includes/Parser.php

   1 <?php
   2
   3 // require_once('Tokenizer.php');
   4
   5 /**
   6  * PHP Parser
   7  *
   8  * Processes wiki markup
   9  *
  10  * There are two main entry points into the Parser class:
  11  * parse()
  12  *   produces HTML output
  13  * preSaveTransform().
  14  *   produces altered wiki markup.
  15  *
  16  * Globals used:
  17  *    objects:   $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
  18  *
  19  * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
  20  *
  21  * settings:
  22  *  $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
  23  *  $wgNamespacesWithSubpages, $wgAllowExternalImages*,
  24  *  $wgLocaltimezone
  25  *
  26  *  * only within ParserOptions
  27  *
  28  * @package MediaWiki
  29  */
  30
  31 /**
  32  * Variable substitution O(N^2) attack
  33  *
  34 * Without countermeasures, it would be possible to attack the parser by saving
  35 * a page filled with a large number of inclusions of large pages. The size of
  36 * the generated page would be proportional to the square of the input size.
  37 * Hence, we limit the number of inclusions of any given page, thus bringing any
  38 * attack back to O(N).
  39 */
  40 define( 'MAX_INCLUDE_REPEAT', 100 );
  41 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
  42
  43 # Allowed values for $mOutputType
  44 define( 'OT_HTML', 1 );
  45 define( 'OT_WIKI', 2 );
  46 define( 'OT_MSG' , 3 );
  47
  48 # string parameter for extractTags which will cause it
  49 # to strip HTML comments in addition to regular
  50 # <XML>-style tags. This should not be anything we
  51 # may want to use in wikisyntax
  52 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
  53
  54 # prefix for escaping, used in two functions at least
  55 define( 'UNIQ_PREFIX', 'NaodW29');
  56
  57 # Constants needed for external link processing
  58 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
  59 define( 'HTTP_PROTOCOLS', 'http|https' );
  60 # Everything except bracket, space, or control characters
  61 define( 'EXT_LINK_URL_CLASS', '[^]\\x00-\\x20\\x7F]' );
  62 define( 'INVERSE_EXT_LINK_URL_CLASS', '[\]\\x00-\\x20\\x7F]' );
  63 # Including space
  64 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
  65 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
  66 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
  67 define( 'EXT_LINK_BRACKETED',  '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
  68 define( 'EXT_IMAGE_REGEX',
  69         '/^('.HTTP_PROTOCOLS.':)'.  # Protocol
  70         '('.EXT_LINK_URL_CLASS.'+)\\/'.  # Hostname and path
  71         '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
  72 );
  73
  74 /**
  75  * @todo document
  76  * @package MediaWiki
  77  */
  78 class Parser
  79 {
  80         # Persistent:
  81         var $mTagHooks;
  82
  83         # Cleared with clearState():
  84         var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
  85         var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
  86
  87         # Temporary:
  88         var $mOptions, $mTitle, $mOutputType,
  89             $mTemplates,        // cache of already loaded templates, avoids
  90                                 // multiple SQL queries for the same string
  91             $mTemplatePath;     // stores an unsorted hash of all the templates already loaded
  92                                 // in this path. Used for loop detection.
  93
  94         function Parser() {
  95                 $this->mTemplates = array();
  96                 $this->mTemplatePath = array();
  97                 $this->mTagHooks = array();
  98                 $this->clearState();
  99         }
 100
 101         function clearState() {
 102                 $this->mOutput = new ParserOutput;
 103                 $this->mAutonumber = 0;
 104                 $this->mLastSection = "";
 105                 $this->mDTopen = false;
 106                 $this->mVariables = false;
 107                 $this->mIncludeCount = array();
 108                 $this->mStripState = array();
 109                 $this->mArgStack = array();
 110                 $this->mInPre = false;
 111         }
 112
 113         # First pass--just handle <nowiki> sections, pass the rest off
 114         # to internalParse() which does all the real work.
 115         #
 116         # Returns a ParserOutput
 117         #
 118         function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
 119                 global $wgUseTidy;
 120                 $fname = 'Parser::parse';
 121                 wfProfileIn( $fname );
 122
 123                 if ( $clearState ) {
 124                         $this->clearState();
 125                 }
 126
 127                 $this->mOptions = $options;
 128                 $this->mTitle =& $title;
 129                 $this->mOutputType = OT_HTML;
 130
 131                 $stripState = NULL;
 132                 $text = $this->strip( $text, $this->mStripState );
 133                 $text = $this->internalParse( $text, $linestart );
 134                 $text = $this->unstrip( $text, $this->mStripState );
 135                 # Clean up special characters, only run once, next-to-last before doBlockLevels
 136                 if(!$wgUseTidy) {
 137                         $fixtags = array(
 138                                 # french spaces, last one Guillemet-left
 139                                 # only if there is something before the space
 140                                 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
 141                                 # french spaces, Guillemet-right
 142                                 "/(\\302\\253) /i"=>"\\1&nbsp;",
 143                                 '/<hr *>/i' => '<hr />',
 144                                 '/<br *>/i' => '<br />',
 145                                 '/<center *>/i' => '<div class="center">',
 146                                 '/<\\/center *>/i' => '</div>',
 147                                 # Clean up spare ampersands; note that we probably ought to be
 148                                 # more careful about named entities.
 149                                 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
 150                         );
 151                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 152                 } else {
 153                         $fixtags = array(
 154                                 # french spaces, last one Guillemet-left
 155                                 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
 156                                 # french spaces, Guillemet-right
 157                                 '/(\\302\\253) /i' => '\\1&nbsp;',
 158                                 '/<center *>/i' => '<div class="center">',
 159                                 '/<\\/center *>/i' => '</div>'
 160                         );
 161                         $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
 162                 }
 163                 # only once and last
 164                 $text = $this->doBlockLevels( $text, $linestart );
 165                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 166                 if($wgUseTidy) {
 167                         $text = $this->tidy($text);
 168                 }
 169                 $this->mOutput->setText( $text );
 170                 wfProfileOut( $fname );
 171                 return $this->mOutput;
 172         }
 173
 174         /* static */ function getRandomString() {
 175                 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
 176         }
 177
 178         # Replaces all occurrences of <$tag>content</$tag> in the text
 179         # with a random marker and returns the new text. the output parameter
 180         # $content will be an associative array filled with data on the form
 181         # $unique_marker => content.
 182
 183         # If $content is already set, the additional entries will be appended
 184
 185         # If $tag is set to STRIP_COMMENTS, the function will extract
 186         # <!-- HTML comments -->
 187
 188         /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ''){
 189                 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
 190                 if ( !$content ) {
 191                         $content = array( );
 192                 }
 193                 $n = 1;
 194                 $stripped = '';
 195
 196                 while ( '' != $text ) {
 197                         if($tag==STRIP_COMMENTS) {
 198                                 $p = preg_split( '/<!--/i', $text, 2 );
 199                         } else {
 200                                 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
 201                         }
 202                         $stripped .= $p[0];
 203                         if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
 204                                 $text = '';
 205                         } else {
 206                                 if($tag==STRIP_COMMENTS) {
 207                                         $q = preg_split( '/-->/i', $p[1], 2 );
 208                                 } else {
 209                                         $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
 210                                 }
 211                                 $marker = $rnd . sprintf('%08X', $n++);
 212                                 $content[$marker] = $q[0];
 213                                 $stripped .= $marker;
 214                                 $text = $q[1];
 215                         }
 216                 }
 217                 return $stripped;
 218         }
 219
 220         # Strips and renders <nowiki>, <pre>, <math>, <hiero>
 221         # If $render is set, performs necessary rendering operations on plugins
 222         # Returns the text, and fills an array with data needed in unstrip()
 223         # If the $state is already a valid strip state, it adds to the state
 224
 225         # When $stripcomments is set, HTML comments <!-- like this -->
 226         # will be stripped in addition to other tags. This is important
 227         # for section editing, where these comments cause confusion when
 228         # counting the sections in the wikisource
 229         function strip( $text, &$state, $stripcomments = false ) {
 230                 $render = ($this->mOutputType == OT_HTML);
 231                 $html_content = array();
 232                 $nowiki_content = array();
 233                 $math_content = array();
 234                 $pre_content = array();
 235                 $comment_content = array();
 236                 $ext_content = array();
 237
 238                 # Replace any instances of the placeholders
 239                 $uniq_prefix = UNIQ_PREFIX;
 240                 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
 241
 242                 # html
 243                 global $wgRawHtml, $wgWhitelistEdit;
 244                 if( $wgRawHtml && $wgWhitelistEdit ) {
 245                         $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
 246                         foreach( $html_content as $marker => $content ) {
 247                                 if ($render ) {
 248                                         # Raw and unchecked for validity.
 249                                         $html_content[$marker] = $content;
 250                                 } else {
 251                                         $html_content[$marker] = '<html>'.$content.'</html>';
 252                                 }
 253                         }
 254                 }
 255
 256                 # nowiki
 257                 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
 258                 foreach( $nowiki_content as $marker => $content ) {
 259                         if( $render ){
 260                                 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
 261                         } else {
 262                                 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
 263                         }
 264                 }
 265
 266                 # math
 267                 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
 268                 foreach( $math_content as $marker => $content ){
 269                         if( $render ) {
 270                                 if( $this->mOptions->getUseTeX() ) {
 271                                         $math_content[$marker] = renderMath( $content );
 272                                 } else {
 273                                         $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
 274                                 }
 275                         } else {
 276                                 $math_content[$marker] = '<math>'.$content.'</math>';
 277                         }
 278                 }
 279
 280                 # pre
 281                 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
 282                 foreach( $pre_content as $marker => $content ){
 283                         if( $render ){
 284                                 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
 285                         } else {
 286                                 $pre_content[$marker] = '<pre>'.$content.'</pre>';
 287                         }
 288                 }
 289
 290                 # Comments
 291                 if($stripcomments) {
 292                         $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
 293                         foreach( $comment_content as $marker => $content ){
 294                                 $comment_content[$marker] = '<!--'.$content.'-->';
 295                         }
 296                 }
 297
 298                 # Extensions
 299                 foreach ( $this->mTagHooks as $tag => $callback ) {
 300                         $ext_contents[$tag] = array();
 301                         $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
 302                         foreach( $ext_content[$tag] as $marker => $content ) {
 303                                 if ( $render ) {
 304                                         $ext_content[$tag][$marker] = $callback( $content );
 305                                 } else {
 306                                         $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
 307                                 }
 308                         }
 309                 }
 310
 311                 # Merge state with the pre-existing state, if there is one
 312                 if ( $state ) {
 313                         $state['html'] = $state['html'] + $html_content;
 314                         $state['nowiki'] = $state['nowiki'] + $nowiki_content;
 315                         $state['math'] = $state['math'] + $math_content;
 316                         $state['pre'] = $state['pre'] + $pre_content;
 317                         $state['comment'] = $state['comment'] + $comment_content;
 318
 319                         foreach( $ext_content as $tag => $array ) {
 320                                 if ( array_key_exists( $tag, $state ) ) {
 321                                         $state[$tag] = $state[$tag] + $array;
 322                                 }
 323                         }
 324                 } else {
 325                         $state = array(
 326                           'html' => $html_content,
 327                           'nowiki' => $nowiki_content,
 328                           'math' => $math_content,
 329                           'pre' => $pre_content,
 330                           'comment' => $comment_content,
 331                         ) + $ext_content;
 332                 }
 333                 return $text;
 334         }
 335
 336         # always call unstripNoWiki() after this one
 337         function unstrip( $text, &$state ) {
 338                 # Must expand in reverse order, otherwise nested tags will be corrupted
 339                 $contentDict = end( $state );
 340                 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
 341                         if( key($state) != 'nowiki' && key($state) != 'html') {
 342                                 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
 343                                         $text = str_replace( key( $contentDict ), $content, $text );
 344                                 }
 345                         }
 346                 }
 347
 348                 return $text;
 349         }
 350         # always call this after unstrip() to preserve the order
 351         function unstripNoWiki( $text, &$state ) {
 352                 # Must expand in reverse order, otherwise nested tags will be corrupted
 353                 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
 354                         $text = str_replace( key( $state['nowiki'] ), $content, $text );
 355                 }
 356
 357                 global $wgRawHtml;
 358                 if ($wgRawHtml) {
 359                         for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
 360                                 $text = str_replace( key( $state['html'] ), $content, $text );
 361                         }
 362                 }
 363
 364                 return $text;
 365         }
 366
 367         # Add an item to the strip state
 368         # Returns the unique tag which must be inserted into the stripped text
 369         # The tag will be replaced with the original text in unstrip()
 370         function insertStripItem( $text, &$state ) {
 371                 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
 372                 if ( !$state ) {
 373                         $state = array(
 374                           'html' => array(),
 375                           'nowiki' => array(),
 376                           'math' => array(),
 377                           'pre' => array()
 378                         );
 379                 }
 380                 $state['item'][$rnd] = $text;
 381                 return $rnd;
 382         }
 383
 384         # Return allowed HTML attributes
 385         function getHTMLattrs () {
 386                 $htmlattrs = array( # Allowed attributes--no scripting, etc.
 387                                 'title', 'align', 'lang', 'dir', 'width', 'height',
 388                                 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
 389                                 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
 390                                 /* FONT */ 'type', 'start', 'value', 'compact',
 391                                 /* For various lists, mostly deprecated but safe */
 392                                 'summary', 'width', 'border', 'frame', 'rules',
 393                                 'cellspacing', 'cellpadding', 'valign', 'char',
 394                                 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
 395                                 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
 396                                 'id', 'class', 'name', 'style' /* For CSS */
 397                                 );
 398                 return $htmlattrs ;
 399         }
 400
 401         # Remove non approved attributes and javascript in css
 402         function fixTagAttributes ( $t ) {
 403                 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
 404                 $htmlattrs = $this->getHTMLattrs() ;
 405
 406                 # Strip non-approved attributes from the tag
 407                 $t = preg_replace(
 408                         '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
 409                         "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
 410                         $t);
 411
 412                 $t = str_replace ( '<></>' , '' , $t ) ; # This should fix bug 980557
 413
 414                 # Strip javascript "expression" from stylesheets. Brute force approach:
 415                 # If anythin offensive is found, all attributes of the HTML tag are dropped
 416
 417                 if( preg_match(
 418                         '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
 419                         wfMungeToUtf8( $t ) ) )
 420                 {
 421                         $t='';
 422                 }
 423
 424                 return trim ( $t ) ;
 425         }
 426
 427         # interface with html tidy, used if $wgUseTidy = true
 428         function tidy ( $text ) {
 429                 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
 430                 global $wgInputEncoding, $wgOutputEncoding;
 431                 $fname = 'Parser::tidy';
 432                 wfProfileIn( $fname );
 433
 434                 $cleansource = '';
 435                 switch(strtoupper($wgOutputEncoding)) {
 436                         case 'ISO-8859-1':
 437                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
 438                                 break;
 439                         case 'UTF-8':
 440                                 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
 441                                 break;
 442                         default:
 443                                 $wgTidyOpts .= ' -raw';
 444                         }
 445
 446                 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
 447 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
 448 '<head><title>test</title></head><body>'.$text.'</body></html>';
 449                 $descriptorspec = array(
 450                         0 => array('pipe', 'r'),
 451                         1 => array('pipe', 'w'),
 452                         2 => array('file', '/dev/null', 'a')
 453                 );
 454                 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
 455                 if (is_resource($process)) {
 456                         fwrite($pipes[0], $wrappedtext);
 457                         fclose($pipes[0]);
 458                         while (!feof($pipes[1])) {
 459                                 $cleansource .= fgets($pipes[1], 1024);
 460                         }
 461                         fclose($pipes[1]);
 462                         $return_value = proc_close($process);
 463                 }
 464
 465                 wfProfileOut( $fname );
 466
 467                 if( $cleansource == '' && $text != '') {
 468                         wfDebug( "Tidy error detected!\n" );
 469                         return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
 470                 } else {
 471                         return $cleansource;
 472                 }
 473         }
 474
 475         # parse the wiki syntax used to render tables
 476         function doTableStuff ( $t ) {
 477                 $fname = 'Parser::doTableStuff';
 478                 wfProfileIn( $fname );
 479
 480                 $t = explode ( "\n" , $t ) ;
 481                 $td = array () ; # Is currently a td tag open?
 482                 $ltd = array () ; # Was it TD or TH?
 483                 $tr = array () ; # Is currently a tr tag open?
 484                 $ltr = array () ; # tr attributes
 485                 $indent_level = 0; # indent level of the table
 486                 foreach ( $t AS $k => $x )
 487                 {
 488                         $x = trim ( $x ) ;
 489                         $fc = substr ( $x , 0 , 1 ) ;
 490                         if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
 491                                 $indent_level = strlen( $matches[1] );
 492                                 $t[$k] = "\n" .
 493                                         str_repeat( '<dl><dd>', $indent_level ) .
 494                                         '<table ' . $this->fixTagAttributes ( $matches[2] ) . '>' ;
 495                                 array_push ( $td , false ) ;
 496                                 array_push ( $ltd , '' ) ;
 497                                 array_push ( $tr , false ) ;
 498                                 array_push ( $ltr , '' ) ;
 499                         }
 500                         else if ( count ( $td ) == 0 ) { } # Don't do any of the following
 501                         else if ( '|}' == substr ( $x , 0 , 2 ) ) {
 502                                 $z = "</table>\n" ;
 503                                 $l = array_pop ( $ltd ) ;
 504                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 505                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 506                                 array_pop ( $ltr ) ;
 507                                 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
 508                         }
 509                         else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
 510                                 $x = substr ( $x , 1 ) ;
 511                                 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
 512                                 $z = '' ;
 513                                 $l = array_pop ( $ltd ) ;
 514                                 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
 515                                 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 516                                 array_pop ( $ltr ) ;
 517                                 $t[$k] = $z ;
 518                                 array_push ( $tr , false ) ;
 519                                 array_push ( $td , false ) ;
 520                                 array_push ( $ltd , '' ) ;
 521                                 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
 522                         }
 523                         else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
 524                                 if ( '|+' == substr ( $x , 0 , 2 ) ) {
 525                                         $fc = '+' ;
 526                                         $x = substr ( $x , 1 ) ;
 527                                 }
 528                                 $after = substr ( $x , 1 ) ;
 529                                 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
 530                                 $after = explode ( '||' , $after ) ;
 531                                 $t[$k] = '' ;
 532                                 foreach ( $after AS $theline )
 533                                 {
 534                                         $z = '' ;
 535                                         if ( $fc != '+' )
 536                                         {
 537                                                 $tra = array_pop ( $ltr ) ;
 538                                                 if ( !array_pop ( $tr ) ) $z = '<tr '.$tra.">\n" ;
 539                                                 array_push ( $tr , true ) ;
 540                                                 array_push ( $ltr , '' ) ;
 541                                         }
 542
 543                                         $l = array_pop ( $ltd ) ;
 544                                         if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
 545                                         if ( $fc == '|' ) $l = 'td' ;
 546                                         else if ( $fc == '!' ) $l = 'th' ;
 547                                         else if ( $fc == '+' ) $l = 'caption' ;
 548                                         else $l = '' ;
 549                                         array_push ( $ltd , $l ) ;
 550                                         $y = explode ( '|' , $theline , 2 ) ;
 551                                         if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
 552                                         else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
 553                                         $t[$k] .= $y ;
 554                                         array_push ( $td , true ) ;
 555                                 }
 556                         }
 557                 }
 558
 559                 # Closing open td, tr && table
 560                 while ( count ( $td ) > 0 )
 561                 {
 562                         if ( array_pop ( $td ) ) $t[] = '</td>' ;
 563                         if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
 564                         $t[] = '</table>' ;
 565                 }
 566
 567                 $t = implode ( "\n" , $t ) ;
 568                 #               $t = $this->removeHTMLtags( $t );
 569                 wfProfileOut( $fname );
 570                 return $t ;
 571         }
 572
 573         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 574         global $wgLang;
 575
 576                 $fname = 'Parser::internalParse';
 577                 wfProfileIn( $fname );
 578
 579                 $text = $this->removeHTMLtags( $text );
 580                 $text = $this->replaceVariables( $text, $args );
 581
 582                 $text = $wgLang->convert($text);
 583
 584                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 585
 586                 $text = $this->doHeadings( $text );
 587                 if($this->mOptions->getUseDynamicDates()) {
 588                         global $wgDateFormatter;
 589                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 590                 }
 591                 $text = $this->doAllQuotes( $text );
 592                 $text = $this->replaceExternalLinks( $text );
 593                 $text = $this->doMagicLinks( $text );
 594                 $text = $this->replaceInternalLinks ( $text );
 595                 $text = $this->replaceInternalLinks ( $text );
 596
 597                 $text = $this->unstrip( $text, $this->mStripState );
 598                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 599
 600                 $text = $this->doTableStuff( $text );
 601                 $text = $this->formatHeadings( $text, $isMain );
 602                 $sk =& $this->mOptions->getSkin();
 603                 $text = $sk->transformContent( $text );
 604
 605                 wfProfileOut( $fname );
 606                 return $text;
 607         }
 608
 609         /* private */ function &doMagicLinks( &$text ) {
 610                 global $wgUseGeoMode;
 611                 $text = $this->magicISBN( $text );
 612                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 613                         $text = $this->magicGEO( $text );
 614                 }
 615                 $text = $this->magicRFC( $text );
 616                 return $text;
 617         }
 618
 619         # Parse ^^ tokens and return html
 620         /* private */ function doExponent ( $text ) {
 621                 $fname = 'Parser::doExponent';
 622                 wfProfileIn( $fname);
 623                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 624                 wfProfileOut( $fname);
 625                 return $text;
 626         }
 627
 628         # Parse headers and return html
 629         /* private */ function doHeadings( $text ) {
 630                 $fname = 'Parser::doHeadings';
 631                 wfProfileIn( $fname );
 632                 for ( $i = 6; $i >= 1; --$i ) {
 633                         $h = substr( '======', 0, $i );
 634                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 635                           "<h{$i}>\\1</h{$i}>\\2", $text );
 636                 }
 637                 wfProfileOut( $fname );
 638                 return $text;
 639         }
 640
 641         /* private */ function doAllQuotes( $text ) {
 642                 $fname = 'Parser::doAllQuotes';
 643                 wfProfileIn( $fname );
 644                 $outtext = '';
 645                 $lines = explode( "\n", $text );
 646                 foreach ( $lines as $line ) {
 647                         $outtext .= $this->doQuotes ( $line ) . "\n";
 648                 }
 649                 $outtext = substr($outtext, 0,-1);
 650                 wfProfileOut( $fname );
 651                 return $outtext;
 652         }
 653
 654         /* private */ function doQuotes( $text ) {
 655                 $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
 656                 if (count ($arr) == 1)
 657                         return $text;
 658                 else
 659                 {
 660                         # First, do some preliminary work. This may shift some apostrophes from
 661                         # being mark-up to being text. It also counts the number of occurrences
 662                         # of bold and italics mark-ups.
 663                         $i = 0;
 664                         $numbold = 0;
 665                         $numitalics = 0;
 666                         foreach ($arr as $r)
 667                         {
 668                                 if (($i % 2) == 1)
 669                                 {
 670                                         # If there are ever four apostrophes, assume the first is supposed to
 671                                         # be text, and the remaining three constitute mark-up for bold text.
 672                                         if (strlen ($arr[$i]) == 4)
 673                                         {
 674                                                 $arr[$i-1] .= "'";
 675                                                 $arr[$i] = "'''";
 676                                         }
 677                                         # If there are more than 5 apostrophes in a row, assume they're all
 678                                         # text except for the last 5.
 679                                         else if (strlen ($arr[$i]) > 5)
 680                                         {
 681                                                 $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
 682                                                 $arr[$i] = "'''''";
 683                                         }
 684                                         # Count the number of occurrences of bold and italics mark-ups.
 685                                         # We are not counting sequences of five apostrophes.
 686                                         if (strlen ($arr[$i]) == 2) $numitalics++;  else
 687                                         if (strlen ($arr[$i]) == 3) $numbold++;     else
 688                                         if (strlen ($arr[$i]) == 5) { $numitalics++; $numbold++; }
 689                                 }
 690                                 $i++;
 691                         }
 692
 693                         # If there is an odd number of both bold and italics, it is likely
 694                         # that one of the bold ones was meant to be an apostrophe followed
 695                         # by italics. Which one we cannot know for certain, but it is more
 696                         # likely to be one that has a single-letter word before it.
 697                         if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
 698                         {
 699                                 $i = 0;
 700                                 $firstsingleletterword = -1;
 701                                 $firstmultiletterword = -1;
 702                                 $firstspace = -1;
 703                                 foreach ($arr as $r)
 704                                 {
 705                                         if (($i % 2 == 1) and (strlen ($r) == 3))
 706                                         {
 707                                                 $x1 = substr ($arr[$i-1], -1);
 708                                                 $x2 = substr ($arr[$i-1], -2, 1);
 709                                                 if ($x1 == ' ') {
 710                                                         if ($firstspace == -1) $firstspace = $i;
 711                                                 } else if ($x2 == ' ') {
 712                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 713                                                 } else {
 714                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 715                                                 }
 716                                         }
 717                                         $i++;
 718                                 }
 719
 720                                 # If there is a single-letter word, use it!
 721                                 if ($firstsingleletterword > -1)
 722                                 {
 723                                         $arr [ $firstsingleletterword ] = "''";
 724                                         $arr [ $firstsingleletterword-1 ] .= "'";
 725                                 }
 726                                 # If not, but there's a multi-letter word, use that one.
 727                                 else if ($firstmultiletterword > -1)
 728                                 {
 729                                         $arr [ $firstmultiletterword ] = "''";
 730                                         $arr [ $firstmultiletterword-1 ] .= "'";
 731                                 }
 732                                 # ... otherwise use the first one that has neither.
 733                                 # (notice that it is possible for all three to be -1 if, for example,
 734                                 # there is only one pentuple-apostrophe in the line)
 735                                 else if ($firstspace > -1)
 736                                 {
 737                                         $arr [ $firstspace ] = "''";
 738                                         $arr [ $firstspace-1 ] .= "'";
 739                                 }
 740                         }
 741
 742                         # Now let's actually convert our apostrophic mush to HTML!
 743                         $output = '';
 744                         $buffer = '';
 745                         $state = '';
 746                         $i = 0;
 747                         foreach ($arr as $r)
 748                         {
 749                                 if (($i % 2) == 0)
 750                                 {
 751                                         if ($state == 'both')
 752                                                 $buffer .= $r;
 753                                         else
 754                                                 $output .= $r;
 755                                 }
 756                                 else
 757                                 {
 758                                         if (strlen ($r) == 2)
 759                                         {
 760                                                 if ($state == 'i')
 761                                                 { $output .= '</i>'; $state = ''; }
 762                                                 else if ($state == 'bi')
 763                                                 { $output .= '</i>'; $state = 'b'; }
 764                                                 else if ($state == 'ib')
 765                                                 { $output .= '</b></i><b>'; $state = 'b'; }
 766                                                 else if ($state == 'both')
 767                                                 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
 768                                                 else # $state can be 'b' or ''
 769                                                 { $output .= '<i>'; $state .= 'i'; }
 770                                         }
 771                                         else if (strlen ($r) == 3)
 772                                         {
 773                                                 if ($state == 'b')
 774                                                 { $output .= '</b>'; $state = ''; }
 775                                                 else if ($state == 'bi')
 776                                                 { $output .= '</i></b><i>'; $state = 'i'; }
 777                                                 else if ($state == 'ib')
 778                                                 { $output .= '</b>'; $state = 'i'; }
 779                                                 else if ($state == 'both')
 780                                                 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
 781                                                 else # $state can be 'i' or ''
 782                                                 { $output .= '<b>'; $state .= 'b'; }
 783                                         }
 784                                         else if (strlen ($r) == 5)
 785                                         {
 786                                                 if ($state == 'b')
 787                                                 { $output .= '</b><i>'; $state = 'i'; }
 788                                                 else if ($state == 'i')
 789                                                 { $output .= '</i><b>'; $state = 'b'; }
 790                                                 else if ($state == 'bi')
 791                                                 { $output .= '</i></b>'; $state = ''; }
 792                                                 else if ($state == 'ib')
 793                                                 { $output .= '</b></i>'; $state = ''; }
 794                                                 else if ($state == 'both')
 795                                                 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
 796                                                 else # ($state == '')
 797                                                 { $buffer = ''; $state = 'both'; }
 798                                         }
 799                                 }
 800                                 $i++;
 801                         }
 802                         # Now close all remaining tags.  Notice that the order is important.
 803                         if ($state == 'b' || $state == 'ib')
 804                                 $output .= '</b>';
 805                         if ($state == 'i' || $state == 'bi' || $state == 'ib')
 806                                 $output .= '</i>';
 807                         if ($state == 'bi')
 808                                 $output .= '</b>';
 809                         if ($state == 'both')
 810                                 $output .= '<b><i>'.$buffer.'</i></b>';
 811                         return $output;
 812                 }
 813         }
 814
 815         # Note: we have to do external links before the internal ones,
 816         # and otherwise take great care in the order of things here, so
 817         # that we don't end up interpreting some URLs twice.
 818
 819         /* private */ function replaceExternalLinks( $text ) {
 820                 $fname = 'Parser::replaceExternalLinks';
 821                 wfProfileIn( $fname );
 822
 823                 $sk =& $this->mOptions->getSkin();
 824                 $linktrail = wfMsg('linktrail');
 825                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 826
 827                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 828
 829                 $i = 0;
 830                 while ( $i<count( $bits ) ) {
 831                         $url = $bits[$i++];
 832                         $protocol = $bits[$i++];
 833                         $text = $bits[$i++];
 834                         $trail = $bits[$i++];
 835
 836                         # If the link text is an image URL, replace it with an <img> tag
 837                         # This happened by accident in the original parser, but some people used it extensively
 838                         $img = $this->maybeMakeImageLink( $text );
 839                         if ( $img !== false ) {
 840                                 $text = $img;
 841                         }
 842
 843                         $dtrail = '';
 844
 845                         # No link text, e.g. [http://domain.tld/some.link]
 846                         if ( $text == '' ) {
 847                                 # Autonumber if allowed
 848                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 849                                         $text = '[' . ++$this->mAutonumber . ']';
 850                                 } else {
 851                                         # Otherwise just use the URL
 852                                         $text = htmlspecialchars( $url );
 853                                 }
 854                         } else {
 855                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 856                                 # Check for trail
 857                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 858                                         $dtrail = $m2[1];
 859                                         $trail = $m2[2];
 860                                 }
 861                         }
 862
 863                         $encUrl = htmlspecialchars( $url );
 864                         # Bit in parentheses showing the URL for the printable version
 865                         if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
 866                                 $paren = '';
 867                         } else {
 868                                 # Expand the URL for printable version
 869                                 if ( ! $sk->suppressUrlExpansion() ) {
 870                                         $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
 871                                 } else {
 872                                         $paren = '';
 873                                 }
 874                         }
 875
 876                         # Process the trail (i.e. everything after this link up until start of the next link),
 877                         # replacing any non-bracketed links
 878                         $trail = $this->replaceFreeExternalLinks( $trail );
 879
 880                         $la = $sk->getExternalLinkAttributes( $url, $text );
 881
 882                         # Use the encoded URL
 883                         # This means that users can paste URLs directly into the text
 884                         # Funny characters like &ouml; aren't valid in URLs anyway
 885                         # This was changed in August 2004
 886                         $s .= "<a href=\"{$url}\" {$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
 887                 }
 888
 889                 wfProfileOut( $fname );
 890                 return $s;
 891         }
 892
 893         # Replace anything that looks like a URL with a link
 894         function replaceFreeExternalLinks( $text ) {
 895                 $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 896                 $s = array_shift( $bits );
 897                 $i = 0;
 898
 899                 $sk =& $this->mOptions->getSkin();
 900
 901                 while ( $i < count( $bits ) ){
 902                         $protocol = $bits[$i++];
 903                         $remainder = $bits[$i++];
 904
 905                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
 906                                 # Found some characters after the protocol that look promising
 907                                 $url = $protocol . $m[1];
 908                                 $trail = $m[2];
 909
 910                                 # Move trailing punctuation to $trail
 911                                 $sep = ',;\.:!?';
 912                                 # If there is no left bracket, then consider right brackets fair game too
 913                                 if ( strpos( $url, '(' ) === false ) {
 914                                         $sep .= ')';
 915                                 }
 916
 917                                 $numSepChars = strspn( strrev( $url ), $sep );
 918                                 if ( $numSepChars ) {
 919                                         $trail = substr( $url, -$numSepChars ) . $trail;
 920                                         $url = substr( $url, 0, -$numSepChars );
 921                                 }
 922
 923                                 # Replace &amp; from obsolete syntax with &
 924                                 $url = str_replace( '&amp;', '&', $url );
 925
 926                                 # Is this an external image?
 927                                 $text = $this->maybeMakeImageLink( $url );
 928                                 if ( $text === false ) {
 929                                         # Not an image, make a link
 930                                         $text = $sk->makeExternalLink( $url, $url );
 931                                 }
 932                                 $s .= $text . $trail;
 933                         } else {
 934                                 $s .= $protocol . $remainder;
 935                         }
 936                 }
 937                 return $s;
 938         }
 939
 940         # make an image if it's allowed
 941         function maybeMakeImageLink( $url ) {
 942                 $sk =& $this->mOptions->getSkin();
 943                 $text = false;
 944                 if ( $this->mOptions->getAllowExternalImages() ) {
 945                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
 946                                 # Image found
 947                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
 948                         }
 949                 }
 950                 return $text;
 951         }
 952
 953         # The wikilinks [[ ]] are procedeed here.
 954         /* private */ function replaceInternalLinks( $s ) {
 955                 global $wgLang, $wgLinkCache;
 956                 global $wgNamespacesWithSubpages;
 957                 static $fname = 'Parser::replaceInternalLinks' ;
 958                 wfProfileIn( $fname );
 959
 960                 wfProfileIn( $fname.'-setup' );
 961                 static $tc = FALSE;
 962                 # the % is needed to support urlencoded titles as well
 963                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
 964                 $sk =& $this->mOptions->getSkin();
 965
 966                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
 967
 968                 $a = explode( '[[', ' ' . $s );
 969                 $s = array_shift( $a );
 970                 $s = substr( $s, 1 );
 971
 972                 # Match a link having the form [[namespace:link|alternate]]trail
 973                 static $e1 = FALSE;
 974                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
 975                 # Match the end of a line for a word that's not followed by whitespace,
 976                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
 977                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
 978
 979                 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
 980                 # Special and Media are pseudo-namespaces; no pages actually exist in them
 981
 982                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
 983
 984                 if ( $useLinkPrefixExtension ) {
 985                         if ( preg_match( $e2, $s, $m ) ) {
 986                                 $first_prefix = $m[2];
 987                                 $s = $m[1];
 988                         } else {
 989                                 $first_prefix = false;
 990                         }
 991                 } else {
 992                         $prefix = '';
 993                 }
 994
 995                 wfProfileOut( $fname.'-setup' );
 996
 997                 # start procedeeding each line
 998                 foreach ( $a as $line ) {
 999                         wfProfileIn( $fname.'-prefixhandling' );
1000                         if ( $useLinkPrefixExtension ) {
1001                                 if ( preg_match( $e2, $s, $m ) ) {
1002                                         $prefix = $m[2];
1003                                         $s = $m[1];
1004                                 } else {
1005                                         $prefix='';
1006                                 }
1007                                 # first link
1008                                 if($first_prefix) {
1009                                         $prefix = $first_prefix;
1010                                         $first_prefix = false;
1011                                 }
1012                         }
1013                         wfProfileOut( $fname.'-prefixhandling' );
1014
1015                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1016                                 $text = $m[2];
1017                                 # fix up urlencoded title texts
1018                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1019                                 $trail = $m[3];
1020                         } else { # Invalid form; output directly
1021                                 $s .= $prefix . '[[' . $line ;
1022                                 continue;
1023                         }
1024
1025                         # Valid link forms:
1026                         # Foobar -- normal
1027                         # :Foobar -- override special treatment of prefix (images, language links)
1028                         # /Foobar -- convert to CurrentPage/Foobar
1029                         # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1030
1031                         # Look at the first character
1032                         $c = substr($m[1],0,1);
1033                         $noforce = ($c != ':');
1034
1035                         # subpage
1036                         if( $c == '/' ) {
1037                                 # / at end means we don't want the slash to be shown
1038                                 if(substr($m[1],-1,1)=='/') {
1039                                         $m[1]=substr($m[1],1,strlen($m[1])-2);
1040                                         $noslash=$m[1];
1041                                 } else {
1042                                         $noslash=substr($m[1],1);
1043                                 }
1044
1045                                 # Some namespaces don't allow subpages
1046                                 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) {
1047                                         # subpages allowed here
1048                                         $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1049                                         if( '' == $text ) {
1050                                                 $text= $m[1];
1051                                         } # this might be changed for ugliness reasons
1052                                 } else {
1053                                         # no subpage allowed, use standard link
1054                                         $link = $noslash;
1055                                 }
1056
1057                         } elseif( $noforce ) { # no subpage
1058                                 $link = $m[1];
1059                         } else {
1060                                 # We don't want to keep the first character
1061                                 $link = substr( $m[1], 1 );
1062                         }
1063
1064                         $wasblank = ( '' == $text );
1065                         if( $wasblank ) $text = $link;
1066
1067                         $nt = Title::newFromText( $link );
1068                         if( !$nt ) {
1069                                 $s .= $prefix . '[[' . $line;
1070                                 continue;
1071                         }
1072
1073                         $ns = $nt->getNamespace();
1074                         $iw = $nt->getInterWiki();
1075
1076                         # Link not escaped by : , create the various objects
1077                         if( $noforce ) {
1078
1079                                 # Interwikis
1080                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1081                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1082                                         $tmp = $prefix . $trail ;
1083                                         $s .= (trim($tmp) == '')? '': $tmp;
1084                                         continue;
1085                                 }
1086
1087                                 if ( $ns == NS_IMAGE ) {
1088                                         $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1089                                         $wgLinkCache->addImageLinkObj( $nt );
1090                                         continue;
1091                                 }
1092
1093                                 if ( $ns == NS_CATEGORY ) {
1094                                         $t = $nt->getText() ;
1095                                         $nnt = Title::newFromText ( Namespace::getCanonicalName(NS_CATEGORY).':'.$t ) ;
1096
1097                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1098                                         $pPLC=$sk->postParseLinkColour();
1099                                         $sk->postParseLinkColour( false );
1100                                         $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1101                                         $sk->postParseLinkColour( $pPLC );
1102                                         $wgLinkCache->resume();
1103
1104                                         if ( $wasblank ) {
1105                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1106                                                         $sortkey = $this->mTitle->getText();
1107                                                 } else {
1108                                                         $sortkey = $this->mTitle->getPrefixedText();
1109                                                 }
1110                                         } else {
1111                                                 $sortkey = $text;
1112                                         }
1113                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1114                                         $this->mOutput->mCategoryLinks[] = $t ;
1115                                         $s .= $prefix . $trail ;
1116                                         continue;
1117                                 }
1118                         }
1119
1120                         if( ( $nt->getPrefixedText() === $this->mTitle->getPrefixedText() ) &&
1121                             ( strpos( $link, '#' ) === FALSE ) ) {
1122                                 # Self-links are handled specially; generally de-link and change to bold.
1123                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1124                                 continue;
1125                         }
1126
1127                         if( $ns == NS_MEDIA ) {
1128                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1129                                 $wgLinkCache->addImageLinkObj( $nt );
1130                                 continue;
1131                         } elseif( $ns == NS_SPECIAL ) {
1132                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1133                                 continue;
1134                         }
1135                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1136                 }
1137                 wfProfileOut( $fname );
1138                 return $s;
1139         }
1140
1141         # Some functions here used by doBlockLevels()
1142         #
1143         /* private */ function closeParagraph() {
1144                 $result = '';
1145                 if ( '' != $this->mLastSection ) {
1146                         $result = '</' . $this->mLastSection  . ">\n";
1147                 }
1148                 $this->mInPre = false;
1149                 $this->mLastSection = '';
1150                 return $result;
1151         }
1152         # getCommon() returns the length of the longest common substring
1153         # of both arguments, starting at the beginning of both.
1154         #
1155         /* private */ function getCommon( $st1, $st2 ) {
1156                 $fl = strlen( $st1 );
1157                 $shorter = strlen( $st2 );
1158                 if ( $fl < $shorter ) { $shorter = $fl; }
1159
1160                 for ( $i = 0; $i < $shorter; ++$i ) {
1161                         if ( $st1{$i} != $st2{$i} ) { break; }
1162                 }
1163                 return $i;
1164         }
1165         # These next three functions open, continue, and close the list
1166         # element appropriate to the prefix character passed into them.
1167         #
1168         /* private */ function openList( $char ) {
1169                 $result = $this->closeParagraph();
1170
1171                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1172                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1173                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1174                 else if ( ';' == $char ) {
1175                         $result .= '<dl><dt>';
1176                         $this->mDTopen = true;
1177                 }
1178                 else { $result = '<!-- ERR 1 -->'; }
1179
1180                 return $result;
1181         }
1182
1183         /* private */ function nextItem( $char ) {
1184                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1185                 else if ( ':' == $char || ';' == $char ) {
1186                         $close = '</dd>';
1187                         if ( $this->mDTopen ) { $close = '</dt>'; }
1188                         if ( ';' == $char ) {
1189                                 $this->mDTopen = true;
1190                                 return $close . '<dt>';
1191                         } else {
1192                                 $this->mDTopen = false;
1193                                 return $close . '<dd>';
1194                         }
1195                 }
1196                 return '<!-- ERR 2 -->';
1197         }
1198
1199         /* private */ function closeList( $char ) {
1200                 if ( '*' == $char ) { $text = '</li></ul>'; }
1201                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1202                 else if ( ':' == $char ) {
1203                         if ( $this->mDTopen ) {
1204                                 $this->mDTopen = false;
1205                                 $text = '</dt></dl>';
1206                         } else {
1207                                 $text = '</dd></dl>';
1208                         }
1209                 }
1210                 else {  return '<!-- ERR 3 -->'; }
1211                 return $text."\n";
1212         }
1213
1214         /* private */ function doBlockLevels( $text, $linestart ) {
1215                 $fname = 'Parser::doBlockLevels';
1216                 wfProfileIn( $fname );
1217
1218                 # Parsing through the text line by line.  The main thing
1219                 # happening here is handling of block-level elements p, pre,
1220                 # and making lists from lines starting with * # : etc.
1221                 #
1222                 $textLines = explode( "\n", $text );
1223
1224                 $lastPrefix = $output = $lastLine = '';
1225                 $this->mDTopen = $inBlockElem = false;
1226                 $prefixLength = 0;
1227                 $paragraphStack = false;
1228
1229                 if ( !$linestart ) {
1230                         $output .= array_shift( $textLines );
1231                 }
1232                 foreach ( $textLines as $oLine ) {
1233                         $lastPrefixLength = strlen( $lastPrefix );
1234                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1235                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1236                         if ( !$this->mInPre ) {
1237                                 # Multiple prefixes may abut each other for nested lists.
1238                                 $prefixLength = strspn( $oLine, '*#:;' );
1239                                 $pref = substr( $oLine, 0, $prefixLength );
1240
1241                                 # eh?
1242                                 $pref2 = str_replace( ';', ':', $pref );
1243                                 $t = substr( $oLine, $prefixLength );
1244                                 $this->mInPre = !empty($preOpenMatch);
1245                         } else {
1246                                 # Don't interpret any other prefixes in preformatted text
1247                                 $prefixLength = 0;
1248                                 $pref = $pref2 = '';
1249                                 $t = $oLine;
1250                         }
1251
1252                         # List generation
1253                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1254                                 # Same as the last item, so no need to deal with nesting or opening stuff
1255                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1256                                 $paragraphStack = false;
1257
1258                                 if ( substr( $pref, -1 ) == ';') {
1259                                         # The one nasty exception: definition lists work like this:
1260                                         # ; title : definition text
1261                                         # So we check for : in the remainder text to split up the
1262                                         # title and definition, without b0rking links.
1263                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1264                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1265                                                 $term = $match[1];
1266                                                 $output .= $term . $this->nextItem( ':' );
1267                                                 $t = $match[2];
1268                                         }
1269                                 }
1270                         } elseif( $prefixLength || $lastPrefixLength ) {
1271                                 # Either open or close a level...
1272                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1273                                 $paragraphStack = false;
1274
1275                                 while( $commonPrefixLength < $lastPrefixLength ) {
1276                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1277                                         --$lastPrefixLength;
1278                                 }
1279                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1280                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1281                                 }
1282                                 while ( $prefixLength > $commonPrefixLength ) {
1283                                         $char = substr( $pref, $commonPrefixLength, 1 );
1284                                         $output .= $this->openList( $char );
1285
1286                                         if ( ';' == $char ) {
1287                                                 # FIXME: This is dupe of code above
1288                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1289                                                         $term = $match[1];
1290                                                         $output .= $term . $this->nextItem( ':' );
1291                                                         $t = $match[2];
1292                                                 }
1293                                         }
1294                                         ++$commonPrefixLength;
1295                                 }
1296                                 $lastPrefix = $pref2;
1297                         }
1298                         if( 0 == $prefixLength ) {
1299                                 # No prefix (not in list)--go to paragraph mode
1300                                 $uniq_prefix = UNIQ_PREFIX;
1301                                 // XXX: use a stack for nestable elements like span, table and div
1302                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1303                                 $closematch = preg_match(
1304                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1305                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1306                                 if ( $openmatch or $closematch ) {
1307                                         $paragraphStack = false;
1308                                         $output .= $this->closeParagraph();
1309                                         if($preOpenMatch and !$preCloseMatch) {
1310                                                 $this->mInPre = true;
1311                                         }
1312                                         if ( $closematch ) {
1313                                                 $inBlockElem = false;
1314                                         } else {
1315                                                 $inBlockElem = true;
1316                                         }
1317                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1318                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1319                                                 // pre
1320                                                 if ($this->mLastSection != 'pre') {
1321                                                         $paragraphStack = false;
1322                                                         $output .= $this->closeParagraph().'<pre>';
1323                                                         $this->mLastSection = 'pre';
1324                                                 }
1325                                                 $t = substr( $t, 1 );
1326                                         } else {
1327                                                 // paragraph
1328                                                 if ( '' == trim($t) ) {
1329                                                         if ( $paragraphStack ) {
1330                                                                 $output .= $paragraphStack.'<br />';
1331                                                                 $paragraphStack = false;
1332                                                                 $this->mLastSection = 'p';
1333                                                         } else {
1334                                                                 if ($this->mLastSection != 'p' ) {
1335                                                                         $output .= $this->closeParagraph();
1336                                                                         $this->mLastSection = '';
1337                                                                         $paragraphStack = '<p>';
1338                                                                 } else {
1339                                                                         $paragraphStack = '</p><p>';
1340                                                                 }
1341                                                         }
1342                                                 } else {
1343                                                         if ( $paragraphStack ) {
1344                                                                 $output .= $paragraphStack;
1345                                                                 $paragraphStack = false;
1346                                                                 $this->mLastSection = 'p';
1347                                                         } else if ($this->mLastSection != 'p') {
1348                                                                 $output .= $this->closeParagraph().'<p>';
1349                                                                 $this->mLastSection = 'p';
1350                                                         }
1351                                                 }
1352                                         }
1353                                 }
1354                         }
1355                         if ($paragraphStack === false) {
1356                                 $output .= $t."\n";
1357                         }
1358                 }
1359                 while ( $prefixLength ) {
1360                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1361                         --$prefixLength;
1362                 }
1363                 if ( '' != $this->mLastSection ) {
1364                         $output .= '</' . $this->mLastSection . '>';
1365                         $this->mLastSection = '';
1366                 }
1367
1368                 wfProfileOut( $fname );
1369                 return $output;
1370         }
1371
1372         # Return value of a magic variable (like PAGENAME)
1373         function getVariableValue( $index ) {
1374                 global $wgLang, $wgSitename, $wgServer;
1375
1376                 switch ( $index ) {
1377                         case MAG_CURRENTMONTH:
1378                                 return $wgLang->formatNum( date( 'm' ) );
1379                         case MAG_CURRENTMONTHNAME:
1380                                 return $wgLang->getMonthName( date('n') );
1381                         case MAG_CURRENTMONTHNAMEGEN:
1382                                 return $wgLang->getMonthNameGen( date('n') );
1383                         case MAG_CURRENTDAY:
1384                                 return $wgLang->formatNum( date('j') );
1385                         case MAG_PAGENAME:
1386                                 return $this->mTitle->getText();
1387                         case MAG_PAGENAMEE:
1388                                 return $this->mTitle->getPartialURL();
1389                         case MAG_NAMESPACE:
1390                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1391                                 return $wgLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1392                         case MAG_CURRENTDAYNAME:
1393                                 return $wgLang->getWeekdayName( date('w')+1 );
1394                         case MAG_CURRENTYEAR:
1395                                 return $wgLang->formatNum( date( 'Y' ) );
1396                         case MAG_CURRENTTIME:
1397                                 return $wgLang->time( wfTimestampNow(), false );
1398                         case MAG_NUMBEROFARTICLES:
1399                                 return $wgLang->formatNum( wfNumberOfArticles() );
1400                         case MAG_SITENAME:
1401                                 return $wgSitename;
1402                         case MAG_SERVER:
1403                                 return $wgServer;
1404                         default:
1405                                 return NULL;
1406                 }
1407         }
1408
1409         # initialise the magic variables (like CURRENTMONTHNAME)
1410         function initialiseVariables() {
1411                 global $wgVariableIDs;
1412                 $this->mVariables = array();
1413                 foreach ( $wgVariableIDs as $id ) {
1414                         $mw =& MagicWord::get( $id );
1415                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1416                 }
1417         }
1418
1419         /* private */ function replaceVariables( $text, $args = array() ) {
1420                 global $wgLang, $wgScript, $wgArticlePath;
1421
1422                 # Prevent too big inclusions
1423                 if(strlen($text)> MAX_INCLUDE_SIZE)
1424                 return $text;
1425
1426                 $fname = 'Parser::replaceVariables';
1427                 wfProfileIn( $fname );
1428
1429                 $titleChars = Title::legalChars();
1430
1431                 # This function is called recursively. To keep track of arguments we need a stack:
1432                 array_push( $this->mArgStack, $args );
1433
1434                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1435                 $GLOBALS['wgCurParser'] =& $this;
1436
1437                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_MSG ) {
1438                         # Variable substitution
1439                         $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", 'wfVariableSubstitution', $text );
1440                 }
1441
1442                 if ( $this->mOutputType == OT_HTML ) {
1443                         # Argument substitution
1444                         $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1445                 }
1446                 # Template substitution
1447                 $regex = '/{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1448                 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1449
1450                 array_pop( $this->mArgStack );
1451
1452                 wfProfileOut( $fname );
1453                 return $text;
1454         }
1455
1456         function variableSubstitution( $matches ) {
1457                 if ( !$this->mVariables ) {
1458                         $this->initialiseVariables();
1459                 }
1460                 if ( array_key_exists( $matches[1], $this->mVariables ) ) {
1461                         $text = $this->mVariables[$matches[1]];
1462                         $this->mOutput->mContainsOldMagic = true;
1463                 } else {
1464                         $text = $matches[0];
1465                 }
1466                 return $text;
1467         }
1468
1469         # Split template arguments
1470         function getTemplateArgs( $argsString ) {
1471                 if ( $argsString === '' ) {
1472                         return array();
1473                 }
1474
1475                 $args = explode( '|', substr( $argsString, 1 ) );
1476
1477                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1478                 # merged with the next arg because the '|' character between belongs
1479                 # to the link syntax and not the template parameter syntax.
1480                 $argc = count($args);
1481                 $i = 0;
1482                 for ( $i = 0; $i < $argc-1; $i++ ) {
1483                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1484                                 $args[$i] .= '|'.$args[$i+1];
1485                                 array_splice($args, $i+1, 1);
1486                                 $i--;
1487                                 $argc--;
1488                         }
1489                 }
1490
1491                 return $args;
1492         }
1493
1494         function braceSubstitution( $matches ) {
1495                 global $wgLinkCache, $wgLang;
1496                 $fname = 'Parser::braceSubstitution';
1497                 $found = false;
1498                 $nowiki = false;
1499                 $noparse = false;
1500
1501                 $title = NULL;
1502
1503                 # $part1 is the bit before the first |, and must contain only title characters
1504                 # $args is a list of arguments, starting from index 0, not including $part1
1505
1506                 $part1 = $matches[1];
1507                 # If the second subpattern matched anything, it will start with |
1508
1509                 $args = $this->getTemplateArgs($matches[2]);
1510                 $argc = count( $args );
1511
1512                 # {{{}}}
1513                 if ( strpos( $matches[0], '{{{' ) !== false ) {
1514                         $text = $matches[0];
1515                         $found = true;
1516                         $noparse = true;
1517                 }
1518
1519                 # SUBST
1520                 if ( !$found ) {
1521                         $mwSubst =& MagicWord::get( MAG_SUBST );
1522                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1523                                 if ( $this->mOutputType != OT_WIKI ) {
1524                                         # Invalid SUBST not replaced at PST time
1525                                         # Return without further processing
1526                                         $text = $matches[0];
1527                                         $found = true;
1528                                         $noparse= true;
1529                                 }
1530                         } elseif ( $this->mOutputType == OT_WIKI ) {
1531                                 # SUBST not found in PST pass, do nothing
1532                                 $text = $matches[0];
1533                                 $found = true;
1534                         }
1535                 }
1536
1537                 # MSG, MSGNW and INT
1538                 if ( !$found ) {
1539                         # Check for MSGNW:
1540                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1541                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1542                                 $nowiki = true;
1543                         } else {
1544                                 # Remove obsolete MSG:
1545                                 $mwMsg =& MagicWord::get( MAG_MSG );
1546                                 $mwMsg->matchStartAndRemove( $part1 );
1547                         }
1548
1549                         # Check if it is an internal message
1550                         $mwInt =& MagicWord::get( MAG_INT );
1551                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1552                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1553                                         $text = wfMsgReal( $part1, $args, true );
1554                                         $found = true;
1555                                 }
1556                         }
1557                 }
1558
1559                 # NS
1560                 if ( !$found ) {
1561                         # Check for NS: (namespace expansion)
1562                         $mwNs = MagicWord::get( MAG_NS );
1563                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1564                                 if ( intval( $part1 ) ) {
1565                                         $text = $wgLang->getNsText( intval( $part1 ) );
1566                                         $found = true;
1567                                 } else {
1568                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1569                                         if ( !is_null( $index ) ) {
1570                                                 $text = $wgLang->getNsText( $index );
1571                                                 $found = true;
1572                                         }
1573                                 }
1574                         }
1575                 }
1576
1577                 # LOCALURL and LOCALURLE
1578                 if ( !$found ) {
1579                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1580                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1581
1582                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1583                                 $func = 'getLocalURL';
1584                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1585                                 $func = 'escapeLocalURL';
1586                         } else {
1587                                 $func = '';
1588                         }
1589
1590                         if ( $func !== '' ) {
1591                                 $title = Title::newFromText( $part1 );
1592                                 if ( !is_null( $title ) ) {
1593                                         if ( $argc > 0 ) {
1594                                                 $text = $title->$func( $args[0] );
1595                                         } else {
1596                                                 $text = $title->$func();
1597                                         }
1598                                         $found = true;
1599                                 }
1600                         }
1601                 }
1602
1603                 # Internal variables
1604                 if ( !$this->mVariables ) {
1605                         $this->initialiseVariables();
1606                 }
1607                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1608                         $text = $this->mVariables[$part1];
1609                         $found = true;
1610                         $this->mOutput->mContainsOldMagic = true;
1611                 }
1612
1613                 # GRAMMAR
1614                 if ( !$found && $argc == 1 ) {
1615                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1616                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1617                                 $text = $wgLang->convertGrammar( $args[0], $part1 );
1618                                 $found = true;
1619                         }
1620                 }
1621
1622                 # Template table test
1623
1624                 # Did we encounter this template already? If yes, it is in the cache
1625                 # and we need to check for loops.
1626                 if ( isset( $this->mTemplates[$part1] ) ) {
1627                         # Infinite loop test
1628                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1629                                 $noparse = true;
1630                                 $found = true;
1631                         }
1632                         # set $text to cached message.
1633                         $text = $this->mTemplates[$part1];
1634                         $found = true;
1635                 }
1636
1637                 # Load from database
1638                 if ( !$found ) {
1639                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1640                         if ( !is_null( $title ) && !$title->isExternal() ) {
1641                                 # Check for excessive inclusion
1642                                 $dbk = $title->getPrefixedDBkey();
1643                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1644                                         # This should never be reached.
1645                                         $article = new Article( $title );
1646                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1647                                         if ( $articleContent !== false ) {
1648                                                 $found = true;
1649                                                 $text = $articleContent;
1650                                         }
1651                                 }
1652
1653                                 # If the title is valid but undisplayable, make a link to it
1654                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1655                                         $text = '[['.$title->getPrefixedText().']]';
1656                                         $found = true;
1657                                 }
1658
1659                                 # Template cache array insertion
1660                                 $this->mTemplates[$part1] = $text;
1661                         }
1662                 }
1663
1664                 # Recursive parsing, escaping and link table handling
1665                 # Only for HTML output
1666                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1667                         $text = wfEscapeWikiText( $text );
1668                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1669                         # Clean up argument array
1670                         $assocArgs = array();
1671                         $index = 1;
1672                         foreach( $args as $arg ) {
1673                                 $eqpos = strpos( $arg, '=' );
1674                                 if ( $eqpos === false ) {
1675                                         $assocArgs[$index++] = $arg;
1676                                 } else {
1677                                         $name = trim( substr( $arg, 0, $eqpos ) );
1678                                         $value = trim( substr( $arg, $eqpos+1 ) );
1679                                         if ( $value === false ) {
1680                                                 $value = '';
1681                                         }
1682                                         if ( $name !== false ) {
1683                                                 $assocArgs[$name] = $value;
1684                                         }
1685                                 }
1686                         }
1687
1688                         # Do not enter included links in link table
1689                         if ( !is_null( $title ) ) {
1690                                 $wgLinkCache->suspend();
1691                         }
1692
1693                         # Add a new element to the templace recursion path
1694                         $this->mTemplatePath[$part1] = 1;
1695
1696                         $text = $this->strip( $text, $this->mStripState );
1697                         $text = $this->removeHTMLtags( $text );
1698                         $text = $this->replaceVariables( $text, $assocArgs );
1699
1700                         # Resume the link cache and register the inclusion as a link
1701                         if ( !is_null( $title ) ) {
1702                                 $wgLinkCache->resume();
1703                                 $wgLinkCache->addLinkObj( $title );
1704                         }
1705                 }
1706
1707                 # Empties the template path
1708                 $this->mTemplatePath = array();
1709
1710                 if ( !$found ) {
1711                         return $matches[0];
1712                 } else {
1713                         return $text;
1714                 }
1715         }
1716
1717         # Triple brace replacement -- used for template arguments
1718         function argSubstitution( $matches ) {
1719                 $arg = trim( $matches[1] );
1720                 $text = $matches[0];
1721                 $inputArgs = end( $this->mArgStack );
1722
1723                 if ( array_key_exists( $arg, $inputArgs ) ) {
1724                         $text = $this->strip( $inputArgs[$arg], $this->mStripState );
1725                         $text = $this->removeHTMLtags( $text );
1726                         $text = $this->replaceVariables( $text, array() );
1727                 }
1728
1729                 return $text;
1730         }
1731
1732         # Returns true if the function is allowed to include this entity
1733         function incrementIncludeCount( $dbk ) {
1734                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1735                         $this->mIncludeCount[$dbk] = 0;
1736                 }
1737                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1738                         return true;
1739                 } else {
1740                         return false;
1741                 }
1742         }
1743
1744
1745         # Cleans up HTML, removes dangerous tags and attributes
1746         /* private */ function removeHTMLtags( $text ) {
1747                 global $wgUseTidy, $wgUserHtml;
1748                 $fname = 'Parser::removeHTMLtags';
1749                 wfProfileIn( $fname );
1750
1751                 if( $wgUserHtml ) {
1752                         $htmlpairs = array( # Tags that must be closed
1753                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1754                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1755                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
1756                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1757                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
1758                         );
1759                         $htmlsingle = array(
1760                                 'br', 'hr', 'li', 'dt', 'dd'
1761                         );
1762                         $htmlnest = array( # Tags that can be nested--??
1763                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1764                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
1765                         );
1766                         $tabletags = array( # Can only appear inside table
1767                                 'td', 'th', 'tr'
1768                         );
1769                 } else {
1770                         $htmlpairs = array();
1771                         $htmlsingle = array();
1772                         $htmlnest = array();
1773                         $tabletags = array();
1774                 }
1775
1776                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1777                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1778
1779                 $htmlattrs = $this->getHTMLattrs () ;
1780
1781                 # Remove HTML comments
1782                 $text = $this->removeHTMLcomments( $text );
1783
1784                 $bits = explode( '<', $text );
1785                 $text = array_shift( $bits );
1786                 if(!$wgUseTidy) {
1787                         $tagstack = array(); $tablestack = array();
1788                         foreach ( $bits as $x ) {
1789                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1790                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1791                                 $x, $regs );
1792                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1793                                 error_reporting( $prev );
1794
1795                                 $badtag = 0 ;
1796                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1797                                         # Check our stack
1798                                         if ( $slash ) {
1799                                                 # Closing a tag...
1800                                                 if ( ! in_array( $t, $htmlsingle ) &&
1801                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
1802                                                         @array_push( $tagstack, $ot );
1803                                                         $badtag = 1;
1804                                                 } else {
1805                                                         if ( $t == 'table' ) {
1806                                                                 $tagstack = array_pop( $tablestack );
1807                                                         }
1808                                                         $newparams = '';
1809                                                 }
1810                                         } else {
1811                                                 # Keep track for later
1812                                                 if ( in_array( $t, $tabletags ) &&
1813                                                 ! in_array( 'table', $tagstack ) ) {
1814                                                         $badtag = 1;
1815                                                 } else if ( in_array( $t, $tagstack ) &&
1816                                                 ! in_array ( $t , $htmlnest ) ) {
1817                                                         $badtag = 1 ;
1818                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1819                                                         if ( $t == 'table' ) {
1820                                                                 array_push( $tablestack, $tagstack );
1821                                                                 $tagstack = array();
1822                                                         }
1823                                                         array_push( $tagstack, $t );
1824                                                 }
1825                                                 # Strip non-approved attributes from the tag
1826                                                 $newparams = $this->fixTagAttributes($params);
1827
1828                                         }
1829                                         if ( ! $badtag ) {
1830                                                 $rest = str_replace( '>', '&gt;', $rest );
1831                                                 $text .= "<$slash$t $newparams$brace$rest";
1832                                                 continue;
1833                                         }
1834                                 }
1835                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1836                         }
1837                         # Close off any remaining tags
1838                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
1839                                 $text .= "</$t>\n";
1840                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
1841                         }
1842                 } else {
1843                         # this might be possible using tidy itself
1844                         foreach ( $bits as $x ) {
1845                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1846                                 $x, $regs );
1847                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1848                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1849                                         $newparams = $this->fixTagAttributes($params);
1850                                         $rest = str_replace( '>', '&gt;', $rest );
1851                                         $text .= "<$slash$t $newparams$brace$rest";
1852                                 } else {
1853                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1854                                 }
1855                         }
1856                 }
1857                 wfProfileOut( $fname );
1858                 return $text;
1859         }
1860
1861         /**
1862          * Remove '<!--', '-->', and everything between.
1863          * To avoid leaving blank lines, when a comment is both preceded
1864          * and followed by a newline (ignoring spaces), trim leading and
1865          * trailing spaces and one of the newlines.
1866          *
1867          * @access private
1868          */
1869         function removeHTMLcomments( $text ) {
1870                 $fname='Parser::removeHTMLcomments';
1871                 wfProfileIn( $fname );
1872                 while (($start = strpos($text, '<!--')) !== false) {
1873                         $end = strpos($text, '-->', $start + 4);
1874                         if ($end === false) {
1875                                 # Unterminated comment; bail out
1876                                 break;
1877                         }
1878
1879                         $end += 3;
1880
1881                         # Trim space and newline if the comment is both
1882                         # preceded and followed by a newline
1883                         $spaceStart = max($start - 1, 0);
1884                         $spaceLen = $end - $spaceStart;
1885                         while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
1886                                 $spaceStart--;
1887                                 $spaceLen++;
1888                         }
1889                         while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
1890                                 $spaceLen++;
1891                         if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
1892                                 # Remove the comment, leading and trailing
1893                                 # spaces, and leave only one newline.
1894                                 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
1895                         }
1896                         else {
1897                                 # Remove just the comment.
1898                                 $text = substr_replace($text, '', $start, $end - $start);
1899                         }
1900                 }
1901                 wfProfileOut( $fname );
1902                 return $text;
1903         }
1904
1905         # This function accomplishes several tasks:
1906         # 1) Auto-number headings if that option is enabled
1907         # 2) Add an [edit] link to sections for logged in users who have enabled the option
1908         # 3) Add a Table of contents on the top for users who have enabled the option
1909         # 4) Auto-anchor headings
1910         #
1911         # It loops through all headlines, collects the necessary data, then splits up the
1912         # string and re-inserts the newly formatted headlines.
1913         /* private */ function formatHeadings( $text, $isMain=true ) {
1914                 global $wgInputEncoding, $wgMaxTocLevel, $wgLang;
1915
1916                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1917                 $doShowToc = $this->mOptions->getShowToc();
1918                 $forceTocHere = false;
1919                 if( !$this->mTitle->userCanEdit() ) {
1920                         $showEditLink = 0;
1921                         $rightClickHack = 0;
1922                 } else {
1923                         $showEditLink = $this->mOptions->getEditSection();
1924                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1925                 }
1926
1927                 # Inhibit editsection links if requested in the page
1928                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1929                 if( $esw->matchAndRemove( $text ) ) {
1930                         $showEditLink = 0;
1931                 }
1932                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1933                 # do not add TOC
1934                 $mw =& MagicWord::get( MAG_NOTOC );
1935                 if( $mw->matchAndRemove( $text ) ) {
1936                         $doShowToc = 0;
1937                 }
1938
1939                 # never add the TOC to the Main Page. This is an entry page that should not
1940                 # be more than 1-2 screens large anyway
1941                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
1942                         $doShowToc = 0;
1943                 }
1944
1945                 # Get all headlines for numbering them and adding funky stuff like [edit]
1946                 # links - this is for later, but we need the number of headlines right now
1947                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
1948
1949                 # if there are fewer than 4 headlines in the article, do not show TOC
1950                 if( $numMatches < 4 ) {
1951                         $doShowToc = 0;
1952                 }
1953
1954                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
1955                 # override above conditions and always show TOC at that place
1956                 $mw =& MagicWord::get( MAG_TOC );
1957                 if ($mw->match( $text ) ) {
1958                         $doShowToc = 1;
1959                         $forceTocHere = true;
1960                 } else {
1961                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1962                         # override above conditions and always show TOC above first header
1963                         $mw =& MagicWord::get( MAG_FORCETOC );
1964                         if ($mw->matchAndRemove( $text ) ) {
1965                                 $doShowToc = 1;
1966                         }
1967                 }
1968
1969
1970
1971                 # We need this to perform operations on the HTML
1972                 $sk =& $this->mOptions->getSkin();
1973
1974                 # headline counter
1975                 $headlineCount = 0;
1976
1977                 # Ugh .. the TOC should have neat indentation levels which can be
1978                 # passed to the skin functions. These are determined here
1979                 $toclevel = 0;
1980                 $toc = '';
1981                 $full = '';
1982                 $head = array();
1983                 $sublevelCount = array();
1984                 $level = 0;
1985                 $prevlevel = 0;
1986                 foreach( $matches[3] as $headline ) {
1987                         $numbering = '';
1988                         if( $level ) {
1989                                 $prevlevel = $level;
1990                         }
1991                         $level = $matches[1][$headlineCount];
1992                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
1993                                 # reset when we enter a new level
1994                                 $sublevelCount[$level] = 0;
1995                                 $toc .= $sk->tocIndent( $level - $prevlevel );
1996                                 $toclevel += $level - $prevlevel;
1997                         }
1998                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
1999                                 # reset when we step back a level
2000                                 $sublevelCount[$level+1]=0;
2001                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
2002                                 $toclevel -= $prevlevel - $level;
2003                         }
2004                         # count number of headlines for each level
2005                         @$sublevelCount[$level]++;
2006                         if( $doNumberHeadings || $doShowToc ) {
2007                                 $dot = 0;
2008                                 for( $i = 1; $i <= $level; $i++ ) {
2009                                         if( !empty( $sublevelCount[$i] ) ) {
2010                                                 if( $dot ) {
2011                                                         $numbering .= '.';
2012                                                 }
2013                                                 $numbering .= $wgLang->formatNum( $sublevelCount[$i] );
2014                                                 $dot = 1;
2015                                         }
2016                                 }
2017                         }
2018
2019                         # The canonized header is a version of the header text safe to use for links
2020                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2021                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2022                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2023
2024                         # Remove link placeholders by the link text.
2025                         #     <!--LINK namespace page_title link text with suffix-->
2026                         # turns into
2027                         #     link text with suffix
2028                         $canonized_headline = preg_replace( '/<!--LINK [0-9]* [^ ]* *(.*?)-->/','$1', $canonized_headline );
2029                         # strip out HTML
2030                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2031                         $tocline = trim( $canonized_headline );
2032                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2033                         $replacearray = array(
2034                                 '%3A' => ':',
2035                                 '%' => '.'
2036                         );
2037                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2038                         $refer[$headlineCount] = $canonized_headline;
2039
2040                         # count how many in assoc. array so we can track dupes in anchors
2041                         @$refers[$canonized_headline]++;
2042                         $refcount[$headlineCount]=$refers[$canonized_headline];
2043
2044                         # Prepend the number to the heading text
2045
2046                         if( $doNumberHeadings || $doShowToc ) {
2047                                 $tocline = $numbering . ' ' . $tocline;
2048
2049                                 # Don't number the heading if it is the only one (looks silly)
2050                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2051                                         # the two are different if the line contains a link
2052                                         $headline=$numbering . ' ' . $headline;
2053                                 }
2054                         }
2055
2056                         # Create the anchor for linking from the TOC to the section
2057                         $anchor = $canonized_headline;
2058                         if($refcount[$headlineCount] > 1 ) {
2059                                 $anchor .= '_' . $refcount[$headlineCount];
2060                         }
2061                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2062                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2063                         }
2064                         if( $showEditLink ) {
2065                                 if ( empty( $head[$headlineCount] ) ) {
2066                                         $head[$headlineCount] = '';
2067                                 }
2068                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
2069                         }
2070
2071                         # Add the edit section span
2072                         if( $rightClickHack ) {
2073                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
2074                         }
2075
2076                         # give headline the correct <h#> tag
2077                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2078
2079                         $headlineCount++;
2080                 }
2081
2082                 if( $doShowToc ) {
2083                         $toclines = $headlineCount;
2084                         $toc .= $sk->tocUnindent( $toclevel );
2085                         $toc = $sk->tocTable( $toc );
2086                 }
2087
2088                 # split up and insert constructed headlines
2089
2090                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2091                 $i = 0;
2092
2093                 foreach( $blocks as $block ) {
2094                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2095                                 # This is the [edit] link that appears for the top block of text when
2096                                 # section editing is enabled
2097
2098                                 # Disabled because it broke block formatting
2099                                 # For example, a bullet point in the top line
2100                                 # $full .= $sk->editSectionLink(0);
2101                         }
2102                         $full .= $block;
2103                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2104                         # Top anchor now in skin
2105                                 $full = $full.$toc;
2106                         }
2107
2108                         if( !empty( $head[$i] ) ) {
2109                                 $full .= $head[$i];
2110                         }
2111                         $i++;
2112                 }
2113                 if($forceTocHere) {
2114                         $mw =& MagicWord::get( MAG_TOC );
2115                         return $mw->replace( $toc, $full );
2116                 } else {
2117                         return $full;
2118                 }
2119         }
2120
2121         # Return an HTML link for the "ISBN 123456" text
2122         /* private */ function magicISBN( $text ) {
2123                 global $wgLang;
2124                 $fname = 'Parser::magicISBN';
2125                 wfProfileIn( $fname );
2126
2127                 $a = split( 'ISBN ', ' '.$text );
2128                 if ( count ( $a ) < 2 ) {
2129                         wfProfileOut( $fname );
2130                         return $text;
2131                 }
2132                 $text = substr( array_shift( $a ), 1);
2133                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2134
2135                 foreach ( $a as $x ) {
2136                         $isbn = $blank = '' ;
2137                         while ( ' ' == $x{0} ) {
2138                                 $blank .= ' ';
2139                                 $x = substr( $x, 1 );
2140                         }
2141                         if ( $x == '' ) { # blank isbn
2142                                 $text .= "ISBN $blank";
2143                                 continue;
2144                         }
2145                         while ( strstr( $valid, $x{0} ) != false ) {
2146                                 $isbn .= $x{0};
2147                                 $x = substr( $x, 1 );
2148                         }
2149                         $num = str_replace( '-', '', $isbn );
2150                         $num = str_replace( ' ', '', $num );
2151
2152                         if ( '' == $num ) {
2153                                 $text .= "ISBN $blank$x";
2154                         } else {
2155                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2156                                 $text .= '<a href="' .
2157                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2158                                         "\" class=\"internal\">ISBN $isbn</a>";
2159                                 $text .= $x;
2160                         }
2161                 }
2162                 wfProfileOut( $fname );
2163                 return $text;
2164         }
2165
2166         # Return an HTML link for the "GEO ..." text
2167         /* private */ function magicGEO( $text ) {
2168                 global $wgLang, $wgUseGeoMode;
2169                 $fname = 'Parser::magicGEO';
2170                 wfProfileIn( $fname );
2171
2172                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2173                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2174                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2175                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2176                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2177                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2178
2179                 $a = split( 'GEO ', ' '.$text );
2180                 if ( count ( $a ) < 2 ) {
2181                         wfProfileOut( $fname );
2182                         return $text;
2183                 }
2184                 $text = substr( array_shift( $a ), 1);
2185                 $valid = '0123456789.+-:';
2186
2187                 foreach ( $a as $x ) {
2188                         $geo = $blank = '' ;
2189                         while ( ' ' == $x{0} ) {
2190                                 $blank .= ' ';
2191                                 $x = substr( $x, 1 );
2192                         }
2193                         while ( strstr( $valid, $x{0} ) != false ) {
2194                                 $geo .= $x{0};
2195                                 $x = substr( $x, 1 );
2196                         }
2197                         $num = str_replace( '+', '', $geo );
2198                         $num = str_replace( ' ', '', $num );
2199
2200                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2201                                 $text .= "GEO $blank$x";
2202                         } else {
2203                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2204                                 $text .= '<a href="' .
2205                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2206                                         "\" class=\"internal\">GEO $geo</a>";
2207                                 $text .= $x;
2208                         }
2209                 }
2210                 wfProfileOut( $fname );
2211                 return $text;
2212         }
2213
2214         /**
2215          * Return an HTML link for the "RFC 1234" text
2216          * @access private
2217          * @param string $text text to be processed
2218          */
2219         function magicRFC( $text ) {
2220                 global $wgLang;
2221
2222                 $valid = '0123456789';
2223                 $internal = false;
2224
2225                 $a = split( 'RFC ', ' '.$text );
2226                 if ( count ( $a ) < 2 ) return $text;
2227                 $text = substr( array_shift( $a ), 1);
2228
2229                 /* Check if RFC keyword is preceed by [[.
2230                  * This test is made here cause of the array_shift above
2231                  * that prevent the test to be done in the foreach.
2232                  */
2233                 if(substr($text, -2) == '[[') { $internal = true; }
2234
2235                 foreach ( $a as $x ) {
2236                         /* token might be empty if we have RFC RFC 1234 */
2237                         if($x=='') {
2238                                 $text.='RFC ';
2239                                 continue;
2240                                 }
2241
2242                         $rfc = $blank = '' ;
2243
2244                         /** remove and save whitespaces in $blank */
2245                         while ( $x{0} == ' ' ) {
2246                                 $blank .= ' ';
2247                                 $x = substr( $x, 1 );
2248                         }
2249
2250                         /** remove and save the rfc number in $rfc */
2251                         while ( strstr( $valid, $x{0} ) != false ) {
2252                                 $rfc .= $x{0};
2253                                 $x = substr( $x, 1 );
2254                         }
2255
2256                         if ( $rfc == '') {
2257                                 /* call back stripped spaces*/
2258                                 $text .= "RFC $blank$x";
2259                         } elseif( $internal) {
2260                                 /* normal link */
2261                                 $text .= "RFC $rfc$x";
2262                         } else {
2263                                 /* build the external link*/
2264                                 $url = wfmsg( 'rfcurl' );
2265                                 $url = str_replace( '$1', $rfc, $url);
2266                                 $sk =& $this->mOptions->getSkin();
2267                                 $la = $sk->getExternalLinkAttributes( $url, 'RFC '.$rfc );
2268                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2269                         }
2270
2271                         /* Check if the next RFC keyword is preceed by [[ */
2272                         $internal = (substr($x,-2) == '[[');
2273                 }
2274                 return $text;
2275         }
2276
2277         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2278                 $this->mOptions = $options;
2279                 $this->mTitle =& $title;
2280                 $this->mOutputType = OT_WIKI;
2281
2282                 if ( $clearState ) {
2283                         $this->clearState();
2284                 }
2285
2286                 $stripState = false;
2287                 $pairs = array(
2288                         "\r\n" => "\n",
2289                         );
2290                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2291                 // now with regexes
2292                 /*
2293                 $pairs = array(
2294                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2295                         "/<br *?>/i" => "<br />",
2296                 );
2297                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2298                 */
2299                 $text = $this->strip( $text, $stripState, false );
2300                 $text = $this->pstPass2( $text, $user );
2301                 $text = $this->unstrip( $text, $stripState );
2302                 $text = $this->unstripNoWiki( $text, $stripState );
2303                 return $text;
2304         }
2305
2306         /* private */ function pstPass2( $text, &$user ) {
2307                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2308
2309                 # Variable replacement
2310                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2311                 $text = $this->replaceVariables( $text );
2312
2313                 # Signatures
2314                 #
2315                 $n = $user->getName();
2316                 $k = $user->getOption( 'nickname' );
2317                 if ( '' == $k ) { $k = $n; }
2318                 if(isset($wgLocaltimezone)) {
2319                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2320                 }
2321                 /* Note: this is an ugly timezone hack for the European wikis */
2322                 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2323                   ' (' . date( 'T' ) . ')';
2324                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2325
2326                 $text = preg_replace( '/~~~~~/', $d, $text );
2327                 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2328                 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2329
2330                 # Context links: [[|name]] and [[name (context)|]]
2331                 #
2332                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2333                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2334                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2335                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2336
2337                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2338                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2339                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2340                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2341                 $context = '';
2342                 $t = $this->mTitle->getText();
2343                 if ( preg_match( $conpat, $t, $m ) ) {
2344                         $context = $m[2];
2345                 }
2346                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2347                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2348                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2349
2350                 if ( '' == $context ) {
2351                         $text = preg_replace( $p2, '[[\\1]]', $text );
2352                 } else {
2353                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2354                 }
2355
2356                 /*
2357                 $mw =& MagicWord::get( MAG_SUBST );
2358                 $wgCurParser = $this->fork();
2359                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2360                 $this->merge( $wgCurParser );
2361                 */
2362
2363                 # Trim trailing whitespace
2364                 # MAG_END (__END__) tag allows for trailing
2365                 # whitespace to be deliberately included
2366                 $text = rtrim( $text );
2367                 $mw =& MagicWord::get( MAG_END );
2368                 $mw->matchAndRemove( $text );
2369
2370                 return $text;
2371         }
2372
2373         # Set up some variables which are usually set up in parse()
2374         # so that an external function can call some class members with confidence
2375         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2376                 $this->mTitle =& $title;
2377                 $this->mOptions = $options;
2378                 $this->mOutputType = $outputType;
2379                 if ( $clearState ) {
2380                         $this->clearState();
2381                 }
2382         }
2383
2384         function transformMsg( $text, $options ) {
2385                 global $wgTitle;
2386                 static $executing = false;
2387
2388                 # Guard against infinite recursion
2389                 if ( $executing ) {
2390                         return $text;
2391                 }
2392                 $executing = true;
2393
2394                 $this->mTitle = $wgTitle;
2395                 $this->mOptions = $options;
2396                 $this->mOutputType = OT_MSG;
2397                 $this->clearState();
2398                 $text = $this->replaceVariables( $text );
2399
2400                 $executing = false;
2401                 return $text;
2402         }
2403
2404         # Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2405         # Callback will be called with the text within
2406         # Transform and return the text within
2407         function setHook( $tag, $callback ) {
2408                 $oldVal = @$this->mTagHooks[$tag];
2409                 $this->mTagHooks[$tag] = $callback;
2410                 return $oldVal;
2411         }
2412 }
2413
2414 /**
2415  * @todo document
2416  * @package MediaWiki
2417  */
2418 class ParserOutput
2419 {
2420         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2421         var $mCacheTime; # Used in ParserCache
2422
2423         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2424                 $containsOldMagic = false )
2425         {
2426                 $this->mText = $text;
2427                 $this->mLanguageLinks = $languageLinks;
2428                 $this->mCategoryLinks = $categoryLinks;
2429                 $this->mContainsOldMagic = $containsOldMagic;
2430                 $this->mCacheTime = '';
2431         }
2432
2433         function getText() { return $this->mText; }
2434         function getLanguageLinks() { return $this->mLanguageLinks; }
2435         function getCategoryLinks() { return $this->mCategoryLinks; }
2436         function getCacheTime() { return $this->mCacheTime; }
2437         function containsOldMagic() { return $this->mContainsOldMagic; }
2438         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2439         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2440         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2441         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2442         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2443
2444         function merge( $other ) {
2445                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2446                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2447                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2448         }
2449
2450 }
2451
2452 /**
2453  * Set options of the Parser
2454  * @todo document
2455  * @package MediaWiki
2456  */
2457 class ParserOptions
2458 {
2459         # All variables are private
2460         var $mUseTeX;                    # Use texvc to expand <math> tags
2461         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2462         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2463         var $mAllowExternalImages;       # Allow external images inline
2464         var $mSkin;                      # Reference to the preferred skin
2465         var $mDateFormat;                # Date format index
2466         var $mEditSection;               # Create "edit section" links
2467         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2468         var $mNumberHeadings;            # Automatically number headings
2469         var $mShowToc;                   # Show table of contents
2470
2471         function getUseTeX()                        { return $this->mUseTeX; }
2472         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2473         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2474         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2475         function getSkin()                          { return $this->mSkin; }
2476         function getDateFormat()                    { return $this->mDateFormat; }
2477         function getEditSection()                   { return $this->mEditSection; }
2478         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
2479         function getNumberHeadings()                { return $this->mNumberHeadings; }
2480         function getShowToc()                       { return $this->mShowToc; }
2481
2482         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
2483         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
2484         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
2485         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
2486         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
2487         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
2488         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2489         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
2490         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
2491
2492         function setSkin( &$x ) { $this->mSkin =& $x; }
2493
2494         # Get parser options
2495         /* static */ function newFromUser( &$user ) {
2496                 $popts = new ParserOptions;
2497                 $popts->initialiseFromUser( $user );
2498                 return $popts;
2499         }
2500
2501         # Get user options
2502         function initialiseFromUser( &$userInput ) {
2503                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2504
2505                 $fname = 'ParserOptions::initialiseFromUser';
2506                 wfProfileIn( $fname );
2507                 if ( !$userInput ) {
2508                         $user = new User;
2509                         $user->setLoaded( true );
2510                 } else {
2511                         $user =& $userInput;
2512                 }
2513
2514                 $this->mUseTeX = $wgUseTeX;
2515                 $this->mUseDynamicDates = $wgUseDynamicDates;
2516                 $this->mInterwikiMagic = $wgInterwikiMagic;
2517                 $this->mAllowExternalImages = $wgAllowExternalImages;
2518                 wfProfileIn( $fname.'-skin' );
2519                 $this->mSkin =& $user->getSkin();
2520                 wfProfileOut( $fname.'-skin' );
2521                 $this->mDateFormat = $user->getOption( 'date' );
2522                 $this->mEditSection = $user->getOption( 'editsection' );
2523                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2524                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2525                 $this->mShowToc = $user->getOption( 'showtoc' );
2526                 wfProfileOut( $fname );
2527         }
2528
2529
2530 }
2531
2532 # Regex callbacks, used in Parser::replaceVariables
2533 function wfBraceSubstitution( $matches ) {
2534         global $wgCurParser;
2535         return $wgCurParser->braceSubstitution( $matches );
2536 }
2537
2538 function wfArgSubstitution( $matches ) {
2539         global $wgCurParser;
2540         return $wgCurParser->argSubstitution( $matches );
2541 }
2542
2543 function wfVariableSubstitution( $matches ) {
2544         global $wgCurParser;
2545         return $wgCurParser->variableSubstitution( $matches );
2546 }
2547
2548 /**
2549  * Return the total number of articles
2550  */
2551 function wfNumberOfArticles() {
2552         global $wgNumberOfArticles;
2553
2554         wfLoadSiteStats();
2555         return $wgNumberOfArticles;
2556 }
2557
2558 /**
2559  * Get various statistics from the database
2560  * @private
2561  */
2562 function wfLoadSiteStats() {
2563         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
2564         $fname = 'wfLoadSiteStats';
2565
2566         if ( -1 != $wgNumberOfArticles ) return;
2567         $dbr =& wfGetDB( DB_SLAVE );
2568         $s = $dbr->getArray( 'site_stats',
2569                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
2570                 array( 'ss_row_id' => 1 ), $fname
2571         );
2572
2573         if ( $s === false ) {
2574                 return;
2575         } else {
2576                 $wgTotalViews = $s->ss_total_views;
2577                 $wgTotalEdits = $s->ss_total_edits;
2578                 $wgNumberOfArticles = $s->ss_good_articles;
2579         }
2580 }
2581
2582 function wfEscapeHTMLTagsOnly( $in ) {
2583         return str_replace(
2584                 array( '"', '>', '<' ),
2585                 array( '&quot;', '&gt;', '&lt;' ),
2586                 $in );
2587 }
2588
2589
2590 ?>