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, $wgLanguageCode, $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;
 244                 if( $wgRawHtml ) {
 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         # Parses the text and adds the result to the strip state
 574         # Returns the strip tag
 575         function stripParse( $text, $newline, $args ) {
 576                 $text = $this->strip( $text, $this->mStripState );
 577                 $text = $this->internalParse( $text, (bool)$newline, $args, false );
 578                 return $newline.$this->insertStripItem( $text, $this->mStripState );
 579         }
 580
 581         function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
 582                 $fname = 'Parser::internalParse';
 583                 wfProfileIn( $fname );
 584
 585                 $text = $this->removeHTMLtags( $text );
 586                 $text = $this->replaceVariables( $text, $args );
 587
 588                 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
 589
 590                 $text = $this->doHeadings( $text );
 591                 if($this->mOptions->getUseDynamicDates()) {
 592                         global $wgDateFormatter;
 593                         $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
 594                 }
 595                 $text = $this->doAllQuotes( $text );
 596                 $text = $this->replaceExternalLinks( $text );
 597                 $text = $this->doMagicLinks( $text );
 598                 $text = $this->replaceInternalLinks ( $text );
 599                 $text = $this->replaceInternalLinks ( $text );
 600
 601                 $text = $this->unstrip( $text, $this->mStripState );
 602                 $text = $this->unstripNoWiki( $text, $this->mStripState );
 603
 604                 $text = $this->doTableStuff( $text );
 605                 $text = $this->formatHeadings( $text, $isMain );
 606                 $sk =& $this->mOptions->getSkin();
 607                 $text = $sk->transformContent( $text );
 608
 609                 wfProfileOut( $fname );
 610                 return $text;
 611         }
 612
 613         /* private */ function &doMagicLinks( &$text ) {
 614                 global $wgUseGeoMode;
 615                 $text = $this->magicISBN( $text );
 616                 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
 617                         $text = $this->magicGEO( $text );
 618                 }
 619                 $text = $this->magicRFC( $text );
 620                 return $text;
 621         }
 622
 623         # Parse ^^ tokens and return html
 624         /* private */ function doExponent ( $text ) {
 625                 $fname = 'Parser::doExponent';
 626                 wfProfileIn( $fname);
 627                 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
 628                 wfProfileOut( $fname);
 629                 return $text;
 630         }
 631
 632         # Parse headers and return html
 633         /* private */ function doHeadings( $text ) {
 634                 $fname = 'Parser::doHeadings';
 635                 wfProfileIn( $fname );
 636                 for ( $i = 6; $i >= 1; --$i ) {
 637                         $h = substr( '======', 0, $i );
 638                         $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
 639                           "<h{$i}>\\1</h{$i}>\\2", $text );
 640                 }
 641                 wfProfileOut( $fname );
 642                 return $text;
 643         }
 644
 645         /* private */ function doAllQuotes( $text ) {
 646                 $fname = 'Parser::doAllQuotes';
 647                 wfProfileIn( $fname );
 648                 $outtext = '';
 649                 $lines = explode( "\n", $text );
 650                 foreach ( $lines as $line ) {
 651                         $outtext .= $this->doQuotes ( $line ) . "\n";
 652                 }
 653                 $outtext = substr($outtext, 0,-1);
 654                 wfProfileOut( $fname );
 655                 return $outtext;
 656         }
 657
 658         /* private */ function doQuotes( $text ) {
 659                 $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
 660                 if (count ($arr) == 1)
 661                         return $text;
 662                 else
 663                 {
 664                         # First, do some preliminary work. This may shift some apostrophes from
 665                         # being mark-up to being text. It also counts the number of occurrences
 666                         # of bold and italics mark-ups.
 667                         $i = 0;
 668                         $numbold = 0;
 669                         $numitalics = 0;
 670                         foreach ($arr as $r)
 671                         {
 672                                 if (($i % 2) == 1)
 673                                 {
 674                                         # If there are ever four apostrophes, assume the first is supposed to
 675                                         # be text, and the remaining three constitute mark-up for bold text.
 676                                         if (strlen ($arr[$i]) == 4)
 677                                         {
 678                                                 $arr[$i-1] .= "'";
 679                                                 $arr[$i] = "'''";
 680                                         }
 681                                         # If there are more than 5 apostrophes in a row, assume they're all
 682                                         # text except for the last 5.
 683                                         else if (strlen ($arr[$i]) > 5)
 684                                         {
 685                                                 $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
 686                                                 $arr[$i] = "'''''";
 687                                         }
 688                                         # Count the number of occurrences of bold and italics mark-ups.
 689                                         # We are not counting sequences of five apostrophes.
 690                                         if (strlen ($arr[$i]) == 2) $numitalics++;  else
 691                                         if (strlen ($arr[$i]) == 3) $numbold++;     else
 692                                         if (strlen ($arr[$i]) == 5) { $numitalics++; $numbold++; }
 693                                 }
 694                                 $i++;
 695                         }
 696
 697                         # If there is an odd number of both bold and italics, it is likely
 698                         # that one of the bold ones was meant to be an apostrophe followed
 699                         # by italics. Which one we cannot know for certain, but it is more
 700                         # likely to be one that has a single-letter word before it.
 701                         if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
 702                         {
 703                                 $i = 0;
 704                                 $firstsingleletterword = -1;
 705                                 $firstmultiletterword = -1;
 706                                 $firstspace = -1;
 707                                 foreach ($arr as $r)
 708                                 {
 709                                         if (($i % 2 == 1) and (strlen ($r) == 3))
 710                                         {
 711                                                 $x1 = substr ($arr[$i-1], -1);
 712                                                 $x2 = substr ($arr[$i-1], -2, 1);
 713                                                 if ($x1 == ' ') {
 714                                                         if ($firstspace == -1) $firstspace = $i;
 715                                                 } else if ($x2 == ' ') {
 716                                                         if ($firstsingleletterword == -1) $firstsingleletterword = $i;
 717                                                 } else {
 718                                                         if ($firstmultiletterword == -1) $firstmultiletterword = $i;
 719                                                 }
 720                                         }
 721                                         $i++;
 722                                 }
 723
 724                                 # If there is a single-letter word, use it!
 725                                 if ($firstsingleletterword > -1)
 726                                 {
 727                                         $arr [ $firstsingleletterword ] = "''";
 728                                         $arr [ $firstsingleletterword-1 ] .= "'";
 729                                 }
 730                                 # If not, but there's a multi-letter word, use that one.
 731                                 else if ($firstmultiletterword > -1)
 732                                 {
 733                                         $arr [ $firstmultiletterword ] = "''";
 734                                         $arr [ $firstmultiletterword-1 ] .= "'";
 735                                 }
 736                                 # ... otherwise use the first one that has neither.
 737                                 # (notice that it is possible for all three to be -1 if, for example,
 738                                 # there is only one pentuple-apostrophe in the line)
 739                                 else if ($firstspace > -1)
 740                                 {
 741                                         $arr [ $firstspace ] = "''";
 742                                         $arr [ $firstspace-1 ] .= "'";
 743                                 }
 744                         }
 745
 746                         # Now let's actually convert our apostrophic mush to HTML!
 747                         $output = '';
 748                         $buffer = '';
 749                         $state = '';
 750                         $i = 0;
 751                         foreach ($arr as $r)
 752                         {
 753                                 if (($i % 2) == 0)
 754                                 {
 755                                         if ($state == 'both')
 756                                                 $buffer .= $r;
 757                                         else
 758                                                 $output .= $r;
 759                                 }
 760                                 else
 761                                 {
 762                                         if (strlen ($r) == 2)
 763                                         {
 764                                                 if ($state == 'em')
 765                                                 { $output .= '</em>'; $state = ''; }
 766                                                 else if ($state == 'strongem')
 767                                                 { $output .= '</em>'; $state = 'strong'; }
 768                                                 else if ($state == 'emstrong')
 769                                                 { $output .= '</strong></em><strong>'; $state = 'strong'; }
 770                                                 else if ($state == 'both')
 771                                                 { $output .= '<strong><em>'.$buffer.'</em>'; $state = 'strong'; }
 772                                                 else # $state can be 'strong' or ''
 773                                                 { $output .= '<em>'; $state .= 'em'; }
 774                                         }
 775                                         else if (strlen ($r) == 3)
 776                                         {
 777                                                 if ($state == 'strong')
 778                                                 { $output .= '</strong>'; $state = ''; }
 779                                                 else if ($state == 'strongem')
 780                                                 { $output .= '</em></strong><em>'; $state = 'em'; }
 781                                                 else if ($state == 'emstrong')
 782                                                 { $output .= '</strong>'; $state = 'em'; }
 783                                                 else if ($state == 'both')
 784                                                 { $output .= '<em><strong>'.$buffer.'</strong>'; $state = 'em'; }
 785                                                 else # $state can be 'em' or ''
 786                                                 { $output .= '<strong>'; $state .= 'strong'; }
 787                                         }
 788                                         else if (strlen ($r) == 5)
 789                                         {
 790                                                 if ($state == 'strong')
 791                                                 { $output .= '</strong><em>'; $state = 'em'; }
 792                                                 else if ($state == 'em')
 793                                                 { $output .= '</em><strong>'; $state = 'strong'; }
 794                                                 else if ($state == 'strongem')
 795                                                 { $output .= '</em></strong>'; $state = ''; }
 796                                                 else if ($state == 'emstrong')
 797                                                 { $output .= '</strong></em>'; $state = ''; }
 798                                                 else if ($state == 'both')
 799                                                 { $output .= '<em><strong>'.$buffer.'</strong></em>'; $state = ''; }
 800                                                 else # ($state == '')
 801                                                 { $buffer = ''; $state = 'both'; }
 802                                         }
 803                                 }
 804                                 $i++;
 805                         }
 806                         # Now close all remaining tags.  Notice that the order is important.
 807                         if ($state == 'strong' || $state == 'emstrong')
 808                                 $output .= '</strong>';
 809                         if ($state == 'em' || $state == 'strongem' || $state == 'emstrong')
 810                                 $output .= '</em>';
 811                         if ($state == 'strongem')
 812                                 $output .= '</strong>';
 813                         if ($state == 'both')
 814                                 $output .= '<strong><em>'.$buffer.'</em></strong>';
 815                         return $output;
 816                 }
 817         }
 818
 819         # Note: we have to do external links before the internal ones,
 820         # and otherwise take great care in the order of things here, so
 821         # that we don't end up interpreting some URLs twice.
 822
 823         /* private */ function replaceExternalLinks( $text ) {
 824                 $fname = 'Parser::replaceExternalLinks';
 825                 wfProfileIn( $fname );
 826
 827                 $sk =& $this->mOptions->getSkin();
 828                 $linktrail = wfMsg('linktrail');
 829                 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 830
 831                 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
 832
 833                 $i = 0;
 834                 while ( $i<count( $bits ) ) {
 835                         $url = $bits[$i++];
 836                         $protocol = $bits[$i++];
 837                         $text = $bits[$i++];
 838                         $trail = $bits[$i++];
 839
 840                         # If the link text is an image URL, replace it with an <img> tag
 841                         # This happened by accident in the original parser, but some people used it extensively
 842                         $img = $this->maybeMakeImageLink( $text );
 843                         if ( $img !== false ) {
 844                                 $text = $img;
 845                         }
 846
 847                         $dtrail = '';
 848
 849                         # No link text, e.g. [http://domain.tld/some.link]
 850                         if ( $text == '' ) {
 851                                 # Autonumber if allowed
 852                                 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
 853                                         $text = '[' . ++$this->mAutonumber . ']';
 854                                 } else {
 855                                         # Otherwise just use the URL
 856                                         $text = htmlspecialchars( $url );
 857                                 }
 858                         } else {
 859                                 # Have link text, e.g. [http://domain.tld/some.link text]s
 860                                 # Check for trail
 861                                 if ( preg_match( $linktrail, $trail, $m2 ) ) {
 862                                         $dtrail = $m2[1];
 863                                         $trail = $m2[2];
 864                                 }
 865                         }
 866
 867                         $encUrl = htmlspecialchars( $url );
 868                         # Bit in parentheses showing the URL for the printable version
 869                         if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
 870                                 $paren = '';
 871                         } else {
 872                                 # Expand the URL for printable version
 873                                 if ( ! $sk->suppressUrlExpansion() ) {
 874                                         $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
 875                                 } else {
 876                                         $paren = '';
 877                                 }
 878                         }
 879
 880                         # Process the trail (i.e. everything after this link up until start of the next link),
 881                         # replacing any non-bracketed links
 882                         $trail = $this->replaceFreeExternalLinks( $trail );
 883
 884                         $la = $sk->getExternalLinkAttributes( $url, $text );
 885
 886                         # Use the encoded URL
 887                         # This means that users can paste URLs directly into the text
 888                         # Funny characters like &ouml; aren't valid in URLs anyway
 889                         # This was changed in August 2004
 890                         $s .= "<a href=\"{$url}\" {$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
 891                 }
 892
 893                 wfProfileOut( $fname );
 894                 return $s;
 895         }
 896
 897         # Replace anything that looks like a URL with a link
 898         function replaceFreeExternalLinks( $text ) {
 899                 $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
 900                 $s = array_shift( $bits );
 901                 $i = 0;
 902
 903                 $sk =& $this->mOptions->getSkin();
 904
 905                 while ( $i < count( $bits ) ){
 906                         $protocol = $bits[$i++];
 907                         $remainder = $bits[$i++];
 908
 909                         if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
 910                                 # Found some characters after the protocol that look promising
 911                                 $url = $protocol . $m[1];
 912                                 $trail = $m[2];
 913
 914                                 # Move trailing punctuation to $trail
 915                                 $sep = ',;\.:!?';
 916                                 # If there is no left bracket, then consider right brackets fair game too
 917                                 if ( strpos( $url, '(' ) === false ) {
 918                                         $sep .= ')';
 919                                 }
 920
 921                                 $numSepChars = strspn( strrev( $url ), $sep );
 922                                 if ( $numSepChars ) {
 923                                         $trail = substr( $url, -$numSepChars ) . $trail;
 924                                         $url = substr( $url, 0, -$numSepChars );
 925                                 }
 926
 927                                 # Replace &amp; from obsolete syntax with &
 928                                 $url = str_replace( '&amp;', '&', $url );
 929
 930                                 # Is this an external image?
 931                                 $text = $this->maybeMakeImageLink( $url );
 932                                 if ( $text === false ) {
 933                                         # Not an image, make a link
 934                                         $text = $sk->makeExternalLink( $url, $url );
 935                                 }
 936                                 $s .= $text . $trail;
 937                         } else {
 938                                 $s .= $protocol . $remainder;
 939                         }
 940                 }
 941                 return $s;
 942         }
 943
 944         # make an image if it's allowed
 945         function maybeMakeImageLink( $url ) {
 946                 $sk =& $this->mOptions->getSkin();
 947                 $text = false;
 948                 if ( $this->mOptions->getAllowExternalImages() ) {
 949                         if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
 950                                 # Image found
 951                                 $text = $sk->makeImage( htmlspecialchars( $url ) );
 952                         }
 953                 }
 954                 return $text;
 955         }
 956
 957         # The wikilinks [[ ]] are procedeed here.
 958         /* private */ function replaceInternalLinks( $s ) {
 959                 global $wgLang, $wgLinkCache;
 960                 global $wgNamespacesWithSubpages, $wgLanguageCode;
 961                 static $fname = 'Parser::replaceInternalLinks' ;
 962                 wfProfileIn( $fname );
 963
 964                 wfProfileIn( $fname.'-setup' );
 965                 static $tc = FALSE;
 966                 # the % is needed to support urlencoded titles as well
 967                 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
 968                 $sk =& $this->mOptions->getSkin();
 969
 970                 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
 971
 972                 $a = explode( '[[', ' ' . $s );
 973                 $s = array_shift( $a );
 974                 $s = substr( $s, 1 );
 975
 976                 # Match a link having the form [[namespace:link|alternate]]trail
 977                 static $e1 = FALSE;
 978                 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
 979                 # Match the end of a line for a word that's not followed by whitespace,
 980                 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
 981                 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
 982
 983                 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
 984                 # Special and Media are pseudo-namespaces; no pages actually exist in them
 985
 986                 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
 987
 988                 if ( $useLinkPrefixExtension ) {
 989                         if ( preg_match( $e2, $s, $m ) ) {
 990                                 $first_prefix = $m[2];
 991                                 $s = $m[1];
 992                         } else {
 993                                 $first_prefix = false;
 994                         }
 995                 } else {
 996                         $prefix = '';
 997                 }
 998
 999                 wfProfileOut( $fname.'-setup' );
1000
1001                 # start procedeeding each line
1002                 foreach ( $a as $line ) {
1003                         wfProfileIn( $fname.'-prefixhandling' );
1004                         if ( $useLinkPrefixExtension ) {
1005                                 if ( preg_match( $e2, $s, $m ) ) {
1006                                         $prefix = $m[2];
1007                                         $s = $m[1];
1008                                 } else {
1009                                         $prefix='';
1010                                 }
1011                                 # first link
1012                                 if($first_prefix) {
1013                                         $prefix = $first_prefix;
1014                                         $first_prefix = false;
1015                                 }
1016                         }
1017                         wfProfileOut( $fname.'-prefixhandling' );
1018
1019                         if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1020                                 $text = $m[2];
1021                                 # fix up urlencoded title texts
1022                                 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1023                                 $trail = $m[3];
1024                         } else { # Invalid form; output directly
1025                                 $s .= $prefix . '[[' . $line ;
1026                                 continue;
1027                         }
1028
1029                         # Valid link forms:
1030                         # Foobar -- normal
1031                         # :Foobar -- override special treatment of prefix (images, language links)
1032                         # /Foobar -- convert to CurrentPage/Foobar
1033                         # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1034
1035                         # Look at the first character
1036                         $c = substr($m[1],0,1);
1037                         $noforce = ($c != ':');
1038
1039                         # subpage
1040                         if( $c == '/' ) {
1041                                 # / at end means we don't want the slash to be shown
1042                                 if(substr($m[1],-1,1)=='/') {
1043                                         $m[1]=substr($m[1],1,strlen($m[1])-2);
1044                                         $noslash=$m[1];
1045                                 } else {
1046                                         $noslash=substr($m[1],1);
1047                                 }
1048
1049                                 # Some namespaces don't allow subpages
1050                                 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) {
1051                                         # subpages allowed here
1052                                         $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1053                                         if( '' == $text ) {
1054                                                 $text= $m[1];
1055                                         } # this might be changed for ugliness reasons
1056                                 } else {
1057                                         # no subpage allowed, use standard link
1058                                         $link = $noslash;
1059                                 }
1060
1061                         } elseif( $noforce ) { # no subpage
1062                                 $link = $m[1];
1063                         } else {
1064                                 # We don't want to keep the first character
1065                                 $link = substr( $m[1], 1 );
1066                         }
1067
1068                         $wasblank = ( '' == $text );
1069                         if( $wasblank ) $text = $link;
1070
1071                         $nt = Title::newFromText( $link );
1072                         if( !$nt ) {
1073                                 $s .= $prefix . '[[' . $line;
1074                                 continue;
1075                         }
1076
1077                         $ns = $nt->getNamespace();
1078                         $iw = $nt->getInterWiki();
1079
1080                         # Link not escaped by : , create the various objects
1081                         if( $noforce ) {
1082
1083                                 # Interwikis
1084                                 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1085                                         array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1086                                         $tmp = $prefix . $trail ;
1087                                         $s .= (trim($tmp) == '')? '': $tmp;
1088                                         continue;
1089                                 }
1090
1091                                 if ( $ns == NS_IMAGE ) {
1092                                         $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1093                                         $wgLinkCache->addImageLinkObj( $nt );
1094                                         continue;
1095                                 }
1096
1097                                 if ( $ns == NS_CATEGORY ) {
1098                                         $t = $nt->getText() ;
1099                                         $nnt = Title::newFromText ( Namespace::getCanonicalName(NS_CATEGORY).':'.$t ) ;
1100
1101                                         $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1102                                         $pPLC=$sk->postParseLinkColour();
1103                                         $sk->postParseLinkColour( false );
1104                                         $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1105                                         $sk->postParseLinkColour( $pPLC );
1106                                         $wgLinkCache->resume();
1107
1108                                         if ( $wasblank ) {
1109                                                 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1110                                                         $sortkey = $this->mTitle->getText();
1111                                                 } else {
1112                                                         $sortkey = $this->mTitle->getPrefixedText();
1113                                                 }
1114                                         } else {
1115                                                 $sortkey = $text;
1116                                         }
1117                                         $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1118                                         $this->mOutput->mCategoryLinks[] = $t ;
1119                                         $s .= $prefix . $trail ;
1120                                         continue;
1121                                 }
1122                         }
1123
1124                         if( ( $nt->getPrefixedText() === $this->mTitle->getPrefixedText() ) &&
1125                             ( strpos( $link, '#' ) === FALSE ) ) {
1126                                 # Self-links are handled specially; generally de-link and change to bold.
1127                                 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1128                                 continue;
1129                         }
1130
1131                         if( $ns == NS_MEDIA ) {
1132                                 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1133                                 $wgLinkCache->addImageLinkObj( $nt );
1134                                 continue;
1135                         } elseif( $ns == NS_SPECIAL ) {
1136                                 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1137                                 continue;
1138                         }
1139                         $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1140                 }
1141                 wfProfileOut( $fname );
1142                 return $s;
1143         }
1144
1145         # Some functions here used by doBlockLevels()
1146         #
1147         /* private */ function closeParagraph() {
1148                 $result = '';
1149                 if ( '' != $this->mLastSection ) {
1150                         $result = '</' . $this->mLastSection  . ">\n";
1151                 }
1152                 $this->mInPre = false;
1153                 $this->mLastSection = '';
1154                 return $result;
1155         }
1156         # getCommon() returns the length of the longest common substring
1157         # of both arguments, starting at the beginning of both.
1158         #
1159         /* private */ function getCommon( $st1, $st2 ) {
1160                 $fl = strlen( $st1 );
1161                 $shorter = strlen( $st2 );
1162                 if ( $fl < $shorter ) { $shorter = $fl; }
1163
1164                 for ( $i = 0; $i < $shorter; ++$i ) {
1165                         if ( $st1{$i} != $st2{$i} ) { break; }
1166                 }
1167                 return $i;
1168         }
1169         # These next three functions open, continue, and close the list
1170         # element appropriate to the prefix character passed into them.
1171         #
1172         /* private */ function openList( $char ) {
1173                 $result = $this->closeParagraph();
1174
1175                 if ( '*' == $char ) { $result .= '<ul><li>'; }
1176                 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1177                 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1178                 else if ( ';' == $char ) {
1179                         $result .= '<dl><dt>';
1180                         $this->mDTopen = true;
1181                 }
1182                 else { $result = '<!-- ERR 1 -->'; }
1183
1184                 return $result;
1185         }
1186
1187         /* private */ function nextItem( $char ) {
1188                 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1189                 else if ( ':' == $char || ';' == $char ) {
1190                         $close = '</dd>';
1191                         if ( $this->mDTopen ) { $close = '</dt>'; }
1192                         if ( ';' == $char ) {
1193                                 $this->mDTopen = true;
1194                                 return $close . '<dt>';
1195                         } else {
1196                                 $this->mDTopen = false;
1197                                 return $close . '<dd>';
1198                         }
1199                 }
1200                 return '<!-- ERR 2 -->';
1201         }
1202
1203         /* private */ function closeList( $char ) {
1204                 if ( '*' == $char ) { $text = '</li></ul>'; }
1205                 else if ( '#' == $char ) { $text = '</li></ol>'; }
1206                 else if ( ':' == $char ) {
1207                         if ( $this->mDTopen ) {
1208                                 $this->mDTopen = false;
1209                                 $text = '</dt></dl>';
1210                         } else {
1211                                 $text = '</dd></dl>';
1212                         }
1213                 }
1214                 else {  return '<!-- ERR 3 -->'; }
1215                 return $text."\n";
1216         }
1217
1218         /* private */ function doBlockLevels( $text, $linestart ) {
1219                 $fname = 'Parser::doBlockLevels';
1220                 wfProfileIn( $fname );
1221
1222                 # Parsing through the text line by line.  The main thing
1223                 # happening here is handling of block-level elements p, pre,
1224                 # and making lists from lines starting with * # : etc.
1225                 #
1226                 $textLines = explode( "\n", $text );
1227
1228                 $lastPrefix = $output = $lastLine = '';
1229                 $this->mDTopen = $inBlockElem = false;
1230                 $prefixLength = 0;
1231                 $paragraphStack = false;
1232
1233                 if ( !$linestart ) {
1234                         $output .= array_shift( $textLines );
1235                 }
1236                 foreach ( $textLines as $oLine ) {
1237                         $lastPrefixLength = strlen( $lastPrefix );
1238                         $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1239                         $preOpenMatch = preg_match('/<pre/i', $oLine );
1240                         if ( !$this->mInPre ) {
1241                                 # Multiple prefixes may abut each other for nested lists.
1242                                 $prefixLength = strspn( $oLine, '*#:;' );
1243                                 $pref = substr( $oLine, 0, $prefixLength );
1244
1245                                 # eh?
1246                                 $pref2 = str_replace( ';', ':', $pref );
1247                                 $t = substr( $oLine, $prefixLength );
1248                                 $this->mInPre = !empty($preOpenMatch);
1249                         } else {
1250                                 # Don't interpret any other prefixes in preformatted text
1251                                 $prefixLength = 0;
1252                                 $pref = $pref2 = '';
1253                                 $t = $oLine;
1254                         }
1255
1256                         # List generation
1257                         if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1258                                 # Same as the last item, so no need to deal with nesting or opening stuff
1259                                 $output .= $this->nextItem( substr( $pref, -1 ) );
1260                                 $paragraphStack = false;
1261
1262                                 if ( substr( $pref, -1 ) == ';') {
1263                                         # The one nasty exception: definition lists work like this:
1264                                         # ; title : definition text
1265                                         # So we check for : in the remainder text to split up the
1266                                         # title and definition, without b0rking links.
1267                                         # FIXME: This is not foolproof. Something better in Tokenizer might help.
1268                                         if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1269                                                 $term = $match[1];
1270                                                 $output .= $term . $this->nextItem( ':' );
1271                                                 $t = $match[2];
1272                                         }
1273                                 }
1274                         } elseif( $prefixLength || $lastPrefixLength ) {
1275                                 # Either open or close a level...
1276                                 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1277                                 $paragraphStack = false;
1278
1279                                 while( $commonPrefixLength < $lastPrefixLength ) {
1280                                         $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1281                                         --$lastPrefixLength;
1282                                 }
1283                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1284                                         $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1285                                 }
1286                                 while ( $prefixLength > $commonPrefixLength ) {
1287                                         $char = substr( $pref, $commonPrefixLength, 1 );
1288                                         $output .= $this->openList( $char );
1289
1290                                         if ( ';' == $char ) {
1291                                                 # FIXME: This is dupe of code above
1292                                                 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1293                                                         $term = $match[1];
1294                                                         $output .= $term . $this->nextItem( ':' );
1295                                                         $t = $match[2];
1296                                                 }
1297                                         }
1298                                         ++$commonPrefixLength;
1299                                 }
1300                                 $lastPrefix = $pref2;
1301                         }
1302                         if( 0 == $prefixLength ) {
1303                                 # No prefix (not in list)--go to paragraph mode
1304                                 $uniq_prefix = UNIQ_PREFIX;
1305                                 // XXX: use a stack for nestable elements like span, table and div
1306                                 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1307                                 $closematch = preg_match(
1308                                         '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1309                                         '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1310                                 if ( $openmatch or $closematch ) {
1311                                         $paragraphStack = false;
1312                                         $output .= $this->closeParagraph();
1313                                         if($preOpenMatch and !$preCloseMatch) {
1314                                                 $this->mInPre = true;
1315                                         }
1316                                         if ( $closematch ) {
1317                                                 $inBlockElem = false;
1318                                         } else {
1319                                                 $inBlockElem = true;
1320                                         }
1321                                 } else if ( !$inBlockElem && !$this->mInPre ) {
1322                                         if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1323                                                 // pre
1324                                                 if ($this->mLastSection != 'pre') {
1325                                                         $paragraphStack = false;
1326                                                         $output .= $this->closeParagraph().'<pre>';
1327                                                         $this->mLastSection = 'pre';
1328                                                 }
1329                                         } else {
1330                                                 // paragraph
1331                                                 if ( '' == trim($t) ) {
1332                                                         if ( $paragraphStack ) {
1333                                                                 $output .= $paragraphStack.'<br />';
1334                                                                 $paragraphStack = false;
1335                                                                 $this->mLastSection = 'p';
1336                                                         } else {
1337                                                                 if ($this->mLastSection != 'p' ) {
1338                                                                         $output .= $this->closeParagraph();
1339                                                                         $this->mLastSection = '';
1340                                                                         $paragraphStack = '<p>';
1341                                                                 } else {
1342                                                                         $paragraphStack = '</p><p>';
1343                                                                 }
1344                                                         }
1345                                                 } else {
1346                                                         if ( $paragraphStack ) {
1347                                                                 $output .= $paragraphStack;
1348                                                                 $paragraphStack = false;
1349                                                                 $this->mLastSection = 'p';
1350                                                         } else if ($this->mLastSection != 'p') {
1351                                                                 $output .= $this->closeParagraph().'<p>';
1352                                                                 $this->mLastSection = 'p';
1353                                                         }
1354                                                 }
1355                                         }
1356                                 }
1357                         }
1358                         if ($paragraphStack === false) {
1359                                 $output .= $t."\n";
1360                         }
1361                 }
1362                 while ( $prefixLength ) {
1363                         $output .= $this->closeList( $pref2{$prefixLength-1} );
1364                         --$prefixLength;
1365                 }
1366                 if ( '' != $this->mLastSection ) {
1367                         $output .= '</' . $this->mLastSection . '>';
1368                         $this->mLastSection = '';
1369                 }
1370
1371                 wfProfileOut( $fname );
1372                 return $output;
1373         }
1374
1375         # Return value of a magic variable (like PAGENAME)
1376         function getVariableValue( $index ) {
1377                 global $wgLang, $wgSitename, $wgServer;
1378
1379                 switch ( $index ) {
1380                         case MAG_CURRENTMONTH:
1381                                 return $wgLang->formatNum( date( 'm' ) );
1382                         case MAG_CURRENTMONTHNAME:
1383                                 return $wgLang->getMonthName( date('n') );
1384                         case MAG_CURRENTMONTHNAMEGEN:
1385                                 return $wgLang->getMonthNameGen( date('n') );
1386                         case MAG_CURRENTDAY:
1387                                 return $wgLang->formatNum( date('j') );
1388                         case MAG_PAGENAME:
1389                                 return $this->mTitle->getText();
1390                         case MAG_PAGENAMEE:
1391                                 return $this->mTitle->getPartialURL();
1392                         case MAG_NAMESPACE:
1393                                 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1394                                 return $wgLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1395                         case MAG_CURRENTDAYNAME:
1396                                 return $wgLang->getWeekdayName( date('w')+1 );
1397                         case MAG_CURRENTYEAR:
1398                                 return $wgLang->formatNum( date( 'Y' ) );
1399                         case MAG_CURRENTTIME:
1400                                 return $wgLang->time( wfTimestampNow(), false );
1401                         case MAG_NUMBEROFARTICLES:
1402                                 return $wgLang->formatNum( wfNumberOfArticles() );
1403                         case MAG_SITENAME:
1404                                 return $wgSitename;
1405                         case MAG_SERVER:
1406                                 return $wgServer;
1407                         default:
1408                                 return NULL;
1409                 }
1410         }
1411
1412         # initialise the magic variables (like CURRENTMONTHNAME)
1413         function initialiseVariables() {
1414                 global $wgVariableIDs;
1415                 $this->mVariables = array();
1416                 foreach ( $wgVariableIDs as $id ) {
1417                         $mw =& MagicWord::get( $id );
1418                         $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1419                 }
1420         }
1421
1422         /* private */ function replaceVariables( $text, $args = array() ) {
1423                 global $wgLang, $wgScript, $wgArticlePath;
1424
1425                 # Prevent too big inclusions
1426                 if(strlen($text)> MAX_INCLUDE_SIZE)
1427                 return $text;
1428
1429                 $fname = 'Parser::replaceVariables';
1430                 wfProfileIn( $fname );
1431
1432                 $bail = false;
1433                 $titleChars = Title::legalChars();
1434                 $nonBraceChars = str_replace( array( '{', '}' ), array( '', '' ), $titleChars );
1435
1436                 # This function is called recursively. To keep track of arguments we need a stack:
1437                 array_push( $this->mArgStack, $args );
1438
1439                 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1440                 $GLOBALS['wgCurParser'] =& $this;
1441
1442                 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_MSG ) {
1443                         # Variable substitution
1444                         $text = preg_replace_callback( "/{{([$nonBraceChars]*?)}}/", 'wfVariableSubstitution', $text );
1445                 }
1446
1447                 if ( $this->mOutputType == OT_HTML ) {
1448                         # Argument substitution
1449                         $text = preg_replace_callback( "/(\\n?){{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1450                 }
1451                 # Template substitution
1452                 $regex = '/(\\n?){{(['.$nonBraceChars.']*)(\\|.*?|)}}/s';
1453                 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1454
1455                 array_pop( $this->mArgStack );
1456
1457                 wfProfileOut( $fname );
1458                 return $text;
1459         }
1460
1461         function variableSubstitution( $matches ) {
1462                 if ( !$this->mVariables ) {
1463                         $this->initialiseVariables();
1464                 }
1465                 if ( array_key_exists( $matches[1], $this->mVariables ) ) {
1466                         $text = $this->mVariables[$matches[1]];
1467                         $this->mOutput->mContainsOldMagic = true;
1468                 } else {
1469                         $text = $matches[0];
1470                 }
1471                 return $text;
1472         }
1473
1474         # Split template arguments
1475         function getTemplateArgs( $argsString ) {
1476                 if ( $argsString === '' ) {
1477                         return array();
1478                 }
1479
1480                 $args = explode( '|', substr( $argsString, 1 ) );
1481
1482                 # If any of the arguments contains a '[[' but no ']]', it needs to be
1483                 # merged with the next arg because the '|' character between belongs
1484                 # to the link syntax and not the template parameter syntax.
1485                 $argc = count($args);
1486                 $i = 0;
1487                 for ( $i = 0; $i < $argc-1; $i++ ) {
1488                         if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1489                                 $args[$i] .= '|'.$args[$i+1];
1490                                 array_splice($args, $i+1, 1);
1491                                 $i--;
1492                                 $argc--;
1493                         }
1494                 }
1495
1496                 return $args;
1497         }
1498
1499         function braceSubstitution( $matches ) {
1500                 global $wgLinkCache, $wgLang;
1501                 $fname = 'Parser::braceSubstitution';
1502                 $found = false;
1503                 $nowiki = false;
1504                 $noparse = false;
1505
1506                 $title = NULL;
1507
1508                 # $newline is an optional newline character before the braces
1509                 # $part1 is the bit before the first |, and must contain only title characters
1510                 # $args is a list of arguments, starting from index 0, not including $part1
1511
1512                 $newline = $matches[1];
1513                 $part1 = $matches[2];
1514                 # If the third subpattern matched anything, it will start with |
1515
1516                 $args = $this->getTemplateArgs($matches[3]);
1517                 $argc = count( $args );
1518
1519                 # {{{}}}
1520                 if ( strpos( $matches[0], '{{{' ) !== false ) {
1521                         $text = $matches[0];
1522                         $found = true;
1523                         $noparse = true;
1524                 }
1525
1526                 # SUBST
1527                 if ( !$found ) {
1528                         $mwSubst =& MagicWord::get( MAG_SUBST );
1529                         if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1530                                 if ( $this->mOutputType != OT_WIKI ) {
1531                                         # Invalid SUBST not replaced at PST time
1532                                         # Return without further processing
1533                                         $text = $matches[0];
1534                                         $found = true;
1535                                         $noparse= true;
1536                                 }
1537                         } elseif ( $this->mOutputType == OT_WIKI ) {
1538                                 # SUBST not found in PST pass, do nothing
1539                                 $text = $matches[0];
1540                                 $found = true;
1541                         }
1542                 }
1543
1544                 # MSG, MSGNW and INT
1545                 if ( !$found ) {
1546                         # Check for MSGNW:
1547                         $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1548                         if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1549                                 $nowiki = true;
1550                         } else {
1551                                 # Remove obsolete MSG:
1552                                 $mwMsg =& MagicWord::get( MAG_MSG );
1553                                 $mwMsg->matchStartAndRemove( $part1 );
1554                         }
1555
1556                         # Check if it is an internal message
1557                         $mwInt =& MagicWord::get( MAG_INT );
1558                         if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1559                                 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1560                                         $text = wfMsgReal( $part1, $args, true );
1561                                         $found = true;
1562                                 }
1563                         }
1564                 }
1565
1566                 # NS
1567                 if ( !$found ) {
1568                         # Check for NS: (namespace expansion)
1569                         $mwNs = MagicWord::get( MAG_NS );
1570                         if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1571                                 if ( intval( $part1 ) ) {
1572                                         $text = $wgLang->getNsText( intval( $part1 ) );
1573                                         $found = true;
1574                                 } else {
1575                                         $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1576                                         if ( !is_null( $index ) ) {
1577                                                 $text = $wgLang->getNsText( $index );
1578                                                 $found = true;
1579                                         }
1580                                 }
1581                         }
1582                 }
1583
1584                 # LOCALURL and LOCALURLE
1585                 if ( !$found ) {
1586                         $mwLocal = MagicWord::get( MAG_LOCALURL );
1587                         $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1588
1589                         if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1590                                 $func = 'getLocalURL';
1591                         } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1592                                 $func = 'escapeLocalURL';
1593                         } else {
1594                                 $func = '';
1595                         }
1596
1597                         if ( $func !== '' ) {
1598                                 $title = Title::newFromText( $part1 );
1599                                 if ( !is_null( $title ) ) {
1600                                         if ( $argc > 0 ) {
1601                                                 $text = $title->$func( $args[0] );
1602                                         } else {
1603                                                 $text = $title->$func();
1604                                         }
1605                                         $found = true;
1606                                 }
1607                         }
1608                 }
1609
1610                 # Internal variables
1611                 if ( !$this->mVariables ) {
1612                         $this->initialiseVariables();
1613                 }
1614                 if ( !$found && array_key_exists( $part1, $this->mVariables ) ) {
1615                         $text = $this->mVariables[$part1];
1616                         $found = true;
1617                         $this->mOutput->mContainsOldMagic = true;
1618                 }
1619
1620                 # GRAMMAR
1621                 if ( !$found && $argc == 1 ) {
1622                         $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1623                         if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1624                                 $text = $wgLang->convertGrammar( $args[0], $part1 );
1625                                 $found = true;
1626                         }
1627                 }
1628
1629                 # Template table test
1630
1631                 # Did we encounter this template already? If yes, it is in the cache
1632                 # and we need to check for loops.
1633                 if ( isset( $this->mTemplates[$part1] ) ) {
1634                         # Infinite loop test
1635                         if ( isset( $this->mTemplatePath[$part1] ) ) {
1636                                 $noparse = true;
1637                                 $found = true;
1638                         }
1639                         # set $text to cached message.
1640                         $text = $this->mTemplates[$part1];
1641                         $found = true;
1642                 }
1643
1644                 # Load from database
1645                 if ( !$found ) {
1646                         $title = Title::newFromText( $part1, NS_TEMPLATE );
1647                         if ( !is_null( $title ) && !$title->isExternal() ) {
1648                                 # Check for excessive inclusion
1649                                 $dbk = $title->getPrefixedDBkey();
1650                                 if ( $this->incrementIncludeCount( $dbk ) ) {
1651                                         # This should never be reached.
1652                                         $article = new Article( $title );
1653                                         $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1654                                         if ( $articleContent !== false ) {
1655                                                 $found = true;
1656                                                 $text = $articleContent;
1657                                         }
1658                                 }
1659
1660                                 # If the title is valid but undisplayable, make a link to it
1661                                 if ( $this->mOutputType == OT_HTML && !$found ) {
1662                                         $text = '[['.$title->getPrefixedText().']]';
1663                                         $found = true;
1664                                 }
1665
1666                                 # Template cache array insertion
1667                                 $this->mTemplates[$part1] = $text;
1668                         }
1669                 }
1670
1671                 # Recursive parsing, escaping and link table handling
1672                 # Only for HTML output
1673                 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1674                         $text = wfEscapeWikiText( $text );
1675                 } elseif ( $this->mOutputType == OT_HTML && $found && !$noparse) {
1676                         # Clean up argument array
1677                         $assocArgs = array();
1678                         $index = 1;
1679                         foreach( $args as $arg ) {
1680                                 $eqpos = strpos( $arg, '=' );
1681                                 if ( $eqpos === false ) {
1682                                         $assocArgs[$index++] = $arg;
1683                                 } else {
1684                                         $name = trim( substr( $arg, 0, $eqpos ) );
1685                                         $value = trim( substr( $arg, $eqpos+1 ) );
1686                                         if ( $value === false ) {
1687                                                 $value = '';
1688                                         }
1689                                         if ( $name !== false ) {
1690                                                 $assocArgs[$name] = $value;
1691                                         }
1692                                 }
1693                         }
1694
1695                         # Do not enter included links in link table
1696                         if ( !is_null( $title ) ) {
1697                                 $wgLinkCache->suspend();
1698                         }
1699
1700                         # Add a new element to the templace recursion path
1701                         $this->mTemplatePath[$part1] = 1;
1702
1703                         $text = $this->stripParse( $text, $newline, $assocArgs );
1704
1705                         # Resume the link cache and register the inclusion as a link
1706                         if ( !is_null( $title ) ) {
1707                                 $wgLinkCache->resume();
1708                                 $wgLinkCache->addLinkObj( $title );
1709                         }
1710                 }
1711
1712                 # Empties the template path
1713                 $this->mTemplatePath = array();
1714
1715                 if ( !$found ) {
1716                         return $matches[0];
1717                 } else {
1718                         return $text;
1719                 }
1720         }
1721
1722         # Triple brace replacement -- used for template arguments
1723         function argSubstitution( $matches ) {
1724                 $newline = $matches[1];
1725                 $arg = trim( $matches[2] );
1726                 $text = $matches[0];
1727                 $inputArgs = end( $this->mArgStack );
1728
1729                 if ( array_key_exists( $arg, $inputArgs ) ) {
1730                         $text = $this->stripParse( $inputArgs[$arg], $newline, array() );
1731                 }
1732
1733                 return $text;
1734         }
1735
1736         # Returns true if the function is allowed to include this entity
1737         function incrementIncludeCount( $dbk ) {
1738                 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1739                         $this->mIncludeCount[$dbk] = 0;
1740                 }
1741                 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1742                         return true;
1743                 } else {
1744                         return false;
1745                 }
1746         }
1747
1748
1749         # Cleans up HTML, removes dangerous tags and attributes
1750         /* private */ function removeHTMLtags( $text ) {
1751                 global $wgUseTidy, $wgUserHtml;
1752                 $fname = 'Parser::removeHTMLtags';
1753                 wfProfileIn( $fname );
1754
1755                 if( $wgUserHtml ) {
1756                         $htmlpairs = array( # Tags that must be closed
1757                                 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1758                                 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1759                                 'strike', 'strong', 'tt', 'var', 'div', 'center',
1760                                 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1761                                 'ruby', 'rt' , 'rb' , 'rp', 'p'
1762                         );
1763                         $htmlsingle = array(
1764                                 'br', 'hr', 'li', 'dt', 'dd'
1765                         );
1766                         $htmlnest = array( # Tags that can be nested--??
1767                                 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1768                                 'dl', 'font', 'big', 'small', 'sub', 'sup'
1769                         );
1770                         $tabletags = array( # Can only appear inside table
1771                                 'td', 'th', 'tr'
1772                         );
1773                 } else {
1774                         $htmlpairs = array();
1775                         $htmlsingle = array();
1776                         $htmlnest = array();
1777                         $tabletags = array();
1778                 }
1779
1780                 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1781                 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1782
1783                 $htmlattrs = $this->getHTMLattrs () ;
1784
1785                 # Remove HTML comments
1786                 $text = preg_replace( '/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU', '$2', $text );
1787
1788                 $bits = explode( '<', $text );
1789                 $text = array_shift( $bits );
1790                 if(!$wgUseTidy) {
1791                         $tagstack = array(); $tablestack = array();
1792                         foreach ( $bits as $x ) {
1793                                 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1794                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1795                                 $x, $regs );
1796                                 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1797                                 error_reporting( $prev );
1798
1799                                 $badtag = 0 ;
1800                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1801                                         # Check our stack
1802                                         if ( $slash ) {
1803                                                 # Closing a tag...
1804                                                 if ( ! in_array( $t, $htmlsingle ) &&
1805                                                 ( $ot = @array_pop( $tagstack ) ) != $t ) {
1806                                                         @array_push( $tagstack, $ot );
1807                                                         $badtag = 1;
1808                                                 } else {
1809                                                         if ( $t == 'table' ) {
1810                                                                 $tagstack = array_pop( $tablestack );
1811                                                         }
1812                                                         $newparams = '';
1813                                                 }
1814                                         } else {
1815                                                 # Keep track for later
1816                                                 if ( in_array( $t, $tabletags ) &&
1817                                                 ! in_array( 'table', $tagstack ) ) {
1818                                                         $badtag = 1;
1819                                                 } else if ( in_array( $t, $tagstack ) &&
1820                                                 ! in_array ( $t , $htmlnest ) ) {
1821                                                         $badtag = 1 ;
1822                                                 } else if ( ! in_array( $t, $htmlsingle ) ) {
1823                                                         if ( $t == 'table' ) {
1824                                                                 array_push( $tablestack, $tagstack );
1825                                                                 $tagstack = array();
1826                                                         }
1827                                                         array_push( $tagstack, $t );
1828                                                 }
1829                                                 # Strip non-approved attributes from the tag
1830                                                 $newparams = $this->fixTagAttributes($params);
1831
1832                                         }
1833                                         if ( ! $badtag ) {
1834                                                 $rest = str_replace( '>', '&gt;', $rest );
1835                                                 $text .= "<$slash$t $newparams$brace$rest";
1836                                                 continue;
1837                                         }
1838                                 }
1839                                 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1840                         }
1841                         # Close off any remaining tags
1842                         while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
1843                                 $text .= "</$t>\n";
1844                                 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
1845                         }
1846                 } else {
1847                         # this might be possible using tidy itself
1848                         foreach ( $bits as $x ) {
1849                                 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1850                                 $x, $regs );
1851                                 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1852                                 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1853                                         $newparams = $this->fixTagAttributes($params);
1854                                         $rest = str_replace( '>', '&gt;', $rest );
1855                                         $text .= "<$slash$t $newparams$brace$rest";
1856                                 } else {
1857                                         $text .= '&lt;' . str_replace( '>', '&gt;', $x);
1858                                 }
1859                         }
1860                 }
1861                 wfProfileOut( $fname );
1862                 return $text;
1863         }
1864
1865
1866         # This function accomplishes several tasks:
1867         # 1) Auto-number headings if that option is enabled
1868         # 2) Add an [edit] link to sections for logged in users who have enabled the option
1869         # 3) Add a Table of contents on the top for users who have enabled the option
1870         # 4) Auto-anchor headings
1871         #
1872         # It loops through all headlines, collects the necessary data, then splits up the
1873         # string and re-inserts the newly formatted headlines.
1874         /* private */ function formatHeadings( $text, $isMain=true ) {
1875                 global $wgInputEncoding, $wgMaxTocLevel, $wgLang;
1876
1877                 $doNumberHeadings = $this->mOptions->getNumberHeadings();
1878                 $doShowToc = $this->mOptions->getShowToc();
1879                 $forceTocHere = false;
1880                 if( !$this->mTitle->userCanEdit() ) {
1881                         $showEditLink = 0;
1882                         $rightClickHack = 0;
1883                 } else {
1884                         $showEditLink = $this->mOptions->getEditSection();
1885                         $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
1886                 }
1887
1888                 # Inhibit editsection links if requested in the page
1889                 $esw =& MagicWord::get( MAG_NOEDITSECTION );
1890                 if( $esw->matchAndRemove( $text ) ) {
1891                         $showEditLink = 0;
1892                 }
1893                 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1894                 # do not add TOC
1895                 $mw =& MagicWord::get( MAG_NOTOC );
1896                 if( $mw->matchAndRemove( $text ) ) {
1897                         $doShowToc = 0;
1898                 }
1899
1900                 # never add the TOC to the Main Page. This is an entry page that should not
1901                 # be more than 1-2 screens large anyway
1902                 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
1903                         $doShowToc = 0;
1904                 }
1905
1906                 # Get all headlines for numbering them and adding funky stuff like [edit]
1907                 # links - this is for later, but we need the number of headlines right now
1908                 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
1909
1910                 # if there are fewer than 4 headlines in the article, do not show TOC
1911                 if( $numMatches < 4 ) {
1912                         $doShowToc = 0;
1913                 }
1914
1915                 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
1916                 # override above conditions and always show TOC at that place
1917                 $mw =& MagicWord::get( MAG_TOC );
1918                 if ($mw->match( $text ) ) {
1919                         $doShowToc = 1;
1920                         $forceTocHere = true;
1921                 } else {
1922                         # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1923                         # override above conditions and always show TOC above first header
1924                         $mw =& MagicWord::get( MAG_FORCETOC );
1925                         if ($mw->matchAndRemove( $text ) ) {
1926                                 $doShowToc = 1;
1927                         }
1928                 }
1929
1930
1931
1932                 # We need this to perform operations on the HTML
1933                 $sk =& $this->mOptions->getSkin();
1934
1935                 # headline counter
1936                 $headlineCount = 0;
1937
1938                 # Ugh .. the TOC should have neat indentation levels which can be
1939                 # passed to the skin functions. These are determined here
1940                 $toclevel = 0;
1941                 $toc = '';
1942                 $full = '';
1943                 $head = array();
1944                 $sublevelCount = array();
1945                 $level = 0;
1946                 $prevlevel = 0;
1947                 foreach( $matches[3] as $headline ) {
1948                         $numbering = '';
1949                         if( $level ) {
1950                                 $prevlevel = $level;
1951                         }
1952                         $level = $matches[1][$headlineCount];
1953                         if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
1954                                 # reset when we enter a new level
1955                                 $sublevelCount[$level] = 0;
1956                                 $toc .= $sk->tocIndent( $level - $prevlevel );
1957                                 $toclevel += $level - $prevlevel;
1958                         }
1959                         if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
1960                                 # reset when we step back a level
1961                                 $sublevelCount[$level+1]=0;
1962                                 $toc .= $sk->tocUnindent( $prevlevel - $level );
1963                                 $toclevel -= $prevlevel - $level;
1964                         }
1965                         # count number of headlines for each level
1966                         @$sublevelCount[$level]++;
1967                         if( $doNumberHeadings || $doShowToc ) {
1968                                 $dot = 0;
1969                                 for( $i = 1; $i <= $level; $i++ ) {
1970                                         if( !empty( $sublevelCount[$i] ) ) {
1971                                                 if( $dot ) {
1972                                                         $numbering .= '.';
1973                                                 }
1974                                                 $numbering .= $wgLang->formatNum( $sublevelCount[$i] );
1975                                                 $dot = 1;
1976                                         }
1977                                 }
1978                         }
1979
1980                         # The canonized header is a version of the header text safe to use for links
1981                         # Avoid insertion of weird stuff like <math> by expanding the relevant sections
1982                         $canonized_headline = $this->unstrip( $headline, $this->mStripState );
1983                         $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
1984
1985                         # Remove link placeholders by the link text.
1986                         #     <!--LINK namespace page_title link text with suffix-->
1987                         # turns into
1988                         #     link text with suffix
1989                         $canonized_headline = preg_replace( '/<!--LINK [0-9]* [^ ]* *(.*?)-->/','$1', $canonized_headline );
1990                         # strip out HTML
1991                         $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
1992                         $tocline = trim( $canonized_headline );
1993                         $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
1994                         $replacearray = array(
1995                                 '%3A' => ':',
1996                                 '%' => '.'
1997                         );
1998                         $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
1999                         $refer[$headlineCount] = $canonized_headline;
2000
2001                         # count how many in assoc. array so we can track dupes in anchors
2002                         @$refers[$canonized_headline]++;
2003                         $refcount[$headlineCount]=$refers[$canonized_headline];
2004
2005                         # Prepend the number to the heading text
2006
2007                         if( $doNumberHeadings || $doShowToc ) {
2008                                 $tocline = $numbering . ' ' . $tocline;
2009
2010                                 # Don't number the heading if it is the only one (looks silly)
2011                                 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2012                                         # the two are different if the line contains a link
2013                                         $headline=$numbering . ' ' . $headline;
2014                                 }
2015                         }
2016
2017                         # Create the anchor for linking from the TOC to the section
2018                         $anchor = $canonized_headline;
2019                         if($refcount[$headlineCount] > 1 ) {
2020                                 $anchor .= '_' . $refcount[$headlineCount];
2021                         }
2022                         if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2023                                 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2024                         }
2025                         if( $showEditLink ) {
2026                                 if ( empty( $head[$headlineCount] ) ) {
2027                                         $head[$headlineCount] = '';
2028                                 }
2029                                 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+1);
2030                         }
2031
2032                         # Add the edit section span
2033                         if( $rightClickHack ) {
2034                                 $headline = $sk->editSectionScript($headlineCount+1,$headline);
2035                         }
2036
2037                         # give headline the correct <h#> tag
2038                         @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2039
2040                         $headlineCount++;
2041                 }
2042
2043                 if( $doShowToc ) {
2044                         $toclines = $headlineCount;
2045                         $toc .= $sk->tocUnindent( $toclevel );
2046                         $toc = $sk->tocTable( $toc );
2047                 }
2048
2049                 # split up and insert constructed headlines
2050
2051                 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2052                 $i = 0;
2053
2054                 foreach( $blocks as $block ) {
2055                         if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2056                                 # This is the [edit] link that appears for the top block of text when
2057                                 # section editing is enabled
2058
2059                                 # Disabled because it broke block formatting
2060                                 # For example, a bullet point in the top line
2061                                 # $full .= $sk->editSectionLink(0);
2062                         }
2063                         $full .= $block;
2064                         if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2065                         # Top anchor now in skin
2066                                 $full = $full.$toc;
2067                         }
2068
2069                         if( !empty( $head[$i] ) ) {
2070                                 $full .= $head[$i];
2071                         }
2072                         $i++;
2073                 }
2074                 if($forceTocHere) {
2075                         $mw =& MagicWord::get( MAG_TOC );
2076                         return $mw->replace( $toc, $full );
2077                 } else {
2078                         return $full;
2079                 }
2080         }
2081
2082         # Return an HTML link for the "ISBN 123456" text
2083         /* private */ function magicISBN( $text ) {
2084                 global $wgLang;
2085                 $fname = 'Parser::magicISBN';
2086                 wfProfileIn( $fname );
2087
2088                 $a = split( 'ISBN ', ' '.$text );
2089                 if ( count ( $a ) < 2 ) {
2090                         wfProfileOut( $fname );
2091                         return $text;
2092                 }
2093                 $text = substr( array_shift( $a ), 1);
2094                 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2095
2096                 foreach ( $a as $x ) {
2097                         if ( $x == '' ) continue;
2098                         $isbn = $blank = '' ;
2099                         while ( ' ' == $x{0} ) {
2100                                 $blank .= ' ';
2101                                 $x = substr( $x, 1 );
2102                         }
2103                         while ( strstr( $valid, $x{0} ) != false ) {
2104                                 $isbn .= $x{0};
2105                                 $x = substr( $x, 1 );
2106                         }
2107                         $num = str_replace( '-', '', $isbn );
2108                         $num = str_replace( ' ', '', $num );
2109
2110                         if ( '' == $num ) {
2111                                 $text .= "ISBN $blank$x";
2112                         } else {
2113                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2114                                 $text .= '<a href="' .
2115                                 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2116                                         "\" class=\"internal\">ISBN $isbn</a>";
2117                                 $text .= $x;
2118                         }
2119                 }
2120                 wfProfileOut( $fname );
2121                 return $text;
2122         }
2123
2124         # Return an HTML link for the "GEO ..." text
2125         /* private */ function magicGEO( $text ) {
2126                 global $wgLang, $wgUseGeoMode;
2127                 $fname = 'Parser::magicGEO';
2128                 wfProfileIn( $fname );
2129
2130                 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2131                 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2132                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2133                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2134                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2135                 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2136
2137                 $a = split( 'GEO ', ' '.$text );
2138                 if ( count ( $a ) < 2 ) {
2139                         wfProfileOut( $fname );
2140                         return $text;
2141                 }
2142                 $text = substr( array_shift( $a ), 1);
2143                 $valid = '0123456789.+-:';
2144
2145                 foreach ( $a as $x ) {
2146                         $geo = $blank = '' ;
2147                         while ( ' ' == $x{0} ) {
2148                                 $blank .= ' ';
2149                                 $x = substr( $x, 1 );
2150                         }
2151                         while ( strstr( $valid, $x{0} ) != false ) {
2152                                 $geo .= $x{0};
2153                                 $x = substr( $x, 1 );
2154                         }
2155                         $num = str_replace( '+', '', $geo );
2156                         $num = str_replace( ' ', '', $num );
2157
2158                         if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2159                                 $text .= "GEO $blank$x";
2160                         } else {
2161                                 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2162                                 $text .= '<a href="' .
2163                                 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2164                                         "\" class=\"internal\">GEO $geo</a>";
2165                                 $text .= $x;
2166                         }
2167                 }
2168                 wfProfileOut( $fname );
2169                 return $text;
2170         }
2171
2172         # Return an HTML link for the "RFC 1234" text
2173         /* private */ function magicRFC( $text ) {
2174                 global $wgLang;
2175
2176                 $a = split( 'RFC ', ' '.$text );
2177                 if ( count ( $a ) < 2 ) return $text;
2178                 $text = substr( array_shift( $a ), 1);
2179                 $valid = '0123456789';
2180
2181                 foreach ( $a as $x ) {
2182                         $rfc = $blank = '' ;
2183                         while ( ' ' == $x{0} ) {
2184                                 $blank .= ' ';
2185                                 $x = substr( $x, 1 );
2186                         }
2187                         while ( strstr( $valid, $x{0} ) != false ) {
2188                                 $rfc .= $x{0};
2189                                 $x = substr( $x, 1 );
2190                         }
2191
2192                         if ( '' == $rfc ) {
2193                                 $text .= "RFC $blank$x";
2194                         } else {
2195                                 $url = wfmsg( 'rfcurl' );
2196                                 $url = str_replace( '$1', $rfc, $url);
2197                                 $sk =& $this->mOptions->getSkin();
2198                                 $la = $sk->getExternalLinkAttributes( $url, 'RFC '.$rfc );
2199                                 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2200                         }
2201                 }
2202                 return $text;
2203         }
2204
2205         function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2206                 $this->mOptions = $options;
2207                 $this->mTitle =& $title;
2208                 $this->mOutputType = OT_WIKI;
2209
2210                 if ( $clearState ) {
2211                         $this->clearState();
2212                 }
2213
2214                 $stripState = false;
2215                 $pairs = array(
2216                         "\r\n" => "\n",
2217                         );
2218                 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2219                 // now with regexes
2220                 /*
2221                 $pairs = array(
2222                         "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2223                         "/<br *?>/i" => "<br />",
2224                 );
2225                 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2226                 */
2227                 $text = $this->strip( $text, $stripState, false );
2228                 $text = $this->pstPass2( $text, $user );
2229                 $text = $this->unstrip( $text, $stripState );
2230                 $text = $this->unstripNoWiki( $text, $stripState );
2231                 return $text;
2232         }
2233
2234         /* private */ function pstPass2( $text, &$user ) {
2235                 global $wgLang, $wgLocaltimezone, $wgCurParser;
2236
2237                 # Variable replacement
2238                 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2239                 $text = $this->replaceVariables( $text );
2240
2241                 # Signatures
2242                 #
2243                 $n = $user->getName();
2244                 $k = $user->getOption( 'nickname' );
2245                 if ( '' == $k ) { $k = $n; }
2246                 if(isset($wgLocaltimezone)) {
2247                         $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2248                 }
2249                 /* Note: this is an ugly timezone hack for the European wikis */
2250                 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2251                   ' (' . date( 'T' ) . ')';
2252                 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2253
2254                 $text = preg_replace( '/~~~~~/', $d, $text );
2255                 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2256                 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2257
2258                 # Context links: [[|name]] and [[name (context)|]]
2259                 #
2260                 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2261                 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2262                 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2263                 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2264
2265                 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/";             # [[page (context)|]]
2266                 $p2 = "/\[\[\\|({$tc}+)]]/";                                    # [[|page]]
2267                 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/";                # [[namespace:page|]] and [[:namespace:page|]]
2268                 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2269                 $context = '';
2270                 $t = $this->mTitle->getText();
2271                 if ( preg_match( $conpat, $t, $m ) ) {
2272                         $context = $m[2];
2273                 }
2274                 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2275                 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2276                 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2277
2278                 if ( '' == $context ) {
2279                         $text = preg_replace( $p2, '[[\\1]]', $text );
2280                 } else {
2281                         $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2282                 }
2283
2284                 /*
2285                 $mw =& MagicWord::get( MAG_SUBST );
2286                 $wgCurParser = $this->fork();
2287                 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2288                 $this->merge( $wgCurParser );
2289                 */
2290
2291                 # Trim trailing whitespace
2292                 # MAG_END (__END__) tag allows for trailing
2293                 # whitespace to be deliberately included
2294                 $text = rtrim( $text );
2295                 $mw =& MagicWord::get( MAG_END );
2296                 $mw->matchAndRemove( $text );
2297
2298                 return $text;
2299         }
2300
2301         # Set up some variables which are usually set up in parse()
2302         # so that an external function can call some class members with confidence
2303         function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2304                 $this->mTitle =& $title;
2305                 $this->mOptions = $options;
2306                 $this->mOutputType = $outputType;
2307                 if ( $clearState ) {
2308                         $this->clearState();
2309                 }
2310         }
2311
2312         function transformMsg( $text, $options ) {
2313                 global $wgTitle;
2314                 static $executing = false;
2315
2316                 # Guard against infinite recursion
2317                 if ( $executing ) {
2318                         return $text;
2319                 }
2320                 $executing = true;
2321
2322                 $this->mTitle = $wgTitle;
2323                 $this->mOptions = $options;
2324                 $this->mOutputType = OT_MSG;
2325                 $this->clearState();
2326                 $text = $this->replaceVariables( $text );
2327
2328                 $executing = false;
2329                 return $text;
2330         }
2331
2332         # Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2333         # Callback will be called with the text within
2334         # Transform and return the text within
2335         function setHook( $tag, $callback ) {
2336                 $oldVal = @$this->mTagHooks[$tag];
2337                 $this->mTagHooks[$tag] = $callback;
2338                 return $oldVal;
2339         }
2340 }
2341
2342 /**
2343  * @todo document
2344  * @package MediaWiki
2345  */
2346 class ParserOutput
2347 {
2348         var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2349         var $mCacheTime; # Used in ParserCache
2350
2351         function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2352                 $containsOldMagic = false )
2353         {
2354                 $this->mText = $text;
2355                 $this->mLanguageLinks = $languageLinks;
2356                 $this->mCategoryLinks = $categoryLinks;
2357                 $this->mContainsOldMagic = $containsOldMagic;
2358                 $this->mCacheTime = '';
2359         }
2360
2361         function getText() { return $this->mText; }
2362         function getLanguageLinks() { return $this->mLanguageLinks; }
2363         function getCategoryLinks() { return $this->mCategoryLinks; }
2364         function getCacheTime() { return $this->mCacheTime; }
2365         function containsOldMagic() { return $this->mContainsOldMagic; }
2366         function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2367         function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2368         function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2369         function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2370         function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2371
2372         function merge( $other ) {
2373                 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2374                 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2375                 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2376         }
2377
2378 }
2379
2380 /**
2381  * Set options of the Parser
2382  * @todo document
2383  * @package MediaWiki
2384  */
2385 class ParserOptions
2386 {
2387         # All variables are private
2388         var $mUseTeX;                    # Use texvc to expand <math> tags
2389         var $mUseDynamicDates;           # Use $wgDateFormatter to format dates
2390         var $mInterwikiMagic;            # Interlanguage links are removed and returned in an array
2391         var $mAllowExternalImages;       # Allow external images inline
2392         var $mSkin;                      # Reference to the preferred skin
2393         var $mDateFormat;                # Date format index
2394         var $mEditSection;               # Create "edit section" links
2395         var $mEditSectionOnRightClick;   # Generate JavaScript to edit section on right click
2396         var $mNumberHeadings;            # Automatically number headings
2397         var $mShowToc;                   # Show table of contents
2398
2399         function getUseTeX()                        { return $this->mUseTeX; }
2400         function getUseDynamicDates()               { return $this->mUseDynamicDates; }
2401         function getInterwikiMagic()                { return $this->mInterwikiMagic; }
2402         function getAllowExternalImages()           { return $this->mAllowExternalImages; }
2403         function getSkin()                          { return $this->mSkin; }
2404         function getDateFormat()                    { return $this->mDateFormat; }
2405         function getEditSection()                   { return $this->mEditSection; }
2406         function getEditSectionOnRightClick()       { return $this->mEditSectionOnRightClick; }
2407         function getNumberHeadings()                { return $this->mNumberHeadings; }
2408         function getShowToc()                       { return $this->mShowToc; }
2409
2410         function setUseTeX( $x )                    { return wfSetVar( $this->mUseTeX, $x ); }
2411         function setUseDynamicDates( $x )           { return wfSetVar( $this->mUseDynamicDates, $x ); }
2412         function setInterwikiMagic( $x )            { return wfSetVar( $this->mInterwikiMagic, $x ); }
2413         function setAllowExternalImages( $x )       { return wfSetVar( $this->mAllowExternalImages, $x ); }
2414         function setDateFormat( $x )                { return wfSetVar( $this->mDateFormat, $x ); }
2415         function setEditSection( $x )               { return wfSetVar( $this->mEditSection, $x ); }
2416         function setEditSectionOnRightClick( $x )   { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2417         function setNumberHeadings( $x )            { return wfSetVar( $this->mNumberHeadings, $x ); }
2418         function setShowToc( $x )                   { return wfSetVar( $this->mShowToc, $x ); }
2419
2420         function setSkin( &$x ) { $this->mSkin =& $x; }
2421
2422         # Get parser options
2423         /* static */ function newFromUser( &$user ) {
2424                 $popts = new ParserOptions;
2425                 $popts->initialiseFromUser( $user );
2426                 return $popts;
2427         }
2428
2429         # Get user options
2430         function initialiseFromUser( &$userInput ) {
2431                 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2432
2433                 $fname = 'ParserOptions::initialiseFromUser';
2434                 wfProfileIn( $fname );
2435                 if ( !$userInput ) {
2436                         $user = new User;
2437                         $user->setLoaded( true );
2438                 } else {
2439                         $user =& $userInput;
2440                 }
2441
2442                 $this->mUseTeX = $wgUseTeX;
2443                 $this->mUseDynamicDates = $wgUseDynamicDates;
2444                 $this->mInterwikiMagic = $wgInterwikiMagic;
2445                 $this->mAllowExternalImages = $wgAllowExternalImages;
2446                 wfProfileIn( $fname.'-skin' );
2447                 $this->mSkin =& $user->getSkin();
2448                 wfProfileOut( $fname.'-skin' );
2449                 $this->mDateFormat = $user->getOption( 'date' );
2450                 $this->mEditSection = $user->getOption( 'editsection' );
2451                 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2452                 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2453                 $this->mShowToc = $user->getOption( 'showtoc' );
2454                 wfProfileOut( $fname );
2455         }
2456
2457
2458 }
2459
2460 # Regex callbacks, used in Parser::replaceVariables
2461 function wfBraceSubstitution( $matches ) {
2462         global $wgCurParser;
2463         return $wgCurParser->braceSubstitution( $matches );
2464 }
2465
2466 function wfArgSubstitution( $matches ) {
2467         global $wgCurParser;
2468         return $wgCurParser->argSubstitution( $matches );
2469 }
2470
2471 function wfVariableSubstitution( $matches ) {
2472         global $wgCurParser;
2473         return $wgCurParser->variableSubstitution( $matches );
2474 }
2475
2476 /**
2477  * Return the total number of articles
2478  */
2479 function wfNumberOfArticles() {
2480         global $wgNumberOfArticles;
2481
2482         wfLoadSiteStats();
2483         return $wgNumberOfArticles;
2484 }
2485
2486 /**
2487  * Get various statistics from the database
2488  * @private
2489  */
2490 function wfLoadSiteStats() {
2491         global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
2492         $fname = 'wfLoadSiteStats';
2493
2494         if ( -1 != $wgNumberOfArticles ) return;
2495         $dbr =& wfGetDB( DB_SLAVE );
2496         $s = $dbr->getArray( 'site_stats',
2497                 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
2498                 array( 'ss_row_id' => 1 ), $fname
2499         );
2500
2501         if ( $s === false ) {
2502                 return;
2503         } else {
2504                 $wgTotalViews = $s->ss_total_views;
2505                 $wgTotalEdits = $s->ss_total_edits;
2506                 $wgNumberOfArticles = $s->ss_good_articles;
2507         }
2508 }
2509
2510 function wfEscapeHTMLTagsOnly( $in ) {
2511         return str_replace(
2512                 array( '"', '>', '<' ),
2513                 array( '&quot;', '&gt;', '&lt;' ),
2514                 $in );
2515 }
2516
2517
2518 ?>