trim( $canonized_headline );
3096                         # Save headline for section edit hint before it's escaped
3097                         $headline_hint = trim( $canonized_headline );
3098                         $canonized_headline = Sanitizer::escapeId( $tocline );
3099                         $refers[$headlineCount] = $canonized_headline;
3100
3101                         # count how many in assoc. array so we can track dupes in anchors
3102                         @$refers[$canonized_headline]++;
3103                         $refcount[$headlineCount]=$refers[$canonized_headline];
3104
3105                         # Don't number the heading if it is the only one (looks silly)
3106                         if( $doNumberHeadings && count( $matches[3] ) > 1) {
3107                                 # the two are different if the line contains a link
3108                                 $headline=$numbering . ' ' . $headline;
3109                         }
3110
3111                         # Create the anchor for linking from the TOC to the section
3112                         $anchor = $canonized_headline;
3113                         if($refcount[$headlineCount] > 1 ) {
3114                                 $anchor .= '_' . $refcount[$headlineCount];
3115                         }
3116                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
3117                                 $toc .= $sk->tocLine($anchor, $tocline, $numbering, $toclevel);
3118                         }
3119                         if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
3120                                 if ( empty( $head[$headlineCount] ) ) {
3121                                         $head[$headlineCount] = '';
3122                                 }
3123                                 if( $istemplate )
3124                                         $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
3125                                 else
3126                                         $head[$headlineCount] .= $sk->editSectionLink($this->mTitle, $sectionCount+1, $headline_hint);
3127                         }
3128
3129                         # give headline the correct <h#> tag
3130                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
3131
3132                         $headlineCount++;
3133                         if( !$istemplate )
3134                                 $sectionCount++;
3135                 }
3136
3137                 if( $doShowToc ) {
3138                         $toc .= $sk->tocUnindent( $toclevel - 1 );
3139                         $toc = $sk->tocList( $toc );
3140                 }
3141
3142                 # split up and insert constructed headlines
3143
3144                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
3145                 $i = 0;
3146
3147                 foreach( $blocks as $block ) {
3148                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
3149                                 # This is the [edit] link that appears for the top block of text when
3150                                 # section editing is enabled
3151
3152                                 # Disabled because it broke block formatting
3153                                 # For example, a bullet point in the top line
3154                                 # $full .= $sk->editSectionLink(0);
3155                         }
3156                         $full .= $block;
3157                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
3158                         # Top anchor now in skin
3159                                 $full = $full.$toc;
3160                         }
3161
3162                         if( !empty( $head[$i] ) ) {
3163                                 $full .= $head[$i];
3164                         }
3165                         $i++;
3166                 }
3167                 if($forceTocHere) {
3168                         $mw =& MagicWord::get( MAG_TOC );
3169                         return $mw->replace( $toc, $full );
3170                 } else {
3171                         return $full;
3172                 }
3173         }
3174
3175         /**
3176          * Return an HTML link for the "ISBN 123456" text
3177          * @access private
3178          */
3179         function magicISBN( $text ) {
3180                 $fname = 'Parser::magicISBN';
3181                 wfProfileIn( $fname );
3182
3183                 $a = split( 'ISBN ', ' '.$text );
3184                 if ( count ( $a ) < 2 ) {
3185                         wfProfileOut( $fname );
3186                         return $text;
3187                 }
3188                 $text = substr( array_shift( $a ), 1);
3189                 $valid = '0123456789-Xx';
3190
3191                 foreach ( $a as $x ) {
3192                         # hack: don't replace inside thumbnail title/alt
3193                         # attributes
3194                         if(preg_match('/<[^>]+(alt|title)="[^">]*$/', $text)) {
3195                                 $text .= "ISBN $x";
3196                                 continue;
3197                         }
3198
3199                         $isbn = $blank = '' ;
3200                         while ( ' ' == $x{0} ) {
3201                                 $blank .= ' ';
3202                                 $x = substr( $x, 1 );
3203                         }
3204                         if ( $x == '' ) { # blank isbn
3205                                 $text .= "ISBN $blank";
3206                                 continue;
3207                         }
3208                         while ( strstr( $valid, $x{0} ) != false ) {
3209                                 $isbn .= $x{0};
3210                                 $x = substr( $x, 1 );
3211                         }
3212                         $num = str_replace( '-', '', $isbn );
3213                         $num = str_replace( ' ', '', $num );
3214                         $num = str_replace( 'x', 'X', $num );
3215
3216                         if ( '' == $num ) {
3217                                 $text .= "ISBN $blank$x";
3218                         } else {
3219                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
3220                                 $text .= '<a href="' .
3221                                         $titleObj->escapeLocalUrl( 'isbn='.$num ) .
3222                                         "\" class=\"internal\">ISBN $isbn</a>";
3223                                 $text .= $x;
3224                         }
3225                 }
3226                 wfProfileOut( $fname );
3227                 return $text;
3228         }
3229
3230         /**
3231          * Return an HTML link for the "RFC 1234" text
3232          *
3233          * @access private
3234          * @param string $text     Text to be processed
3235          * @param string $keyword  Magic keyword to use (default RFC)
3236          * @param string $urlmsg   Interface message to use (default rfcurl)
3237          * @return string
3238          */
3239         function magicRFC( $text, $keyword='RFC ', $urlmsg='rfcurl'  ) {
3240
3241                 $valid = '0123456789';
3242                 $internal = false;
3243
3244                 $a = split( $keyword, ' '.$text );
3245                 if ( count ( $a ) < 2 ) {
3246                         return $text;
3247                 }
3248                 $text = substr( array_shift( $a ), 1);
3249
3250                 /* Check if keyword is preceed by [[.
3251                  * This test is made here cause of the array_shift above
3252                  * that prevent the test to be done in the foreach.
3253                  */
3254                 if ( substr( $text, -2 ) == '[[' ) {
3255                         $internal = true;
3256                 }
3257
3258                 foreach ( $a as $x ) {
3259                         /* token might be empty if we have RFC RFC 1234 */
3260                         if ( $x=='' ) {
3261                                 $text.=$keyword;
3262                                 continue;
3263                                 }
3264
3265                         # hack: don't replace inside thumbnail title/alt
3266                         # attributes
3267                         if(preg_match('/<[^>]+(alt|title)="[^">]*$/', $text)) {
3268                                 $text .= $keyword . $x;
3269                                 continue;
3270                         }
3271
3272                         $id = $blank = '' ;
3273
3274                         /** remove and save whitespaces in $blank */
3275                         while ( $x{0} == ' ' ) {
3276                                 $blank .= ' ';
3277                                 $x = substr( $x, 1 );
3278                         }
3279
3280                         /** remove and save the rfc number in $id */
3281                         while ( strstr( $valid, $x{0} ) != false ) {
3282                                 $id .= $x{0};
3283                                 $x = substr( $x, 1 );
3284                         }
3285
3286                         if ( $id == '' ) {
3287                                 /* call back stripped spaces*/
3288                                 $text .= $keyword.$blank.$x;
3289                         } elseif( $internal ) {
3290                                 /* normal link */
3291                                 $text .= $keyword.$id.$x;
3292                         } else {
3293                                 /* build the external link*/
3294                                 $url = wfMsg( $urlmsg, $id);
3295                                 $sk =& $this->mOptions->getSkin();
3296                                 $la = $sk->getExternalLinkAttributes( $url, $keyword.$id );
3297                                 $text .= "<a href='{$url}'{$la}>{$keyword}{$id}</a>{$x}";
3298                         }
3299
3300                         /* Check if the next RFC keyword is preceed by [[ */
3301                         $internal = ( substr($x,-2) == '[[' );
3302                 }
3303                 return $text;
3304         }
3305
3306         /**
3307          * Transform wiki markup when saving a page by doing \r\n -> \n
3308          * conversion, substitting signatures, {{subst:}} templates, etc.
3309          *
3310          * @param string $text the text to transform
3311          * @param Title &$title the Title object for the current article
3312          * @param User &$user the User object describing the current user
3313          * @param ParserOptions $options parsing options
3314          * @param bool $clearState whether to clear the parser state first
3315          * @return string the altered wiki markup
3316          * @access public
3317          */
3318         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
3319                 $this->mOptions = $options;
3320                 $this->mTitle =& $title;
3321                 $this->mOutputType = OT_WIKI;
3322
3323                 if ( $clearState ) {
3324                         $this->clearState();
3325                 }
3326
3327                 $stripState = false;
3328                 $pairs = array(
3329                         "\r\n" => "\n",
3330                 );
3331                 $text = str_replace( array_keys( $pairs ), array_values( $pairs ), $text );
3332                 $text = $this->strip( $text, $stripState, true );
3333                 $text = $this->pstPass2( $text, $user );
3334                 $text = $this->unstrip( $text, $stripState );
3335                 $text = $this->unstripNoWiki( $text, $stripState );
3336                 return $text;
3337         }
3338
3339         /**
3340          * Pre-save transform helper function
3341          * @access private
3342          */
3343         function pstPass2( $text, &$user ) {
3344                 global $wgContLang, $wgLocaltimezone;
3345
3346                 /* Note: This is the timestamp saved as hardcoded wikitext to
3347                  * the database, we use $wgContLang here in order to give
3348                  * everyone the same signiture and use the default one rather
3349                  * than the one selected in each users preferences.
3350                  */
3351                 if ( isset( $wgLocaltimezone ) ) {
3352                         $oldtz = getenv( 'TZ' );
3353                         putenv( 'TZ='.$wgLocaltimezone );
3354                 }
3355                 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false, false) .
3356                   ' (' . date( 'T' ) . ')';
3357                 if ( isset( $wgLocaltimezone ) ) {
3358                         putenv( 'TZ='.$oldtz );
3359                 }
3360
3361                 # Variable replacement
3362                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
3363                 $text = $this->replaceVariables( $text );
3364
3365                 # Signatures
3366                 $sigText = $this->getUserSig( $user );
3367                 $text = strtr( $text, array(
3368                         '~~~~~' => $d,
3369                         '~~~~' => "$sigText $d",
3370                         '~~~' => $sigText
3371                 ) );
3372
3373                 # Context links: [[|name]] and [[name (context)|]]
3374                 #
3375                 global $wgLegalTitleChars;
3376                 $tc = "[$wgLegalTitleChars]";
3377                 $np = str_replace( array( '(', ')' ), array( '', '' ), $tc ); # No parens
3378
3379                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
3380                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
3381
3382                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
3383                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
3384                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
3385                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
3386                 $context = '';
3387                 $t = $this->mTitle->getText();
3388                 if ( preg_match( $conpat, $t, $m ) ) {
3389                         $context = $m[2];
3390                 }
3391                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
3392                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
3393                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
3394
3395                 if ( '' == $context ) {
3396                         $text = preg_replace( $p2, '[[\\1]]', $text );
3397                 } else {
3398                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
3399                 }
3400
3401                 # Trim trailing whitespace
3402                 # MAG_END (__END__) tag allows for trailing
3403                 # whitespace to be deliberately included
3404                 $text = rtrim( $text );
3405                 $mw =& MagicWord::get( MAG_END );
3406                 $mw->matchAndRemove( $text );
3407
3408                 return $text;
3409         }
3410
3411         /**
3412          * Fetch the user's signature text, if any, and normalize to
3413          * validated, ready-to-insert wikitext.
3414          *
3415          * @param User $user
3416          * @return string
3417          * @access private
3418          */
3419         function getUserSig( &$user ) {
3420                 $username = $user->getName();
3421                 $nickname = $user->getOption( 'nickname' );
3422                 $nickname = $nickname === '' ? $username : $nickname;
3423
3424                 if( $user->getBoolOption( 'fancysig' ) !== false ) {
3425                         # Sig. might contain markup; validate this
3426                         if( $this->validateSig( $nickname ) !== false ) {
3427                                 # Validated; clean up (if needed) and return it
3428                                 return( $this->cleanSig( $nickname ) );
3429                         } else {
3430                                 # Failed to validate; fall back to the default
3431                                 $nickname = $username;
3432                                 wfDebug( "Parser::getUserSig: $username has bad XML tags in signature.\n" );
3433                         }
3434                 }
3435
3436                 # If we're still here, make it a link to the user page
3437                 $userpage = $user->getUserPage();
3438                 return( '[[' . $userpage->getPrefixedText() . '|' . wfEscapeWikiText( $nickname ) . ']]' );
3439         }
3440
3441         /**
3442          * Check that the user's signature contains no bad XML
3443          *
3444          * @param string $text
3445          * @return mixed An expanded string, or false if invalid.
3446          */
3447         function validateSig( $text ) {
3448                 return( wfIsWellFormedXmlFragment( $text ) ? $text : false );
3449         }
3450
3451         /**
3452          * Clean up signature text
3453          *
3454          * 1) Strip ~~~, ~~~~ and ~~~~~ out of signatures
3455          * 2) Substitute all transclusions
3456          *
3457          * @param string $text
3458          * @return string Signature text
3459          */
3460         function cleanSig( $text ) {
3461                 $substWord = MagicWord::get( MAG_SUBST );
3462                 $substRegex = '/\{\{(?!(?:' . $substWord->getBaseRegex() . '))/x' . $substWord->getRegexCase();
3463                 $substText = '{{' . $substWord->getSynonym( 0 );
3464
3465                 $text = preg_replace( $substRegex, $substText, $text );
3466                 $text = preg_replace( '/~{3,5}/', '', $text );
3467                 $text = $this->replaceVariables( $text );
3468
3469                 return $text;
3470         }
3471
3472         /**
3473          * Set up some variables which are usually set up in parse()
3474          * so that an external function can call some class members with confidence
3475          * @access public
3476          */
3477         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
3478                 $this->mTitle =& $title;
3479                 $this->mOptions = $options;
3480                 $this->mOutputType = $outputType;
3481                 if ( $clearState ) {
3482                         $this->clearState();
3483                 }
3484         }
3485
3486         /**
3487          * Transform a MediaWiki message by replacing magic variables.
3488          *
3489          * @param string $text the text to transform
3490          * @param ParserOptions $options  options
3491          * @return string the text with variables substituted
3492          * @access public
3493          */
3494         function transformMsg( $text, $options ) {
3495                 global $wgTitle;
3496                 static $executing = false;
3497
3498                 $fname = "Parser::transformMsg";
3499
3500                 # Guard against infinite recursion
3501                 if ( $executing ) {
3502                         return $text;
3503                 }
3504                 $executing = true;
3505
3506                 wfProfileIn($fname);
3507
3508                 $this->mTitle = $wgTitle;
3509                 $this->mOptions = $options;
3510                 $this->mOutputType = OT_MSG;
3511                 $this->clearState();
3512                 $text = $this->replaceVariables( $text );
3513
3514                 $executing = false;
3515                 wfProfileOut($fname);
3516                 return $text;
3517         }
3518
3519         /**
3520          * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
3521          * The callback should have the following form:
3522          *    function myParserHook( $text, $params, &$parser ) { ... }
3523          *
3524          * Transform and return $text. Use $parser for any required context, e.g. use
3525          * $parser->getTitle() and $parser->getOptions() not $wgTitle or $wgOut->mParserOptions
3526          *
3527          * @access public
3528          *
3529          * @param mixed $tag The tag to use, e.g. 'hook' for <hook>
3530          * @param mixed $callback The callback function (and object) to use for the tag
3531          *
3532          * @return The old value of the mTagHooks array associated with the hook
3533          */
3534         function setHook( $tag, $callback ) {
3535                 $oldVal = @$this->mTagHooks[$tag];
3536                 $this->mTagHooks[$tag] = $callback;
3537
3538                 return $oldVal;
3539         }
3540
3541         /**
3542          * Replace <!--LINK--> link placeholders with actual links, in the buffer
3543          * Placeholders created in Skin::makeLinkObj()
3544          * Returns an array of links found, indexed by PDBK:
3545          *  0 - broken
3546          *  1 - normal link
3547          *  2 - stub
3548          * $options is a bit field, RLH_FOR_UPDATE to select for update
3549          */
3550         function replaceLinkHolders( &$text, $options = 0 ) {
3551                 global $wgUser;
3552                 global $wgOutputReplace;
3553
3554                 $fname = 'Parser::replaceLinkHolders';
3555                 wfProfileIn( $fname );
3556
3557                 $pdbks = array();
3558                 $colours = array();
3559                 $sk =& $this->mOptions->getSkin();
3560                 $linkCache =& LinkCache::singleton();
3561
3562                 if ( !empty( $this->mLinkHolders['namespaces'] ) ) {
3563                         wfProfileIn( $fname.'-check' );
3564                         $dbr =& wfGetDB( DB_SLAVE );
3565                         $page = $dbr->tableName( 'page' );
3566                         $threshold = $wgUser->getOption('stubthreshold');
3567
3568                         # Sort by namespace
3569                         asort( $this->mLinkHolders['namespaces'] );
3570
3571                         # Generate query
3572                         $query = false;
3573                         foreach ( $this->mLinkHolders['namespaces'] as $key => $ns ) {
3574                                 # Make title object
3575                                 $title = $this->mLinkHolders['titles'][$key];
3576
3577                                 # Skip invalid entries.
3578                                 # Result will be ugly, but prevents crash.
3579                                 if ( is_null( $title ) ) {
3580                                         continue;
3581                                 }
3582                                 $pdbk = $pdbks[$key] = $title->getPrefixedDBkey();
3583
3584                                 # Check if it's a static known link, e.g. interwiki
3585                                 if ( $title->isAlwaysKnown() ) {
3586                                         $colours[$pdbk] = 1;
3587                                 } elseif ( ( $id = $linkCache->getGoodLinkID( $pdbk ) ) != 0 ) {
3588                                         $colours[$pdbk] = 1;
3589                                         $this->mOutput->addLink( $title, $id );
3590                                 } elseif ( $linkCache->isBadLink( $pdbk ) ) {
3591                                         $colours[$pdbk] = 0;
3592                                 } else {
3593                                         # Not in the link cache, add it to the query
3594                                         if ( !isset( $current ) ) {
3595                                                 $current = $ns;
3596                                                 $query =  "SELECT page_id, page_namespace, page_title";
3597                                                 if ( $threshold > 0 ) {
3598                                                         $query .= ', page_len, page_is_redirect';
3599                                                 }
3600                                                 $query .= " FROM $page WHERE (page_namespace=$ns AND page_title IN(";
3601                                         } elseif ( $current != $ns ) {
3602                                                 $current = $ns;
3603                                                 $query .= ")) OR (page_namespace=$ns AND page_title IN(";
3604                                         } else {
3605                                                 $query .= ', ';
3606                                         }
3607
3608                                         $query .= $dbr->addQuotes( $this->mLinkHolders['dbkeys'][$key] );
3609                                 }
3610                         }
3611                         if ( $query ) {
3612                                 $query .= '))';
3613                                 if ( $options & RLH_FOR_UPDATE ) {
3614                                         $query .= ' FOR UPDATE';
3615                                 }
3616
3617                                 $res = $dbr->query( $query, $fname );
3618
3619                                 # Fetch data and form into an associative array
3620                                 # non-existent = broken
3621                                 # 1 = known
3622                                 # 2 = stub
3623                                 while ( $s = $dbr->fetchObject($res) ) {
3624                                         $title = Title::makeTitle( $s->page_namespace, $s->page_title );
3625                                         $pdbk = $title->getPrefixedDBkey();
3626                                         $linkCache->addGoodLinkObj( $s->page_id, $title );
3627                                         $this->mOutput->addLink( $title, $s->page_id );
3628
3629                                         if ( $threshold >  0 ) {
3630                                                 $size = $s->page_len;
3631                                                 if ( $s->page_is_redirect || $s->page_namespace != 0 || $size >= $threshold ) {
3632                                                         $colours[$pdbk] = 1;
3633                                                 } else {
3634                                                         $colours[$pdbk] = 2;
3635                                                 }
3636                                         } else {
3637                                                 $colours[$pdbk] = 1;
3638                                         }
3639                                 }
3640                         }
3641                         wfProfileOut( $fname.'-check' );
3642
3643                         # Construct search and replace arrays
3644                         wfProfileIn( $fname.'-construct' );
3645                         $wgOutputReplace = array();
3646                         foreach ( $this->mLinkHolders['namespaces'] as $key => $ns ) {
3647                                 $pdbk = $pdbks[$key];
3648                                 $searchkey = "<!--LINK $key-->";
3649                                 $title = $this->mLinkHolders['titles'][$key];
3650                                 if ( empty( $colours[$pdbk] ) ) {
3651                                         $linkCache->addBadLinkObj( $title );
3652                                         $colours[$pdbk] = 0;
3653                                         $this->mOutput->addLink( $title, 0 );
3654                                         $wgOutputReplace[$searchkey] = $sk->makeBrokenLinkObj( $title,
3655                                                                         $this->mLinkHolders['texts'][$key],
3656                                                                         $this->mLinkHolders['queries'][$key] );
3657                                 } elseif ( $colours[$pdbk] == 1 ) {
3658                                         $wgOutputReplace[$searchkey] = $sk->makeKnownLinkObj( $title,
3659                                                                         $this->mLinkHolders['texts'][$key],
3660                                                                         $this->mLinkHolders['queries'][$key] );
3661                                 } elseif ( $colours[$pdbk] == 2 ) {
3662                                         $wgOutputReplace[$searchkey] = $sk->makeStubLinkObj( $title,
3663                                                                         $this->mLinkHolders['texts'][$key],
3664                                                                         $this->mLinkHolders['queries'][$key] );
3665                                 }
3666                         }
3667                         wfProfileOut( $fname.'-construct' );
3668
3669                         # Do the thing
3670                         wfProfileIn( $fname.'-replace' );
3671
3672                         $text = preg_replace_callback(
3673                                 '/(<!--LINK .*?-->)/',
3674                                 "wfOutputReplaceMatches",
3675                                 $text);
3676
3677                         wfProfileOut( $fname.'-replace' );
3678                 }
3679
3680                 # Now process interwiki link holders
3681                 # This is quite a bit simpler than internal links
3682                 if ( !empty( $this->mInterwikiLinkHolders['texts'] ) ) {
3683                         wfProfileIn( $fname.'-interwiki' );
3684                         # Make interwiki link HTML
3685                         $wgOutputReplace = array();
3686                         foreach( $this->mInterwikiLinkHolders['texts'] as $key => $link ) {
3687                                 $title = $this->mInterwikiLinkHolders['titles'][$key];
3688                                 $wgOutputReplace[$key] = $sk->makeLinkObj( $title, $link );
3689                         }
3690
3691                         $text = preg_replace_callback(
3692                                 '/<!--IWLINK (.*?)-->/',
3693                                 "wfOutputReplaceMatches",
3694                                 $text );
3695                         wfProfileOut( $fname.'-interwiki' );
3696                 }
3697
3698                 wfProfileOut( $fname );
3699                 return $colours;
3700         }
3701
3702         /**
3703          * Replace <!--LINK--> link placeholders with plain text of links
3704          * (not HTML-formatted).
3705          * @param string $text
3706          * @return string
3707          */
3708         function replaceLinkHoldersText( $text ) {
3709                 $fname = 'Parser::replaceLinkHoldersText';
3710                 wfProfileIn( $fname );
3711
3712                 $text = preg_replace_callback(
3713                         '/<!--(LINK|IWLINK) (.*?)-->/',
3714                         array( &$this, 'replaceLinkHoldersTextCallback' ),
3715                         $text );
3716
3717                 wfProfileOut( $fname );
3718                 return $text;
3719         }
3720
3721         /**
3722          * @param array $matches
3723          * @return string
3724          * @access private
3725          */
3726         function replaceLinkHoldersTextCallback( $matches ) {
3727                 $type = $matches[1];
3728                 $key  = $matches[2];
3729                 if( $type == 'LINK' ) {
3730                         if( isset( $this->mLinkHolders['texts'][$key] ) ) {
3731                                 return $this->mLinkHolders['texts'][$key];
3732                         }
3733                 } elseif( $type == 'IWLINK' ) {
3734                         if( isset( $this->mInterwikiLinkHolders['texts'][$key] ) ) {
3735                                 return $this->mInterwikiLinkHolders['texts'][$key];
3736                         }
3737                 }
3738                 return $matches[0];
3739         }
3740
3741         /**
3742          * Renders an image gallery from a text with one line per image.
3743          * text labels may be given by using |-style alternative text. E.g.
3744          *   Image:one.jpg|The number "1"
3745          *   Image:tree.jpg|A tree
3746          * given as text will return the HTML of a gallery with two images,
3747          * labeled 'The number "1"' and
3748          * 'A tree'.
3749          */
3750         function renderImageGallery( $text ) {
3751                 # Setup the parser
3752                 $parserOptions = new ParserOptions;
3753                 $localParser = new Parser();
3754
3755                 $ig = new ImageGallery();
3756                 $ig->setShowBytes( false );
3757                 $ig->setShowFilename( false );
3758                 $lines = explode( "\n", $text );
3759
3760                 foreach ( $lines as $line ) {
3761                         # match lines like these:
3762                         # Image:someimage.jpg|This is some image
3763                         preg_match( "/^([^|]+)(\\|(.*))?$/", $line, $matches );
3764                         # Skip empty lines
3765                         if ( count( $matches ) == 0 ) {
3766                                 continue;
3767                         }
3768                         $nt =& Title::newFromText( $matches[1] );
3769                         if( is_null( $nt ) ) {
3770                                 # Bogus title. Ignore these so we don't bomb out later.
3771                                 continue;
3772                         }
3773                         if ( isset( $matches[3] ) ) {
3774                                 $label = $matches[3];
3775                         } else {
3776                                 $label = '';
3777                         }
3778
3779                         $pout = $localParser->parse( $label , $this->mTitle, $parserOptions );
3780                         $html = $pout->getText();
3781
3782                         $ig->add( new Image( $nt ), $html );
3783                         $this->mOutput->addImage( $nt->getDBkey() );
3784                 }
3785                 return $ig->toHTML();
3786         }
3787
3788         /**
3789          * Parse image options text and use it to make an image
3790          */
3791         function makeImage( &$nt, $options ) {
3792                 global $wgUseImageResize;
3793
3794                 $align = '';
3795
3796                 # Check if the options text is of the form "options|alt text"
3797                 # Options are:
3798                 #  * thumbnail          make a thumbnail with enlarge-icon and caption, alignment depends on lang
3799                 #  * left               no resizing, just left align. label is used for alt= only
3800                 #  * right              same, but right aligned
3801                 #  * none               same, but not aligned
3802                 #  * ___px              scale to ___ pixels width, no aligning. e.g. use in taxobox
3803                 #  * center             center the image
3804                 #  * framed             Keep original image size, no magnify-button.
3805
3806                 $part = explode( '|', $options);
3807
3808                 $mwThumb  =& MagicWord::get( MAG_IMG_THUMBNAIL );
3809                 $mwManualThumb =& MagicWord::get( MAG_IMG_MANUALTHUMB );
3810                 $mwLeft   =& MagicWord::get( MAG_IMG_LEFT );
3811                 $mwRight  =& MagicWord::get( MAG_IMG_RIGHT );
3812                 $mwNone   =& MagicWord::get( MAG_IMG_NONE );
3813                 $mwWidth  =& MagicWord::get( MAG_IMG_WIDTH );
3814                 $mwCenter =& MagicWord::get( MAG_IMG_CENTER );
3815                 $mwFramed =& MagicWord::get( MAG_IMG_FRAMED );
3816                 $caption = '';
3817
3818                 $width = $height = $framed = $thumb = false;
3819                 $manual_thumb = '' ;
3820
3821                 foreach( $part as $key => $val ) {
3822                         if ( $wgUseImageResize && ! is_null( $mwThumb->matchVariableStartToEnd($val) ) ) {
3823                                 $thumb=true;
3824                         } elseif ( ! is_null( $match = $mwManualThumb->matchVariableStartToEnd($val) ) ) {
3825                                 # use manually specified thumbnail
3826                                 $thumb=true;
3827                                 $manual_thumb = $match;
3828                         } elseif ( ! is_null( $mwRight->matchVariableStartToEnd($val) ) ) {
3829                                 # remember to set an alignment, don't render immediately
3830                                 $align = 'right';
3831                         } elseif ( ! is_null( $mwLeft->matchVariableStartToEnd($val) ) ) {
3832                                 # remember to set an alignment, don't render immediately
3833                                 $align = 'left';
3834                         } elseif ( ! is_null( $mwCenter->matchVariableStartToEnd($val) ) ) {
3835                                 # remember to set an alignment, don't render immediately
3836                                 $align = 'center';
3837                         } elseif ( ! is_null( $mwNone->matchVariableStartToEnd($val) ) ) {
3838                                 # remember to set an alignment, don't render immediately
3839                                 $align = 'none';
3840                         } elseif ( $wgUseImageResize && ! is_null( $match = $mwWidth->matchVariableStartToEnd($val) ) ) {
3841                                 wfDebug( "MAG_IMG_WIDTH match: $match\n" );
3842                                 # $match is the image width in pixels
3843                                 if ( preg_match( '/^([0-9]*)x([0-9]*)$/', $match, $m ) ) {
3844                                         $width = intval( $m[1] );
3845                                         $height = intval( $m[2] );
3846                                 } else {
3847                                         $width = intval($match);
3848                                 }
3849                         } elseif ( ! is_null( $mwFramed->matchVariableStartToEnd($val) ) ) {
3850                                 $framed=true;
3851                         } else {
3852                                 $caption = $val;
3853                         }
3854                 }
3855                 # Strip bad stuff out of the alt text
3856                 $alt = $this->replaceLinkHoldersText( $caption );
3857
3858                 # make sure there are no placeholders in thumbnail attributes
3859                 # that are later expanded to html- so expand them now and
3860                 # remove the tags
3861                 $alt = $this->unstrip($alt, $this->mStripState);
3862                 $alt = Sanitizer::stripAllTags( $alt );
3863
3864                 # Linker does the rest
3865                 $sk =& $this->mOptions->getSkin();
3866                 return $sk->makeImageLinkObj( $nt, $caption, $alt, $align, $width, $height, $framed, $thumb, $manual_thumb );
3867         }
3868
3869         /**
3870          * Set a flag in the output object indicating that the content is dynamic and
3871          * shouldn't be cached.
3872          */
3873         function disableCache() {
3874                 $this->mOutput->mCacheTime = -1;
3875         }
3876
3877         /**#@+
3878          * Callback from the Sanitizer for expanding items found in HTML attribute
3879          * values, so they can be safely tested and escaped.
3880          * @param string $text
3881          * @param array $args
3882          * @return string
3883          * @access private
3884          */
3885         function attributeStripCallback( &$text, $args ) {
3886                 $text = $this->replaceVariables( $text, $args );
3887                 $text = $this->unstripForHTML( $text );
3888                 return $text;
3889         }
3890
3891         function unstripForHTML( $text ) {
3892                 $text = $this->unstrip( $text, $this->mStripState );
3893                 $text = $this->unstripNoWiki( $text, $this->mStripState );
3894                 return $text;
3895         }
3896         /**#@-*/
3897
3898         /**#@+
3899          * Accessor/mutator
3900          */
3901         function Title( $x = NULL ) { return wfSetVar( $this->mTitle, $x ); }
3902         function Options( $x = NULL ) { return wfSetVar( $this->mOptions, $x ); }
3903         function OutputType( $x = NULL ) { return wfSetVar( $this->mOutputType, $x ); }
3904         /**#@-*/
3905
3906         /**#@+
3907          * Accessor
3908          */
3909         function getTags() { return array_keys( $this->mTagHooks ); }
3910         /**#@-*/
3911 }
3912
3913 /**
3914  * @todo document
3915  * @package MediaWiki
3916  */
3917 class ParserOutput
3918 {
3919         var $mText,             # The output text
3920                 $mLanguageLinks,    # List of the full text of language links, in the order they appear
3921                 $mCategories,       # Map of category names to sort keys
3922                 $mContainsOldMagic, # Boolean variable indicating if the input contained variables like {{CURRENTDAY}}
3923                 $mCacheTime,        # Time when this object was generated, or -1 for uncacheable. Used in ParserCache.
3924                 $mVersion,          # Compatibility check
3925                 $mTitleText,        # title text of the chosen language variant
3926                 $mLinks,            # 2-D map of NS/DBK to ID for the links in the document. ID=zero for broken.
3927                 $mTemplates,        # 2-D map of NS/DBK to ID for the template references. ID=zero for broken.
3928                 $mImages,           # DB keys of the images used, in the array key only
3929                 $mExternalLinks;    # External link URLs, in the key only
3930
3931         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
3932                 $containsOldMagic = false, $titletext = '' )
3933         {
3934                 $this->mText = $text;
3935                 $this->mLanguageLinks = $languageLinks;
3936                 $this->mCategories = $categoryLinks;
3937                 $this->mContainsOldMagic = $containsOldMagic;
3938                 $this->mCacheTime = '';
3939                 $this->mVersion = MW_PARSER_VERSION;
3940                 $this->mTitleText = $titletext;
3941                 $this->mLinks = array();
3942                 $this->mTemplates = array();
3943                 $this->mImages = array();
3944                 $this->mExternalLinks = array();
3945         }
3946
3947         function getText()                   { return $this->mText; }
3948         function getLanguageLinks()          { return $this->mLanguageLinks; }
3949         function getCategoryLinks()          { return array_keys( $this->mCategories ); }
3950         function &getCategories()            { return $this->mCategories; }
3951         function getCacheTime()              { return $this->mCacheTime; }
3952         function getTitleText()              { return $this->mTitleText; }
3953         function &getLinks()                 { return $this->mLinks; }
3954         function &getTemplates()             { return $this->mTemplates; }
3955         function &getImages()                { return $this->mImages; }
3956         function &getExternalLinks()         { return $this->mExternalLinks; }
3957
3958         function containsOldMagic()          { return $this->mContainsOldMagic; }
3959         function setText( $text )            { return wfSetVar( $this->mText, $text ); }
3960         function setLanguageLinks( $ll )     { return wfSetVar( $this->mLanguageLinks, $ll ); }
3961         function setCategoryLinks( $cl )     { return wfSetVar( $this->mCategories, $cl ); }
3962         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
3963         function setCacheTime( $t )          { return wfSetVar( $this->mCacheTime, $t ); }
3964         function setTitleText( $t )          { return wfSetVar ($this->mTitleText, $t); }
3965
3966         function addCategory( $c, $sort )    { $this->mCategories[$c] = $sort; }
3967         function addImage( $name )           { $this->mImages[$name] = 1; }
3968         function addLanguageLink( $t )       { $this->mLanguageLinks[] = $t; }
3969         function addExternalLink( $url )     { $this->mExternalLinks[$url] = 1; }
3970
3971         function addLink( $title, $id ) {
3972                 $ns = $title->getNamespace();
3973                 $dbk = $title->getDBkey();
3974                 if ( !isset( $this->mLinks[$ns] ) ) {
3975                         $this->mLinks[$ns] = array();
3976                 }
3977                 $this->mLinks[$ns][$dbk] = $id;
3978         }
3979
3980         function addTemplate( $title, $id ) {
3981                 $ns = $title->getNamespace();
3982                 $dbk = $title->getDBkey();
3983                 if ( !isset( $this->mTemplates[$ns] ) ) {
3984                         $this->mTemplates[$ns] = array();
3985                 }
3986                 $this->mTemplates[$ns][$dbk] = $id;
3987         }
3988
3989         /**
3990          * @deprecated
3991          */
3992         /*
3993         function merge( $other ) {
3994                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
3995                 $this->mCategories = array_merge( $this->mCategories, $this->mLanguageLinks );
3996                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
3997         }*/
3998
3999         /**
4000          * Return true if this cached output object predates the global or
4001          * per-article cache invalidation timestamps, or if it comes from
4002          * an incompatible older version.
4003          *
4004          * @param string $touched the affected article's last touched timestamp
4005          * @return bool
4006          * @access public
4007          */
4008         function expired( $touched ) {
4009                 global $wgCacheEpoch;
4010                 return $this->getCacheTime() == -1 || // parser says it's uncacheable
4011                        $this->getCacheTime() < $touched ||
4012                        $this->getCacheTime() <= $wgCacheEpoch ||
4013                        !isset( $this->mVersion ) ||
4014                        version_compare( $this->mVersion, MW_PARSER_VERSION, "lt" );
4015         }
4016 }
4017
4018 /**
4019  * Set options of the Parser
4020  * @todo document
4021  * @package MediaWiki
4022  */
4023 class ParserOptions
4024 {
4025         # All variables are private
4026         var $mUseTeX;                    # Use texvc to expand <math> tags
4027         var $mUseDynamicDates;           # Use DateFormatter to format dates
4028         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
4029         var $mAllowExternalImages;       # Allow external images inline
4030         var $mAllowExternalImagesFrom;   # If not, any exception?
4031         var $mSkin;                      # Reference to the preferred skin
4032         var $mDateFormat;                # Date format index
4033         var $mEditSection;               # Create "edit section" links
4034         var $mNumberHeadings;            # Automatically number headings
4035         var $mAllowSpecialInclusion;     # Allow inclusion of special pages
4036         var $mTidy;                      # Ask for tidy cleanup
4037
4038         function getUseTeX()                        { return $this->mUseTeX; }
4039         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
4040         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
4041         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
4042         function getAllowExternalImagesFrom()       { return $this->mAllowExternalImagesFrom; }
4043         function &getSkin()                         { return $this->mSkin; }
4044         function getDateFormat()                    { return $this->mDateFormat; }
4045         function getEditSection()                   { return $this->mEditSection; }
4046         function getNumberHeadings()                { return $this->mNumberHeadings; }
4047         function getAllowSpecialInclusion()         { return $this->mAllowSpecialInclusion; }
4048         function getTidy()                          { return $this->mTidy; }
4049
4050         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
4051         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
4052         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
4053         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
4054         function setAllowExternalImagesFrom( $x )   { return wfSetVar( $this->mAllowExternalImagesFrom, $x ); }
4055         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
4056         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
4057         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
4058         function setAllowSpecialInclusion( $x )     { return wfSetVar( $this->mAllowSpecialInclusion, $x ); }
4059         function setTidy( $x )                      { return wfSetVar( $this->mTidy, $x); }
4060         function setSkin( &$x ) { $this->mSkin =& $x; }
4061
4062         function ParserOptions() {
4063                 global $wgUser;
4064                 $this->initialiseFromUser( $wgUser );
4065         }
4066
4067         /**
4068          * Get parser options
4069          * @static
4070          */
4071         function newFromUser( &$user ) {
4072                 $popts = new ParserOptions;
4073                 $popts->initialiseFromUser( $user );
4074                 return $popts;
4075         }
4076
4077         /** Get user options */
4078         function initialiseFromUser( &$userInput ) {
4079                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
4080                 global $wgAllowExternalImagesFrom, $wgAllowSpecialInclusion;
4081                 $fname = 'ParserOptions::initialiseFromUser';
4082                 wfProfileIn( $fname );
4083                 if ( !$userInput ) {
4084                         $user = new User;
4085                         $user->setLoaded( true );
4086                 } else {
4087                         $user =& $userInput;
4088                 }
4089
4090                 $this->mUseTeX = $wgUseTeX;
4091                 $this->mUseDynamicDates = $wgUseDynamicDates;
4092                 $this->mInterwikiMagic = $wgInterwikiMagic;
4093                 $this->mAllowExternalImages = $wgAllowExternalImages;
4094                 $this->mAllowExternalImagesFrom = $wgAllowExternalImagesFrom;
4095                 wfProfileIn( $fname.'-skin' );
4096                 $this->mSkin =& $user->getSkin();
4097                 wfProfileOut( $fname.'-skin' );
4098                 $this->mDateFormat = $user->getOption( 'date' );
4099                 $this->mEditSection = true;
4100                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
4101                 $this->mAllowSpecialInclusion = $wgAllowSpecialInclusion;
4102                 $this->mTidy = false;
4103                 wfProfileOut( $fname );
4104         }
4105 }
4106
4107 /**
4108  * Callback function used by Parser::replaceLinkHolders()
4109  * to substitute link placeholders.
4110  */
4111 function &wfOutputReplaceMatches( $matches ) {
4112         global $wgOutputReplace;
4113         return $wgOutputReplace[$matches[1]];
4114 }
4115
4116 /**
4117  * Return the total number of articles
4118  */
4119 function wfNumberOfArticles() {
4120         global $wgNumberOfArticles;
4121
4122         wfLoadSiteStats();
4123         return $wgNumberOfArticles;
4124 }
4125
4126 /**
4127  * Return the number of files
4128  */
4129 function wfNumberOfFiles() {
4130         $fname = 'wfNumberOfFiles';
4131
4132         wfProfileIn( $fname );
4133         $dbr =& wfGetDB( DB_SLAVE );
4134         $numImages = $dbr->selectField('site_stats', 'ss_images', array(), $fname );
4135         wfProfileOut( $fname );
4136
4137         return $numImages;
4138 }
4139
4140 /**
4141  * Get various statistics from the database
4142  * @access private
4143  */
4144 function wfLoadSiteStats() {
4145         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
4146         $fname = 'wfLoadSiteStats';
4147
4148         if ( -1 != $wgNumberOfArticles ) return;
4149         $dbr =& wfGetDB( DB_SLAVE );
4150         $s = $dbr->selectRow( 'site_stats',
4151                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
4152                 array( 'ss_row_id' => 1 ), $fname
4153         );
4154
4155         if ( $s === false ) {
4156                 return;
4157         } else {
4158                 $wgTotalViews = $s->ss_total_views;
4159                 $wgTotalEdits = $s->ss_total_edits;
4160                 $wgNumberOfArticles = $s->ss_good_articles;
4161         }
4162 }
4163
4164 /**
4165  * Escape html tags
4166  * Basically replacing " > and < with HTML entities ( &quot;, &gt;, &lt;)
4167  *
4168  * @param string $in Text that might contain HTML tags
4169  * @return string Escaped string
4170  */
4171 function wfEscapeHTMLTagsOnly( $in ) {
4172         return str_replace(
4173                 array( '"', '>', '<' ),
4174                 array( '&quot;', '&gt;', '&lt;' ),
4175                 $in );
4176 }
4177
4178 ?>