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
  51 # Constants needed for external link processing
  52
  53 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  54 define( 'HTTP_PROTOCOLS', 'http|https' );
  55 # Everything except bracket, space, or control characters
  56 define( 'EXT_LINK_URL_CLASS', '[^]\\x00-\\x20\\x7F]' );
  57 define( 'INVERSE_EXT_LINK_URL_CLASS', '[\]\\x00-\\x20\\x7F]' );
  58 # Including space
  59 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  60 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  61 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  62 define( 'EXT_LINK_BRACKETED',  '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  63 define( 'EXT_IMAGE_REGEX',
  64         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  65         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  66         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  67 );
  68
  69 class Parser
  70 {
  71         # Persistent:
  72         var $mTagHooks;
  73
  74         # Cleared with clearState():
  75         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  76         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  77
  78         # Temporary:
  79         var $mOptions, $mTitle, $mOutputType,
  80             $mTemplates,        // cache of already loaded templates, avoids
  81                                 // multiple SQL queries for the same string
  82             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
  83                                 // in this path. Used for loop detection.
  84
  85         function Parser() {
  86                 $this->mTemplates = array();
  87                 $this->mTemplatePath = array();
  88                 $this->mTagHooks = array();
  89                 $this->clearState();
  90         }
  91
  92         function clearState() {
  93                 $this->mOutput = new ParserOutput;
  94                 $this->mAutonumber = 0;
  95                 $this->mLastSection = "";
  96                 $this->mDTopen = false;
  97                 $this->mVariables = false;
  98                 $this->mIncludeCount = array();
  99                 $this->mStripState = array();
 100                 $this->mArgStack = array();
 101                 $this->mInPre = false;
 102         }
 103
 104         # First pass--just handle <nowiki> sections, pass the rest off
 105         # to internalParse() which does all the real work.
 106         #
 107         # Returns a ParserOutput
 108         #
 109         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 110                 global $wgUseTidy;
 111                 $fname = "Parser::parse";
 112                 wfProfileIn( $fname );
 113
 114                 if ( $clearState ) {
 115                         $this->clearState();
 116                 }
 117
 118                 $this->mOptions = $options;
 119                 $this->mTitle =& $title;
 120                 $this->mOutputType = OT_HTML;
 121
 122                 $stripState = NULL;
 123                 $text = $this->strip( $text, $this->mStripState );
 124                 $text = $this->internalParse( $text, $linestart );
 125                 $text = $this->unstrip( $text, $this->mStripState );
 126                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 127                 if(!$wgUseTidy) {
 128                         $fixtags = array(
 129                                 # french spaces, last one Guillemet-left
 130                                 # only if there is something before the space
 131                                 '/(.) (\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 132                                 # french spaces, Guillemet-right
 133                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 134                                 '/<hr *>/i' => '<hr />',
 135                                 '/<br *>/i' => '<br />',
 136                                 '/<center *>/i' => '<div class="center">',
 137                                 '/<\\/center *>/i' => '</div>',
 138                                 # Clean up spare ampersands; note that we probably ought to be
 139                                 # more careful about named entities.
 140                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 141                         );
 142                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 143                 } else {
 144                         $fixtags = array(
 145                                 # french spaces, last one Guillemet-left
 146                                 '/ (\\?|:|!|\\302\\273)/i' => '&nbsp;\\1',
 147                                 # french spaces, Guillemet-right
 148                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 149                                 '/([^> ]+(&#x30(1|3|9);)[^< ]*)/i' => '<span class="diacrit">\\1</span>',
 150                                 '/<center *>/i' => '<div class="center">',
 151                                 '/<\\/center *>/i' => '</div>'
 152                         );
 153                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 154                 }
 155                 # only once and last
 156                 $text = $this->doBlockLevels( $text, $linestart );
 157                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 158                 if($wgUseTidy) {
 159                         $text = $this->tidy($text);
 160                 }
 161                 $this->mOutput->setText( $text );
 162                 wfProfileOut( $fname );
 163                 return $this->mOutput;
 164         }
 165
 166         /* static */ function getRandomString() {
 167                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 168         }
 169
 170         # Replaces all occurrences of <$tag>content</$tag> in the text
 171         # with a random marker and returns the new text. the output parameter
 172         # $content will be an associative array filled with data on the form
 173         # $unique_marker => content.
 174
 175         # If $content is already set, the additional entries will be appended
 176
 177         # If $tag is set to STRIP_COMMENTS, the function will extract
 178         # <!-- HTML comments -->
 179
 180         /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ""){
 181                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 182                 if ( !$content ) {
 183                         $content = array( );
 184                 }
 185                 $n = 1;
 186                 $stripped = '';
 187
 188                 while ( '' != $text ) {
 189                         if($tag==STRIP_COMMENTS) {
 190                                 $p = preg_split( '/<!--/i', $text, 2 );
 191                         } else {
 192                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 193                         }
 194                         $stripped .= $p[0];
 195                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 196                                 $text = '';
 197                         } else {
 198                                 if($tag==STRIP_COMMENTS) {
 199                                         $q = preg_split( '/-->/i', $p[1], 2 );
 200                                 } else {
 201                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 202                                 }
 203                                 $marker = $rnd . sprintf('%08X', $n++);
 204                                 $content[$marker] = $q[0];
 205                                 $stripped .= $marker;
 206                                 $text = $q[1];
 207                         }
 208                 }
 209                 return $stripped;
 210         }
 211
 212         # Strips and renders <nowiki>, <pre>, <math>, <hiero>
 213         # If $render is set, performs necessary rendering operations on plugins
 214         # Returns the text, and fills an array with data needed in unstrip()
 215         # If the $state is already a valid strip state, it adds to the state
 216
 217         # When $stripcomments is set, HTML comments <!-- like this -->
 218         # will be stripped in addition to other tags. This is important
 219         # for section editing, where these comments cause confusion when
 220         # counting the sections in the wikisource
 221         function strip( $text, &$state, $stripcomments = false ) {
 222                 $render = ($this->mOutputType == OT_HTML);
 223                 $html_content = array();
 224                 $nowiki_content = array();
 225                 $math_content = array();
 226                 $pre_content = array();
 227                 $comment_content = array();
 228                 $ext_content = array();
 229
 230                 # Replace any instances of the placeholders
 231                 $uniq_prefix = UNIQ_PREFIX;
 232                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 233
 234                 # html
 235                 global $wgRawHtml;
 236                 if( $wgRawHtml ) {
 237                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 238                         foreach( $html_content as $marker => $content ) {
 239                                 if ($render ) {
 240                                         # Raw and unchecked for validity.
 241                                         $html_content[$marker] = $content;
 242                                 } else {
 243                                         $html_content[$marker] = "<html>$content</html>";
 244                                 }
 245                         }
 246                 }
 247
 248                 # nowiki
 249                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 250                 foreach( $nowiki_content as $marker => $content ) {
 251                         if( $render ){
 252                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 253                         } else {
 254                                 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
 255                         }
 256                 }
 257
 258                 # math
 259                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 260                 foreach( $math_content as $marker => $content ){
 261                         if( $render ) {
 262                                 if( $this->mOptions->getUseTeX() ) {
 263                                         $math_content[$marker] = renderMath( $content );
 264                                 } else {
 265                                         $math_content[$marker] = "&lt;math&gt;$content&lt;math&gt;";
 266                                 }
 267                         } else {
 268                                 $math_content[$marker] = "<math>$content</math>";
 269                         }
 270                 }
 271
 272                 # pre
 273                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 274                 foreach( $pre_content as $marker => $content ){
 275                         if( $render ){
 276                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 277                         } else {
 278                                 $pre_content[$marker] = "<pre>$content</pre>";
 279                         }
 280                 }
 281
 282                 # Comments
 283                 if($stripcomments) {
 284                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 285                         foreach( $comment_content as $marker => $content ){
 286                                 $comment_content[$marker] = "<!--$content-->";
 287                         }
 288                 }
 289
 290                 # Extensions
 291                 foreach ( $this->mTagHooks as $tag => $callback ) {
 292                         $ext_contents[$tag] = array();
 293                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 294                         foreach( $ext_content[$tag] as $marker => $content ) {
 295                                 if ( $render ) {
 296                                         $ext_content[$tag][$marker] = $callback( $content );
 297                                 } else {
 298                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 299                                 }
 300                         }
 301                 }
 302
 303                 # Merge state with the pre-existing state, if there is one
 304                 if ( $state ) {
 305                         $state['html'] = $state['html'] + $html_content;
 306                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 307                         $state['math'] = $state['math'] + $math_content;
 308                         $state['pre'] = $state['pre'] + $pre_content;
 309                         $state['comment'] = $state['comment'] + $comment_content;
 310
 311                         foreach( $ext_content as $tag => $array ) {
 312                                 if ( array_key_exists( $tag, $state ) ) {
 313                                         $state[$tag] = $state[$tag] + $array;
 314                                 }
 315                         }
 316                 } else {
 317                         $state = array(
 318                           'html' => $html_content,
 319                           'nowiki' => $nowiki_content,
 320                           'math' => $math_content,
 321                           'pre' => $pre_content,
 322                           'comment' => $comment_content,
 323                         ) + $ext_content;
 324                 }
 325                 return $text;
 326         }
 327
 328         # always call unstripNoWiki() after this one
 329         function unstrip( $text, &$state ) {
 330                 # Must expand in reverse order, otherwise nested tags will be corrupted
 331                 $contentDict = end( $state );
 332                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 333                         if( key($state) != 'nowiki' && key($state) != 'html') {
 334                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 335                                         $text = str_replace( key( $contentDict ), $content, $text );
 336                                 }
 337                         }
 338                 }
 339
 340                 return $text;
 341         }
 342         # always call this after unstrip() to preserve the order
 343         function unstripNoWiki( $text, &$state ) {
 344                 # Must expand in reverse order, otherwise nested tags will be corrupted
 345                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 346                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 347                 }
 348
 349                 global $wgRawHtml;
 350                 if ($wgRawHtml) {
 351                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 352                                 $text = str_replace( key( $state['html'] ), $content, $text );
 353                         }
 354                 }
 355
 356                 return $text;
 357         }
 358
 359         # Add an item to the strip state
 360         # Returns the unique tag which must be inserted into the stripped text
 361         # The tag will be replaced with the original text in unstrip()
 362
 363         function insertStripItem( $text, &$state ) {
 364                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 365                 if ( !$state ) {
 366                         $state = array(
 367                           'html' => array(),
 368                           'nowiki' => array(),
 369                           'math' => array(),
 370                           'pre' => array()
 371                         );
 372                 }
 373                 $state['item'][$rnd] = $text;
 374                 return $rnd;
 375         }
 376
 377         # categoryMagic
 378         # generate a list of subcategories and pages for a category
 379         # depending on wfMsg("usenewcategorypage") it either calls the new
 380         # or the old code. The new code will not work properly for some
 381         # languages due to sorting issues, so they might want to turn it
 382         # off.
 383         function categoryMagic() {
 384                 $msg = wfMsg('usenewcategorypage');
 385                 if ( '0' == @$msg[0] )
 386                 {
 387                         return $this->oldCategoryMagic();
 388                 } else {
 389                         return $this->newCategoryMagic();
 390                 }
 391         }
 392
 393         # This method generates the list of subcategories and pages for a category
 394         function oldCategoryMagic () {
 395                 global $wgLang ;
 396                 $fname = 'Parser::oldCategoryMagic';
 397
 398                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 399
 400                 if ( $this->mTitle->getNamespace() != NS_CATEGORY ) return "" ; # This ain't a category page
 401
 402                 $r = "<br style=\"clear:both;\"/>\n";
 403
 404
 405                 $sk =& $this->mOptions->getSkin() ;
 406
 407                 $articles = array() ;
 408                 $children = array() ;
 409                 $data = array () ;
 410                 $id = $this->mTitle->getArticleID() ;
 411
 412                 # FIXME: add limits
 413                 $dbr =& wfGetDB( DB_SLAVE );
 414                 $cur = $dbr->tableName( 'cur' );
 415                 $categorylinks = $dbr->tableName( 'categorylinks' );
 416
 417                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 418                 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM $cur,$categorylinks " .
 419                         "WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 420                 $res = $dbr->query( $sql, $fname ) ;
 421                 while ( $x = $dbr->fetchObject ( $res ) ) $data[] = $x ;
 422
 423                 # For all pages that link to this category
 424                 foreach ( $data AS $x )
 425                 {
 426                         $t = $wgLang->getNsText ( $x->cur_namespace ) ;
 427                         if ( $t != "" ) $t .= ":" ;
 428                         $t .= $x->cur_title ;
 429
 430                         if ( $x->cur_namespace == NS_CATEGORY ) {
 431                                 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
 432                         } else {
 433                                 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
 434                         }
 435                 }
 436                 $dbr->freeResult ( $res ) ;
 437
 438                 # Showing subcategories
 439                 if ( count ( $children ) > 0 ) {
 440                         $r .= '<h2>'.wfMsg('subcategories')."</h2>\n" ;
 441                         $r .= implode ( ', ' , $children ) ;
 442                 }
 443
 444                 # Showing pages in this category
 445                 if ( count ( $articles ) > 0 ) {
 446                         $ti = $this->mTitle->getText() ;
 447                         $h =  wfMsg( 'category_header', $ti );
 448                         $r .= "<h2>{$h}</h2>\n" ;
 449                         $r .= implode ( ', ' , $articles ) ;
 450                 }
 451
 452                 return $r ;
 453         }
 454
 455
 456
 457         function newCategoryMagic () {
 458                 global $wgLang;
 459                 if ( !$this->mOptions->getUseCategoryMagic() ) return ; # Doesn't use categories at all
 460
 461                 if ( $this->mTitle->getNamespace() != NS_CATEGORY ) return '' ; # This ain't a category page
 462
 463                 $r = "<br style=\"clear:both;\"/>\n";
 464
 465
 466                 $sk =& $this->mOptions->getSkin() ;
 467
 468                 $articles = array() ;
 469                 $articles_start_char = array();
 470                 $children = array() ;
 471                 $children_start_char = array();
 472                 $data = array () ;
 473                 $id = $this->mTitle->getArticleID() ;
 474
 475                 # FIXME: add limits
 476                 $dbr =& wfGetDB( DB_SLAVE );
 477                 $cur = $dbr->tableName( 'cur' );
 478                 $categorylinks = $dbr->tableName( 'categorylinks' );
 479
 480                 $t = $dbr->strencode( $this->mTitle->getDBKey() );
 481                 $sql = "SELECT DISTINCT cur_title,cur_namespace,cl_sortkey FROM " .
 482                         "$cur,$categorylinks WHERE cl_to='$t' AND cl_from=cur_id ORDER BY cl_sortkey" ;
 483                 $res = $dbr->query ( $sql ) ;
 484                 while ( $x = $dbr->fetchObject ( $res ) )
 485                 {
 486                         $t = $ns = $wgLang->getNsText ( $x->cur_namespace ) ;
 487                         if ( $t != '' ) $t .= ':' ;
 488                         $t .= $x->cur_title ;
 489
 490                         if ( $x->cur_namespace == NS_CATEGORY ) {
 491                                 $ctitle = str_replace( '_',' ',$x->cur_title );
 492                                 array_push ( $children, $sk->makeKnownLink ( $t, $ctitle ) ) ; # Subcategory
 493
 494                                 // If there's a link from Category:A to Category:B, the sortkey of the resulting
 495                                 // entry in the categorylinks table is Category:A, not A, which it SHOULD be.
 496                                 // Workaround: If sortkey == "Category:".$title, than use $title for sorting,
 497                                 // else use sortkey...
 498                                 if ( ($ns.":".$ctitle) ==  $x->cl_sortkey ) {
 499                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cur_title ) );
 500                                 } else {
 501                                         array_push ( $children_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 502                                 }
 503                         } else {
 504                                 array_push ( $articles , $sk->makeKnownLink ( $t ) ) ; # Page in this category
 505                                 array_push ( $articles_start_char, $wgLang->firstChar( $x->cl_sortkey ) ) ;
 506                         }
 507                 }
 508                 $dbr->freeResult ( $res ) ;
 509
 510                 $ti = $this->mTitle->getText() ;
 511
 512                 # Don't show subcategories section if there are none.
 513                 if ( count ( $children ) > 0 )
 514                 {
 515                         # Showing subcategories
 516                         $r .= '<h2>' . wfMsg( 'subcategories' ) . "</h2>\n"
 517                                 . wfMsg( 'subcategorycount', count( $children ) );
 518                         if ( count ( $children ) > 6 ) {
 519
 520                                 // divide list into three equal chunks
 521                                 $chunk = (int) (count ( $children ) / 3);
 522
 523                                 // get and display header
 524                                 $r .= '<table width="100%"><tr valign="top">';
 525
 526                                 $startChunk = 0;
 527                                 $endChunk = $chunk;
 528
 529                                 // loop through the chunks
 530                                 for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 531                                         $chunkIndex < 3;
 532                                         $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 533                                 {
 534
 535                                         $r .= '<td><ul>';
 536                                         // output all subcategories to category
 537                                         for ($index = $startChunk ;
 538                                                 $index < $endChunk && $index < count($children);
 539                                                 $index++ )
 540                                         {
 541                                                 // check for change of starting letter or begging of chunk
 542                                                 if ( ($children_start_char[$index] != $children_start_char[$index - 1])
 543                                                         || ($index == $startChunk) )
 544                                                 {
 545                                                         $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 546                                                 }
 547
 548                                                 $r .= "<li>{$children[$index]}</li>";
 549                                         }
 550                                         $r .= '</ul></td>';
 551
 552
 553                                 }
 554                                 $r .= '</tr></table>';
 555                         } else {
 556                                 // for short lists of subcategories to category.
 557
 558                                 $r .= "<h3>{$children_start_char[0]}</h3>\n";
 559                                 $r .= '<ul><li>'.$children[0].'</li>';
 560                                 for ($index = 1; $index < count($children); $index++ )
 561                                 {
 562                                         if ($children_start_char[$index] != $children_start_char[$index - 1])
 563                                         {
 564                                                 $r .= "</ul><h3>{$children_start_char[$index]}</h3>\n<ul>";
 565                                         }
 566
 567                                         $r .= "<li>{$children[$index]}</li>";
 568                                 }
 569                                 $r .= '</ul>';
 570                         }
 571                 } # END of if ( count($children) > 0 )
 572
 573                 $r .= '<h2>' . wfMsg( 'category_header', $ti ) . "</h2>\n" .
 574                         wfMsg( 'categoryarticlecount', count( $articles ) );
 575
 576                 # Showing articles in this category
 577                 if ( count ( $articles ) > 6) {
 578                         $ti = $this->mTitle->getText() ;
 579
 580                         // divide list into three equal chunks
 581                         $chunk = (int) (count ( $articles ) / 3);
 582
 583                         // get and display header
 584                         $r .= '<table width="100%"><tr valign="top">';
 585
 586                         // loop through the chunks
 587                         for($startChunk = 0, $endChunk = $chunk, $chunkIndex = 0;
 588                                 $chunkIndex < 3;
 589                                 $chunkIndex++, $startChunk = $endChunk, $endChunk += $chunk + 1)
 590                         {
 591
 592                                 $r .= '<td><ul>';
 593
 594                                 // output all articles in category
 595                                 for ($index = $startChunk ;
 596                                         $index < $endChunk && $index < count($articles);
 597                                         $index++ )
 598                                 {
 599                                         // check for change of starting letter or begging of chunk
 600                                         if ( ($articles_start_char[$index] != $articles_start_char[$index - 1])
 601                                                 || ($index == $startChunk) )
 602                                         {
 603                                                 $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 604                                         }
 605
 606                                         $r .= "<li>{$articles[$index]}</li>";
 607                                 }
 608                                 $r .= '</ul></td>';
 609
 610
 611                         }
 612                         $r .= '</tr></table>';
 613                 } elseif ( count ( $articles )  > 0) {
 614                         // for short lists of articles in categories.
 615                         $ti = $this->mTitle->getText() ;
 616
 617                         $r .= '<h3>'.$articles_start_char[0]."</h3>\n";
 618                         $r .= '<ul><li>'.$articles[0].'</li>';
 619                         for ($index = 1; $index < count($articles); $index++ )
 620                         {
 621                                 if ($articles_start_char[$index] != $articles_start_char[$index - 1])
 622                                 {
 623                                         $r .= "</ul><h3>{$articles_start_char[$index]}</h3>\n<ul>";
 624                                 }
 625
 626                                 $r .= "<li>{$articles[$index]}</li>";
 627                         }
 628                         $r .= '</ul>';
 629                 }
 630
 631
 632                 return $r ;
 633         }
 634
 635         # Return allowed HTML attributes
 636         function getHTMLattrs () {
 637                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 638                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 639                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 640                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 641                                 /* FONT */ 'type', 'start', 'value', 'compact',
 642                                 /* For various lists, mostly deprecated but safe */
 643                                 'summary', 'width', 'border', 'frame', 'rules',
 644                                 'cellspacing', 'cellpadding', 'valign', 'char',
 645                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 646                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 647                                 'id', 'class', 'name', 'style' /* For CSS */
 648                                 );
 649                 return $htmlattrs ;
 650         }
 651
 652         # Remove non approved attributes and javascript in css
 653         function fixTagAttributes ( $t ) {
 654                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 655                 $htmlattrs = $this->getHTMLattrs() ;
 656
 657                 # Strip non-approved attributes from the tag
 658                 $t = preg_replace(
 659                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 660                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 661                         $t);
 662
 663                 $t = str_replace ( "<></>" , "" , $t ) ; # This should fix bug 980557
 664
 665                 # Strip javascript "expression" from stylesheets. Brute force approach:
 666                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 667
 668                 if( preg_match(
 669                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 670                         wfMungeToUtf8( $t ) ) )
 671                 {
 672                         $t='';
 673                 }
 674
 675                 return trim ( $t ) ;
 676         }
 677
 678         # interface with html tidy, used if $wgUseTidy = true
 679         function tidy ( $text ) {
 680                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 681                 global $wgInputEncoding, $wgOutputEncoding;
 682                 $fname = 'Parser::tidy';
 683                 wfProfileIn( $fname );
 684
 685                 $cleansource = '';
 686                 switch(strtoupper($wgOutputEncoding)) {
 687                         case 'ISO-8859-1':
 688                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 689                                 break;
 690                         case 'UTF-8':
 691                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 692                                 break;
 693                         default:
 694                                 $wgTidyOpts .= ' -raw';
 695                         }
 696
 697                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 698 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 699 '<head><title>test</title></head><body>'.$text.'</body></html>';
 700                 $descriptorspec = array(
 701                         0 => array('pipe', 'r'),
 702                         1 => array('pipe', 'w'),
 703                         2 => array('file', '/dev/null', 'a')
 704                 );
 705                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 706                 if (is_resource($process)) {
 707                         fwrite($pipes[0], $wrappedtext);
 708                         fclose($pipes[0]);
 709                         while (!feof($pipes[1])) {
 710                                 $cleansource .= fgets($pipes[1], 1024);
 711                         }
 712                         fclose($pipes[1]);
 713                         $return_value = proc_close($process);
 714                 }
 715
 716                 wfProfileOut( $fname );
 717
 718                 if( $cleansource == '' && $text != '') {
 719                         wfDebug( "Tidy error detected!\n" );
 720                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 721                 } else {
 722                         return $cleansource;
 723                 }
 724         }
 725
 726         # parse the wiki syntax used to render tables
 727         function doTableStuff ( $t ) {
 728                 $fname = 'Parser::doTableStuff';
 729                 wfProfileIn( $fname );
 730
 731                 $t = explode ( "\n" , $t ) ;
 732                 $td = array () ; # Is currently a td tag open?
 733                 $ltd = array () ; # Was it TD or TH?
 734                 $tr = array () ; # Is currently a tr tag open?
 735                 $ltr = array () ; # tr attributes
 736                 $indent_level = 0; # indent level of the table
 737                 foreach ( $t AS $k => $x )
 738                 {
 739                         $x = trim ( $x ) ;
 740                         $fc = substr ( $x , 0 , 1 ) ;
 741                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) )
 742                         {
 743                                 $indent_level = strlen( $matches[1] );
 744                                 $t[$k] = "\n" .
 745                                         str_repeat( "<dl><dd>", $indent_level ) .
 746                                         "<table " . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 747                                 array_push ( $td , false ) ;
 748                                 array_push ( $ltd , '' ) ;
 749                                 array_push ( $tr , false ) ;
 750                                 array_push ( $ltr , '' ) ;
 751                         }
 752                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 753                         else if ( '|}' == substr ( $x , 0 , 2 ) )
 754                         {
 755                                 $z = "</table>\n" ;
 756                                 $l = array_pop ( $ltd ) ;
 757                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 758                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 759                                 array_pop ( $ltr ) ;
 760                                 $t[$k] = $z . str_repeat( "</dd></dl>", $indent_level );
 761                         }
 762                         else if ( '|-' == substr ( $x , 0 , 2 ) ) # Allows for |---------------
 763                         {
 764                                 $x = substr ( $x , 1 ) ;
 765                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 766                                 $z = '' ;
 767                                 $l = array_pop ( $ltd ) ;
 768                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 769                                 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 770                                 array_pop ( $ltr ) ;
 771                                 $t[$k] = $z ;
 772                                 array_push ( $tr , false ) ;
 773                                 array_push ( $td , false ) ;
 774                                 array_push ( $ltd , '' ) ;
 775                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 776                         }
 777                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) # Caption
 778                         {
 779                                 if ( '|+' == substr ( $x , 0 , 2 ) )
 780                                 {
 781                                         $fc = '+' ;
 782                                         $x = substr ( $x , 1 ) ;
 783                                 }
 784                                 $after = substr ( $x , 1 ) ;
 785                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 786                                 $after = explode ( '||' , $after ) ;
 787                                 $t[$k] = '' ;
 788                                 foreach ( $after AS $theline )
 789                                 {
 790                                         $z = '' ;
 791                                         if ( $fc != '+' )
 792                                         {
 793                                                 $tra = array_pop ( $ltr ) ;
 794                                                 if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
 795                                                 array_push ( $tr , true ) ;
 796                                                 array_push ( $ltr , '' ) ;
 797                                         }
 798
 799                                         $l = array_pop ( $ltd ) ;
 800                                         if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
 801                                         if ( $fc == '|' ) $l = 'td' ;
 802                                         else if ( $fc == '!' ) $l = 'th' ;
 803                                         else if ( $fc == '+' ) $l = 'caption' ;
 804                                         else $l = '' ;
 805                                         array_push ( $ltd , $l ) ;
 806                                         $y = explode ( '|' , $theline , 2 ) ;
 807                                         if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 808                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 809                                         $t[$k] .= $y ;
 810                                         array_push ( $td , true ) ;
 811                                 }
 812                         }
 813                 }
 814
 815                 # Closing open td, tr && table
 816                 while ( count ( $td ) > 0 )
 817                 {
 818                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 819                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 820                         $t[] = '</table>' ;
 821                 }
 822
 823                 $t = implode ( "\n" , $t ) ;
 824                 #               $t = $this->removeHTMLtags( $t );
 825                 wfProfileOut( $fname );
 826                 return $t ;
 827         }
 828
 829         # Parses the text and adds the result to the strip state
 830         # Returns the strip tag
 831         function stripParse( $text, $newline, $args )
 832         {
 833                 $text = $this->strip( $text, $this->mStripState );
 834                 $text = $this->internalParse( $text, (bool)$newline, $args, false );
 835                 return $newline.$this->insertStripItem( $text, $this->mStripState );
 836         }
 837
 838         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 839                 $fname = 'Parser::internalParse';
 840                 wfProfileIn( $fname );
 841
 842                 $text = $this->removeHTMLtags( $text );
 843                 $text = $this->replaceVariables( $text, $args );
 844
 845                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 846
 847                 $text = $this->doHeadings( $text );
 848                 if($this->mOptions->getUseDynamicDates()) {
 849                         global $wgDateFormatter;
 850                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 851                 }
 852                 $text = $this->doAllQuotes( $text );
 853                 $text = $this->replaceExternalLinks( $text );
 854                 $text = $this->doMagicLinks( $text );
 855                 $text = $this->replaceInternalLinks ( $text );
 856                 $text = $this->replaceInternalLinks ( $text );
 857
 858                 $text = $this->unstrip( $text, $this->mStripState );
 859                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 860
 861                 $text = $this->doTableStuff( $text );
 862                 $text = $this->formatHeadings( $text, $isMain );
 863                 $sk =& $this->mOptions->getSkin();
 864                 $text = $sk->transformContent( $text );
 865
 866                 if ( $isMain && !isset ( $this->categoryMagicDone ) ) {
 867                         $text .= $this->categoryMagic () ;
 868                         $this->categoryMagicDone = true ;
 869                 }
 870
 871                 wfProfileOut( $fname );
 872                 return $text;
 873         }
 874
 875         /* private */ function &doMagicLinks( &$text ) {
 876                 $text = $this->magicISBN( $text );
 877                 $text = $this->magicGEO( $text );
 878                 $text = $this->magicRFC( $text );
 879                 return $text;
 880         }
 881
 882         # Parse ^^ tokens and return html
 883         /* private */ function doExponent ( $text )
 884         {
 885                 $fname = 'Parser::doExponent';
 886                 wfProfileIn( $fname);
 887                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 888                 wfProfileOut( $fname);
 889                 return $text;
 890         }
 891
 892     # Parse headers and return html
 893         /* private */ function doHeadings( $text ) {
 894                 $fname = 'Parser::doHeadings';
 895                 wfProfileIn( $fname );
 896                 for ( $i = 6; $i >= 1; --$i ) {
 897                         $h = substr( '======', 0, $i );
 898                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 899                           "<h{$i}>\\1</h{$i}>\\2", $text );
 900                 }
 901                 wfProfileOut( $fname );
 902                 return $text;
 903         }
 904
 905         /* private */ function doAllQuotes( $text ) {
 906                 $fname = 'Parser::doAllQuotes';
 907                 wfProfileIn( $fname );
 908                 $outtext = '';
 909                 $lines = explode( "\n", $text );
 910                 foreach ( $lines as $line ) {
 911                         $outtext .= $this->doQuotes ( $line ) . "\n";
 912                 }
 913                 $outtext = substr($outtext, 0,-1);
 914                 wfProfileOut( $fname );
 915                 return $outtext;
 916         }
 917
 918         /* private */ function doQuotes( $text ) {
 919                 $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
 920                 if (count ($arr) == 1)
 921                         return $text;
 922                 else
 923                 {
 924                         # First, do some preliminary work. This may shift some apostrophes from
 925                         # being mark-up to being text. It also counts the number of occurrences
 926                         # of bold and italics mark-ups.
 927                         $i = 0;
 928                         $numbold = 0;
 929                         $numitalics = 0;
 930                         foreach ($arr as $r)
 931                         {
 932                                 if (($i % 2) == 1)
 933                                 {
 934                                         # If there are ever four apostrophes, assume the first is supposed to
 935                                         # be text, and the remaining three constitute mark-up for bold text.
 936                                         if (strlen ($arr[$i]) == 4)
 937                                         {
 938                                                 $arr[$i-1] .= "'";
 939                                                 $arr[$i] = "'''";
 940                                         }
 941                                         # If there are more than 5 apostrophes in a row, assume they're all
 942                                         # text except for the last 5.
 943                                         else if (strlen ($arr[$i]) > 5)
 944                                         {
 945                                                 $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
 946                                                 $arr[$i] = "'''''";
 947                                         }
 948                                         # Count the number of occurrences of bold and italics mark-ups.
 949                                         # We are not counting sequences of five apostrophes.
 950                                         if (strlen ($arr[$i]) == 2) $numitalics++;  else
 951                                         if (strlen ($arr[$i]) == 3) $numbold++;
 952                                 }
 953                                 $i++;
 954                         }
 955
 956                         # If there is an odd number of both bold and italics, it is likely
 957                         # that one of the bold ones was meant to be an apostrophe followed
 958                         # by italics. Which one we cannot know for certain, but it is more
 959                         # likely to be one that has a single-letter word before it.
 960                         if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
 961                         {
 962                                 $i = 0;
 963                                 $firstsingleletterword = -1;
 964                                 $firstmultiletterword = -1;
 965                                 $firstspace = -1;
 966                                 foreach ($arr as $r)
 967                                 {
 968                                         if (($i % 2 == 1) and (strlen ($r) == 3))
 969                                         {
 970                                                 $x1 = substr ($arr[$i-1], -1);
 971                                                 $x2 = substr ($arr[$i-1], -2, 1);
 972                                                 if ($x1 == " ") {
 973                                                         if ($firstspace == -1) $firstspace = $i;
 974                                                 } else if ($x2 == " ") {
 975                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 976                                                 } else {
 977                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 978                                                 }
 979                                         }
 980                                         $i++;
 981                                 }
 982
 983                                 # If there is a single-letter word, use it!
 984                                 if ($firstsingleletterword > -1)
 985                                 {
 986                                         $arr [ $firstsingleletterword ] = "''";
 987                                         $arr [ $firstsingleletterword-1 ] .= "'";
 988                                 }
 989                                 # If not, but there's a multi-letter word, use that one.
 990                                 else if ($firstmultiletterword > -1)
 991                                 {
 992                                         $arr [ $firstmultiletterword ] = "''";
 993                                         $arr [ $firstmultiletterword-1 ] .= "'";
 994                                 }
 995                                 # ... otherwise use the first one that has neither.
 996                                 else
 997                                 {
 998                                         $arr [ $firstspace ] = "''";
 999                                         $arr [ $firstspace-1 ] .= "'";
1000                                 }
1001                         }
1002
1003                         # Now let's actually convert our apostrophic mush to HTML!
1004                         $output = '';
1005                         $buffer = '';
1006                         $state = '';
1007                         $i = 0;
1008                         foreach ($arr as $r)
1009                         {
1010                                 if (($i % 2) == 0)
1011                                 {
1012                                         if ($state == 'both')
1013                                                 $buffer .= $r;
1014                                         else
1015                                                 $output .= $r;
1016                                 }
1017                                 else
1018                                 {
1019                                         if (strlen ($r) == 2)
1020                                         {
1021                                                 if ($state == 'em')
1022                                                 { $output .= "</em>"; $state = ''; }
1023                                                 else if ($state == 'strongem')
1024                                                 { $output .= "</em>"; $state = 'strong'; }
1025                                                 else if ($state == 'emstrong')
1026                                                 { $output .= "</strong></em><strong>"; $state = 'strong'; }
1027                                                 else if ($state == 'both')
1028                                                 { $output .= "<strong><em>{$buffer}</em>"; $state = 'strong'; }
1029                                                 else # $state can be 'strong' or ''
1030                                                 { $output .= "<em>"; $state .= 'em'; }
1031                                         }
1032                                         else if (strlen ($r) == 3)
1033                                         {
1034                                                 if ($state == 'strong')
1035                                                 { $output .= "</strong>"; $state = ''; }
1036                                                 else if ($state == 'strongem')
1037                                                 { $output .= "</em></strong><em>"; $state = 'em'; }
1038                                                 else if ($state == 'emstrong')
1039                                                 { $output .= "</strong>"; $state = 'em'; }
1040                                                 else if ($state == 'both')
1041                                                 { $output .= "<em><strong>{$buffer}</strong>"; $state = 'em'; }
1042                                                 else # $state can be 'em' or ''
1043                                                 { $output .= "<strong>"; $state .= 'strong'; }
1044                                         }
1045                                         else if (strlen ($r) == 5)
1046                                         {
1047                                                 if ($state == 'strong')
1048                                                 { $output .= "</strong><em>"; $state = 'em'; }
1049                                                 else if ($state == 'em')
1050                                                 { $output .= "</em><strong>"; $state = 'strong'; }
1051                                                 else if ($state == 'strongem')
1052                                                 { $output .= "</em></strong>"; $state = ''; }
1053                                                 else if ($state == 'emstrong')
1054                                                 { $output .= "</strong></em>"; $state = ''; }
1055                                                 else if ($state == 'both')
1056                                                 { $output .= "<em><strong>{$buffer}</strong></em>"; $state = ''; }
1057                                                 else # ($state == '')
1058                                                 { $buffer = ''; $state = 'both'; }
1059                                         }
1060                                 }
1061                                 $i++;
1062                         }
1063                         # Now close all remaining tags.  Notice that the order is important.
1064                         if ($state == 'strong' || $state == 'emstrong')
1065                                 $output .= "</strong>";
1066                         if ($state == 'em' || $state == 'strongem' || $state == 'emstrong')
1067                                 $output .= "</em>";
1068                         if ($state == 'strongem')
1069                                 $output .= "</strong>";
1070                         if ($state == 'both')
1071                                 $output .= "<strong><em>{$buffer}</em></strong>";
1072                         return $output;
1073                 }
1074         }
1075
1076         # Note: we have to do external links before the internal ones,
1077         # and otherwise take great care in the order of things here, so
1078         # that we don't end up interpreting some URLs twice.
1079
1080         /* private */ function replaceExternalLinks( $text ) {
1081                 $fname = 'Parser::replaceExternalLinks';
1082                 wfProfileIn( $fname );
1083
1084                 $sk =& $this->mOptions->getSkin();
1085                 $linktrail = wfMsg('linktrail');
1086                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1087
1088                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
1089
1090                 $i = 0;
1091                 while ( $i<count( $bits ) ) {
1092                         $url = $bits[$i++];
1093                         $protocol = $bits[$i++];
1094                         $text = $bits[$i++];
1095                         $trail = $bits[$i++];
1096
1097                         # If the link text is an image URL, replace it with an <img> tag
1098                         # This happened by accident in the original parser, but some people used it extensively
1099                         $img = $this->maybeMakeImageLink( $text );
1100                         if ( $img !== false ) {
1101                                 $text = $img;
1102                         }
1103
1104                         $dtrail = '';
1105
1106                         # No link text, e.g. [http://domain.tld/some.link]
1107                         if ( $text == '' ) {
1108                                 # Autonumber if allowed
1109                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
1110                                         $text = "[" . ++$this->mAutonumber . "]";
1111                                 } else {
1112                                         # Otherwise just use the URL
1113                                         $text = wfEscapeHTML( $url );
1114                                 }
1115                         } else {
1116                                 # Have link text, e.g. [http://domain.tld/some.link text]s
1117                                 # Check for trail
1118                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
1119                                         $dtrail = $m2[1];
1120                                         $trail = $m2[2];
1121                                 }
1122                         }
1123
1124                         $encUrl = htmlspecialchars( $url );
1125                         # Bit in parentheses showing the URL for the printable version
1126                         if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $url ) ) {
1127                                 $paren = '';
1128                         } else {
1129                                 # Expand the URL for printable version
1130                                 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
1131                         }
1132
1133                         # Process the trail (i.e. everything after this link up until start of the next link),
1134                         # replacing any non-bracketed links
1135                         $trail = $this->replaceFreeExternalLinks( $trail );
1136
1137                         $la = $sk->getExternalLinkAttributes( $url, $text );
1138
1139                         # Use the encoded URL
1140                         # This means that users can paste URLs directly into the text
1141                         # Funny characters like &ouml; aren't valid in URLs anyway
1142                         # This was changed in August 2004
1143                         $s .= "<a href=\"{$url}\" {$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
1144                 }
1145
1146                 wfProfileOut( $fname );
1147                 return $s;
1148         }
1149
1150         # Replace anything that looks like a URL with a link
1151         function replaceFreeExternalLinks( $text ) {
1152                 $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
1153                 $s = array_shift( $bits );
1154                 $i = 0;
1155
1156                 $sk =& $this->mOptions->getSkin();
1157
1158                 while ( $i < count( $bits ) ){
1159                         $protocol = $bits[$i++];
1160                         $remainder = $bits[$i++];
1161
1162                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1163                                 # Found some characters after the protocol that look promising
1164                                 $url = $protocol . $m[1];
1165                                 $trail = $m[2];
1166
1167                                 # Move trailing punctuation to $trail
1168                                 $sep = ',;\.:!?';
1169                                 # If there is no left bracket, then consider right brackets fair game too
1170                                 if ( strpos( $url, '(' ) === false ) {
1171                                         $sep .= ')';
1172                                 }
1173
1174                                 $numSepChars = strspn( strrev( $url ), $sep );
1175                                 if ( $numSepChars ) {
1176                                         $trail = substr( $url, -$numSepChars ) . $trail;
1177                                         $url = substr( $url, 0, -$numSepChars );
1178                                 }
1179
1180                                 # Is this an external image?
1181                                 $text = $this->maybeMakeImageLink( $url );
1182                                 if ( $text === false ) {
1183                                         # Not an image, make a link
1184                                         $text = $sk->makeExternalLink( $url, $url );
1185                                 }
1186                                 $s .= $text . $trail;
1187                         } else {
1188                                 $s .= $protocol . $remainder;
1189                         }
1190                 }
1191                 return $s;
1192         }
1193
1194         function maybeMakeImageLink( $url ) {
1195                 $sk =& $this->mOptions->getSkin();
1196                 $text = false;
1197                 if ( $this->mOptions->getAllowExternalImages() ) {
1198                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1199                                 # Image found
1200                                 $text = $sk->makeImage( $url );
1201                         }
1202                 }
1203                 return $text;
1204         }
1205
1206         /* private */ function replaceInternalLinks( $s ) {
1207                 global $wgLang, $wgLinkCache;
1208                 global $wgNamespacesWithSubpages, $wgLanguageCode;
1209                 static $fname = 'Parser::replaceInternalLinks' ;
1210                 wfProfileIn( $fname );
1211
1212                 wfProfileIn( $fname.'-setup' );
1213                 static $tc = FALSE;
1214                 # the % is needed to support urlencoded titles as well
1215                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1216                 $sk =& $this->mOptions->getSkin();
1217
1218                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1219
1220                 $a = explode( '[[', ' ' . $s );
1221                 $s = array_shift( $a );
1222                 $s = substr( $s, 1 );
1223
1224                 # Match a link having the form [[namespace:link|alternate]]trail
1225                 static $e1 = FALSE;
1226                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1227                 # Match the end of a line for a word that's not followed by whitespace,
1228                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1229                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1230
1231                 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
1232                 # Special and Media are pseudo-namespaces; no pages actually exist in them
1233
1234                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1235
1236                 if ( $useLinkPrefixExtension ) {
1237                         if ( preg_match( $e2, $s, $m ) ) {
1238                                 $first_prefix = $m[2];
1239                                 $s = $m[1];
1240                         } else {
1241                                 $first_prefix = false;
1242                         }
1243                 } else {
1244                         $prefix = '';
1245                 }
1246
1247                 wfProfileOut( $fname.'-setup' );
1248
1249                 foreach ( $a as $line ) {
1250                         wfProfileIn( $fname.'-prefixhandling' );
1251                         if ( $useLinkPrefixExtension ) {
1252                                 if ( preg_match( $e2, $s, $m ) ) {
1253                                         $prefix = $m[2];
1254                                         $s = $m[1];
1255                                 } else {
1256                                         $prefix='';
1257                                 }
1258                                 # first link
1259                                 if($first_prefix) {
1260                                         $prefix = $first_prefix;
1261                                         $first_prefix = false;
1262                                 }
1263                         }
1264                         wfProfileOut( $fname.'-prefixhandling' );
1265
1266                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1267                                 $text = $m[2];
1268                                 # fix up urlencoded title texts
1269                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1270                                 $trail = $m[3];
1271                         } else { # Invalid form; output directly
1272                                 $s .= $prefix . '[[' . $line ;
1273                                 continue;
1274                         }
1275
1276                         /* Valid link forms:
1277                         Foobar -- normal
1278                         :Foobar -- override special treatment of prefix (images, language links)
1279                         /Foobar -- convert to CurrentPage/Foobar
1280                         /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1281                         */
1282                         $c = substr($m[1],0,1);
1283                         $noforce = ($c != ':');
1284                         if( $c == '/' ) { # subpage
1285                                 if(substr($m[1],-1,1)=='/') {                 # / at end means we don't want the slash to be shown
1286                                         $m[1]=substr($m[1],1,strlen($m[1])-2);
1287                                         $noslash=$m[1];
1288                                 } else {
1289                                         $noslash=substr($m[1],1);
1290                                 }
1291                                 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) { # subpages allowed here
1292                                         $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1293                                         if( '' == $text ) {
1294                                                 $text= $m[1];
1295                                         } # this might be changed for ugliness reasons
1296                                 } else {
1297                                         $link = $noslash; # no subpage allowed, use standard link
1298                                 }
1299                         } elseif( $noforce ) { # no subpage
1300                                 $link = $m[1];
1301                         } else {
1302                                 $link = substr( $m[1], 1 );
1303                         }
1304                         $wasblank = ( '' == $text );
1305                         if( $wasblank )
1306                         $text = $link;
1307
1308                         $nt = Title::newFromText( $link );
1309                         if( !$nt ) {
1310                                 $s .= $prefix . '[[' . $line;
1311                                 continue;
1312                         }
1313                         $ns = $nt->getNamespace();
1314                         $iw = $nt->getInterWiki();
1315                         if( $noforce ) {
1316                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1317                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1318                                         $tmp = $prefix . $trail ;
1319                                         $s .= (trim($tmp) == '')? '': $tmp;
1320                                         continue;
1321                                 }
1322                                 if ( $ns == NS_IMAGE ) {
1323                                         $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1324                                         $wgLinkCache->addImageLinkObj( $nt );
1325                                         continue;
1326                                 }
1327                                 if ( $ns == NS_CATEGORY ) {
1328                                         $t = $nt->getText() ;
1329                                         $nnt = Title::newFromText ( Namespace::getCanonicalName(NS_CATEGORY).":".$t ) ;
1330
1331                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1332                                         $pPLC=$sk->postParseLinkColour();
1333                                         $sk->postParseLinkColour( false );
1334                                         $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1335                                         $sk->postParseLinkColour( $pPLC );
1336                                         $wgLinkCache->resume();
1337
1338                                         $sortkey = $wasblank ? $this->mTitle->getPrefixedText() : $text;
1339                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1340                                         $this->mOutput->mCategoryLinks[] = $t ;
1341                                         $s .= $prefix . $trail ;
1342                                         continue;
1343                                 }
1344                         }
1345                         if( ( $nt->getPrefixedText() == $this->mTitle->getPrefixedText() ) &&
1346                         ( strpos( $link, '#' ) == FALSE ) ) {
1347                                 # Self-links are handled specially; generally de-link and change to bold.
1348                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1349                                 continue;
1350                         }
1351
1352                         if( $ns == NS_MEDIA ) {
1353                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1354                                 $wgLinkCache->addImageLinkObj( $nt );
1355                                 continue;
1356                         } elseif( $ns == NS_SPECIAL ) {
1357                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1358                                 continue;
1359                         }
1360                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1361                 }
1362                 wfProfileOut( $fname );
1363                 return $s;
1364         }
1365
1366         # Some functions here used by doBlockLevels()
1367         #
1368         /* private */ function closeParagraph() {
1369                 $result = '';
1370                 if ( '' != $this->mLastSection ) {
1371                         $result = '</' . $this->mLastSection  . ">\n";
1372                 }
1373                 $this->mInPre = false;
1374                 $this->mLastSection = '';
1375                 return $result;
1376         }
1377         # getCommon() returns the length of the longest common substring
1378         # of both arguments, starting at the beginning of both.
1379         #
1380         /* private */ function getCommon( $st1, $st2 ) {
1381                 $fl = strlen( $st1 );
1382                 $shorter = strlen( $st2 );
1383                 if ( $fl < $shorter ) { $shorter = $fl; }
1384
1385                 for ( $i = 0; $i < $shorter; ++$i ) {
1386                         if ( $st1{$i} != $st2{$i} ) { break; }
1387                 }
1388                 return $i;
1389         }
1390         # These next three functions open, continue, and close the list
1391         # element appropriate to the prefix character passed into them.
1392         #
1393         /* private */ function openList( $char )
1394     {
1395                 $result = $this->closeParagraph();
1396
1397                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1398                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1399                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1400                 else if ( ';' == $char ) {
1401                         $result .= '<dl><dt>';
1402                         $this->mDTopen = true;
1403                 }
1404                 else { $result = '<!-- ERR 1 -->'; }
1405
1406                 return $result;
1407         }
1408
1409         /* private */ function nextItem( $char ) {
1410                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1411                 else if ( ':' == $char || ';' == $char ) {
1412                         $close = "</dd>";
1413                         if ( $this->mDTopen ) { $close = '</dt>'; }
1414                         if ( ';' == $char ) {
1415                                 $this->mDTopen = true;
1416                                 return $close . '<dt>';
1417                         } else {
1418                                 $this->mDTopen = false;
1419                                 return $close . '<dd>';
1420                         }
1421                 }
1422                 return '<!-- ERR 2 -->';
1423         }
1424
1425         /* private */function closeList( $char ) {
1426                 if ( '*' == $char ) { $text = '</li></ul>'; }
1427                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1428                 else if ( ':' == $char ) {
1429                         if ( $this->mDTopen ) {
1430                                 $this->mDTopen = false;
1431                                 $text = '</dt></dl>';
1432                         } else {
1433                                 $text = '</dd></dl>';
1434                         }
1435                 }
1436                 else {  return '<!-- ERR 3 -->'; }
1437                 return $text."\n";
1438         }
1439
1440         /* private */ function doBlockLevels( $text, $linestart ) {
1441                 $fname = 'Parser::doBlockLevels';
1442                 wfProfileIn( $fname );
1443
1444                 # Parsing through the text line by line.  The main thing
1445                 # happening here is handling of block-level elements p, pre,
1446                 # and making lists from lines starting with * # : etc.
1447                 #
1448                 $textLines = explode( "\n", $text );
1449
1450                 $lastPrefix = $output = $lastLine = '';
1451                 $this->mDTopen = $inBlockElem = false;
1452                 $prefixLength = 0;
1453                 $paragraphStack = false;
1454
1455                 if ( !$linestart ) {
1456                         $output .= array_shift( $textLines );
1457                 }
1458                 foreach ( $textLines as $oLine ) {
1459                         $lastPrefixLength = strlen( $lastPrefix );
1460                         $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1461                         $preOpenMatch = preg_match("/<pre/i", $oLine );
1462                         if ( !$this->mInPre ) {
1463                                 # Multiple prefixes may abut each other for nested lists.
1464                                 $prefixLength = strspn( $oLine, '*#:;' );
1465                                 $pref = substr( $oLine, 0, $prefixLength );
1466
1467                                 # eh?
1468                                 $pref2 = str_replace( ';', ':', $pref );
1469                                 $t = substr( $oLine, $prefixLength );
1470                                 $this->mInPre = !empty($preOpenMatch);
1471                         } else {
1472                                 # Don't interpret any other prefixes in preformatted text
1473                                 $prefixLength = 0;
1474                                 $pref = $pref2 = '';
1475                                 $t = $oLine;
1476                         }
1477
1478                         # List generation
1479                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1480                                 # Same as the last item, so no need to deal with nesting or opening stuff
1481                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1482                                 $paragraphStack = false;
1483
1484                                 if ( ";" == substr( $pref, -1 ) ) {
1485                                         # The one nasty exception: definition lists work like this:
1486                                         # ; title : definition text
1487                                         # So we check for : in the remainder text to split up the
1488                                         # title and definition, without b0rking links.
1489                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1490                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1491                                                 $term = $match[1];
1492                                                 $output .= $term . $this->nextItem( ':' );
1493                                                 $t = $match[2];
1494                                         }
1495                                 }
1496                         } elseif( $prefixLength || $lastPrefixLength ) {
1497                                 # Either open or close a level...
1498                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1499                                 $paragraphStack = false;
1500
1501                                 while( $commonPrefixLength < $lastPrefixLength ) {
1502                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1503                                         --$lastPrefixLength;
1504                                 }
1505                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1506                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1507                                 }
1508                                 while ( $prefixLength > $commonPrefixLength ) {
1509                                         $char = substr( $pref, $commonPrefixLength, 1 );
1510                                         $output .= $this->openList( $char );
1511
1512                                         if ( ';' == $char ) {
1513                                                 # FIXME: This is dupe of code above
1514                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1515                                                         $term = $match[1];
1516                                                         $output .= $term . $this->nextItem( ":" );
1517                                                         $t = $match[2];
1518                                                 }
1519                                         }
1520                                         ++$commonPrefixLength;
1521                                 }
1522                                 $lastPrefix = $pref2;
1523                         }
1524                         if( 0 == $prefixLength ) {
1525                                 # No prefix (not in list)--go to paragraph mode
1526                                 $uniq_prefix = UNIQ_PREFIX;
1527                                 // XXX: use a stack for nestable elements like span, table and div
1528                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1529                                 $closematch = preg_match(
1530                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1531                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1532                                 if ( $openmatch or $closematch ) {
1533                                         $paragraphStack = false;
1534                                         $output .= $this->closeParagraph();
1535                                         if($preOpenMatch and !$preCloseMatch) {
1536                                                 $this->mInPre = true;
1537                                         }
1538                                         if ( $closematch  ) {
1539                                                 $inBlockElem = false;
1540                                         } else {
1541                                                 $inBlockElem = true;
1542                                         }
1543                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1544                                         if ( " " == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1545                                                 // pre
1546                                                 if ($this->mLastSection != 'pre') {
1547                                                         $paragraphStack = false;
1548                                                         $output .= $this->closeParagraph().'<pre>';
1549                                                         $this->mLastSection = 'pre';
1550                                                 }
1551                                         } else {
1552                                                 // paragraph
1553                                                 if ( '' == trim($t) ) {
1554                                                         if ( $paragraphStack ) {
1555                                                                 $output .= $paragraphStack.'<br />';
1556                                                                 $paragraphStack = false;
1557                                                                 $this->mLastSection = 'p';
1558                                                         } else {
1559                                                                 if ($this->mLastSection != 'p' ) {
1560                                                                         $output .= $this->closeParagraph();
1561                                                                         $this->mLastSection = '';
1562                                                                         $paragraphStack = '<p>';
1563                                                                 } else {
1564                                                                         $paragraphStack = '</p><p>';
1565                                                                 }
1566                                                         }
1567                                                 } else {
1568                                                         if ( $paragraphStack ) {
1569                                                                 $output .= $paragraphStack;
1570                                                                 $paragraphStack = false;
1571                                                                 $this->mLastSection = 'p';
1572                                                         } else if ($this->mLastSection != 'p') {
1573                                                                 $output .= $this->closeParagraph().'<p>';
1574                                                                 $this->mLastSection = 'p';
1575                                                         }
1576                                                 }
1577                                         }
1578                                 }
1579                         }
1580                         if ($paragraphStack === false) {
1581                                 $output .= $t."\n";
1582                         }
1583                 }
1584                 while ( $prefixLength ) {
1585                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1586                         --$prefixLength;
1587                 }
1588                 if ( '' != $this->mLastSection ) {
1589                         $output .= '</' . $this->mLastSection . '>';
1590                         $this->mLastSection = '';
1591                 }
1592
1593                 wfProfileOut( $fname );
1594                 return $output;
1595         }
1596
1597         # Return value of a magic variable (like PAGENAME)
1598         function getVariableValue( $index ) {
1599                 global $wgLang, $wgSitename, $wgServer;
1600
1601                 switch ( $index ) {
1602                         case MAG_CURRENTMONTH:
1603                                 return $wgLang->formatNum( date( 'm' ) );
1604                         case MAG_CURRENTMONTHNAME:
1605                                 return $wgLang->getMonthName( date('n') );
1606                         case MAG_CURRENTMONTHNAMEGEN:
1607                                 return $wgLang->getMonthNameGen( date('n') );
1608                         case MAG_CURRENTDAY:
1609                                 return $wgLang->formatNum( date('j') );
1610                         case MAG_PAGENAME:
1611                                 return $this->mTitle->getText();
1612                         case MAG_NAMESPACE:
1613                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1614                                 return $wgLang->getNsText($this->mTitle->getNamespace()); // Patch  by Dori
1615                         case MAG_CURRENTDAYNAME:
1616                                 return $wgLang->getWeekdayName( date('w')+1 );
1617                         case MAG_CURRENTYEAR:
1618                                 return $wgLang->formatNum( date( 'Y' ) );
1619                         case MAG_CURRENTTIME:
1620                                 return $wgLang->time( wfTimestampNow(), false );
1621                         case MAG_NUMBEROFARTICLES:
1622                                 return $wgLang->formatNum( wfNumberOfArticles() );
1623                         case MAG_SITENAME:
1624                                 return $wgSitename;
1625                         case MAG_SERVER:
1626                                 return $wgServer;
1627                         default:
1628                                 return NULL;
1629                 }
1630         }
1631
1632         # initialise the magic variables (like CURRENTMONTHNAME)
1633         function initialiseVariables() {
1634                 global $wgVariableIDs;
1635                 $this->mVariables = array();
1636                 foreach ( $wgVariableIDs as $id ) {
1637                         $mw =& MagicWord::get( $id );
1638                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1639                 }
1640         }
1641
1642         /* private */ function replaceVariables( $text, $args = array() ) {
1643                 global $wgLang, $wgScript, $wgArticlePath;
1644
1645                 # Prevent too big inclusions
1646                 if(strlen($text)> MAX_INCLUDE_SIZE)
1647                    return $text;
1648
1649                 $fname = 'Parser::replaceVariables';
1650                 wfProfileIn( $fname );
1651
1652                 $bail = false;
1653                 $titleChars = Title::legalChars();
1654                 $nonBraceChars = str_replace( array( '{', '}' ), array( '', '' ), $titleChars );
1655
1656                 # This function is called recursively. To keep track of arguments we need a stack:
1657                 array_push( $this->mArgStack, $args );
1658
1659                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1660                 $GLOBALS['wgCurParser'] =& $this;
1661
1662
1663                 if ( $this->mOutputType == OT_HTML ) {
1664                         # Variable substitution
1665                         $text = preg_replace_callback( "/{{([$nonBraceChars]*?)}}/", 'wfVariableSubstitution', $text );
1666
1667                         # Argument substitution
1668                         $text = preg_replace_callback( "/(\\n?){{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1669                 }
1670                 # Template substitution
1671                 $regex = '/(\\n?){{(['.$nonBraceChars.']*)(\\|.*?|)}}/s';
1672                 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1673
1674                 array_pop( $this->mArgStack );
1675
1676                 wfProfileOut( $fname );
1677                 return $text;
1678         }
1679
1680         function variableSubstitution( $matches ) {
1681                 if ( !$this->mVariables ) {
1682                         $this->initialiseVariables();
1683                 }
1684                 if ( array_key_exists( $matches[1], $this->mVariables ) ) {
1685                         $text = $this->mVariables[$matches[1]];
1686                         $this->mOutput->mContainsOldMagic = true;
1687                 } else {
1688                         $text = $matches[0];
1689                 }
1690                 return $text;
1691         }
1692
1693         # Split template arguments
1694         function getTemplateArgs( $argsString ) {
1695                 if ( $argsString === '' ) {
1696                         return array();
1697                 }
1698
1699                 $args = explode( '|', substr( $argsString, 1 ) );
1700
1701                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1702                 # merged with the next arg because the '|' character between belongs
1703                 # to the link syntax and not the template parameter syntax.
1704                 $argc = count($args);
1705                 $i = 0;
1706                 for ( $i = 0; $i < $argc-1; $i++ ) {
1707                         if ( substr_count ( $args[$i], "[[" ) != substr_count ( $args[$i], "]]" ) ) {
1708                                 $args[$i] .= "|".$args[$i+1];
1709                                 array_splice($args, $i+1, 1);
1710                                 $i--;
1711                                 $argc--;
1712                         }
1713                 }
1714
1715                 return $args;
1716         }
1717
1718         function braceSubstitution( $matches ) {
1719                 global $wgLinkCache, $wgLang;
1720                 $fname = 'Parser::braceSubstitution';
1721                 $found = false;
1722                 $nowiki = false;
1723                 $noparse = false;
1724
1725                 $title = NULL;
1726
1727                 # $newline is an optional newline character before the braces
1728                 # $part1 is the bit before the first |, and must contain only title characters
1729                 # $args is a list of arguments, starting from index 0, not including $part1
1730
1731                 $newline = $matches[1];
1732                 $part1 = $matches[2];
1733                 # If the third subpattern matched anything, it will start with |
1734
1735                 $args = $this->getTemplateArgs($matches[3]);
1736                 $argc = count( $args );
1737
1738                 # {{{}}}
1739                 if ( strpos( $matches[0], '{{{' ) !== false ) {
1740                         $text = $matches[0];
1741                         $found = true;
1742                         $noparse = true;
1743                 }
1744
1745                 # SUBST
1746                 if ( !$found ) {
1747                         $mwSubst =& MagicWord::get( MAG_SUBST );
1748                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1749                                 if ( $this->mOutputType != OT_WIKI ) {
1750                                         # Invalid SUBST not replaced at PST time
1751                                         # Return without further processing
1752                                         $text = $matches[0];
1753                                         $found = true;
1754                                         $noparse= true;
1755                                 }
1756                         } elseif ( $this->mOutputType == OT_WIKI ) {
1757                                 # SUBST not found in PST pass, do nothing
1758                                 $text = $matches[0];
1759                                 $found = true;
1760                         }
1761                 }
1762
1763                 # MSG, MSGNW and INT
1764                 if ( !$found ) {
1765                         # Check for MSGNW:
1766                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1767                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1768                                 $nowiki = true;
1769                         } else {
1770                                 # Remove obsolete MSG:
1771                                 $mwMsg =& MagicWord::get( MAG_MSG );
1772                                 $mwMsg->matchStartAndRemove( $part1 );
1773                         }
1774
1775                         # Check if it is an internal message
1776                         $mwInt =& MagicWord::get( MAG_INT );
1777                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1778                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1779                                         $text = wfMsgReal( $part1, $args, true );
1780                                         $found = true;
1781                                 }
1782                         }
1783                 }
1784
1785                 # NS
1786                 if ( !$found ) {
1787                         # Check for NS: (namespace expansion)
1788                         $mwNs = MagicWord::get( MAG_NS );
1789                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1790                                 if ( intval( $part1 ) ) {
1791                                         $text = $wgLang->getNsText( intval( $part1 ) );
1792                                         $found = true;
1793                                 } else {
1794                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1795                                         if ( !is_null( $index ) ) {
1796                                                 $text = $wgLang->getNsText( $index );
1797                                                 $found = true;
1798                                         }
1799                                 }
1800                         }
1801                 }
1802
1803                 # LOCALURL and LOCALURLE
1804                 if ( !$found ) {
1805                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1806                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1807
1808                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1809                                 $func = 'getLocalURL';
1810                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1811                                 $func = 'escapeLocalURL';
1812                         } else {
1813                                 $func = '';
1814                         }
1815
1816                         if ( $func !== '' ) {
1817                                 $title = Title::newFromText( $part1 );
1818                                 if ( !is_null( $title ) ) {
1819                                         if ( $argc > 0 ) {
1820                                                 $text = $title->$func( $args[0] );
1821                                         } else {
1822                                                 $text = $title->$func();
1823                                         }
1824                                         $found = true;
1825                                 }
1826                         }
1827                 }
1828
1829                 # Internal variables
1830                 if ( !$this->mVariables ) {
1831                         $this->initialiseVariables();
1832                 }
1833                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1834                         $text = $this->mVariables[$part1];
1835                         $found = true;
1836                         $this->mOutput->mContainsOldMagic = true;
1837                 }
1838
1839                 # Template table test
1840
1841                 # Did we encounter this template already? If yes, it is in the cache
1842                 # and we need to check for loops.
1843                 if ( isset( $this->mTemplates[$part1] ) ) {
1844                         # Infinite loop test
1845                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1846                                 $noparse = true;
1847                                 $found = true;
1848                         }
1849                         # set $text to cached message.
1850                         $text = $this->mTemplates[$part1];
1851                         $found = true;
1852                 }
1853
1854                 # Load from database
1855                 if ( !$found ) {
1856                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1857                         if ( !is_null( $title ) && !$title->isExternal() ) {
1858                                 # Check for excessive inclusion
1859                                 $dbk = $title->getPrefixedDBkey();
1860                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1861                                         # This should never be reached.
1862                                         $article = new Article( $title );
1863                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1864                                         if ( $articleContent !== false ) {
1865                                                 $found = true;
1866                                                 $text = $articleContent;
1867
1868                                         }
1869                                 }
1870
1871                                 # If the title is valid but undisplayable, make a link to it
1872                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1873                                         $text = '[[' . $title->getPrefixedText() . ']]';
1874                                         $found = true;
1875                                 }
1876
1877                                 # Template cache array insertion
1878                                 $this->mTemplates[$part1] = $text;
1879                         }
1880                 }
1881
1882                 # Recursive parsing, escaping and link table handling
1883                 # Only for HTML output
1884                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1885                         $text = wfEscapeWikiText( $text );
1886                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1887                         # Clean up argument array
1888                         $assocArgs = array();
1889                         $index = 1;
1890                         foreach( $args as $arg ) {
1891                                 $eqpos = strpos( $arg, '=' );
1892                                 if ( $eqpos === false ) {
1893                                         $assocArgs[$index++] = $arg;
1894                                 } else {
1895                                         $name = trim( substr( $arg, 0, $eqpos ) );
1896                                         $value = trim( substr( $arg, $eqpos+1 ) );
1897                                         if ( $value === false ) {
1898                                                 $value = '';
1899                                         }
1900                                         if ( $name !== false ) {
1901                                                 $assocArgs[$name] = $value;
1902                                         }
1903                                 }
1904                         }
1905
1906                         # Do not enter included links in link table
1907                         if ( !is_null( $title ) ) {
1908                                 $wgLinkCache->suspend();
1909                         }
1910
1911                         # Add a new element to the templace recursion path
1912                         $this->mTemplatePath[$part1] = 1;
1913
1914                         $text = $this->stripParse( $text, $newline, $assocArgs );
1915
1916                         # Resume the link cache and register the inclusion as a link
1917                         if ( !is_null( $title ) ) {
1918                                 $wgLinkCache->resume();
1919                                 $wgLinkCache->addLinkObj( $title );
1920                         }
1921                 }
1922                 # Empties the template path
1923                 $this->mTemplatePath = array();
1924
1925                 if ( !$found ) {
1926                         return $matches[0];
1927                 } else {
1928                         return $text;
1929                 }
1930         }
1931
1932         # Triple brace replacement -- used for template arguments
1933         function argSubstitution( $matches ) {
1934                 $newline = $matches[1];
1935                 $arg = trim( $matches[2] );
1936                 $text = $matches[0];
1937                 $inputArgs = end( $this->mArgStack );
1938
1939                 if ( array_key_exists( $arg, $inputArgs ) ) {
1940                         $text = $this->stripParse( $inputArgs[$arg], $newline, array() );
1941                 }
1942
1943                 return $text;
1944         }
1945
1946         # Returns true if the function is allowed to include this entity
1947         function incrementIncludeCount( $dbk ) {
1948                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1949                         $this->mIncludeCount[$dbk] = 0;
1950                 }
1951                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1952                         return true;
1953                 } else {
1954                         return false;
1955                 }
1956         }
1957
1958
1959         # Cleans up HTML, removes dangerous tags and attributes
1960         /* private */ function removeHTMLtags( $text ) {
1961                 global $wgUseTidy, $wgUserHtml;
1962                 $fname = 'Parser::removeHTMLtags';
1963                 wfProfileIn( $fname );
1964
1965                 if( $wgUserHtml ) {
1966                         $htmlpairs = array( # Tags that must be closed
1967                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1968                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1969                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
1970                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1971                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
1972                         );
1973                         $htmlsingle = array(
1974                                 'br', 'hr', 'li', 'dt', 'dd'
1975                         );
1976                         $htmlnest = array( # Tags that can be nested--??
1977                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1978                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
1979                         );
1980                         $tabletags = array( # Can only appear inside table
1981                                 'td', 'th', 'tr'
1982                         );
1983                 } else {
1984                         $htmlpairs = array();
1985                         $htmlsingle = array();
1986                         $htmlnest = array();
1987                         $tabletags = array();
1988                 }
1989
1990                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1991                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1992
1993                 $htmlattrs = $this->getHTMLattrs () ;
1994
1995                 # Remove HTML comments
1996                 $text = preg_replace( '/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU', '$2', $text );
1997
1998                 $bits = explode( '<', $text );
1999                 $text = array_shift( $bits );
2000                 if(!$wgUseTidy) {
2001                         $tagstack = array(); $tablestack = array();
2002                         foreach ( $bits as $x ) {
2003                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2004                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2005                                 $x, $regs );
2006                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2007                                 error_reporting( $prev );
2008
2009                                 $badtag = 0 ;
2010                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2011                                         # Check our stack
2012                                         if ( $slash ) {
2013                                                 # Closing a tag...
2014                                                 if ( ! in_array( $t, $htmlsingle ) &&
2015                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2016                                                         @array_push( $tagstack, $ot );
2017                                                         $badtag = 1;
2018                                                 } else {
2019                                                         if ( $t == 'table' ) {
2020                                                                 $tagstack = array_pop( $tablestack );
2021                                                         }
2022                                                         $newparams = '';
2023                                                 }
2024                                         } else {
2025                                                 # Keep track for later
2026                                                 if ( in_array( $t, $tabletags ) &&
2027                                                 ! in_array( 'table', $tagstack ) ) {
2028                                                         $badtag = 1;
2029                                                 } else if ( in_array( $t, $tagstack ) &&
2030                                                 ! in_array ( $t , $htmlnest ) ) {
2031                                                         $badtag = 1 ;
2032                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
2033                                                         if ( $t == 'table' ) {
2034                                                                 array_push( $tablestack, $tagstack );
2035                                                                 $tagstack = array();
2036                                                         }
2037                                                         array_push( $tagstack, $t );
2038                                                 }
2039                                                 # Strip non-approved attributes from the tag
2040                                                 $newparams = $this->fixTagAttributes($params);
2041
2042                                         }
2043                                         if ( ! $badtag ) {
2044                                                 $rest = str_replace( '>', '&gt;', $rest );
2045                                                 $text .= "<$slash$t $newparams$brace$rest";
2046                                                 continue;
2047                                         }
2048                                 }
2049                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2050                         }
2051                         # Close off any remaining tags
2052                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2053                                 $text .= "</$t>\n";
2054                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2055                         }
2056                 } else {
2057                         # this might be possible using tidy itself
2058                         foreach ( $bits as $x ) {
2059                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2060                                 $x, $regs );
2061                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2062                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2063                                         $newparams = $this->fixTagAttributes($params);
2064                                         $rest = str_replace( '>', '&gt;', $rest );
2065                                         $text .= "<$slash$t $newparams$brace$rest";
2066                                 } else {
2067                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2068                                 }
2069                         }
2070                 }
2071                 wfProfileOut( $fname );
2072                 return $text;
2073         }
2074
2075
2076 /*
2077  *
2078  * This function accomplishes several tasks:
2079  * 1) Auto-number headings if that option is enabled
2080  * 2) Add an [edit] link to sections for logged in users who have enabled the option
2081  * 3) Add a Table of contents on the top for users who have enabled the option
2082  * 4) Auto-anchor headings
2083  *
2084  * It loops through all headlines, collects the necessary data, then splits up the
2085  * string and re-inserts the newly formatted headlines.
2086  *
2087  */
2088
2089         /* private */ function formatHeadings( $text, $isMain=true ) {
2090                 global $wgInputEncoding, $wgMaxTocLevel;
2091
2092                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2093                 $doShowToc = $this->mOptions->getShowToc();
2094                 $forceTocHere = false;
2095                 if( !$this->mTitle->userCanEdit() ) {
2096                         $showEditLink = 0;
2097                         $rightClickHack = 0;
2098                 } else {
2099                         $showEditLink = $this->mOptions->getEditSection();
2100                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2101                 }
2102
2103                 # Inhibit editsection links if requested in the page
2104                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2105                 if( $esw->matchAndRemove( $text ) ) {
2106                         $showEditLink = 0;
2107                 }
2108                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2109                 # do not add TOC
2110                 $mw =& MagicWord::get( MAG_NOTOC );
2111                 if( $mw->matchAndRemove( $text ) ) {
2112                         $doShowToc = 0;
2113                 }
2114
2115                 # never add the TOC to the Main Page. This is an entry page that should not
2116                 # be more than 1-2 screens large anyway
2117                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2118                         $doShowToc = 0;
2119                 }
2120
2121                 # Get all headlines for numbering them and adding funky stuff like [edit]
2122                 # links - this is for later, but we need the number of headlines right now
2123                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
2124
2125                 # if there are fewer than 4 headlines in the article, do not show TOC
2126                 if( $numMatches < 4 ) {
2127                         $doShowToc = 0;
2128                 }
2129
2130                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2131                 # override above conditions and always show TOC at that place
2132                 $mw =& MagicWord::get( MAG_TOC );
2133                 if ($mw->match( $text ) ) {
2134                         $doShowToc = 1;
2135                         $forceTocHere = true;
2136                 } else {
2137                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2138                         # override above conditions and always show TOC above first header
2139                         $mw =& MagicWord::get( MAG_FORCETOC );
2140                         if ($mw->matchAndRemove( $text ) ) {
2141                                 $doShowToc = 1;
2142                         }
2143                 }
2144
2145
2146
2147                 # We need this to perform operations on the HTML
2148                 $sk =& $this->mOptions->getSkin();
2149
2150                 # headline counter
2151                 $headlineCount = 0;
2152
2153                 # Ugh .. the TOC should have neat indentation levels which can be
2154                 # passed to the skin functions. These are determined here
2155                 $toclevel = 0;
2156                 $toc = '';
2157                 $full = '';
2158                 $head = array();
2159                 $sublevelCount = array();
2160                 $level = 0;
2161                 $prevlevel = 0;
2162                 foreach( $matches[3] as $headline ) {
2163                         $numbering = '';
2164                         if( $level ) {
2165                                 $prevlevel = $level;
2166                         }
2167                         $level = $matches[1][$headlineCount];
2168                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2169                                 # reset when we enter a new level
2170                                 $sublevelCount[$level] = 0;
2171                                 $toc .= $sk->tocIndent( $level - $prevlevel );
2172                                 $toclevel += $level - $prevlevel;
2173                         }
2174                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2175                                 # reset when we step back a level
2176                                 $sublevelCount[$level+1]=0;
2177                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2178                                 $toclevel -= $prevlevel - $level;
2179                         }
2180                         # count number of headlines for each level
2181                         @$sublevelCount[$level]++;
2182                         if( $doNumberHeadings || $doShowToc ) {
2183                                 $dot = 0;
2184                                 for( $i = 1; $i <= $level; $i++ ) {
2185                                         if( !empty( $sublevelCount[$i] ) ) {
2186                                                 if( $dot ) {
2187                                                         $numbering .= '.';
2188                                                 }
2189                                                 $numbering .= $sublevelCount[$i];
2190                                                 $dot = 1;
2191                                         }
2192                                 }
2193                         }
2194
2195                         # The canonized header is a version of the header text safe to use for links
2196                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2197                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2198                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2199
2200                         # strip out HTML
2201                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2202                         $tocline = trim( $canonized_headline );
2203                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2204                         $replacearray = array(
2205                                 '%3A' => ':',
2206                                 '%' => '.'
2207                         );
2208                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2209                         $refer[$headlineCount] = $canonized_headline;
2210
2211                         # count how many in assoc. array so we can track dupes in anchors
2212                         @$refers[$canonized_headline]++;
2213                         $refcount[$headlineCount]=$refers[$canonized_headline];
2214
2215                         # Prepend the number to the heading text
2216
2217                         if( $doNumberHeadings || $doShowToc ) {
2218                                 $tocline = $numbering . ' ' . $tocline;
2219
2220                                 # Don't number the heading if it is the only one (looks silly)
2221                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2222                                         # the two are different if the line contains a link
2223                                         $headline=$numbering . ' ' . $headline;
2224                                 }
2225                         }
2226
2227                         # Create the anchor for linking from the TOC to the section
2228                         $anchor = $canonized_headline;
2229                         if($refcount[$headlineCount] > 1 ) {
2230                                 $anchor .= '_' . $refcount[$headlineCount];
2231                         }
2232                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2233                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2234                         }
2235                         if( $showEditLink ) {
2236                                 if ( empty( $head[$headlineCount] ) ) {
2237                                         $head[$headlineCount] = '';
2238                                 }
2239                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
2240                         }
2241
2242                         # Add the edit section span
2243                         if( $rightClickHack ) {
2244                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
2245                         }
2246
2247                         # give headline the correct <h#> tag
2248                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
2249
2250                         $headlineCount++;
2251                 }
2252
2253                 if( $doShowToc ) {
2254                         $toclines = $headlineCount;
2255                         $toc .= $sk->tocUnindent( $toclevel );
2256                         $toc = $sk->tocTable( $toc );
2257                 }
2258
2259                 # split up and insert constructed headlines
2260
2261                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2262                 $i = 0;
2263
2264                 foreach( $blocks as $block ) {
2265                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2266                             # This is the [edit] link that appears for the top block of text when
2267                                 # section editing is enabled
2268
2269                                 # Disabled because it broke block formatting
2270                                 # For example, a bullet point in the top line
2271                                 # $full .= $sk->editSectionLink(0);
2272                         }
2273                         $full .= $block;
2274                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2275                         # Top anchor now in skin
2276                                 $full = $full.$toc;
2277                         }
2278
2279                         if( !empty( $head[$i] ) ) {
2280                                 $full .= $head[$i];
2281                         }
2282                         $i++;
2283                 }
2284                 if($forceTocHere) {
2285                         $mw =& MagicWord::get( MAG_TOC );
2286                         return $mw->replace( $toc, $full );
2287                 } else {
2288                         return $full;
2289                 }
2290         }
2291
2292         # Return an HTML link for the "ISBN 123456" text
2293         /* private */ function magicISBN( $text ) {
2294                 global $wgLang;
2295                 $fname = 'Parser::magicISBN';
2296                 wfProfileIn( $fname );
2297
2298                 $a = split( 'ISBN ', " $text" );
2299                 if ( count ( $a ) < 2 ) {
2300                         wfProfileOut( $fname );
2301                         return $text;
2302                 }
2303                 $text = substr( array_shift( $a ), 1);
2304                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2305
2306                 foreach ( $a as $x ) {
2307                         $isbn = $blank = '' ;
2308                         while ( ' ' == $x{0} ) {
2309                                 $blank .= ' ';
2310                                 $x = substr( $x, 1 );
2311                         }
2312                         while ( strstr( $valid, $x{0} ) != false ) {
2313                                 $isbn .= $x{0};
2314                                 $x = substr( $x, 1 );
2315                         }
2316                         $num = str_replace( '-', '', $isbn );
2317                         $num = str_replace( ' ', '', $num );
2318
2319                         if ( '' == $num ) {
2320                                 $text .= "ISBN $blank$x";
2321                         } else {
2322                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2323                                 $text .= '<a href="' .
2324                                 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
2325                                         "\" class=\"internal\">ISBN $isbn</a>";
2326                                 $text .= $x;
2327                         }
2328                 }
2329                 wfProfileOut( $fname );
2330                 return $text;
2331         }
2332
2333         # Return an HTML link for the "GEO ..." text
2334         /* private */ function magicGEO( $text ) {
2335                 global $wgLang, $wgUseGeoMode;
2336                 if ( !isset ( $wgUseGeoMode ) || !$wgUseGeoMode ) return $text ;
2337                 $fname = 'Parser::magicGEO';
2338                 wfProfileIn( $fname );
2339
2340                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2341                 $directions = array ( "N" => "North" , "S" => "South" , "E" => "East" , "W" => "West" ) ;
2342                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2343                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2344                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2345                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2346
2347                 $a = split( 'GEO ', " $text" );
2348                 if ( count ( $a ) < 2 ) {
2349                         wfProfileOut( $fname );
2350                         return $text;
2351                 }
2352                 $text = substr( array_shift( $a ), 1);
2353                 $valid = '0123456789.+-:';
2354
2355                 foreach ( $a as $x ) {
2356                         $geo = $blank = '' ;
2357                         while ( ' ' == $x{0} ) {
2358                                 $blank .= ' ';
2359                                 $x = substr( $x, 1 );
2360                         }
2361                         while ( strstr( $valid, $x{0} ) != false ) {
2362                                 $geo .= $x{0};
2363                                 $x = substr( $x, 1 );
2364                         }
2365                         $num = str_replace( '+', '', $geo );
2366                         $num = str_replace( ' ', '', $num );
2367
2368                         if ( '' == $num || count ( explode ( ":" , $num , 3 ) ) < 2 ) {
2369                                 $text .= "GEO $blank$x";
2370                         } else {
2371                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2372                                 $text .= '<a href="' .
2373                                 $titleObj->escapeLocalUrl( "coordinates={$num}" ) .
2374                                         "\" class=\"internal\">GEO $geo</a>";
2375                                 $text .= $x;
2376                         }
2377                 }
2378                 wfProfileOut( $fname );
2379                 return $text;
2380         }
2381
2382         # Return an HTML link for the "RFC 1234" text
2383         /* private */ function magicRFC( $text ) {
2384                 global $wgLang;
2385
2386                 $a = split( 'RFC ', ' '.$text );
2387                 if ( count ( $a ) < 2 ) return $text;
2388                 $text = substr( array_shift( $a ), 1);
2389                 $valid = '0123456789';
2390
2391                 foreach ( $a as $x ) {
2392                         $rfc = $blank = '' ;
2393                         while ( ' ' == $x{0} ) {
2394                                 $blank .= ' ';
2395                                 $x = substr( $x, 1 );
2396                         }
2397                         while ( strstr( $valid, $x{0} ) != false ) {
2398                                 $rfc .= $x{0};
2399                                 $x = substr( $x, 1 );
2400                         }
2401
2402                         if ( '' == $rfc ) {
2403                                 $text .= "RFC $blank$x";
2404                         } else {
2405                                 $url = wfmsg( 'rfcurl' );
2406                                 $url = str_replace( '$1', $rfc, $url);
2407                                 $sk =& $this->mOptions->getSkin();
2408                                 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
2409                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2410                         }
2411                 }
2412                 return $text;
2413         }
2414
2415         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2416                 $this->mOptions = $options;
2417                 $this->mTitle =& $title;
2418                 $this->mOutputType = OT_WIKI;
2419
2420                 if ( $clearState ) {
2421                         $this->clearState();
2422                 }
2423
2424                 $stripState = false;
2425                 $pairs = array(
2426                         "\r\n" => "\n",
2427                         );
2428                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2429                 // now with regexes
2430                 /*
2431                 $pairs = array(
2432                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2433                         "/<br *?>/i" => "<br />",
2434                 );
2435                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2436                 */
2437                 $text = $this->strip( $text, $stripState, false );
2438                 $text = $this->pstPass2( $text, $user );
2439                 $text = $this->unstrip( $text, $stripState );
2440                 $text = $this->unstripNoWiki( $text, $stripState );
2441                 return $text;
2442         }
2443
2444         /* private */ function pstPass2( $text, &$user ) {
2445                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2446
2447                 # Variable replacement
2448                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2449                 $text = $this->replaceVariables( $text );
2450
2451                 # Signatures
2452                 #
2453                 $n = $user->getName();
2454                 $k = $user->getOption( 'nickname' );
2455                 if ( '' == $k ) { $k = $n; }
2456                 if(isset($wgLocaltimezone)) {
2457                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2458                 }
2459                 /* Note: this is an ugly timezone hack for the European wikis */
2460                 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2461                   ' (' . date( 'T' ) . ')';
2462                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2463
2464                 $text = preg_replace( '/~~~~~/', $d, $text );
2465                 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2466                 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2467
2468                 # Context links: [[|name]] and [[name (context)|]]
2469                 #
2470                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2471                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2472                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2473                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2474
2475                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2476                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2477                 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/";          # [[namespace:page|]]
2478                 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2479                                                                                                                 # [[ns:page (cont)|]]
2480                 $context = "";
2481                 $t = $this->mTitle->getText();
2482                 if ( preg_match( $conpat, $t, $m ) ) {
2483                         $context = $m[2];
2484                 }
2485                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2486                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2487                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2488
2489                 if ( '' == $context ) {
2490                         $text = preg_replace( $p2, '[[\\1]]', $text );
2491                 } else {
2492                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2493                 }
2494
2495                 /*
2496                 $mw =& MagicWord::get( MAG_SUBST );
2497                 $wgCurParser = $this->fork();
2498                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2499                 $this->merge( $wgCurParser );
2500                 */
2501
2502                 # Trim trailing whitespace
2503                 # MAG_END (__END__) tag allows for trailing
2504                 # whitespace to be deliberately included
2505                 $text = rtrim( $text );
2506                 $mw =& MagicWord::get( MAG_END );
2507                 $mw->matchAndRemove( $text );
2508
2509                 return $text;
2510         }
2511
2512         # Set up some variables which are usually set up in parse()
2513         # so that an external function can call some class members with confidence
2514         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2515                 $this->mTitle =& $title;
2516                 $this->mOptions = $options;
2517                 $this->mOutputType = $outputType;
2518                 if ( $clearState ) {
2519                         $this->clearState();
2520                 }
2521         }
2522
2523         function transformMsg( $text, $options ) {
2524                 global $wgTitle;
2525                 static $executing = false;
2526
2527                 # Guard against infinite recursion
2528                 if ( $executing ) {
2529                         return $text;
2530                 }
2531                 $executing = true;
2532
2533                 $this->mTitle = $wgTitle;
2534                 $this->mOptions = $options;
2535                 $this->mOutputType = OT_MSG;
2536                 $this->clearState();
2537                 $text = $this->replaceVariables( $text );
2538
2539                 $executing = false;
2540                 return $text;
2541         }
2542
2543         # Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2544         # Callback will be called with the text within
2545         # Transform and return the text within
2546         function setHook( $tag, $callback ) {
2547                 $oldVal = @$this->mTagHooks[$tag];
2548                 $this->mTagHooks[$tag] = $callback;
2549                 return $oldVal;
2550         }
2551 }
2552
2553 class ParserOutput
2554 {
2555         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2556         var $mCacheTime; # Used in ParserCache
2557
2558         function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2559                 $containsOldMagic = false )
2560         {
2561                 $this->mText = $text;
2562                 $this->mLanguageLinks = $languageLinks;
2563                 $this->mCategoryLinks = $categoryLinks;
2564                 $this->mContainsOldMagic = $containsOldMagic;
2565                 $this->mCacheTime = "";
2566         }
2567
2568         function getText() { return $this->mText; }
2569         function getLanguageLinks() { return $this->mLanguageLinks; }
2570         function getCategoryLinks() { return $this->mCategoryLinks; }
2571         function getCacheTime() { return $this->mCacheTime; }
2572         function containsOldMagic() { return $this->mContainsOldMagic; }
2573         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2574         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2575         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2576         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2577         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2578
2579         function merge( $other ) {
2580                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2581                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2582                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2583         }
2584
2585 }
2586
2587 class ParserOptions
2588 {
2589         # All variables are private
2590         var $mUseTeX;                    # Use texvc to expand <math> tags
2591         var $mUseCategoryMagic;          # Treat [[Category:xxxx]] tags specially
2592         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2593         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2594         var $mAllowExternalImages;       # Allow external images inline
2595         var $mSkin;                      # Reference to the preferred skin
2596         var $mDateFormat;                # Date format index
2597         var $mEditSection;               # Create "edit section" links
2598         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2599         var $mNumberHeadings;            # Automatically number headings
2600         var $mShowToc;                   # Show table of contents
2601
2602         function getUseTeX()                        { return $this->mUseTeX; }
2603         function getUseCategoryMagic()              { return $this->mUseCategoryMagic; }
2604         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2605         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2606         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2607         function getSkin()                          { return $this->mSkin; }
2608         function getDateFormat()                    { return $this->mDateFormat; }
2609         function getEditSection()                   { return $this->mEditSection; }
2610         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
2611         function getNumberHeadings()                { return $this->mNumberHeadings; }
2612         function getShowToc()                       { return $this->mShowToc; }
2613
2614         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
2615         function setUseCategoryMagic( $x )          { return wfSetVar( $this->mUseCategoryMagic, $x ); }
2616         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
2617         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
2618         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
2619         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
2620         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
2621         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2622         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
2623         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
2624
2625     function setSkin( &$x ) { $this->mSkin =& $x; }
2626
2627         /* static */ function newFromUser( &$user ) {
2628                 $popts = new ParserOptions;
2629                 $popts->initialiseFromUser( $user );
2630                 return $popts;
2631         }
2632
2633         function initialiseFromUser( &$userInput ) {
2634                 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2635
2636                 if ( !$userInput ) {
2637                         $user = new User;
2638                         $user->setLoaded( true );
2639                 } else {
2640                         $user =& $userInput;
2641                 }
2642
2643                 $this->mUseTeX = $wgUseTeX;
2644                 $this->mUseCategoryMagic = $wgUseCategoryMagic;
2645                 $this->mUseDynamicDates = $wgUseDynamicDates;
2646                 $this->mInterwikiMagic = $wgInterwikiMagic;
2647                 $this->mAllowExternalImages = $wgAllowExternalImages;
2648                 $this->mSkin =& $user->getSkin();
2649                 $this->mDateFormat = $user->getOption( 'date' );
2650                 $this->mEditSection = $user->getOption( 'editsection' );
2651                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2652                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2653                 $this->mShowToc = $user->getOption( 'showtoc' );
2654         }
2655
2656
2657 }
2658
2659 # Regex callbacks, used in Parser::replaceVariables
2660 function wfBraceSubstitution( $matches )
2661 {
2662         global $wgCurParser;
2663         return $wgCurParser->braceSubstitution( $matches );
2664 }
2665
2666 function wfArgSubstitution( $matches )
2667 {
2668         global $wgCurParser;
2669         return $wgCurParser->argSubstitution( $matches );
2670 }
2671
2672 function wfVariableSubstitution( $matches )
2673 {
2674         global $wgCurParser;
2675         return $wgCurParser->variableSubstitution( $matches );
2676 }
2677
2678 ?>