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