includes/Parser.php

   1 <?php
   2
   3 // require_once('Tokenizer.php');
   4
   5 # PHP Parser
   6 #
   7 # Processes wiki markup
   8 #
   9 # There are two main entry points into the Parser class: parse() and preSaveTransform().
  10 # The parse() function produces HTML output, preSaveTransform() produces altered wiki markup.
  11 #
  12 # Globals used:
  13 #    objects:   $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
  14 #
  15 # NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  16 #
  17 #    settings:  $wgUseTex*, $wgUseCategoryMagic*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  18 #               $wgNamespacesWithSubpages, $wgLanguageCode, $wgAllowExternalImages*,
  19 #               $wgLocaltimezone
  20 #
  21 #      * only within ParserOptions
  22 #
  23 #
  24 #----------------------------------------
  25 #    Variable substitution O(N^2) attack
  26 #-----------------------------------------
  27 # Without countermeasures, it would be possible to attack the parser by saving a page
  28 # filled with a large number of inclusions of large pages. The size of the generated
  29 # page would be proportional to the square of the input size. Hence, we limit the number
  30 # of inclusions of any given page, thus bringing any attack back to O(N).
  31 #
  32
  33 define( "MAX_INCLUDE_REPEAT", 5 );
  34 define( "MAX_INCLUDE_SIZE", 1000000 ); // 1 Million
  35
  36 # Allowed values for $mOutputType
  37 define( "OT_HTML", 1 );
  38 define( "OT_WIKI", 2 );
  39 define( "OT_MSG", 3 );
  40
  41 # string parameter for extractTags which will cause it
  42 # to strip HTML comments in addition to regular
  43 # <XML>-style tags. This should not be anything we
  44 # may want to use in wikisyntax
  45 define( "STRIP_COMMENTS", "HTMLCommentStrip" );
  46
  47 # prefix for escaping, used in two functions at least
  48 define( "UNIQ_PREFIX", "NaodW29");
  49
  50 class Parser
  51 {
  52         # Persistent:
  53         var $mTagHooks;
  54
  55         # Cleared with clearState():
  56         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  57         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  58
  59         # Temporary:
  60         var $mOptions, $mTitle, $mOutputType,
  61             $mTemplates,        // cache of already loaded templates, avoids
  62                                 // multiple SQL queries for the same string
  63             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
  64                                 // in this path. Used for loop detection.
  65
  66         function Parser() {
  67                 $this->mTemplates = array();
  68                 $this->mTemplatePath = array();
  69                 $this->mTagHooks = array();
  70                 $this->clearState();
  71         }
  72
  73         function clearState() {
  74                 $this->mOutput = new ParserOutput;
  75                 $this->mAutonumber = 0;
  76                 $this->mLastSection = "";
  77                 $this->mDTopen = false;
  78                 $this->mVariables = false;
  79                 $this->mIncludeCount = array();
  80                 $this->mStripState = array();
  81                 $this->mArgStack = array();
  82                 $this->mInPre = false;
  83         }
  84
  85         # First pass--just handle <nowiki> sections, pass the rest off
  86         # to internalParse() which does all the real work.
  87         #
  88         # Returns a ParserOutput
  89         #
  90         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
  91                 global $wgUseTidy;
  92                 $fname = "Parser::parse";
  93                 wfProfileIn( $fname );
  94
  95                 if ( $clearState ) {
  96                         $this->clearState();
  97                 }
  98
  99                 $this->mOptions = $options;
 100                 $this->mTitle =& $title;
 101                 $this->mOutputType = OT_HTML;
 102
 103                 $stripState = NULL;
 104                 $text = $this->strip( $text, $this->mStripState );
 105                 $text = $this->internalParse( $text, $linestart );
 106                 $text = $this->unstrip( $text, $this->mStripState );
 107                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 108                 if(!$wgUseTidy) {
 109                         $fixtags = array(
 110                                 # french spaces, last one Guillemet-left
 111                                 # only if there is something before the space
 112                                 '/(.) (\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 113                                 # french spaces, Guillemet-right
 114                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 115                                 '/<hr *>/i' => '<hr />',
 116                                 '/<br *>/i' => '<br />',
 117                                 '/<center *>/i' => '<div class="center">',
 118                                 '/<\\/center *>/i' => '</div>',
 119                                 # Clean up spare ampersands; note that we probably ought to be
 120                                 # more careful about named entities.
 121                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 122                         );
 123                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 124                 } else {
 125                         $fixtags = array(
 126                                 # french spaces, last one Guillemet-left
 127                                 '/ (\\?|:|!|\\302\\273)/i' => '&nbsp;\\1',
 128                                 # french spaces, Guillemet-right
 129                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 130                                 '/([^> ]+(&#x30(1|3|9);)[^< ]*)/i' => '<span class="diacrit">\\1</span>',
 131                                 '/<center *>/i' => '<div class="center">',
 132                                 '/<\\/center *>/i' => '</div>'
 133                         );
 134                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 135                 }
 136                 # only once and last
 137                 $text = $this->doBlockLevels( $text, $linestart );
 138                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 139                 if($wgUseTidy) {
 140                         $text = $this->tidy($text);
 141                 }
 142                 $this->mOutput->setText( $text );
 143                 wfProfileOut( $fname );
 144                 return $this->mOutput;
 145         }
 146
 147         /* static */ function getRandomString() {
 148                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 149         }
 150
 151         # Replaces all occurrences of <$tag>content</$tag> in the text
 152         # with a random marker and returns the new text. the output parameter
 153         # $content will be an associative array filled with data on the form
 154         # $unique_marker => content.
 155
 156         # If $content is already set, the additional entries will be appended
 157
 158         # If $tag is set to STRIP_COMMENTS, the function will extract
 159         # <!-- HTML comments -->
 160
 161         /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ""){
 162                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 163                 if ( !$content ) {
 164                         $content = array( );
 165                 }
 166                 $n = 1;
 167                 $stripped = '';
 168
 169                 while ( '' != $text ) {
 170                         if($tag==STRIP_COMMENTS) {
 171                                 $p = preg_split( '/<!--/i', $text, 2 );
 172                         } else {
 173                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 174                         }
 175                         $stripped .= $p[0];
 176                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 177                                 $text = '';
 178                         } else {
 179                                 if($tag==STRIP_COMMENTS) {
 180                                         $q = preg_split( '/-->/i', $p[1], 2 );
 181                                 } else {
 182                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 183                                 }
 184                                 $marker = $rnd . sprintf('%08X', $n++);
 185                                 $content[$marker] = $q[0];
 186                                 $stripped .= $marker;
 187                                 $text = $q[1];
 188                         }
 189                 }
 190                 return $stripped;
 191         }
 192
 193         # Strips and renders <nowiki>, <pre>, <math>, <hiero>
 194         # If $render is set, performs necessary rendering operations on plugins
 195         # Returns the text, and fills an array with data needed in unstrip()
 196         # If the $state is already a valid strip state, it adds to the state
 197
 198         # When $stripcomments is set, HTML comments <!-- like this -->
 199         # will be stripped in addition to other tags. This is important
 200         # for section editing, where these comments cause confusion when
 201         # counting the sections in the wikisource
 202         function strip( $text, &$state, $stripcomments = false ) {
 203                 $render = ($this->mOutputType == OT_HTML);
 204                 $html_content = array();
 205                 $nowiki_content = array();
 206                 $math_content = array();
 207                 $pre_content = array();
 208                 $comment_content = array();
 209                 $ext_content = array();
 210
 211                 # Replace any instances of the placeholders
 212                 $uniq_prefix = UNIQ_PREFIX;
 213                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 214
 215                 # html
 216                 global $wgRawHtml;
 217                 if( $wgRawHtml ) {
 218                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 219                         foreach( $html_content as $marker => $content ) {
 220                                 if ($render ) {
 221                                         # Raw and unchecked for validity.
 222                                         $html_content[$marker] = $content;
 223                                 } else {
 224                                         $html_content[$marker] = "<html>$content</html>";
 225                                 }
 226                         }
 227                 }
 228
 229                 # nowiki
 230                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 231                 foreach( $nowiki_content as $marker => $content ) {
 232                         if( $render ){
 233                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 234                         } else {
 235                                 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
 236                         }
 237                 }
 238
 239                 # math
 240                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 241                 foreach( $math_content as $marker => $content ){
 242                         if( $render ) {
 243                                 if( $this->mOptions->getUseTeX() ) {
 244                                         $math_content[$marker] = renderMath( $content );
 245                                 } else {
 246                                         $math_content[$marker] = "&lt;math&gt;$content&lt;math&gt;";
 247                                 }
 248                         } else {
 249                                 $math_content[$marker] = "<math>$content</math>";
 250                         }
 251                 }
 252
 253                 # pre
 254                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 255                 foreach( $pre_content as $marker => $content ){
 256                         if( $render ){
 257                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 258                         } else {
 259                                 $pre_content[$marker] = "<pre>$content</pre>";
 260                         }
 261                 }
 262
 263                 # Comments
 264                 if($stripcomments) {
 265                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 266                         foreach( $comment_content as $marker => $content ){
 267                                 $comment_content[$marker] = "<!--$content-->";
 268                         }
 269                 }
 270
 271                 # Extensions
 272                 foreach ( $this->mTagHooks as $tag => $callback ) {
 273                         $ext_contents[$tag] = array();
 274                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 275                         foreach( $ext_content[$tag] as $marker => $content ) {
 276                                 if ( $render ) {
 277                                         $ext_content[$tag][$marker] = $callback( $content );
 278                                 } else {
 279                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 280                                 }
 281                         }
 282                 }
 283
 284                 # Merge state with the pre-existing state, if there is one
 285                 if ( $state ) {
 286                         $state['html'] = $state['html'] + $html_content;
 287                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 288                         $state['math'] = $state['math'] + $math_content;
 289                         $state['pre'] = $state['pre'] + $pre_content;
 290                         $state['comment'] = $state['comment'] + $comment_content;
 291
 292                         foreach( $ext_content as $tag => $array ) {
 293                                 if ( array_key_exists( $tag, $state ) ) {
 294                                         $state[$tag] = $state[$tag] + $array;
 295                                 }
 296                         }
 297                 } else {
 298                         $state = array(
 299                           'html' => $html_content,
 300                           'nowiki' => $nowiki_content,
 301                           'math' => $math_content,
 302                           'pre' => $pre_content,
 303                           'comment' => $comment_content,
 304                         ) + $ext_content;
 305                 }
 306                 return $text;
 307         }
 308
 309         # always call unstripNoWiki() after this one
 310         function unstrip( $text, &$state ) {
 311                 # Must expand in reverse order, otherwise nested tags will be corrupted
 312                 $contentDict = end( $state );
 313                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 314                         if( key($state) != 'nowiki' && key($state) != 'html') {
 315                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 316                                         $text = str_replace( key( $contentDict ), $content, $text );
 317                                 }
 318                         }
 319                 }
 320
 321                 return $text;
 322         }
 323         # always call this after unstrip() to preserve the order
 324         function unstripNoWiki( $text, &$state ) {
 325                 # Must expand in reverse order, otherwise nested tags will be corrupted
 326                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 327                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 328                 }
 329
 330                 global $wgRawHtml;
 331                 if ($wgRawHtml) {
 332                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 333                                 $text = str_replace( key( $state['html'] ), $content, $text );
 334                         }
 335                 }
 336
 337                 return $text;
 338         }
 339
 340         # Add an item to the strip state
 341         # Returns the unique tag which must be inserted into the stripped text
 342         # The tag will be replaced with the original text in unstrip()
 343
 344         function insertStripItem( $text, &$state ) {
 345                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 346                 if ( !$state ) {
 347                         $state = array(
 348                           'html' => array(),
 349                           'nowiki' => array(),
 350                           'math' => array(),
 351                           'pre' => array()
 352                         );
 353                 }
 354                 $state['item'][$rnd] = $text;
 355                 return $rnd;
 356         }
 357
 358         # categoryMagic
 359         # generate a list of subcategories and pages for a category
 360         # depending on wfMsg("usenewcategorypage") it either calls the new
 361         # or the old code. The new code will not work properly for some
 362         # languages due to sorting issues, so they might want to turn it
 363         # off.
 364         function categoryMagic() {
 365                 $msg = wfMsg('usenewcategorypage');
 366                 if ( '0' == @$msg[0] )
 367                 {
 368                         return $this->oldCategoryMagic();
 369                 } else {
 370                         return $this->newCategoryMagic();
 371                 }
 372         }
 373
 374         # This method generates the list of subcategories and pages for a category
 375         function oldCategoryMagic () {
 376                 global $wgLang , $wgUser ;
 377                 $fname = 'Parser::oldCategoryMagic';
 378
 379                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 380
 381                 $cns = Namespace::getCategory() ;
 382                 if ( $this->mTitle->getNamespace() != $cns ) return "" ; # This ain't a category page
 383
 384                 $r = "<br style=\"clear:both;\"/>\n";
 385
 386
 387                 $sk =& $wgUser->getSkin() ;
 388
 389                 $articles = array() ;
 390                 $children = array() ;
 391                 $data = array () ;
 392                 $id = $this->mTitle->getArticleID() ;
 393
 394                 # FIXME: add limits
 395                 $dbr =& wfGetDB( DB_SLAVE );
 396                 $cur = $dbr->tableName( 'cur' );
 397                 $categorylinks = $dbr->tableName( 'categorylinks' );
 398
 399                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 400                 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM $cur,$categorylinks " .
 401                         "WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 402                 $res = $dbr->query( $sql, $fname ) ;
 403                 while ( $x = $dbr->fetchObject ( $res ) ) $data[] = $x ;
 404
 405                 # For all pages that link to this category
 406                 foreach ( $data AS $x )
 407                 {
 408                         $t = $wgLang->getNsText ( $x->cur_namespace ) ;
 409                         if ( $t != "" ) $t .= ":" ;
 410                         $t .= $x->cur_title ;
 411
 412                         if ( $x->cur_namespace == $cns ) {
 413                                 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
 414                         } else {
 415                                 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
 416                         }
 417                 }
 418                 $dbr->freeResult ( $res ) ;
 419
 420                 # Showing subcategories
 421                 if ( count ( $children ) > 0 ) {
 422                         $r .= '<h2>'.wfMsg('subcategories')."</h2>\n" ;
 423                         $r .= implode ( ', ' , $children ) ;
 424                 }
 425
 426                 # Showing pages in this category
 427                 if ( count ( $articles ) > 0 ) {
 428                         $ti = $this->mTitle->getText() ;
 429                         $h =  wfMsg( 'category_header', $ti );
 430                         $r .= "<h2>{$h}</h2>\n" ;
 431                         $r .= implode ( ', ' , $articles ) ;
 432                 }
 433
 434                 return $r ;
 435         }
 436
 437
 438
 439         function newCategoryMagic () {
 440                 global $wgLang , $wgUser ;
 441                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 442
 443                 $cns = Namespace::getCategory() ;
 444                 if ( $this->mTitle->getNamespace() != $cns ) return '' ; # This ain't a category page
 445
 446                 $r = "<br style=\"clear:both;\"/>\n";
 447
 448
 449                 $sk =& $wgUser->getSkin() ;
 450
 451                 $articles = array() ;
 452                 $articles_start_char = array();
 453                 $children = array() ;
 454                 $children_start_char = array();
 455                 $data = array () ;
 456                 $id = $this->mTitle->getArticleID() ;
 457
 458                 # FIXME: add limits
 459                 $dbr =& wfGetDB( DB_SLAVE );
 460                 $cur = $dbr->tableName( 'cur' );
 461                 $categorylinks = $dbr->tableName( 'categorylinks' );
 462
 463                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 464                 $sql = "SELECT DISTINCT cur_title,cur_namespace,cl_sortkey FROM " .
 465                         "$cur,$categorylinks WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 466                 $res = $dbr->query ( $sql ) ;
 467                 while ( $x = $dbr->fetchObject ( $res ) )
 468                 {
 469                         $t = $ns = $wgLang->getNsText ( $x->cur_namespace ) ;
 470                         if ( $t != '' ) $t .= ':' ;
 471                         $t .= $x->cur_title ;
 472
 473                         if ( $x->cur_namespace == $cns ) {
 474                                 $ctitle = str_replace( '_',' ',$x->cur_title );
 475                                 array_push ( $children, $sk->makeKnownLink ( $t, $ctitle ) ) ; # Subcategory
 476
 477                                 // If there's a link from Category:A to Category:B, the sortkey of the resulting
 478                                 // entry in the categorylinks table is Category:A, not A, which it SHOULD be.
 479                                 // Workaround: If sortkey == "Category:".$title, than use $title for sorting,
 480                                 // else use sortkey...
 481                                 if ( ($ns.":".$ctitle) ==  $x->cl_sortkey ) {
 482                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cur_title ) );
 483                                 } else {
 484                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 485                                 }
 486                         } else {
 487                                 array_push ( $articles , $sk->makeKnownLink ( $t ) ) ; # Page in this category
 488                                 array_push ( $articles_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 489                         }
 490                 }
 491                 $dbr->freeResult ( $res ) ;
 492
 493                 $ti = $this->mTitle->getText() ;
 494
 495                 # Don't show subcategories section if there are none.
 496                 if ( count ( $children ) > 0 )
 497                 {
 498                         # Showing subcategories
 499                         $r .= '<h2>' . wfMsg( 'subcategories' ) . "</h2>\n"
 500                                 . wfMsg( 'subcategorycount', count( $children ) );
 501                         if ( count ( $children ) > 6 ) {
 502
 503                                 // divide list into three equal chunks
 504                                 $chunk = (int) (count ( $children ) / 3);
 505
 506                                 // get and display header
 507                                 $r .= '<table width="100%"><tr valign="top">';
 508
 509                                 $startChunk = 0;
 510                                 $endChunk = $chunk;
 511
 512                                 // loop through the chunks
 513                                 for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 514                                         $chunkIndex < 3;
 515                                         $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 516                                 {
 517
 518                                         $r .= '<td><ul>';
 519                                         // output all subcategories to category
 520                                         for ($index = $startChunk ;
 521                                                 $index < $endChunk && $index < count($children);
 522                                                 $index++ )
 523                                         {
 524                                                 // check for change of starting letter or begging of chunk
 525                                                 if ( ($children_start_char[$index] != $children_start_char[$index - 1])
 526                                                         || ($index == $startChunk) )
 527                                                 {
 528                                                         $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 529                                                 }
 530
 531                                                 $r .= "<li>{$children[$index]}</li>";
 532                                         }
 533                                         $r .= '</ul></td>';
 534
 535
 536                                 }
 537                                 $r .= '</tr></table>';
 538                         } else {
 539                                 // for short lists of subcategories to category.
 540
 541                                 $r .= "<h3>{$children_start_char[0]}</h3>\n";
 542                                 $r .= '<ul><li>'.$children[0].'</li>';
 543                                 for ($index = 1; $index < count($children); $index++ )
 544                                 {
 545                                         if ($children_start_char[$index] != $children_start_char[$index - 1])
 546                                         {
 547                                                 $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 548                                         }
 549
 550                                         $r .= "<li>{$children[$index]}</li>";
 551                                 }
 552                                 $r .= '</ul>';
 553                         }
 554                 } # END of if ( count($children) > 0 )
 555
 556                 $r .= '<h2>' . wfMsg( 'category_header', $ti ) . "</h2>\n" .
 557                         wfMsg( 'categoryarticlecount', count( $articles ) );
 558
 559                 # Showing articles in this category
 560                 if ( count ( $articles ) > 6) {
 561                         $ti = $this->mTitle->getText() ;
 562
 563                         // divide list into three equal chunks
 564                         $chunk = (int) (count ( $articles ) / 3);
 565
 566                         // get and display header
 567                         $r .= '<table width="100%"><tr valign="top">';
 568
 569                         // loop through the chunks
 570                         for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 571                                 $chunkIndex < 3;
 572                                 $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 573                         {
 574
 575                                 $r .= '<td><ul>';
 576
 577                                 // output all articles in category
 578                                 for ($index = $startChunk ;
 579                                         $index < $endChunk && $index < count($articles);
 580                                         $index++ )
 581                                 {
 582                                         // check for change of starting letter or begging of chunk
 583                                         if ( ($articles_start_char[$index] != $articles_start_char[$index - 1])
 584                                                 || ($index == $startChunk) )
 585                                         {
 586                                                 $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 587                                         }
 588
 589                                         $r .= "<li>{$articles[$index]}</li>";
 590                                 }
 591                                 $r .= '</ul></td>';
 592
 593
 594                         }
 595                         $r .= '</tr></table>';
 596                 } elseif ( count ( $articles )  > 0) {
 597                         // for short lists of articles in categories.
 598                         $ti = $this->mTitle->getText() ;
 599
 600                         $r .= '<h3>'.$articles_start_char[0]."</h3>\n";
 601                         $r .= '<ul><li>'.$articles[0].'</li>';
 602                         for ($index = 1; $index < count($articles); $index++ )
 603                         {
 604                                 if ($articles_start_char[$index] != $articles_start_char[$index - 1])
 605                                 {
 606                                         $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 607                                 }
 608
 609                                 $r .= "<li>{$articles[$index]}</li>";
 610                         }
 611                         $r .= '</ul>';
 612                 }
 613
 614
 615                 return $r ;
 616         }
 617
 618         # Return allowed HTML attributes
 619         function getHTMLattrs () {
 620                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 621                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 622                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 623                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 624                                 /* FONT */ 'type', 'start', 'value', 'compact',
 625                                 /* For various lists, mostly deprecated but safe */
 626                                 'summary', 'width', 'border', 'frame', 'rules',
 627                                 'cellspacing', 'cellpadding', 'valign', 'char',
 628                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 629                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 630                                 'id', 'class', 'name', 'style' /* For CSS */
 631                                 );
 632                 return $htmlattrs ;
 633         }
 634
 635         # Remove non approved attributes and javascript in css
 636         function fixTagAttributes ( $t ) {
 637                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 638                 $htmlattrs = $this->getHTMLattrs() ;
 639
 640                 # Strip non-approved attributes from the tag
 641                 $t = preg_replace(
 642                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 643                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 644                         $t);
 645                 # Strip javascript "expression" from stylesheets. Brute force approach:
 646                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 647
 648                 if( preg_match(
 649                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 650                         wfMungeToUtf8( $t ) ) )
 651                 {
 652                         $t='';
 653                 }
 654
 655                 return trim ( $t ) ;
 656         }
 657
 658         # interface with html tidy, used if $wgUseTidy = true
 659         function tidy ( $text ) {
 660                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 661                 global $wgInputEncoding, $wgOutputEncoding;
 662                 $fname = 'Parser::tidy';
 663                 wfProfileIn( $fname );
 664
 665                 $cleansource = '';
 666                 switch(strtoupper($wgOutputEncoding)) {
 667                         case 'ISO-8859-1':
 668                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 669                                 break;
 670                         case 'UTF-8':
 671                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 672                                 break;
 673                         default:
 674                                 $wgTidyOpts .= ' -raw';
 675                         }
 676
 677                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 678 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 679 '<head><title>test</title></head><body>'.$text.'</body></html>';
 680                 $descriptorspec = array(
 681                         0 => array('pipe', 'r'),
 682                         1 => array('pipe', 'w'),
 683                         2 => array('file', '/dev/null', 'a')
 684                 );
 685                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 686                 if (is_resource($process)) {
 687                         fwrite($pipes[0], $wrappedtext);
 688                         fclose($pipes[0]);
 689                         while (!feof($pipes[1])) {
 690                                 $cleansource .= fgets($pipes[1], 1024);
 691                         }
 692                         fclose($pipes[1]);
 693                         $return_value = proc_close($process);
 694                 }
 695
 696                 wfProfileOut( $fname );
 697
 698                 if( $cleansource == '' && $text != '') {
 699                         wfDebug( "Tidy error detected!\n" );
 700                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 701                 } else {
 702                         return $cleansource;
 703                 }
 704         }
 705
 706         # parse the wiki syntax used to render tables
 707         function doTableStuff ( $t ) {
 708                 $fname = 'Parser::doTableStuff';
 709                 wfProfileIn( $fname );
 710
 711                 $t = explode ( "\n" , $t ) ;
 712                 $td = array () ; # Is currently a td tag open?
 713                 $ltd = array () ; # Was it TD or TH?
 714                 $tr = array () ; # Is currently a tr tag open?
 715                 $ltr = array () ; # tr attributes
 716                 $indent_level = 0; # indent level of the table
 717                 foreach ( $t AS $k => $x )
 718                 {
 719                         $x = trim ( $x ) ;
 720                         $fc = substr ( $x , 0 , 1 ) ;
 721                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) )
 722                         {
 723                                 $indent_level = strlen( $matches[1] );
 724                                 $t[$k] = "\n" .
 725                                         str_repeat( "<dl><dd>", $indent_level ) .
 726                                         "<table " . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 727                                 array_push ( $td , false ) ;
 728                                 array_push ( $ltd , '' ) ;
 729                                 array_push ( $tr , false ) ;
 730                                 array_push ( $ltr , '' ) ;
 731                         }
 732                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 733                         else if ( '|}' == substr ( $x , 0 , 2 ) )
 734                         {
 735                                 $z = "</table>\n" ;
 736                                 $l = array_pop ( $ltd ) ;
 737                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 738                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 739                                 array_pop ( $ltr ) ;
 740                                 $t[$k] = $z . str_repeat( "</dd></dl>", $indent_level );
 741                         }
 742                         else if ( '|-' == substr ( $x , 0 , 2 ) ) # Allows for |---------------
 743                         {
 744                                 $x = substr ( $x , 1 ) ;
 745                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 746                                 $z = '' ;
 747                                 $l = array_pop ( $ltd ) ;
 748                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 749                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 750                                 array_pop ( $ltr ) ;
 751                                 $t[$k] = $z ;
 752                                 array_push ( $tr , false ) ;
 753                                 array_push ( $td , false ) ;
 754                                 array_push ( $ltd , '' ) ;
 755                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 756                         }
 757                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) # Caption
 758                         {
 759                                 if ( '|+' == substr ( $x , 0 , 2 ) )
 760                                 {
 761                                         $fc = '+' ;
 762                                         $x = substr ( $x , 1 ) ;
 763                                 }
 764                                 $after = substr ( $x , 1 ) ;
 765                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 766                                 $after = explode ( '||' , $after ) ;
 767                                 $t[$k] = '' ;
 768                                 foreach ( $after AS $theline )
 769                                 {
 770                                         $z = '' ;
 771                                         if ( $fc != '+' )
 772                                         {
 773                                                 $tra = array_pop ( $ltr ) ;
 774                                                 if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
 775                                                 array_push ( $tr , true ) ;
 776                                                 array_push ( $ltr , '' ) ;
 777                                         }
 778
 779                                         $l = array_pop ( $ltd ) ;
 780                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 781                                         if ( $fc == '|' ) $l = 'td' ;
 782                                         else if ( $fc == '!' ) $l = 'th' ;
 783                                         else if ( $fc == '+' ) $l = 'caption' ;
 784                                         else $l = '' ;
 785                                         array_push ( $ltd , $l ) ;
 786                                         $y = explode ( '|' , $theline , 2 ) ;
 787                                         if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 788                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 789                                         $t[$k] .= $y ;
 790                                         array_push ( $td , true ) ;
 791                                 }
 792                         }
 793                 }
 794
 795                 # Closing open td, tr && table
 796                 while ( count ( $td ) > 0 )
 797                 {
 798                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 799                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 800                         $t[] = '</table>' ;
 801                 }
 802
 803                 $t = implode ( "\n" , $t ) ;
 804                 #               $t = $this->removeHTMLtags( $t );
 805                 wfProfileOut( $fname );
 806                 return $t ;
 807         }
 808
 809         # Parses the text and adds the result to the strip state
 810         # Returns the strip tag
 811         function stripParse( $text, $newline, $args )
 812         {
 813                 $text = $this->strip( $text, $this->mStripState );
 814                 $text = $this->internalParse( $text, (bool)$newline, $args, false );
 815                 return $newline.$this->insertStripItem( $text, $this->mStripState );
 816         }
 817
 818         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 819                 $fname = 'Parser::internalParse';
 820                 wfProfileIn( $fname );
 821
 822                 $text = $this->removeHTMLtags( $text );
 823                 $text = $this->replaceVariables( $text, $args );
 824
 825                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 826
 827                 $text = $this->doHeadings( $text );
 828                 if($this->mOptions->getUseDynamicDates()) {
 829                         global $wgDateFormatter;
 830                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 831                 }
 832                 $text = $this->doAllQuotes( $text );
 833                 // $text = $this->doExponent( $text );
 834                 $text = $this->replaceExternalLinks( $text );
 835                 $text = $this->doMagicLinks( $text );
 836                 $text = $this->replaceInternalLinks ( $text );
 837                 $text = $this->replaceInternalLinks ( $text );
 838                 //$text = $this->doTokenizedParser ( $text );
 839                 $text = $this->doTableStuff( $text );
 840                 $text = $this->formatHeadings( $text, $isMain );
 841                 $sk =& $this->mOptions->getSkin();
 842                 $text = $sk->transformContent( $text );
 843
 844                 if ( $isMain && !isset ( $this->categoryMagicDone ) ) {
 845                         $text .= $this->categoryMagic () ;
 846                         $this->categoryMagicDone = true ;
 847                 }
 848
 849                 wfProfileOut( $fname );
 850                 return $text;
 851         }
 852
 853         /* private */ function &doMagicLinks( &$text ) {
 854                 $text = $this->magicISBN( $text );
 855                 $text = $this->magicGEO( $text );
 856                 $text = $this->magicRFC( $text );
 857                 return $text;
 858         }
 859
 860         # Parse ^^ tokens and return html
 861         /* private */ function doExponent ( $text )
 862         {
 863                 $fname = 'Parser::doExponent';
 864                 wfProfileIn( $fname);
 865                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 866                 wfProfileOut( $fname);
 867                 return $text;
 868         }
 869
 870     # Parse headers and return html
 871         /* private */ function doHeadings( $text ) {
 872                 $fname = 'Parser::doHeadings';
 873                 wfProfileIn( $fname );
 874                 for ( $i = 6; $i >= 1; --$i ) {
 875                         $h = substr( '======', 0, $i );
 876                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 877                           "<h{$i}>\\1</h{$i}>\\2", $text );
 878                 }
 879                 wfProfileOut( $fname );
 880                 return $text;
 881         }
 882
 883         /* private */ function doAllQuotes( $text ) {
 884                 $fname = 'Parser::doAllQuotes';
 885                 wfProfileIn( $fname );
 886                 $outtext = '';
 887                 $lines = explode( "\n", $text );
 888                 foreach ( $lines as $line ) {
 889                         $outtext .= $this->doQuotes ( '', $line, '' ) . "\n";
 890                 }
 891                 $outtext = substr($outtext, 0,-1);
 892                 wfProfileOut( $fname );
 893                 return $outtext;
 894         }
 895
 896         /* private */ function doQuotes( $pre, $text, $mode ) {
 897                 if ( preg_match( "/^(.*)''(.*)$/sU", $text, $m ) ) {
 898                         $m1_strong = ($m[1] == "") ? "" : "<strong>{$m[1]}</strong>";
 899                         $m1_em = ($m[1] == "") ? "" : "<em>{$m[1]}</em>";
 900                         if ( substr ($m[2], 0, 1) == '\'' ) {
 901                                 $m[2] = substr ($m[2], 1);
 902                                 if ($mode == 'em') {
 903                                         return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'emstrong' );
 904                                 } else if ($mode == 'strong') {
 905                                         return $m1_strong . $this->doQuotes ( '', $m[2], '' );
 906                                 } else if (($mode == 'emstrong') || ($mode == 'both')) {
 907                                         return $this->doQuotes ( '', $pre.$m1_strong.$m[2], 'em' );
 908                                 } else if ($mode == 'strongem') {
 909                                         return "<strong>{$pre}{$m1_em}</strong>" . $this->doQuotes ( '', $m[2], 'em' );
 910                                 } else {
 911                                         return $m[1] . $this->doQuotes ( '', $m[2], 'strong' );
 912                                 }
 913                         } else {
 914                                 if ($mode == 'strong') {
 915                                         return $this->doQuotes ( $m[1], $m[2], ($m[1] == '') ? 'both' : 'strongem' );
 916                                 } else if ($mode == 'em') {
 917                                         return $m1_em . $this->doQuotes ( '', $m[2], '' );
 918                                 } else if ($mode == 'emstrong') {
 919                                         return "<em>{$pre}{$m1_strong}</em>" . $this->doQuotes ( '', $m[2], 'strong' );
 920                                 } else if (($mode == 'strongem') || ($mode == 'both')) {
 921                                         return $this->doQuotes ( '', $pre.$m1_em.$m[2], 'strong' );
 922                                 } else {
 923                                         return $m[1] . $this->doQuotes ( '', $m[2], 'em' );
 924                                 }
 925                         }
 926                 } else {
 927                         $text_strong = ($text == '') ? '' : "<strong>{$text}</strong>";
 928                         $text_em = ($text == '') ? '' : "<em>{$text}</em>";
 929                         if ($mode == '') {
 930                                 return $pre . $text;
 931                         } else if ($mode == 'em') {
 932                                 return $pre . $text_em;
 933                         } else if ($mode == 'strong') {
 934                                 return $pre . $text_strong;
 935                         } else if ($mode == 'strongem') {
 936                                 return (($pre == '') && ($text == '')) ? '' : "<strong>{$pre}{$text_em}</strong>";
 937                         } else {
 938                                 return (($pre == '') && ($text == '')) ? '' : "<em>{$pre}{$text_strong}</em>";
 939                         }
 940                 }
 941         }
 942
 943         # Note: we have to do external links before the internal ones,
 944         # and otherwise take great care in the order of things here, so
 945         # that we don't end up interpreting some URLs twice.
 946
 947         /* private */ function replaceExternalLinks( $text ) {
 948                 $fname = 'Parser::replaceExternalLinks';
 949                 wfProfileIn( $fname );
 950                 $text = $this->subReplaceExternalLinks( $text, 'http', true );
 951                 $text = $this->subReplaceExternalLinks( $text, 'https', true );
 952                 $text = $this->subReplaceExternalLinks( $text, 'ftp', false );
 953                 $text = $this->subReplaceExternalLinks( $text, 'irc', false );
 954                 $text = $this->subReplaceExternalLinks( $text, 'gopher', false );
 955                 $text = $this->subReplaceExternalLinks( $text, 'news', false );
 956                 $text = $this->subReplaceExternalLinks( $text, 'mailto', false );
 957                 wfProfileOut( $fname );
 958                 return $text;
 959         }
 960
 961         /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber ) {
 962                 $unique = '4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3';
 963                 $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
 964
 965                 # this is  the list of separators that should be ignored if they
 966                 # are the last character of an URL but that should be included
 967                 # if they occur within the URL, e.g. "go to www.foo.com, where .."
 968                 # in this case, the last comma should not become part of the URL,
 969                 # but in "www.foo.com/123,2342,32.htm" it should.
 970                 $sep = ",;\.:";
 971                 $fnc = 'A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF';
 972                 $images = 'gif|png|jpg|jpeg';
 973
 974                 # PLEASE NOTE: The curly braces { } are not part of the regex,
 975                 # they are interpreted as part of the string (used to tell PHP
 976                 # that the content of the string should be inserted there).
 977                 $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
 978                   "((?i){$images})([^{$uc}]|$)/";
 979
 980                 $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
 981                 $sk =& $this->mOptions->getSkin();
 982
 983                 if ( $autonumber and $this->mOptions->getAllowExternalImages() ) { # Use img tags only for HTTP urls
 984                         $s = preg_replace( $e1, '\\1' . $sk->makeImage( "{$unique}:\\3" .
 985                           '/\\4.\\5', '\\4.\\5' ) . '\\6', $s );
 986                 }
 987                 $s = preg_replace( $e2, '\\1' . "<a href=\"{$unique}:\\3\"" .
 988                   $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
 989                   "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
 990                   '</a>\\5', $s );
 991                 $s = str_replace( $unique, $protocol, $s );
 992
 993                 $a = explode( "[{$protocol}:", " " . $s );
 994                 $s = array_shift( $a );
 995                 $s = substr( $s, 1 );
 996
 997                 # Regexp for URL in square brackets
 998                 $e1 = "/^([{$uc}{$sep}]+)\\](.*)\$/sD";
 999                 # Regexp for URL with link text in square brackets
1000                 $e2 = "/^([{$uc}{$sep}]+)\\s+([^\\]]+)\\](.*)\$/sD";
1001
1002                 foreach ( $a as $line ) {
1003
1004                         # CASE 1: Link in square brackets, e.g.
1005                         # some text [http://domain.tld/some.link] more text
1006                         if ( preg_match( $e1, $line, $m ) ) {
1007                                 $link = "{$protocol}:{$m[1]}";
1008                                 $trail = $m[2];
1009                                 if ( $autonumber ) { $text = "[" . ++$this->mAutonumber . "]"; }
1010                                 else { $text = wfEscapeHTML( $link ); }
1011                         }
1012
1013                         # CASE 2: Link with link text and text directly following it, e.g.
1014                         # This is a collection of [http://domain.tld/some.link link]s
1015                         else if ( preg_match( $e2, $line, $m ) ) {
1016                                 $link = "{$protocol}:{$m[1]}";
1017                                 $text = $m[2];
1018                                 $dtrail = '';
1019                                 $trail = $m[3];
1020                                 if ( preg_match( wfMsg ('linktrail'), $trail, $m2 ) ) {
1021                                         $dtrail = $m2[1];
1022                                         $trail = $m2[2];
1023                                 }
1024                         }
1025
1026                         # CASE 3: Nothing matches, just output the source text
1027                         else {
1028                                 $s .= "[{$protocol}:" . $line;
1029                                 continue;
1030                         }
1031
1032                         if( $link == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
1033                                 $paren = '';
1034                         } else {
1035                                 # Expand the URL for printable version
1036                                 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
1037                         }
1038                         $la = $sk->getExternalLinkAttributes( $link, $text );
1039                         $s .= "<a href='{$link}'{$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
1040
1041                 }
1042                 return $s;
1043         }
1044
1045
1046         /* private */ function replaceInternalLinks( $s ) {
1047                 global $wgLang, $wgLinkCache;
1048                 global $wgNamespacesWithSubpages, $wgLanguageCode;
1049                 static $fname = 'Parser::replaceInternalLinks' ;
1050                 wfProfileIn( $fname );
1051
1052                 wfProfileIn( $fname.'-setup' );
1053                 static $tc = FALSE;
1054                 # the % is needed to support urlencoded titles as well
1055                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1056                 $sk =& $this->mOptions->getSkin();
1057
1058                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1059                 $isRedirect = $redirect->matchStart ( strtoupper ( substr ( $s , 0 , 10 ) ) ) ;
1060
1061                 $a = explode( '[[', ' ' . $s );
1062                 $s = array_shift( $a );
1063                 $s = substr( $s, 1 );
1064
1065                 # Match a link having the form [[namespace:link|alternate]]trail
1066                 static $e1 = FALSE;
1067                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1068                 # Match the end of a line for a word that's not followed by whitespace,
1069                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1070                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1071
1072                 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
1073                 # Special and Media are pseudo-namespaces; no pages actually exist in them
1074                 static $image = FALSE;
1075                 static $special = FALSE;
1076                 static $media = FALSE;
1077                 static $category = FALSE;
1078                 if ( !$image ) { $image = Namespace::getImage(); }
1079                 if ( !$special ) { $special = Namespace::getSpecial(); }
1080                 if ( !$media ) { $media = Namespace::getMedia(); }
1081                 if ( !$category ) { $category = Namespace::getCategory(); }
1082
1083                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1084
1085                 if ( $useLinkPrefixExtension ) {
1086                         if ( preg_match( $e2, $s, $m ) ) {
1087                                 $first_prefix = $m[2];
1088                                 $s = $m[1];
1089                         } else {
1090                                 $first_prefix = false;
1091                         }
1092                 } else {
1093                         $prefix = '';
1094                 }
1095
1096                 wfProfileOut( $fname.'-setup' );
1097
1098                 foreach ( $a as $line ) {
1099                         wfProfileIn( $fname.'-prefixhandling' );
1100                         if ( $useLinkPrefixExtension ) {
1101                                 if ( preg_match( $e2, $s, $m ) ) {
1102                                         $prefix = $m[2];
1103                                         $s = $m[1];
1104                                 } else {
1105                                         $prefix='';
1106                                 }
1107                                 # first link
1108                                 if($first_prefix) {
1109                                         $prefix = $first_prefix;
1110                                         $first_prefix = false;
1111                                 }
1112                         }
1113                         wfProfileOut( $fname.'-prefixhandling' );
1114
1115                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1116                                 $text = $m[2];
1117                                 # fix up urlencoded title texts
1118                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1119                                 $trail = $m[3];
1120                         } else { # Invalid form; output directly
1121                                 $s .= $prefix . '[[' . $line ;
1122                                 continue;
1123                         }
1124
1125                         /* Valid link forms:
1126                         Foobar -- normal
1127                         :Foobar -- override special treatment of prefix (images, language links)
1128                         /Foobar -- convert to CurrentPage/Foobar
1129                         /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1130                         */
1131                         $c = substr($m[1],0,1);
1132                         $noforce = ($c != ':');
1133                         if( $c == '/' ) { # subpage
1134                                 if(substr($m[1],-1,1)=='/') {                 # / at end means we don't want the slash to be shown
1135                                         $m[1]=substr($m[1],1,strlen($m[1])-2);
1136                                         $noslash=$m[1];
1137                                 } else {
1138                                         $noslash=substr($m[1],1);
1139                                 }
1140                                 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) { # subpages allowed here
1141                                         $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1142                                         if( '' == $text ) {
1143                                                 $text= $m[1];
1144                                         } # this might be changed for ugliness reasons
1145                                 } else {
1146                                         $link = $noslash; # no subpage allowed, use standard link
1147                                 }
1148                         } elseif( $noforce ) { # no subpage
1149                                 $link = $m[1];
1150                         } else {
1151                                 $link = substr( $m[1], 1 );
1152                         }
1153                         $wasblank = ( '' == $text );
1154                         if( $wasblank )
1155                         $text = $link;
1156
1157                         $nt = Title::newFromText( $link );
1158                         if( !$nt ) {
1159                                 $s .= $prefix . '[[' . $line;
1160                                 continue;
1161                         }
1162                         $ns = $nt->getNamespace();
1163                         $iw = $nt->getInterWiki();
1164                         if( $noforce ) {
1165                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1166                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1167                                         $tmp = $prefix . $trail ;
1168                                         $s .= (trim($tmp) == '')? '': $tmp;
1169                                         continue;
1170                                 }
1171                                 if ( $ns == $image ) {
1172                                         $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1173                                         $wgLinkCache->addImageLinkObj( $nt );
1174                                         continue;
1175                                 }
1176                                 if ( $ns == $category && !$isRedirect ) {
1177                                         $t = $nt->getText() ;
1178                                         $nnt = Title::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
1179
1180                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1181                                         $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1182                                         $wgLinkCache->resume();
1183
1184                                         $sortkey = $wasblank ? $this->mTitle->getPrefixedText() : $text;
1185                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1186                                         $this->mOutput->mCategoryLinks[] = $t ;
1187                                         $s .= $prefix . $trail ;
1188                                         continue;
1189                                 }
1190                         }
1191                         if( ( $nt->getPrefixedText() == $this->mTitle->getPrefixedText() ) &&
1192                         ( strpos( $link, '#' ) == FALSE ) ) {
1193                                 # Self-links are handled specially; generally de-link and change to bold.
1194                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1195                                 continue;
1196                         }
1197
1198                         if( $ns == $media ) {
1199                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1200                                 $wgLinkCache->addImageLinkObj( $nt );
1201                                 continue;
1202                         } elseif( $ns == $special ) {
1203                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1204                                 continue;
1205                         }
1206                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1207                 }
1208                 wfProfileOut( $fname );
1209                 return $s;
1210         }
1211
1212         # Some functions here used by doBlockLevels()
1213         #
1214         /* private */ function closeParagraph() {
1215                 $result = '';
1216                 if ( '' != $this->mLastSection ) {
1217                         $result = '</' . $this->mLastSection  . ">\n";
1218                 }
1219                 $this->mInPre = false;
1220                 $this->mLastSection = '';
1221                 return $result;
1222         }
1223         # getCommon() returns the length of the longest common substring
1224         # of both arguments, starting at the beginning of both.
1225         #
1226         /* private */ function getCommon( $st1, $st2 ) {
1227                 $fl = strlen( $st1 );
1228                 $shorter = strlen( $st2 );
1229                 if ( $fl < $shorter ) { $shorter = $fl; }
1230
1231                 for ( $i = 0; $i < $shorter; ++$i ) {
1232                         if ( $st1{$i} != $st2{$i} ) { break; }
1233                 }
1234                 return $i;
1235         }
1236         # These next three functions open, continue, and close the list
1237         # element appropriate to the prefix character passed into them.
1238         #
1239         /* private */ function openList( $char )
1240     {
1241                 $result = $this->closeParagraph();
1242
1243                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1244                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1245                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1246                 else if ( ';' == $char ) {
1247                         $result .= '<dl><dt>';
1248                         $this->mDTopen = true;
1249                 }
1250                 else { $result = '<!-- ERR 1 -->'; }
1251
1252                 return $result;
1253         }
1254
1255         /* private */ function nextItem( $char ) {
1256                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1257                 else if ( ':' == $char || ';' == $char ) {
1258                         $close = "</dd>";
1259                         if ( $this->mDTopen ) { $close = '</dt>'; }
1260                         if ( ';' == $char ) {
1261                                 $this->mDTopen = true;
1262                                 return $close . '<dt>';
1263                         } else {
1264                                 $this->mDTopen = false;
1265                                 return $close . '<dd>';
1266                         }
1267                 }
1268                 return '<!-- ERR 2 -->';
1269         }
1270
1271         /* private */function closeList( $char ) {
1272                 if ( '*' == $char ) { $text = '</li></ul>'; }
1273                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1274                 else if ( ':' == $char ) {
1275                         if ( $this->mDTopen ) {
1276                                 $this->mDTopen = false;
1277                                 $text = '</dt></dl>';
1278                         } else {
1279                                 $text = '</dd></dl>';
1280                         }
1281                 }
1282                 else {  return '<!-- ERR 3 -->'; }
1283                 return $text."\n";
1284         }
1285
1286         /* private */ function doBlockLevels( $text, $linestart ) {
1287                 $fname = 'Parser::doBlockLevels';
1288                 wfProfileIn( $fname );
1289
1290                 # Parsing through the text line by line.  The main thing
1291                 # happening here is handling of block-level elements p, pre,
1292                 # and making lists from lines starting with * # : etc.
1293                 #
1294                 $textLines = explode( "\n", $text );
1295
1296                 $lastPrefix = $output = $lastLine = '';
1297                 $this->mDTopen = $inBlockElem = false;
1298                 $prefixLength = 0;
1299                 $paragraphStack = false;
1300
1301                 if ( !$linestart ) {
1302                         $output .= array_shift( $textLines );
1303                 }
1304                 foreach ( $textLines as $oLine ) {
1305                         $lastPrefixLength = strlen( $lastPrefix );
1306                         $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1307                         $preOpenMatch = preg_match("/<pre/i", $oLine );
1308                         if ( !$this->mInPre ) {
1309                                 # Multiple prefixes may abut each other for nested lists.
1310                                 $prefixLength = strspn( $oLine, '*#:;' );
1311                                 $pref = substr( $oLine, 0, $prefixLength );
1312
1313                                 # eh?
1314                                 $pref2 = str_replace( ';', ':', $pref );
1315                                 $t = substr( $oLine, $prefixLength );
1316                                 $this->mInPre = !empty($preOpenMatch);
1317                         } else {
1318                                 # Don't interpret any other prefixes in preformatted text
1319                                 $prefixLength = 0;
1320                                 $pref = $pref2 = '';
1321                                 $t = $oLine;
1322                         }
1323
1324                         # List generation
1325                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1326                                 # Same as the last item, so no need to deal with nesting or opening stuff
1327                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1328                                 $paragraphStack = false;
1329
1330                                 if ( ";" == substr( $pref, -1 ) ) {
1331                                         # The one nasty exception: definition lists work like this:
1332                                         # ; title : definition text
1333                                         # So we check for : in the remainder text to split up the
1334                                         # title and definition, without b0rking links.
1335                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1336                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1337                                                 $term = $match[1];
1338                                                 $output .= $term . $this->nextItem( ':' );
1339                                                 $t = $match[2];
1340                                         }
1341                                 }
1342                         } elseif( $prefixLength || $lastPrefixLength ) {
1343                                 # Either open or close a level...
1344                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1345                                 $paragraphStack = false;
1346
1347                                 while( $commonPrefixLength < $lastPrefixLength ) {
1348                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1349                                         --$lastPrefixLength;
1350                                 }
1351                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1352                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1353                                 }
1354                                 while ( $prefixLength > $commonPrefixLength ) {
1355                                         $char = substr( $pref, $commonPrefixLength, 1 );
1356                                         $output .= $this->openList( $char );
1357
1358                                         if ( ';' == $char ) {
1359                                                 # FIXME: This is dupe of code above
1360                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1361                                                         $term = $match[1];
1362                                                         $output .= $term . $this->nextItem( ":" );
1363                                                         $t = $match[2];
1364                                                 }
1365                                         }
1366                                         ++$commonPrefixLength;
1367                                 }
1368                                 $lastPrefix = $pref2;
1369                         }
1370                         if( 0 == $prefixLength ) {
1371                                 # No prefix (not in list)--go to paragraph mode
1372                                 $uniq_prefix = UNIQ_PREFIX;
1373                                 // XXX: use a stack for nestable elements like span, table and div
1374                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1375                                 $closematch = preg_match(
1376                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1377                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1378                                 if ( $openmatch or $closematch ) {
1379                                         $paragraphStack = false;
1380                                         $output .= $this->closeParagraph();
1381                                         if($preOpenMatch and !$preCloseMatch) {
1382                                                 $this->mInPre = true;
1383                                         }
1384                                         if ( $closematch  ) {
1385                                                 $inBlockElem = false;
1386                                         } else {
1387                                                 $inBlockElem = true;
1388                                         }
1389                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1390                                         if ( " " == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1391                                                 // pre
1392                                                 if ($this->mLastSection != 'pre') {
1393                                                         $paragraphStack = false;
1394                                                         $output .= $this->closeParagraph().'<pre>';
1395                                                         $this->mLastSection = 'pre';
1396                                                 }
1397                                         } else {
1398                                                 // paragraph
1399                                                 if ( '' == trim($t) ) {
1400                                                         if ( $paragraphStack ) {
1401                                                                 $output .= $paragraphStack.'<br />';
1402                                                                 $paragraphStack = false;
1403                                                                 $this->mLastSection = 'p';
1404                                                         } else {
1405                                                                 if ($this->mLastSection != 'p' ) {
1406                                                                         $output .= $this->closeParagraph();
1407                                                                         $this->mLastSection = '';
1408                                                                         $paragraphStack = '<p>';
1409                                                                 } else {
1410                                                                         $paragraphStack = '</p><p>';
1411                                                                 }
1412                                                         }
1413                                                 } else {
1414                                                         if ( $paragraphStack ) {
1415                                                                 $output .= $paragraphStack;
1416                                                                 $paragraphStack = false;
1417                                                                 $this->mLastSection = 'p';
1418                                                         } else if ($this->mLastSection != 'p') {
1419                                                                 $output .= $this->closeParagraph().'<p>';
1420                                                                 $this->mLastSection = 'p';
1421                                                         }
1422                                                 }
1423                                         }
1424                                 }
1425                         }
1426                         if ($paragraphStack === false) {
1427                                 $output .= $t."\n";
1428                         }
1429                 }
1430                 while ( $prefixLength ) {
1431                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1432                         --$prefixLength;
1433                 }
1434                 if ( '' != $this->mLastSection ) {
1435                         $output .= '</' . $this->mLastSection . '>';
1436                         $this->mLastSection = '';
1437                 }
1438
1439                 wfProfileOut( $fname );
1440                 return $output;
1441         }
1442
1443         # Return value of a magic variable (like PAGENAME)
1444         function getVariableValue( $index ) {
1445                 global $wgLang, $wgSitename, $wgServer;
1446
1447                 switch ( $index ) {
1448                         case MAG_CURRENTMONTH:
1449                                 return $wgLang->formatNum( date( 'm' ) );
1450                         case MAG_CURRENTMONTHNAME:
1451                                 return $wgLang->getMonthName( date('n') );
1452                         case MAG_CURRENTMONTHNAMEGEN:
1453                                 return $wgLang->getMonthNameGen( date('n') );
1454                         case MAG_CURRENTDAY:
1455                                 return $wgLang->formatNum( date('j') );
1456                         case MAG_PAGENAME:
1457                                 return $this->mTitle->getText();
1458                         case MAG_NAMESPACE:
1459                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1460                                 return $wgLang->getNsText($this->mTitle->getNamespace()); // Patch  by Dori
1461                         case MAG_CURRENTDAYNAME:
1462                                 return $wgLang->getWeekdayName( date('w')+1 );
1463                         case MAG_CURRENTYEAR:
1464                                 return $wgLang->formatNum( date( 'Y' ) );
1465                         case MAG_CURRENTTIME:
1466                                 return $wgLang->time( wfTimestampNow(), false );
1467                         case MAG_NUMBEROFARTICLES:
1468                                 return $wgLang->formatNum( wfNumberOfArticles() );
1469                         case MAG_SITENAME:
1470                                 return $wgSitename;
1471                         case MAG_SERVER:
1472                                 return $wgServer;
1473                         default:
1474                                 return NULL;
1475                 }
1476         }
1477
1478         # initialise the magic variables (like CURRENTMONTHNAME)
1479         function initialiseVariables() {
1480                 global $wgVariableIDs;
1481                 $this->mVariables = array();
1482                 foreach ( $wgVariableIDs as $id ) {
1483                         $mw =& MagicWord::get( $id );
1484                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1485                 }
1486         }
1487
1488         /* private */ function replaceVariables( $text, $args = array() ) {
1489                 global $wgLang, $wgScript, $wgArticlePath;
1490
1491                 # Prevent too big inclusions
1492                 if(strlen($text)> MAX_INCLUDE_SIZE)
1493                    return $text;
1494
1495                 $fname = 'Parser::replaceVariables';
1496                 wfProfileIn( $fname );
1497
1498                 $bail = false;
1499                 $titleChars = Title::legalChars();
1500                 $nonBraceChars = str_replace( array( '{', '}' ), array( '', '' ), $titleChars );
1501
1502                 # This function is called recursively. To keep track of arguments we need a stack:
1503                 array_push( $this->mArgStack, $args );
1504
1505                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1506                 $GLOBALS['wgCurParser'] =& $this;
1507
1508
1509                 if ( $this->mOutputType == OT_HTML ) {
1510                         # Variable substitution
1511                         $text = preg_replace_callback( "/{{([$nonBraceChars]*?)}}/", 'wfVariableSubstitution', $text );
1512
1513                         # Argument substitution
1514                         $text = preg_replace_callback( "/(\\n?){{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1515                 }
1516                 # Template substitution
1517                 $regex = '/(\\n?){{(['.$nonBraceChars.']*)(\\|.*?|)}}/s';
1518                 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1519
1520                 array_pop( $this->mArgStack );
1521
1522                 wfProfileOut( $fname );
1523                 return $text;
1524         }
1525
1526         function variableSubstitution( $matches ) {
1527                 if ( !$this->mVariables ) {
1528                         $this->initialiseVariables();
1529                 }
1530                 if ( array_key_exists( $matches[1], $this->mVariables ) ) {
1531                         $text = $this->mVariables[$matches[1]];
1532                         $this->mOutput->mContainsOldMagic = true;
1533                 } else {
1534                         $text = $matches[0];
1535                 }
1536                 return $text;
1537         }
1538
1539         # Split template arguments
1540         function getTemplateArgs( $argsString ) {
1541                 if ( $argsString === '' ) {
1542                         return array();
1543                 }
1544
1545                 $args = explode( '|', substr( $argsString, 1 ) );
1546
1547                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1548                 # merged with the next arg because the '|' character between belongs
1549                 # to the link syntax and not the template parameter syntax.
1550                 $argc = count($args);
1551                 $i = 0;
1552                 for ( $i = 0; $i < $argc-1; $i++ ) {
1553                         if ( substr_count ( $args[$i], "[[" ) != substr_count ( $args[$i], "]]" ) ) {
1554                                 $args[$i] .= "|".$args[$i+1];
1555                                 array_splice($args, $i+1, 1);
1556                                 $i--;
1557                                 $argc--;
1558                         }
1559                 }
1560
1561                 return $args;
1562         }
1563
1564         function braceSubstitution( $matches ) {
1565                 global $wgLinkCache, $wgLang;
1566                 $fname = 'Parser::braceSubstitution';
1567                 $found = false;
1568                 $nowiki = false;
1569                 $noparse = false;
1570
1571                 $title = NULL;
1572
1573                 # $newline is an optional newline character before the braces
1574                 # $part1 is the bit before the first |, and must contain only title characters
1575                 # $args is a list of arguments, starting from index 0, not including $part1
1576
1577                 $newline = $matches[1];
1578                 $part1 = $matches[2];
1579                 # If the third subpattern matched anything, it will start with |
1580
1581                 $args = $this->getTemplateArgs($matches[3]);
1582                 $argc = count( $args );
1583
1584                 # {{{}}}
1585                 if ( strpos( $matches[0], '{{{' ) !== false ) {
1586                         $text = $matches[0];
1587                         $found = true;
1588                         $noparse = true;
1589                 }
1590
1591                 # SUBST
1592                 if ( !$found ) {
1593                         $mwSubst =& MagicWord::get( MAG_SUBST );
1594                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1595                                 if ( $this->mOutputType != OT_WIKI ) {
1596                                         # Invalid SUBST not replaced at PST time
1597                                         # Return without further processing
1598                                         $text = $matches[0];
1599                                         $found = true;
1600                                         $noparse= true;
1601                                 }
1602                         } elseif ( $this->mOutputType == OT_WIKI ) {
1603                                 # SUBST not found in PST pass, do nothing
1604                                 $text = $matches[0];
1605                                 $found = true;
1606                         }
1607                 }
1608
1609                 # MSG, MSGNW and INT
1610                 if ( !$found ) {
1611                         # Check for MSGNW:
1612                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1613                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1614                                 $nowiki = true;
1615                         } else {
1616                                 # Remove obsolete MSG:
1617                                 $mwMsg =& MagicWord::get( MAG_MSG );
1618                                 $mwMsg->matchStartAndRemove( $part1 );
1619                         }
1620
1621                         # Check if it is an internal message
1622                         $mwInt =& MagicWord::get( MAG_INT );
1623                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1624                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1625                                         $text = wfMsgReal( $part1, $args, true );
1626                                         $found = true;
1627                                 }
1628                         }
1629                 }
1630
1631                 # NS
1632                 if ( !$found ) {
1633                         # Check for NS: (namespace expansion)
1634                         $mwNs = MagicWord::get( MAG_NS );
1635                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1636                                 if ( intval( $part1 ) ) {
1637                                         $text = $wgLang->getNsText( intval( $part1 ) );
1638                                         $found = true;
1639                                 } else {
1640                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1641                                         if ( !is_null( $index ) ) {
1642                                                 $text = $wgLang->getNsText( $index );
1643                                                 $found = true;
1644                                         }
1645                                 }
1646                         }
1647                 }
1648
1649                 # LOCALURL and LOCALURLE
1650                 if ( !$found ) {
1651                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1652                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1653
1654                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1655                                 $func = 'getLocalURL';
1656                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1657                                 $func = 'escapeLocalURL';
1658                         } else {
1659                                 $func = '';
1660                         }
1661
1662                         if ( $func !== '' ) {
1663                                 $title = Title::newFromText( $part1 );
1664                                 if ( !is_null( $title ) ) {
1665                                         if ( $argc > 0 ) {
1666                                                 $text = $title->$func( $args[0] );
1667                                         } else {
1668                                                 $text = $title->$func();
1669                                         }
1670                                         $found = true;
1671                                 }
1672                         }
1673                 }
1674
1675                 # Internal variables
1676                 if ( !$this->mVariables ) {
1677                         $this->initialiseVariables();
1678                 }
1679                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1680                         $text = $this->mVariables[$part1];
1681                         $found = true;
1682                         $this->mOutput->mContainsOldMagic = true;
1683                 }
1684
1685                 # Template table test
1686
1687                 # Did we encounter this template already? If yes, it is in the cache
1688                 # and we need to check for loops.
1689                 if ( isset( $this->mTemplates[$part1] ) ) {
1690                         # Infinite loop test
1691                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1692                                 $noparse = true;
1693                                 $found = true;
1694                         }
1695                         # set $text to cached message.
1696                         $text = $this->mTemplates[$part1];
1697                         $found = true;
1698                 }
1699
1700                 # Load from database
1701                 if ( !$found ) {
1702                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1703                         if ( !is_null( $title ) && !$title->isExternal() ) {
1704                                 # Check for excessive inclusion
1705                                 $dbk = $title->getPrefixedDBkey();
1706                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1707                                         # This should never be reached.
1708                                         $article = new Article( $title );
1709                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1710                                         if ( $articleContent !== false ) {
1711                                                 $found = true;
1712                                                 $text = $articleContent;
1713
1714                                         }
1715                                 }
1716
1717                                 # If the title is valid but undisplayable, make a link to it
1718                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1719                                         $text = '[[' . $title->getPrefixedText() . ']]';
1720                                         $found = true;
1721                                 }
1722
1723                                 # Template cache array insertion
1724                                 $this->mTemplates[$part1] = $text;
1725                         }
1726                 }
1727
1728                 # Recursive parsing, escaping and link table handling
1729                 # Only for HTML output
1730                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1731                         $text = wfEscapeWikiText( $text );
1732                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1733                         # Clean up argument array
1734                         $assocArgs = array();
1735                         $index = 1;
1736                         foreach( $args as $arg ) {
1737                                 $eqpos = strpos( $arg, '=' );
1738                                 if ( $eqpos === false ) {
1739                                         $assocArgs[$index++] = $arg;
1740                                 } else {
1741                                         $name = trim( substr( $arg, 0, $eqpos ) );
1742                                         $value = trim( substr( $arg, $eqpos+1 ) );
1743                                         if ( $value === false ) {
1744                                                 $value = '';
1745                                         }
1746                                         if ( $name !== false ) {
1747                                                 $assocArgs[$name] = $value;
1748                                         }
1749                                 }
1750                         }
1751
1752                         # Do not enter included links in link table
1753                         if ( !is_null( $title ) ) {
1754                                 $wgLinkCache->suspend();
1755                         }
1756
1757                         # Add a new element to the templace recursion path
1758                         $this->mTemplatePath[$part1] = 1;
1759
1760                         # Run full parser on the included text
1761                         $text = $this->stripParse( $text, $newline, $assocArgs );
1762
1763                         # Resume the link cache and register the inclusion as a link
1764                         if ( !is_null( $title ) ) {
1765                                 $wgLinkCache->resume();
1766                                 $wgLinkCache->addLinkObj( $title );
1767                         }
1768                 }
1769                 # Empties the template path
1770                 $this->mTemplatePath = array();
1771
1772                 if ( !$found ) {
1773                         return $matches[0];
1774                 } else {
1775                         return $text;
1776                 }
1777         }
1778
1779         # Triple brace replacement -- used for template arguments
1780         function argSubstitution( $matches ) {
1781                 $newline = $matches[1];
1782                 $arg = trim( $matches[2] );
1783                 $text = $matches[0];
1784                 $inputArgs = end( $this->mArgStack );
1785
1786                 if ( array_key_exists( $arg, $inputArgs ) ) {
1787                         $text = $this->stripParse( $inputArgs[$arg], $newline, array() );
1788                 }
1789
1790                 return $text;
1791         }
1792
1793         # Returns true if the function is allowed to include this entity
1794         function incrementIncludeCount( $dbk ) {
1795                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1796                         $this->mIncludeCount[$dbk] = 0;
1797                 }
1798                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1799                         return true;
1800                 } else {
1801                         return false;
1802                 }
1803         }
1804
1805
1806         # Cleans up HTML, removes dangerous tags and attributes
1807         /* private */ function removeHTMLtags( $text ) {
1808                 global $wgUseTidy, $wgUserHtml;
1809                 $fname = 'Parser::removeHTMLtags';
1810                 wfProfileIn( $fname );
1811
1812                 if( $wgUserHtml ) {
1813                         $htmlpairs = array( # Tags that must be closed
1814                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1815                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1816                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
1817                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1818                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
1819                         );
1820                         $htmlsingle = array(
1821                                 'br', 'hr', 'li', 'dt', 'dd'
1822                         );
1823                         $htmlnest = array( # Tags that can be nested--??
1824                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1825                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
1826                         );
1827                         $tabletags = array( # Can only appear inside table
1828                                 'td', 'th', 'tr'
1829                         );
1830                 } else {
1831                         $htmlpairs = array();
1832                         $htmlsingle = array();
1833                         $htmlnest = array();
1834                         $tabletags = array();
1835                 }
1836
1837                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1838                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1839
1840                 $htmlattrs = $this->getHTMLattrs () ;
1841
1842                 # Remove HTML comments
1843                 $text = preg_replace( '/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU', '$2', $text );
1844
1845                 $bits = explode( '<', $text );
1846                 $text = array_shift( $bits );
1847                 if(!$wgUseTidy) {
1848                         $tagstack = array(); $tablestack = array();
1849                         foreach ( $bits as $x ) {
1850                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1851                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1852                                 $x, $regs );
1853                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1854                                 error_reporting( $prev );
1855
1856                                 $badtag = 0 ;
1857                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1858                                         # Check our stack
1859                                         if ( $slash ) {
1860                                                 # Closing a tag...
1861                                                 if ( ! in_array( $t, $htmlsingle ) &&
1862                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
1863                                                         @array_push( $tagstack, $ot );
1864                                                         $badtag = 1;
1865                                                 } else {
1866                                                         if ( $t == 'table' ) {
1867                                                                 $tagstack = array_pop( $tablestack );
1868                                                         }
1869                                                         $newparams = '';
1870                                                 }
1871                                         } else {
1872                                                 # Keep track for later
1873                                                 if ( in_array( $t, $tabletags ) &&
1874                                                 ! in_array( 'table', $tagstack ) ) {
1875                                                         $badtag = 1;
1876                                                 } else if ( in_array( $t, $tagstack ) &&
1877                                                 ! in_array ( $t , $htmlnest ) ) {
1878                                                         $badtag = 1 ;
1879                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1880                                                         if ( $t == 'table' ) {
1881                                                                 array_push( $tablestack, $tagstack );
1882                                                                 $tagstack = array();
1883                                                         }
1884                                                         array_push( $tagstack, $t );
1885                                                 }
1886                                                 # Strip non-approved attributes from the tag
1887                                                 $newparams = $this->fixTagAttributes($params);
1888
1889                                         }
1890                                         if ( ! $badtag ) {
1891                                                 $rest = str_replace( '>', '&gt;', $rest );
1892                                                 $text .= "<$slash$t $newparams$brace$rest";
1893                                                 continue;
1894                                         }
1895                                 }
1896                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1897                         }
1898                         # Close off any remaining tags
1899                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
1900                                 $text .= "</$t>\n";
1901                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
1902                         }
1903                 } else {
1904                         # this might be possible using tidy itself
1905                         foreach ( $bits as $x ) {
1906                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1907                                 $x, $regs );
1908                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1909                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1910                                         $newparams = $this->fixTagAttributes($params);
1911                                         $rest = str_replace( '>', '&gt;', $rest );
1912                                         $text .= "<$slash$t $newparams$brace$rest";
1913                                 } else {
1914                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1915                                 }
1916                         }
1917                 }
1918                 wfProfileOut( $fname );
1919                 return $text;
1920         }
1921
1922
1923 /*
1924  *
1925  * This function accomplishes several tasks:
1926  * 1) Auto-number headings if that option is enabled
1927  * 2) Add an [edit] link to sections for logged in users who have enabled the option
1928  * 3) Add a Table of contents on the top for users who have enabled the option
1929  * 4) Auto-anchor headings
1930  *
1931  * It loops through all headlines, collects the necessary data, then splits up the
1932  * string and re-inserts the newly formatted headlines.
1933  *
1934  */
1935
1936         /* private */ function formatHeadings( $text, $isMain=true ) {
1937                 global $wgInputEncoding, $wgMaxTocLevel;
1938
1939                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1940                 $doShowToc = $this->mOptions->getShowToc();
1941                 $forceTocHere = false;
1942                 if( !$this->mTitle->userCanEdit() ) {
1943                         $showEditLink = 0;
1944                         $rightClickHack = 0;
1945                 } else {
1946                         $showEditLink = $this->mOptions->getEditSection();
1947                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1948                 }
1949
1950                 # Inhibit editsection links if requested in the page
1951                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1952                 if( $esw->matchAndRemove( $text ) ) {
1953                         $showEditLink = 0;
1954                 }
1955                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1956                 # do not add TOC
1957                 $mw =& MagicWord::get( MAG_NOTOC );
1958                 if( $mw->matchAndRemove( $text ) ) {
1959                         $doShowToc = 0;
1960                 }
1961
1962                 # never add the TOC to the Main Page. This is an entry page that should not
1963                 # be more than 1-2 screens large anyway
1964                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
1965                         $doShowToc = 0;
1966                 }
1967
1968                 # Get all headlines for numbering them and adding funky stuff like [edit]
1969                 # links - this is for later, but we need the number of headlines right now
1970                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
1971
1972                 # if there are fewer than 4 headlines in the article, do not show TOC
1973                 if( $numMatches < 4 ) {
1974                         $doShowToc = 0;
1975                 }
1976
1977                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
1978                 # override above conditions and always show TOC at that place
1979                 $mw =& MagicWord::get( MAG_TOC );
1980                 if ($mw->match( $text ) ) {
1981                         $doShowToc = 1;
1982                         $forceTocHere = true;
1983                 } else {
1984                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1985                         # override above conditions and always show TOC above first header
1986                         $mw =& MagicWord::get( MAG_FORCETOC );
1987                         if ($mw->matchAndRemove( $text ) ) {
1988                                 $doShowToc = 1;
1989                         }
1990                 }
1991
1992
1993
1994                 # We need this to perform operations on the HTML
1995                 $sk =& $this->mOptions->getSkin();
1996
1997                 # headline counter
1998                 $headlineCount = 0;
1999
2000                 # Ugh .. the TOC should have neat indentation levels which can be
2001                 # passed to the skin functions. These are determined here
2002                 $toclevel = 0;
2003                 $toc = '';
2004                 $full = '';
2005                 $head = array();
2006                 $sublevelCount = array();
2007                 $level = 0;
2008                 $prevlevel = 0;
2009                 foreach( $matches[3] as $headline ) {
2010                         $numbering = '';
2011                         if( $level ) {
2012                                 $prevlevel = $level;
2013                         }
2014                         $level = $matches[1][$headlineCount];
2015                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2016                                 # reset when we enter a new level
2017                                 $sublevelCount[$level] = 0;
2018                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2019                                 $toclevel += $level - $prevlevel;
2020                         }
2021                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2022                                 # reset when we step back a level
2023                                 $sublevelCount[$level+1]=0;
2024                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2025                                 $toclevel -= $prevlevel - $level;
2026                         }
2027                         # count number of headlines for each level
2028                         @$sublevelCount[$level]++;
2029                         if( $doNumberHeadings || $doShowToc ) {
2030                                 $dot = 0;
2031                                 for( $i = 1; $i <= $level; $i++ ) {
2032                                         if( !empty( $sublevelCount[$i] ) ) {
2033                                                 if( $dot ) {
2034                                                         $numbering .= '.';
2035                                                 }
2036                                                 $numbering .= $sublevelCount[$i];
2037                                                 $dot = 1;
2038                                         }
2039                                 }
2040                         }
2041
2042                         # The canonized header is a version of the header text safe to use for links
2043                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2044                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2045                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2046
2047                         # strip out HTML
2048                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2049                         $tocline = trim( $canonized_headline );
2050                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2051                         $replacearray = array(
2052                                 '%3A' => ':',
2053                                 '%' => '.'
2054                         );
2055                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2056                         $refer[$headlineCount] = $canonized_headline;
2057
2058                         # count how many in assoc. array so we can track dupes in anchors
2059                         @$refers[$canonized_headline]++;
2060                         $refcount[$headlineCount]=$refers[$canonized_headline];
2061
2062                         # Prepend the number to the heading text
2063
2064                         if( $doNumberHeadings || $doShowToc ) {
2065                                 $tocline = $numbering . ' ' . $tocline;
2066
2067                                 # Don't number the heading if it is the only one (looks silly)
2068                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2069                                         # the two are different if the line contains a link
2070                                         $headline=$numbering . ' ' . $headline;
2071                                 }
2072                         }
2073
2074                         # Create the anchor for linking from the TOC to the section
2075                         $anchor = $canonized_headline;
2076                         if($refcount[$headlineCount] > 1 ) {
2077                                 $anchor .= '_' . $refcount[$headlineCount];
2078                         }
2079                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2080                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2081                         }
2082                         if( $showEditLink ) {
2083                                 if ( empty( $head[$headlineCount] ) ) {
2084                                         $head[$headlineCount] = '';
2085                                 }
2086                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
2087                         }
2088
2089                         # Add the edit section span
2090                         if( $rightClickHack ) {
2091                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
2092                         }
2093
2094                         # give headline the correct <h#> tag
2095                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
2096
2097                         $headlineCount++;
2098                 }
2099
2100                 if( $doShowToc ) {
2101                         $toclines = $headlineCount;
2102                         $toc .= $sk->tocUnindent( $toclevel );
2103                         $toc = $sk->tocTable( $toc );
2104                 }
2105
2106                 # split up and insert constructed headlines
2107
2108                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2109                 $i = 0;
2110
2111                 foreach( $blocks as $block ) {
2112                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2113                             # This is the [edit] link that appears for the top block of text when
2114                                 # section editing is enabled
2115
2116                                 # Disabled because it broke block formatting
2117                                 # For example, a bullet point in the top line
2118                                 # $full .= $sk->editSectionLink(0);
2119                         }
2120                         $full .= $block;
2121                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2122                         # Top anchor now in skin
2123                                 $full = $full.$toc;
2124                         }
2125
2126                         if( !empty( $head[$i] ) ) {
2127                                 $full .= $head[$i];
2128                         }
2129                         $i++;
2130                 }
2131                 if($forceTocHere) {
2132                         $mw =& MagicWord::get( MAG_TOC );
2133                         return $mw->replace( $toc, $full );
2134                 } else {
2135                         return $full;
2136                 }
2137         }
2138
2139         # Return an HTML link for the "ISBN 123456" text
2140         /* private */ function magicISBN( $text ) {
2141                 global $wgLang;
2142                 $fname = 'Parser::magicISBN';
2143                 wfProfileIn( $fname );
2144
2145                 $a = split( 'ISBN ', " $text" );
2146                 if ( count ( $a ) < 2 ) {
2147                         wfProfileOut( $fname );
2148                         return $text;
2149                 }
2150                 $text = substr( array_shift( $a ), 1);
2151                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2152
2153                 foreach ( $a as $x ) {
2154                         $isbn = $blank = '' ;
2155                         while ( ' ' == $x{0} ) {
2156                                 $blank .= ' ';
2157                                 $x = substr( $x, 1 );
2158                         }
2159                         while ( strstr( $valid, $x{0} ) != false ) {
2160                                 $isbn .= $x{0};
2161                                 $x = substr( $x, 1 );
2162                         }
2163                         $num = str_replace( '-', '', $isbn );
2164                         $num = str_replace( ' ', '', $num );
2165
2166                         if ( '' == $num ) {
2167                                 $text .= "ISBN $blank$x";
2168                         } else {
2169                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2170                                 $text .= '<a href="' .
2171                                 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
2172                                         "\" class=\"internal\">ISBN $isbn</a>";
2173                                 $text .= $x;
2174                         }
2175                 }
2176                 wfProfileOut( $fname );
2177                 return $text;
2178         }
2179
2180         # Return an HTML link for the "GEO ..." text
2181         /* private */ function magicGEO( $text ) {
2182                 global $wgLang, $wgUseGeoMode;
2183                 if ( !isset ( $wgUseGeoMode ) || !$wgUseGeoMode ) return $text ;
2184                 $fname = 'Parser::magicGEO';
2185                 wfProfileIn( $fname );
2186
2187                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2188                 $directions = array ( "N" => "North" , "S" => "South" , "E" => "East" , "W" => "West" ) ;
2189                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2190                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2191                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2192                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2193
2194                 $a = split( 'GEO ', " $text" );
2195                 if ( count ( $a ) < 2 ) {
2196                         wfProfileOut( $fname );
2197                         return $text;
2198                 }
2199                 $text = substr( array_shift( $a ), 1);
2200                 $valid = '0123456789.+-:';
2201
2202                 foreach ( $a as $x ) {
2203                         $geo = $blank = '' ;
2204                         while ( ' ' == $x{0} ) {
2205                                 $blank .= ' ';
2206                                 $x = substr( $x, 1 );
2207                         }
2208                         while ( strstr( $valid, $x{0} ) != false ) {
2209                                 $geo .= $x{0};
2210                                 $x = substr( $x, 1 );
2211                         }
2212                         $num = str_replace( '+', '', $geo );
2213                         $num = str_replace( ' ', '', $num );
2214
2215                         if ( '' == $num || count ( explode ( ":" , $num , 3 ) ) < 2 ) {
2216                                 $text .= "GEO $blank$x";
2217                         } else {
2218                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2219                                 $text .= '<a href="' .
2220                                 $titleObj->escapeLocalUrl( "coordinates={$num}" ) .
2221                                         "\" class=\"internal\">GEO $geo</a>";
2222                                 $text .= $x;
2223                         }
2224                 }
2225                 wfProfileOut( $fname );
2226                 return $text;
2227         }
2228
2229         # Return an HTML link for the "RFC 1234" text
2230         /* private */ function magicRFC( $text ) {
2231                 global $wgLang;
2232
2233                 $a = split( 'RFC ', ' '.$text );
2234                 if ( count ( $a ) < 2 ) return $text;
2235                 $text = substr( array_shift( $a ), 1);
2236                 $valid = '0123456789';
2237
2238                 foreach ( $a as $x ) {
2239                         $rfc = $blank = '' ;
2240                         while ( ' ' == $x{0} ) {
2241                                 $blank .= ' ';
2242                                 $x = substr( $x, 1 );
2243                         }
2244                         while ( strstr( $valid, $x{0} ) != false ) {
2245                                 $rfc .= $x{0};
2246                                 $x = substr( $x, 1 );
2247                         }
2248
2249                         if ( '' == $rfc ) {
2250                                 $text .= "RFC $blank$x";
2251                         } else {
2252                                 $url = wfmsg( 'rfcurl' );
2253                                 $url = str_replace( '$1', $rfc, $url);
2254                                 $sk =& $this->mOptions->getSkin();
2255                                 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
2256                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2257                         }
2258                 }
2259                 return $text;
2260         }
2261
2262         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2263                 $this->mOptions = $options;
2264                 $this->mTitle =& $title;
2265                 $this->mOutputType = OT_WIKI;
2266
2267                 if ( $clearState ) {
2268                         $this->clearState();
2269                 }
2270
2271                 $stripState = false;
2272                 $pairs = array(
2273                         "\r\n" => "\n",
2274                         );
2275                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2276                 // now with regexes
2277                 /*
2278                 $pairs = array(
2279                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2280                         "/<br *?>/i" => "<br />",
2281                 );
2282                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2283                 */
2284                 $text = $this->strip( $text, $stripState, false );
2285                 $text = $this->pstPass2( $text, $user );
2286                 $text = $this->unstrip( $text, $stripState );
2287                 $text = $this->unstripNoWiki( $text, $stripState );
2288                 return $text;
2289         }
2290
2291         /* private */ function pstPass2( $text, &$user ) {
2292                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2293
2294                 # Variable replacement
2295                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2296                 $text = $this->replaceVariables( $text );
2297
2298                 # Signatures
2299                 #
2300                 $n = $user->getName();
2301                 $k = $user->getOption( 'nickname' );
2302                 if ( '' == $k ) { $k = $n; }
2303                 if(isset($wgLocaltimezone)) {
2304                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2305                 }
2306                 /* Note: this is an ugly timezone hack for the European wikis */
2307                 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2308                   ' (' . date( 'T' ) . ')';
2309                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2310
2311                 $text = preg_replace( '/~~~~~/', $d, $text );
2312                 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText(
2313                   Namespace::getUser() ) . ":$n|$k]] $d", $text );
2314                 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText(
2315                   Namespace::getUser() ) . ":$n|$k]]", $text );
2316
2317                 # Context links: [[|name]] and [[name (context)|]]
2318                 #
2319                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2320                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2321                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2322                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2323
2324                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2325                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2326                 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/";          # [[namespace:page|]]
2327                 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2328                                                                                                                 # [[ns:page (cont)|]]
2329                 $context = "";
2330                 $t = $this->mTitle->getText();
2331                 if ( preg_match( $conpat, $t, $m ) ) {
2332                         $context = $m[2];
2333                 }
2334                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2335                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2336                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2337
2338                 if ( '' == $context ) {
2339                         $text = preg_replace( $p2, '[[\\1]]', $text );
2340                 } else {
2341                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2342                 }
2343
2344                 /*
2345                 $mw =& MagicWord::get( MAG_SUBST );
2346                 $wgCurParser = $this->fork();
2347                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2348                 $this->merge( $wgCurParser );
2349                 */
2350
2351                 # Trim trailing whitespace
2352                 # MAG_END (__END__) tag allows for trailing
2353                 # whitespace to be deliberately included
2354                 $text = rtrim( $text );
2355                 $mw =& MagicWord::get( MAG_END );
2356                 $mw->matchAndRemove( $text );
2357
2358                 return $text;
2359         }
2360
2361         # Set up some variables which are usually set up in parse()
2362         # so that an external function can call some class members with confidence
2363         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2364                 $this->mTitle =& $title;
2365                 $this->mOptions = $options;
2366                 $this->mOutputType = $outputType;
2367                 if ( $clearState ) {
2368                         $this->clearState();
2369                 }
2370         }
2371
2372         function transformMsg( $text, $options ) {
2373                 global $wgTitle;
2374                 static $executing = false;
2375
2376                 # Guard against infinite recursion
2377                 if ( $executing ) {
2378                         return $text;
2379                 }
2380                 $executing = true;
2381
2382                 $this->mTitle = $wgTitle;
2383                 $this->mOptions = $options;
2384                 $this->mOutputType = OT_MSG;
2385                 $this->clearState();
2386                 $text = $this->replaceVariables( $text );
2387
2388                 $executing = false;
2389                 return $text;
2390         }
2391
2392         # Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2393         # Callback will be called with the text within
2394         # Transform and return the text within
2395         function setHook( $tag, $callback ) {
2396                 $oldVal = @$this->mTagHooks[$tag];
2397                 $this->mTagHooks[$tag] = $callback;
2398                 return $oldVal;
2399         }
2400 }
2401
2402 class ParserOutput
2403 {
2404         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2405         var $mCacheTime; # Used in ParserCache
2406
2407         function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2408                 $containsOldMagic = false )
2409         {
2410                 $this->mText = $text;
2411                 $this->mLanguageLinks = $languageLinks;
2412                 $this->mCategoryLinks = $categoryLinks;
2413                 $this->mContainsOldMagic = $containsOldMagic;
2414                 $this->mCacheTime = "";
2415         }
2416
2417         function getText() { return $this->mText; }
2418         function getLanguageLinks() { return $this->mLanguageLinks; }
2419         function getCategoryLinks() { return $this->mCategoryLinks; }
2420         function getCacheTime() { return $this->mCacheTime; }
2421         function containsOldMagic() { return $this->mContainsOldMagic; }
2422         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2423         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2424         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2425         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2426         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2427
2428         function merge( $other ) {
2429                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2430                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2431                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2432         }
2433
2434 }
2435
2436 class ParserOptions
2437 {
2438         # All variables are private
2439         var $mUseTeX;                    # Use texvc to expand <math> tags
2440         var $mUseCategoryMagic;          # Treat [[Category:xxxx]] tags specially
2441         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2442         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2443         var $mAllowExternalImages;       # Allow external images inline
2444         var $mSkin;                      # Reference to the preferred skin
2445         var $mDateFormat;                # Date format index
2446         var $mEditSection;               # Create "edit section" links
2447         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2448         var $mNumberHeadings;            # Automatically number headings
2449         var $mShowToc;                   # Show table of contents
2450
2451         function getUseTeX()                        { return $this->mUseTeX; }
2452         function getUseCategoryMagic()              { return $this->mUseCategoryMagic; }
2453         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2454         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2455         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2456         function getSkin()                          { return $this->mSkin; }
2457         function getDateFormat()                    { return $this->mDateFormat; }
2458         function getEditSection()                   { return $this->mEditSection; }
2459         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
2460         function getNumberHeadings()                { return $this->mNumberHeadings; }
2461         function getShowToc()                       { return $this->mShowToc; }
2462
2463         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
2464         function setUseCategoryMagic( $x )          { return wfSetVar( $this->mUseCategoryMagic, $x ); }
2465         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
2466         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
2467         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
2468         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
2469         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
2470         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2471         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
2472         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
2473
2474     function setSkin( &$x ) { $this->mSkin =& $x; }
2475
2476         /* static */ function newFromUser( &$user ) {
2477                 $popts = new ParserOptions;
2478                 $popts->initialiseFromUser( $user );
2479                 return $popts;
2480         }
2481
2482         function initialiseFromUser( &$userInput ) {
2483                 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2484
2485                 if ( !$userInput ) {
2486                         $user = new User;
2487                         $user->setLoaded( true );
2488                 } else {
2489                         $user =& $userInput;
2490                 }
2491
2492                 $this->mUseTeX = $wgUseTeX;
2493                 $this->mUseCategoryMagic = $wgUseCategoryMagic;
2494                 $this->mUseDynamicDates = $wgUseDynamicDates;
2495                 $this->mInterwikiMagic = $wgInterwikiMagic;
2496                 $this->mAllowExternalImages = $wgAllowExternalImages;
2497                 $this->mSkin =& $user->getSkin();
2498                 $this->mDateFormat = $user->getOption( 'date' );
2499                 $this->mEditSection = $user->getOption( 'editsection' );
2500                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2501                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2502                 $this->mShowToc = $user->getOption( 'showtoc' );
2503         }
2504
2505
2506 }
2507
2508 # Regex callbacks, used in Parser::replaceVariables
2509 function wfBraceSubstitution( $matches )
2510 {
2511         global $wgCurParser;
2512         return $wgCurParser->braceSubstitution( $matches );
2513 }
2514
2515 function wfArgSubstitution( $matches )
2516 {
2517         global $wgCurParser;
2518         return $wgCurParser->argSubstitution( $matches );
2519 }
2520
2521 function wfVariableSubstitution( $matches )
2522 {
2523         global $wgCurParser;
2524         return $wgCurParser->variableSubstitution( $matches );
2525 }
2526
2527 ?>