?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Remove a space for parser checker consistency
[mediawiki.git] / includes / Parser.php
blob1549f823046215427ed18907a62c430edc9698c6
1 <?php
2
3 // require_once('Tokenizer.php');
4
5 /**
6 * File for Parser and related classes
7 *
8 * @package MediaWiki
9 * @version $Id$
10 */
11
12 /**
13 * Variable substitution O(N^2) attack
14 *
15 * Without countermeasures, it would be possible to attack the parser by saving
16 * a page filled with a large number of inclusions of large pages. The size of
17 * the generated page would be proportional to the square of the input size.
18 * Hence, we limit the number of inclusions of any given page, thus bringing any
19 * attack back to O(N).
20 */
21 define( 'MAX_INCLUDE_REPEAT', 100 );
22 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
23
24 # Allowed values for $mOutputType
25 define( 'OT_HTML', 1 );
26 define( 'OT_WIKI', 2 );
27 define( 'OT_MSG' , 3 );
28
29 # string parameter for extractTags which will cause it
30 # to strip HTML comments in addition to regular
31 # <XML>-style tags. This should not be anything we
32 # may want to use in wikisyntax
33 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
34
35 # prefix for escaping, used in two functions at least
36 define( 'UNIQ_PREFIX', 'NaodW29');
37
38 # Constants needed for external link processing
39 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
40 define( 'HTTP_PROTOCOLS', 'http|https' );
41 # Everything except bracket, space, or control characters
42 define( 'EXT_LINK_URL_CLASS', '[^]\\x00-\\x20\\x7F]' );
43 define( 'INVERSE_EXT_LINK_URL_CLASS', '[\]\\x00-\\x20\\x7F]' );
44 # Including space
45 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
46 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
47 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
48 define( 'EXT_LINK_BRACKETED', '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
49 define( 'EXT_IMAGE_REGEX',
50 '/^('.HTTP_PROTOCOLS.':)'. # Protocol
51 '('.EXT_LINK_URL_CLASS.'+)\\/'. # Hostname and path
52 '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
53 );
54
55 /**
56 * PHP Parser
57 *
58 * Processes wiki markup
59 *
60 * <pre>
61 * There are three main entry points into the Parser class:
62 * parse()
63 * produces HTML output
64 * preSaveTransform().
65 * produces altered wiki markup.
66 * transformMsg()
67 * performs brace substitution on MediaWiki messages
68 *
69 * Globals used:
70 * objects: $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
71 *
72 * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
73 *
74 * settings:
75 * $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
76 * $wgNamespacesWithSubpages, $wgAllowExternalImages*,
77 * $wgLocaltimezone
78 *
79 * * only within ParserOptions
80 * </pre>
81 *
82 * @package MediaWiki
83 */
84 class Parser
85 {
86 /**#@+
87 * @access private
88 */
89 # Persistent:
90 var $mTagHooks;
91
92 # Cleared with clearState():
93 var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
94 var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
95
96 # Temporary:
97 var $mOptions, $mTitle, $mOutputType,
98 $mTemplates, // cache of already loaded templates, avoids
99 // multiple SQL queries for the same string
100 $mTemplatePath; // stores an unsorted hash of all the templates already loaded
101 // in this path. Used for loop detection.
102
103 /**#@-*/
104
105 /**
106 * Constructor
107 *
108 * @access public
109 */
110 function Parser() {
111 $this->mTemplates = array();
112 $this->mTemplatePath = array();
113 $this->mTagHooks = array();
114 $this->clearState();
115 }
116
117 /**
118 * Clear Parser state
119 *
120 * @access private
121 */
122 function clearState() {
123 $this->mOutput = new ParserOutput;
124 $this->mAutonumber = 0;
125 $this->mLastSection = "";
126 $this->mDTopen = false;
127 $this->mVariables = false;
128 $this->mIncludeCount = array();
129 $this->mStripState = array();
130 $this->mArgStack = array();
131 $this->mInPre = false;
132 }
133
134 /**
135 * First pass--just handle <nowiki> sections, pass the rest off
136 * to internalParse() which does all the real work.
137 *
138 * @access private
139 * @return ParserOutput a ParserOutput
140 */
141 function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
142 global $wgUseTidy;
143 $fname = 'Parser::parse';
144 wfProfileIn( $fname );
145
146 if ( $clearState ) {
147 $this->clearState();
148 }
149
150 $this->mOptions = $options;
151 $this->mTitle =& $title;
152 $this->mOutputType = OT_HTML;
153
154 $stripState = NULL;
155 $text = $this->strip( $text, $this->mStripState );
156 $text = $this->internalParse( $text, $linestart );
157 $text = $this->unstrip( $text, $this->mStripState );
158 # Clean up special characters, only run once, next-to-last before doBlockLevels
159 if(!$wgUseTidy) {
160 $fixtags = array(
161 # french spaces, last one Guillemet-left
162 # only if there is something before the space
163 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
164 # french spaces, Guillemet-right
165 "/(\\302\\253) /i"=>"\\1&nbsp;",
166 '/<hr *>/i' => '<hr />',
167 '/<br *>/i' => '<br />',
168 '/<center *>/i' => '<div class="center">',
169 '/<\\/center *>/i' => '</div>',
170 # Clean up spare ampersands; note that we probably ought to be
171 # more careful about named entities.
172 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
173 );
174 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
175 } else {
176 $fixtags = array(
177 # french spaces, last one Guillemet-left
178 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
179 # french spaces, Guillemet-right
180 '/(\\302\\253) /i' => '\\1&nbsp;',
181 '/<center *>/i' => '<div class="center">',
182 '/<\\/center *>/i' => '</div>'
183 );
184 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
185 }
186 # only once and last
187 $text = $this->doBlockLevels( $text, $linestart );
188 $text = $this->unstripNoWiki( $text, $this->mStripState );
189 if($wgUseTidy) {
190 $text = $this->tidy($text);
191 }
192 $this->mOutput->setText( $text );
193 wfProfileOut( $fname );
194 return $this->mOutput;
195 }
196
197 /**
198 * Get a random string
199 *
200 * @access private
201 * @static
202 */
203 function getRandomString() {
204 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
205 }
206
207 /**
208 * Replaces all occurrences of <$tag>content</$tag> in the text
209 * with a random marker and returns the new text. the output parameter
210 * $content will be an associative array filled with data on the form
211 * $unique_marker => content.
212 *
213 * If $content is already set, the additional entries will be appended
214 * If $tag is set to STRIP_COMMENTS, the function will extract
215 * <!-- HTML comments -->
216 *
217 * @access private
218 * @static
219 */
220 function extractTags($tag, $text, &$content, $uniq_prefix = ''){
221 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
222 if ( !$content ) {
223 $content = array( );
224 }
225 $n = 1;
226 $stripped = '';
227
228 while ( '' != $text ) {
229 if($tag==STRIP_COMMENTS) {
230 $p = preg_split( '/<!--/i', $text, 2 );
231 } else {
232 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
233 }
234 $stripped .= $p[0];
235 if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
236 $text = '';
237 } else {
238 if($tag==STRIP_COMMENTS) {
239 $q = preg_split( '/-->/i', $p[1], 2 );
240 } else {
241 $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
242 }
243 $marker = $rnd . sprintf('%08X', $n++);
244 $content[$marker] = $q[0];
245 $stripped .= $marker;
246 $text = $q[1];
247 }
248 }
249 return $stripped;
250 }
251
252 /**
253 * Strips and renders nowiki, pre, math, hiero
254 * If $render is set, performs necessary rendering operations on plugins
255 * Returns the text, and fills an array with data needed in unstrip()
256 * If the $state is already a valid strip state, it adds to the state
257 *
258 * @param bool $stripcomments when set, HTML comments <!-- like this -->
259 * will be stripped in addition to other tags. This is important
260 * for section editing, where these comments cause confusion when
261 * counting the sections in the wikisource
262 *
263 * @access private
264 */
265 function strip( $text, &$state, $stripcomments = false ) {
266 $render = ($this->mOutputType == OT_HTML);
267 $html_content = array();
268 $nowiki_content = array();
269 $math_content = array();
270 $pre_content = array();
271 $comment_content = array();
272 $ext_content = array();
273
274 # Replace any instances of the placeholders
275 $uniq_prefix = UNIQ_PREFIX;
276 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
277
278 # html
279 global $wgRawHtml, $wgWhitelistEdit;
280 if( $wgRawHtml && $wgWhitelistEdit ) {
281 $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
282 foreach( $html_content as $marker => $content ) {
283 if ($render ) {
284 # Raw and unchecked for validity.
285 $html_content[$marker] = $content;
286 } else {
287 $html_content[$marker] = '<html>'.$content.'</html>';
288 }
289 }
290 }
291
292 # nowiki
293 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
294 foreach( $nowiki_content as $marker => $content ) {
295 if( $render ){
296 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
297 } else {
298 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
299 }
300 }
301
302 # math
303 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
304 foreach( $math_content as $marker => $content ){
305 if( $render ) {
306 if( $this->mOptions->getUseTeX() ) {
307 $math_content[$marker] = renderMath( $content );
308 } else {
309 $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
310 }
311 } else {
312 $math_content[$marker] = '<math>'.$content.'</math>';
313 }
314 }
315
316 # pre
317 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
318 foreach( $pre_content as $marker => $content ){
319 if( $render ){
320 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
321 } else {
322 $pre_content[$marker] = '<pre>'.$content.'</pre>';
323 }
324 }
325
326 # Comments
327 if($stripcomments) {
328 $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
329 foreach( $comment_content as $marker => $content ){
330 $comment_content[$marker] = '<!--'.$content.'-->';
331 }
332 }
333
334 # Extensions
335 foreach ( $this->mTagHooks as $tag => $callback ) {
336 $ext_contents[$tag] = array();
337 $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
338 foreach( $ext_content[$tag] as $marker => $content ) {
339 if ( $render ) {
340 $ext_content[$tag][$marker] = $callback( $content );
341 } else {
342 $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
343 }
344 }
345 }
346
347 # Merge state with the pre-existing state, if there is one
348 if ( $state ) {
349 $state['html'] = $state['html'] + $html_content;
350 $state['nowiki'] = $state['nowiki'] + $nowiki_content;
351 $state['math'] = $state['math'] + $math_content;
352 $state['pre'] = $state['pre'] + $pre_content;
353 $state['comment'] = $state['comment'] + $comment_content;
354
355 foreach( $ext_content as $tag => $array ) {
356 if ( array_key_exists( $tag, $state ) ) {
357 $state[$tag] = $state[$tag] + $array;
358 }
359 }
360 } else {
361 $state = array(
362 'html' => $html_content,
363 'nowiki' => $nowiki_content,
364 'math' => $math_content,
365 'pre' => $pre_content,
366 'comment' => $comment_content,
367 ) + $ext_content;
368 }
369 return $text;
370 }
371
372 /**
373 * restores pre, math, and heiro removed by strip()
374 *
375 * always call unstripNoWiki() after this one
376 * @access private
377 */
378 function unstrip( $text, &$state ) {
379 # Must expand in reverse order, otherwise nested tags will be corrupted
380 $contentDict = end( $state );
381 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
382 if( key($state) != 'nowiki' && key($state) != 'html') {
383 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
384 $text = str_replace( key( $contentDict ), $content, $text );
385 }
386 }
387 }
388
389 return $text;
390 }
391
392 /**
393 * always call this after unstrip() to preserve the order
394 *
395 * @access private
396 */
397 function unstripNoWiki( $text, &$state ) {
398 # Must expand in reverse order, otherwise nested tags will be corrupted
399 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
400 $text = str_replace( key( $state['nowiki'] ), $content, $text );
401 }
402
403 global $wgRawHtml;
404 if ($wgRawHtml) {
405 for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
406 $text = str_replace( key( $state['html'] ), $content, $text );
407 }
408 }
409
410 return $text;
411 }
412
413 /**
414 * Add an item to the strip state
415 * Returns the unique tag which must be inserted into the stripped text
416 * The tag will be replaced with the original text in unstrip()
417 *
418 * @access private
419 */
420 function insertStripItem( $text, &$state ) {
421 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
422 if ( !$state ) {
423 $state = array(
424 'html' => array(),
425 'nowiki' => array(),
426 'math' => array(),
427 'pre' => array()
428 );
429 }
430 $state['item'][$rnd] = $text;
431 return $rnd;
432 }
433
434 /**
435 * Return allowed HTML attributes
436 *
437 * @access private
438 */
439 function getHTMLattrs () {
440 $htmlattrs = array( # Allowed attributes--no scripting, etc.
441 'title', 'align', 'lang', 'dir', 'width', 'height',
442 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
443 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
444 /* FONT */ 'type', 'start', 'value', 'compact',
445 /* For various lists, mostly deprecated but safe */
446 'summary', 'width', 'border', 'frame', 'rules',
447 'cellspacing', 'cellpadding', 'valign', 'char',
448 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
449 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
450 'id', 'class', 'name', 'style' /* For CSS */
451 );
452 return $htmlattrs ;
453 }
454
455 /**
456 * Remove non approved attributes and javascript in css
457 *
458 * @access private
459 */
460 function fixTagAttributes ( $t ) {
461 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
462 $htmlattrs = $this->getHTMLattrs() ;
463
464 # Strip non-approved attributes from the tag
465 $t = preg_replace(
466 '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
467 "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
468 $t);
469
470 $t = str_replace ( '<></>' , '' , $t ) ; # This should fix bug 980557
471
472 # Strip javascript "expression" from stylesheets. Brute force approach:
473 # If anythin offensive is found, all attributes of the HTML tag are dropped
474
475 if( preg_match(
476 '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
477 wfMungeToUtf8( $t ) ) )
478 {
479 $t='';
480 }
481
482 return trim ( $t ) ;
483 }
484
485 /**
486 * interface with html tidy, used if $wgUseTidy = true
487 *
488 * @access private
489 */
490 function tidy ( $text ) {
491 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
492 global $wgInputEncoding, $wgOutputEncoding;
493 $fname = 'Parser::tidy';
494 wfProfileIn( $fname );
495
496 $cleansource = '';
497 switch(strtoupper($wgOutputEncoding)) {
498 case 'ISO-8859-1':
499 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
500 break;
501 case 'UTF-8':
502 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
503 break;
504 default:
505 $wgTidyOpts .= ' -raw';
506 }
507
508 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
509 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
510 '<head><title>test</title></head><body>'.$text.'</body></html>';
511 $descriptorspec = array(
512 0 => array('pipe', 'r'),
513 1 => array('pipe', 'w'),
514 2 => array('file', '/dev/null', 'a')
515 );
516 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
517 if (is_resource($process)) {
518 fwrite($pipes[0], $wrappedtext);
519 fclose($pipes[0]);
520 while (!feof($pipes[1])) {
521 $cleansource .= fgets($pipes[1], 1024);
522 }
523 fclose($pipes[1]);
524 $return_value = proc_close($process);
525 }
526
527 wfProfileOut( $fname );
528
529 if( $cleansource == '' && $text != '') {
530 wfDebug( "Tidy error detected!\n" );
531 return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
532 } else {
533 return $cleansource;
534 }
535 }
536
537 /**
538 * parse the wiki syntax used to render tables
539 *
540 * @access private
541 */
542 function doTableStuff ( $t ) {
543 $fname = 'Parser::doTableStuff';
544 wfProfileIn( $fname );
545
546 $t = explode ( "\n" , $t ) ;
547 $td = array () ; # Is currently a td tag open?
548 $ltd = array () ; # Was it TD or TH?
549 $tr = array () ; # Is currently a tr tag open?
550 $ltr = array () ; # tr attributes
551 $indent_level = 0; # indent level of the table
552 foreach ( $t AS $k => $x )
553 {
554 $x = trim ( $x ) ;
555 $fc = substr ( $x , 0 , 1 ) ;
556 if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
557 $indent_level = strlen( $matches[1] );
558 $t[$k] = "\n" .
559 str_repeat( '<dl><dd>', $indent_level ) .
560 '<table ' . $this->fixTagAttributes ( $matches[2] ) . '>' ;
561 array_push ( $td , false ) ;
562 array_push ( $ltd , '' ) ;
563 array_push ( $tr , false ) ;
564 array_push ( $ltr , '' ) ;
565 }
566 else if ( count ( $td ) == 0 ) { } # Don't do any of the following
567 else if ( '|}' == substr ( $x , 0 , 2 ) ) {
568 $z = "</table>\n" ;
569 $l = array_pop ( $ltd ) ;
570 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
571 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
572 array_pop ( $ltr ) ;
573 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
574 }
575 else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
576 $x = substr ( $x , 1 ) ;
577 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
578 $z = '' ;
579 $l = array_pop ( $ltd ) ;
580 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
581 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
582 array_pop ( $ltr ) ;
583 $t[$k] = $z ;
584 array_push ( $tr , false ) ;
585 array_push ( $td , false ) ;
586 array_push ( $ltd , '' ) ;
587 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
588 }
589 else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
590 if ( '|+' == substr ( $x , 0 , 2 ) ) {
591 $fc = '+' ;
592 $x = substr ( $x , 1 ) ;
593 }
594 $after = substr ( $x , 1 ) ;
595 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
596 $after = explode ( '||' , $after ) ;
597 $t[$k] = '' ;
598 foreach ( $after AS $theline )
599 {
600 $z = '' ;
601 if ( $fc != '+' )
602 {
603 $tra = array_pop ( $ltr ) ;
604 if ( !array_pop ( $tr ) ) $z = '<tr '.$tra.">\n" ;
605 array_push ( $tr , true ) ;
606 array_push ( $ltr , '' ) ;
607 }
608
609 $l = array_pop ( $ltd ) ;
610 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
611 if ( $fc == '|' ) $l = 'td' ;
612 else if ( $fc == '!' ) $l = 'th' ;
613 else if ( $fc == '+' ) $l = 'caption' ;
614 else $l = '' ;
615 array_push ( $ltd , $l ) ;
616 $y = explode ( '|' , $theline , 2 ) ;
617 if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
618 else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
619 $t[$k] .= $y ;
620 array_push ( $td , true ) ;
621 }
622 }
623 }
624
625 # Closing open td, tr && table
626 while ( count ( $td ) > 0 )
627 {
628 if ( array_pop ( $td ) ) $t[] = '</td>' ;
629 if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
630 $t[] = '</table>' ;
631 }
632
633 $t = implode ( "\n" , $t ) ;
634 # $t = $this->removeHTMLtags( $t );
635 wfProfileOut( $fname );
636 return $t ;
637 }
638
639 /**
640 * Helper function for parse() that transforms wiki markup into
641 * HTML. Only called for $mOutputType == OT_HTML.
642 *
643 * @access private
644 */
645 function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
646 global $wgContLang;
647
648 $fname = 'Parser::internalParse';
649 wfProfileIn( $fname );
650
651 $text = $this->removeHTMLtags( $text );
652 $text = $this->replaceVariables( $text, $args );
653
654 $text = $wgContLang->convert($text);
655
656 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
657
658 $text = $this->doHeadings( $text );
659 if($this->mOptions->getUseDynamicDates()) {
660 global $wgDateFormatter;
661 $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
662 }
663 $text = $this->doAllQuotes( $text );
664 $text = $this->replaceExternalLinks( $text );
665 $text = $this->doMagicLinks( $text );
666 $text = $this->replaceInternalLinks ( $text );
667 # Another call to replace links and images inside captions of images
668 $text = $this->replaceInternalLinks ( $text );
669
670 $text = $this->doTableStuff( $text );
671 $text = $this->formatHeadings( $text, $isMain );
672 $sk =& $this->mOptions->getSkin();
673 $text = $sk->transformContent( $text );
674
675 wfProfileOut( $fname );
676 return $text;
677 }
678
679 /**
680 * Replace special strings like "ISBN xxx" and "RFC xxx" with
681 * magic external links.
682 *
683 * @access private
684 */
685 function &doMagicLinks( &$text ) {
686 global $wgUseGeoMode;
687 $text = $this->magicISBN( $text );
688 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
689 $text = $this->magicGEO( $text );
690 }
691 $text = $this->magicRFC( $text );
692 return $text;
693 }
694
695 /**
696 * Parse ^^ tokens and return html
697 *
698 * @access private
699 */
700 function doExponent ( $text ) {
701 $fname = 'Parser::doExponent';
702 wfProfileIn( $fname);
703 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
704 wfProfileOut( $fname);
705 return $text;
706 }
707
708 /**
709 * Parse headers and return html
710 *
711 * @access private
712 */
713 function doHeadings( $text ) {
714 $fname = 'Parser::doHeadings';
715 wfProfileIn( $fname );
716 for ( $i = 6; $i >= 1; --$i ) {
717 $h = substr( '======', 0, $i );
718 $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
719 "<h{$i}>\\1</h{$i}>\\2", $text );
720 }
721 wfProfileOut( $fname );
722 return $text;
723 }
724
725 /**
726 * Replace single quotes with HTML markup
727 * @access private
728 * @return string the altered text
729 */
730 function doAllQuotes( $text ) {
731 $fname = 'Parser::doAllQuotes';
732 wfProfileIn( $fname );
733 $outtext = '';
734 $lines = explode( "\n", $text );
735 foreach ( $lines as $line ) {
736 $outtext .= $this->doQuotes ( $line ) . "\n";
737 }
738 $outtext = substr($outtext, 0,-1);
739 wfProfileOut( $fname );
740 return $outtext;
741 }
742
743 /**
744 * Helper function for doAllQuotes()
745 * @access private
746 */
747 function doQuotes( $text ) {
748 $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
749 if (count ($arr) == 1)
750 return $text;
751 else
752 {
753 # First, do some preliminary work. This may shift some apostrophes from
754 # being mark-up to being text. It also counts the number of occurrences
755 # of bold and italics mark-ups.
756 $i = 0;
757 $numbold = 0;
758 $numitalics = 0;
759 foreach ($arr as $r)
760 {
761 if (($i % 2) == 1)
762 {
763 # If there are ever four apostrophes, assume the first is supposed to
764 # be text, and the remaining three constitute mark-up for bold text.
765 if (strlen ($arr[$i]) == 4)
766 {
767 $arr[$i-1] .= "'";
768 $arr[$i] = "'''";
769 }
770 # If there are more than 5 apostrophes in a row, assume they're all
771 # text except for the last 5.
772 else if (strlen ($arr[$i]) > 5)
773 {
774 $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
775 $arr[$i] = "'''''";
776 }
777 # Count the number of occurrences of bold and italics mark-ups.
778 # We are not counting sequences of five apostrophes.
779 if (strlen ($arr[$i]) == 2) $numitalics++; else
780 if (strlen ($arr[$i]) == 3) $numbold++; else
781 if (strlen ($arr[$i]) == 5) { $numitalics++; $numbold++; }
782 }
783 $i++;
784 }
785
786 # If there is an odd number of both bold and italics, it is likely
787 # that one of the bold ones was meant to be an apostrophe followed
788 # by italics. Which one we cannot know for certain, but it is more
789 # likely to be one that has a single-letter word before it.
790 if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
791 {
792 $i = 0;
793 $firstsingleletterword = -1;
794 $firstmultiletterword = -1;
795 $firstspace = -1;
796 foreach ($arr as $r)
797 {
798 if (($i % 2 == 1) and (strlen ($r) == 3))
799 {
800 $x1 = substr ($arr[$i-1], -1);
801 $x2 = substr ($arr[$i-1], -2, 1);
802 if ($x1 == ' ') {
803 if ($firstspace == -1) $firstspace = $i;
804 } else if ($x2 == ' ') {
805 if ($firstsingleletterword == -1) $firstsingleletterword = $i;
806 } else {
807 if ($firstmultiletterword == -1) $firstmultiletterword = $i;
808 }
809 }
810 $i++;
811 }
812
813 # If there is a single-letter word, use it!
814 if ($firstsingleletterword > -1)
815 {
816 $arr [ $firstsingleletterword ] = "''";
817 $arr [ $firstsingleletterword-1 ] .= "'";
818 }
819 # If not, but there's a multi-letter word, use that one.
820 else if ($firstmultiletterword > -1)
821 {
822 $arr [ $firstmultiletterword ] = "''";
823 $arr [ $firstmultiletterword-1 ] .= "'";
824 }
825 # ... otherwise use the first one that has neither.
826 # (notice that it is possible for all three to be -1 if, for example,
827 # there is only one pentuple-apostrophe in the line)
828 else if ($firstspace > -1)
829 {
830 $arr [ $firstspace ] = "''";
831 $arr [ $firstspace-1 ] .= "'";
832 }
833 }
834
835 # Now let's actually convert our apostrophic mush to HTML!
836 $output = '';
837 $buffer = '';
838 $state = '';
839 $i = 0;
840 foreach ($arr as $r)
841 {
842 if (($i % 2) == 0)
843 {
844 if ($state == 'both')
845 $buffer .= $r;
846 else
847 $output .= $r;
848 }
849 else
850 {
851 if (strlen ($r) == 2)
852 {
853 if ($state == 'i')
854 { $output .= '</i>'; $state = ''; }
855 else if ($state == 'bi')
856 { $output .= '</i>'; $state = 'b'; }
857 else if ($state == 'ib')
858 { $output .= '</b></i><b>'; $state = 'b'; }
859 else if ($state == 'both')
860 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
861 else # $state can be 'b' or ''
862 { $output .= '<i>'; $state .= 'i'; }
863 }
864 else if (strlen ($r) == 3)
865 {
866 if ($state == 'b')
867 { $output .= '</b>'; $state = ''; }
868 else if ($state == 'bi')
869 { $output .= '</i></b><i>'; $state = 'i'; }
870 else if ($state == 'ib')
871 { $output .= '</b>'; $state = 'i'; }
872 else if ($state == 'both')
873 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
874 else # $state can be 'i' or ''
875 { $output .= '<b>'; $state .= 'b'; }
876 }
877 else if (strlen ($r) == 5)
878 {
879 if ($state == 'b')
880 { $output .= '</b><i>'; $state = 'i'; }
881 else if ($state == 'i')
882 { $output .= '</i><b>'; $state = 'b'; }
883 else if ($state == 'bi')
884 { $output .= '</i></b>'; $state = ''; }
885 else if ($state == 'ib')
886 { $output .= '</b></i>'; $state = ''; }
887 else if ($state == 'both')
888 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
889 else # ($state == '')
890 { $buffer = ''; $state = 'both'; }
891 }
892 }
893 $i++;
894 }
895 # Now close all remaining tags. Notice that the order is important.
896 if ($state == 'b' || $state == 'ib')
897 $output .= '</b>';
898 if ($state == 'i' || $state == 'bi' || $state == 'ib')
899 $output .= '</i>';
900 if ($state == 'bi')
901 $output .= '</b>';
902 if ($state == 'both')
903 $output .= '<b><i>'.$buffer.'</i></b>';
904 return $output;
905 }
906 }
907
908 /**
909 * Replace external links
910 *
911 * Note: we have to do external links before the internal ones,
912 * and otherwise take great care in the order of things here, so
913 * that we don't end up interpreting some URLs twice.
914 *
915 * @access private
916 */
917 function replaceExternalLinks( $text ) {
918 $fname = 'Parser::replaceExternalLinks';
919 wfProfileIn( $fname );
920
921 $sk =& $this->mOptions->getSkin();
922 $linktrail = wfMsg('linktrail');
923 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
924
925 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
926
927 $i = 0;
928 while ( $i<count( $bits ) ) {
929 $url = $bits[$i++];
930 $protocol = $bits[$i++];
931 $text = $bits[$i++];
932 $trail = $bits[$i++];
933
934 # If the link text is an image URL, replace it with an <img> tag
935 # This happened by accident in the original parser, but some people used it extensively
936 $img = $this->maybeMakeImageLink( $text );
937 if ( $img !== false ) {
938 $text = $img;
939 }
940
941 $dtrail = '';
942
943 # No link text, e.g. [http://domain.tld/some.link]
944 if ( $text == '' ) {
945 # Autonumber if allowed
946 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
947 $text = '[' . ++$this->mAutonumber . ']';
948 } else {
949 # Otherwise just use the URL
950 $text = htmlspecialchars( $url );
951 }
952 } else {
953 # Have link text, e.g. [http://domain.tld/some.link text]s
954 # Check for trail
955 if ( preg_match( $linktrail, $trail, $m2 ) ) {
956 $dtrail = $m2[1];
957 $trail = $m2[2];
958 }
959 }
960
961 $encUrl = htmlspecialchars( $url );
962 # Bit in parentheses showing the URL for the printable version
963 if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
964 $paren = '';
965 } else {
966 # Expand the URL for printable version
967 if ( ! $sk->suppressUrlExpansion() ) {
968 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
969 } else {
970 $paren = '';
971 }
972 }
973
974 # Process the trail (i.e. everything after this link up until start of the next link),
975 # replacing any non-bracketed links
976 $trail = $this->replaceFreeExternalLinks( $trail );
977
978 $la = $sk->getExternalLinkAttributes( $url, $text );
979
980 # Use the encoded URL
981 # This means that users can paste URLs directly into the text
982 # Funny characters like &ouml; aren't valid in URLs anyway
983 # This was changed in August 2004
984 $s .= "<a href=\"{$url}\"{$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
985 }
986
987 wfProfileOut( $fname );
988 return $s;
989 }
990
991 /**
992 * Replace anything that looks like a URL with a link
993 * @access private
994 */
995 function replaceFreeExternalLinks( $text ) {
996 $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
997 $s = array_shift( $bits );
998 $i = 0;
999
1000 $sk =& $this->mOptions->getSkin();
1001
1002 while ( $i < count( $bits ) ){
1003 $protocol = $bits[$i++];
1004 $remainder = $bits[$i++];
1005
1006 if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1007 # Found some characters after the protocol that look promising
1008 $url = $protocol . $m[1];
1009 $trail = $m[2];
1010
1011 # Move trailing punctuation to $trail
1012 $sep = ',;\.:!?';
1013 # If there is no left bracket, then consider right brackets fair game too
1014 if ( strpos( $url, '(' ) === false ) {
1015 $sep .= ')';
1016 }
1017
1018 $numSepChars = strspn( strrev( $url ), $sep );
1019 if ( $numSepChars ) {
1020 $trail = substr( $url, -$numSepChars ) . $trail;
1021 $url = substr( $url, 0, -$numSepChars );
1022 }
1023
1024 # Replace &amp; from obsolete syntax with &
1025 $url = str_replace( '&amp;', '&', $url );
1026
1027 # Is this an external image?
1028 $text = $this->maybeMakeImageLink( $url );
1029 if ( $text === false ) {
1030 # Not an image, make a link
1031 $text = $sk->makeExternalLink( $url, $url );
1032 }
1033 $s .= $text . $trail;
1034 } else {
1035 $s .= $protocol . $remainder;
1036 }
1037 }
1038 return $s;
1039 }
1040
1041 /**
1042 * make an image if it's allowed
1043 * @access private
1044 */
1045 function maybeMakeImageLink( $url ) {
1046 $sk =& $this->mOptions->getSkin();
1047 $text = false;
1048 if ( $this->mOptions->getAllowExternalImages() ) {
1049 if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1050 # Image found
1051 $text = $sk->makeImage( htmlspecialchars( $url ) );
1052 }
1053 }
1054 return $text;
1055 }
1056
1057 /**
1058 * Process [[ ]] wikilinks
1059 *
1060 * @access private
1061 */
1062 function replaceInternalLinks( $s ) {
1063 global $wgLang, $wgLinkCache;
1064 global $wgNamespacesWithSubpages;
1065 static $fname = 'Parser::replaceInternalLinks' ;
1066 wfProfileIn( $fname );
1067
1068 wfProfileIn( $fname.'-setup' );
1069 static $tc = FALSE;
1070 # the % is needed to support urlencoded titles as well
1071 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1072 $sk =& $this->mOptions->getSkin();
1073
1074 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1075
1076 $a = explode( '[[', ' ' . $s );
1077 $s = array_shift( $a );
1078 $s = substr( $s, 1 );
1079
1080 # Match a link having the form [[namespace:link|alternate]]trail
1081 static $e1 = FALSE;
1082 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1083 # Match the end of a line for a word that's not followed by whitespace,
1084 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1085 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1086
1087 $useLinkPrefixExtension = $wgLang->linkPrefixExtension();
1088 # Special and Media are pseudo-namespaces; no pages actually exist in them
1089
1090 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1091
1092 if ( $useLinkPrefixExtension ) {
1093 if ( preg_match( $e2, $s, $m ) ) {
1094 $first_prefix = $m[2];
1095 $s = $m[1];
1096 } else {
1097 $first_prefix = false;
1098 }
1099 } else {
1100 $prefix = '';
1101 }
1102
1103 wfProfileOut( $fname.'-setup' );
1104
1105 # start procedeeding each line
1106 foreach ( $a as $line ) {
1107 wfProfileIn( $fname.'-prefixhandling' );
1108 if ( $useLinkPrefixExtension ) {
1109 if ( preg_match( $e2, $s, $m ) ) {
1110 $prefix = $m[2];
1111 $s = $m[1];
1112 } else {
1113 $prefix='';
1114 }
1115 # first link
1116 if($first_prefix) {
1117 $prefix = $first_prefix;
1118 $first_prefix = false;
1119 }
1120 }
1121 wfProfileOut( $fname.'-prefixhandling' );
1122
1123 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1124 $text = $m[2];
1125 # fix up urlencoded title texts
1126 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1127 $trail = $m[3];
1128 } else { # Invalid form; output directly
1129 $s .= $prefix . '[[' . $line ;
1130 continue;
1131 }
1132
1133 # Valid link forms:
1134 # Foobar -- normal
1135 # :Foobar -- override special treatment of prefix (images, language links)
1136 # /Foobar -- convert to CurrentPage/Foobar
1137 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1138
1139 # Look at the first character
1140 $c = substr($m[1],0,1);
1141 $noforce = ($c != ':');
1142
1143 # subpage
1144 if( $c == '/' ) {
1145 # / at end means we don't want the slash to be shown
1146 if(substr($m[1],-1,1)=='/') {
1147 $m[1]=substr($m[1],1,strlen($m[1])-2);
1148 $noslash=$m[1];
1149 } else {
1150 $noslash=substr($m[1],1);
1151 }
1152
1153 # Some namespaces don't allow subpages
1154 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) {
1155 # subpages allowed here
1156 $link = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1157 if( '' == $text ) {
1158 $text= $m[1];
1159 } # this might be changed for ugliness reasons
1160 } else {
1161 # no subpage allowed, use standard link
1162 $link = $noslash;
1163 }
1164
1165 } elseif( $noforce ) { # no subpage
1166 $link = $m[1];
1167 } else {
1168 # We don't want to keep the first character
1169 $link = substr( $m[1], 1 );
1170 }
1171
1172 $wasblank = ( '' == $text );
1173 if( $wasblank ) $text = $link;
1174
1175 $nt = Title::newFromText( $link );
1176 if( !$nt ) {
1177 $s .= $prefix . '[[' . $line;
1178 continue;
1179 }
1180
1181 $ns = $nt->getNamespace();
1182 $iw = $nt->getInterWiki();
1183
1184 # Link not escaped by : , create the various objects
1185 if( $noforce ) {
1186
1187 # Interwikis
1188 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
1189 array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1190 $tmp = $prefix . $trail ;
1191 $s .= (trim($tmp) == '')? '': $tmp;
1192 continue;
1193 }
1194
1195 if ( $ns == NS_IMAGE ) {
1196 $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1197 $wgLinkCache->addImageLinkObj( $nt );
1198 continue;
1199 }
1200
1201 if ( $ns == NS_CATEGORY ) {
1202 $t = $nt->getText() ;
1203 $nnt = Title::newFromText ( Namespace::getCanonicalName(NS_CATEGORY).':'.$t ) ;
1204
1205 $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1206 $pPLC=$sk->postParseLinkColour();
1207 $sk->postParseLinkColour( false );
1208 $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1209 $sk->postParseLinkColour( $pPLC );
1210 $wgLinkCache->resume();
1211
1212 if ( $wasblank ) {
1213 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1214 $sortkey = $this->mTitle->getText();
1215 } else {
1216 $sortkey = $this->mTitle->getPrefixedText();
1217 }
1218 } else {
1219 $sortkey = $text;
1220 }
1221 $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1222 $this->mOutput->mCategoryLinks[] = $t ;
1223 $s .= $prefix . $trail ;
1224 continue;
1225 }
1226 }
1227
1228 if( ( $nt->getPrefixedText() === $this->mTitle->getPrefixedText() ) &&
1229 ( strpos( $link, '#' ) === FALSE ) ) {
1230 # Self-links are handled specially; generally de-link and change to bold.
1231 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1232 continue;
1233 }
1234
1235 if( $ns == NS_MEDIA ) {
1236 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1237 $wgLinkCache->addImageLinkObj( $nt );
1238 continue;
1239 } elseif( $ns == NS_SPECIAL ) {
1240 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1241 continue;
1242 }
1243 $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1244 }
1245 wfProfileOut( $fname );
1246 return $s;
1247 }
1248
1249 /**#@+
1250 * Used by doBlockLevels()
1251 * @access private
1252 */
1253 /* private */ function closeParagraph() {
1254 $result = '';
1255 if ( '' != $this->mLastSection ) {
1256 $result = '</' . $this->mLastSection . ">\n";
1257 }
1258 $this->mInPre = false;
1259 $this->mLastSection = '';
1260 return $result;
1261 }
1262 # getCommon() returns the length of the longest common substring
1263 # of both arguments, starting at the beginning of both.
1264 #
1265 /* private */ function getCommon( $st1, $st2 ) {
1266 $fl = strlen( $st1 );
1267 $shorter = strlen( $st2 );
1268 if ( $fl < $shorter ) { $shorter = $fl; }
1269
1270 for ( $i = 0; $i < $shorter; ++$i ) {
1271 if ( $st1{$i} != $st2{$i} ) { break; }
1272 }
1273 return $i;
1274 }
1275 # These next three functions open, continue, and close the list
1276 # element appropriate to the prefix character passed into them.
1277 #
1278 /* private */ function openList( $char ) {
1279 $result = $this->closeParagraph();
1280
1281 if ( '*' == $char ) { $result .= '<ul><li>'; }
1282 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1283 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1284 else if ( ';' == $char ) {
1285 $result .= '<dl><dt>';
1286 $this->mDTopen = true;
1287 }
1288 else { $result = '<!-- ERR 1 -->'; }
1289
1290 return $result;
1291 }
1292
1293 /* private */ function nextItem( $char ) {
1294 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1295 else if ( ':' == $char || ';' == $char ) {
1296 $close = '</dd>';
1297 if ( $this->mDTopen ) { $close = '</dt>'; }
1298 if ( ';' == $char ) {
1299 $this->mDTopen = true;
1300 return $close . '<dt>';
1301 } else {
1302 $this->mDTopen = false;
1303 return $close . '<dd>';
1304 }
1305 }
1306 return '<!-- ERR 2 -->';
1307 }
1308
1309 /* private */ function closeList( $char ) {
1310 if ( '*' == $char ) { $text = '</li></ul>'; }
1311 else if ( '#' == $char ) { $text = '</li></ol>'; }
1312 else if ( ':' == $char ) {
1313 if ( $this->mDTopen ) {
1314 $this->mDTopen = false;
1315 $text = '</dt></dl>';
1316 } else {
1317 $text = '</dd></dl>';
1318 }
1319 }
1320 else { return '<!-- ERR 3 -->'; }
1321 return $text."\n";
1322 }
1323 /**#@-*/
1324
1325 /**
1326 * Make lists from lines starting with ':', '*', '#', etc.
1327 *
1328 * @access private
1329 * @return string the lists rendered as HTML
1330 */
1331 function doBlockLevels( $text, $linestart ) {
1332 $fname = 'Parser::doBlockLevels';
1333 wfProfileIn( $fname );
1334
1335 # Parsing through the text line by line. The main thing
1336 # happening here is handling of block-level elements p, pre,
1337 # and making lists from lines starting with * # : etc.
1338 #
1339 $textLines = explode( "\n", $text );
1340
1341 $lastPrefix = $output = $lastLine = '';
1342 $this->mDTopen = $inBlockElem = false;
1343 $prefixLength = 0;
1344 $paragraphStack = false;
1345
1346 if ( !$linestart ) {
1347 $output .= array_shift( $textLines );
1348 }
1349 foreach ( $textLines as $oLine ) {
1350 $lastPrefixLength = strlen( $lastPrefix );
1351 $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1352 $preOpenMatch = preg_match('/<pre/i', $oLine );
1353 if ( !$this->mInPre ) {
1354 # Multiple prefixes may abut each other for nested lists.
1355 $prefixLength = strspn( $oLine, '*#:;' );
1356 $pref = substr( $oLine, 0, $prefixLength );
1357
1358 # eh?
1359 $pref2 = str_replace( ';', ':', $pref );
1360 $t = substr( $oLine, $prefixLength );
1361 $this->mInPre = !empty($preOpenMatch);
1362 } else {
1363 # Don't interpret any other prefixes in preformatted text
1364 $prefixLength = 0;
1365 $pref = $pref2 = '';
1366 $t = $oLine;
1367 }
1368
1369 # List generation
1370 if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1371 # Same as the last item, so no need to deal with nesting or opening stuff
1372 $output .= $this->nextItem( substr( $pref, -1 ) );
1373 $paragraphStack = false;
1374
1375 if ( substr( $pref, -1 ) == ';') {
1376 # The one nasty exception: definition lists work like this:
1377 # ; title : definition text
1378 # So we check for : in the remainder text to split up the
1379 # title and definition, without b0rking links.
1380 # FIXME: This is not foolproof. Something better in Tokenizer might help.
1381 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1382 $term = $match[1];
1383 $output .= $term . $this->nextItem( ':' );
1384 $t = $match[2];
1385 }
1386 }
1387 } elseif( $prefixLength || $lastPrefixLength ) {
1388 # Either open or close a level...
1389 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1390 $paragraphStack = false;
1391
1392 while( $commonPrefixLength < $lastPrefixLength ) {
1393 $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1394 --$lastPrefixLength;
1395 }
1396 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1397 $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1398 }
1399 while ( $prefixLength > $commonPrefixLength ) {
1400 $char = substr( $pref, $commonPrefixLength, 1 );
1401 $output .= $this->openList( $char );
1402
1403 if ( ';' == $char ) {
1404 # FIXME: This is dupe of code above
1405 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1406 $term = $match[1];
1407 $output .= $term . $this->nextItem( ':' );
1408 $t = $match[2];
1409 }
1410 }
1411 ++$commonPrefixLength;
1412 }
1413 $lastPrefix = $pref2;
1414 }
1415 if( 0 == $prefixLength ) {
1416 # No prefix (not in list)--go to paragraph mode
1417 $uniq_prefix = UNIQ_PREFIX;
1418 // XXX: use a stack for nestable elements like span, table and div
1419 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1420 $closematch = preg_match(
1421 '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1422 '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1423 if ( $openmatch or $closematch ) {
1424 $paragraphStack = false;
1425 $output .= $this->closeParagraph();
1426 if($preOpenMatch and !$preCloseMatch) {
1427 $this->mInPre = true;
1428 }
1429 if ( $closematch ) {
1430 $inBlockElem = false;
1431 } else {
1432 $inBlockElem = true;
1433 }
1434 } else if ( !$inBlockElem && !$this->mInPre ) {
1435 if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1436 // pre
1437 if ($this->mLastSection != 'pre') {
1438 $paragraphStack = false;
1439 $output .= $this->closeParagraph().'<pre>';
1440 $this->mLastSection = 'pre';
1441 }
1442 $t = substr( $t, 1 );
1443 } else {
1444 // paragraph
1445 if ( '' == trim($t) ) {
1446 if ( $paragraphStack ) {
1447 $output .= $paragraphStack.'<br />';
1448 $paragraphStack = false;
1449 $this->mLastSection = 'p';
1450 } else {
1451 if ($this->mLastSection != 'p' ) {
1452 $output .= $this->closeParagraph();
1453 $this->mLastSection = '';
1454 $paragraphStack = '<p>';
1455 } else {
1456 $paragraphStack = '</p><p>';
1457 }
1458 }
1459 } else {
1460 if ( $paragraphStack ) {
1461 $output .= $paragraphStack;
1462 $paragraphStack = false;
1463 $this->mLastSection = 'p';
1464 } else if ($this->mLastSection != 'p') {
1465 $output .= $this->closeParagraph().'<p>';
1466 $this->mLastSection = 'p';
1467 }
1468 }
1469 }
1470 }
1471 }
1472 if ($paragraphStack === false) {
1473 $output .= $t."\n";
1474 }
1475 }
1476 while ( $prefixLength ) {
1477 $output .= $this->closeList( $pref2{$prefixLength-1} );
1478 --$prefixLength;
1479 }
1480 if ( '' != $this->mLastSection ) {
1481 $output .= '</' . $this->mLastSection . '>';
1482 $this->mLastSection = '';
1483 }
1484
1485 wfProfileOut( $fname );
1486 return $output;
1487 }
1488
1489 /**
1490 * Return value of a magic variable (like PAGENAME)
1491 *
1492 * @access private
1493 */
1494 function getVariableValue( $index ) {
1495 global $wgLang, $wgSitename, $wgServer;
1496
1497 switch ( $index ) {
1498 case MAG_CURRENTMONTH:
1499 return $wgLang->formatNum( date( 'm' ) );
1500 case MAG_CURRENTMONTHNAME:
1501 return $wgLang->getMonthName( date('n') );
1502 case MAG_CURRENTMONTHNAMEGEN:
1503 return $wgLang->getMonthNameGen( date('n') );
1504 case MAG_CURRENTDAY:
1505 return $wgLang->formatNum( date('j') );
1506 case MAG_PAGENAME:
1507 return $this->mTitle->getText();
1508 case MAG_PAGENAMEE:
1509 return $this->mTitle->getPartialURL();
1510 case MAG_NAMESPACE:
1511 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1512 return $wgLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1513 case MAG_CURRENTDAYNAME:
1514 return $wgLang->getWeekdayName( date('w')+1 );
1515 case MAG_CURRENTYEAR:
1516 return $wgLang->formatNum( date( 'Y' ) );
1517 case MAG_CURRENTTIME:
1518 return $wgLang->time( wfTimestampNow(), false );
1519 case MAG_NUMBEROFARTICLES:
1520 return $wgLang->formatNum( wfNumberOfArticles() );
1521 case MAG_SITENAME:
1522 return $wgSitename;
1523 case MAG_SERVER:
1524 return $wgServer;
1525 default:
1526 return NULL;
1527 }
1528 }
1529
1530 /**
1531 * initialise the magic variables (like CURRENTMONTHNAME)
1532 *
1533 * @access private
1534 */
1535 function initialiseVariables() {
1536 global $wgVariableIDs;
1537 $this->mVariables = array();
1538 foreach ( $wgVariableIDs as $id ) {
1539 $mw =& MagicWord::get( $id );
1540 $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1541 }
1542 }
1543
1544 /**
1545 * Replace magic variables, templates, and template arguments
1546 * with the appropriate text. Templates are substituted recursively,
1547 * taking care to avoid infinite loops.
1548 *
1549 * Note that the substitution depends on value of $mOutputType:
1550 * OT_WIKI: only {{subst:}} templates
1551 * OT_MSG: only magic variables
1552 * OT_HTML: all templates and magic variables
1553 *
1554 * @param string $tex The text to transform
1555 * @param array $args Key-value pairs representing template parameters to substitute
1556 * @access private
1557 */
1558 function replaceVariables( $text, $args = array() ) {
1559 global $wgLang, $wgScript, $wgArticlePath;
1560
1561 # Prevent too big inclusions
1562 if(strlen($text)> MAX_INCLUDE_SIZE)
1563 return $text;
1564
1565 $fname = 'Parser::replaceVariables';
1566 wfProfileIn( $fname );
1567
1568 $titleChars = Title::legalChars();
1569
1570 # This function is called recursively. To keep track of arguments we need a stack:
1571 array_push( $this->mArgStack, $args );
1572
1573 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1574 $GLOBALS['wgCurParser'] =& $this;
1575
1576 # Variable substitution
1577 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", 'wfVariableSubstitution', $text );
1578
1579 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1580 # Argument substitution
1581 $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1582 }
1583 # Template substitution
1584 $regex = '/{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1585 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1586
1587 array_pop( $this->mArgStack );
1588
1589 wfProfileOut( $fname );
1590 return $text;
1591 }
1592
1593 /**
1594 * Replace magic variables
1595 * @access private
1596 */
1597 function variableSubstitution( $matches ) {
1598 if ( !$this->mVariables ) {
1599 $this->initialiseVariables();
1600 }
1601 $skip = false;
1602 if ( $this->mOutputType == OT_WIKI ) {
1603 # Do only magic variables prefixed by SUBST
1604 $mwSubst =& MagicWord::get( MAG_SUBST );
1605 if (!$mwSubst->matchStartAndRemove( $matches[1] ))
1606 $skip = true;
1607 # Note that if we don't substitute the variable below,
1608 # we don't remove the {{subst:}} magic word, in case
1609 # it is a template rather than a magic variable.
1610 }
1611 if ( !$skip && array_key_exists( $matches[1], $this->mVariables ) ) {
1612 $text = $this->mVariables[$matches[1]];
1613 $this->mOutput->mContainsOldMagic = true;
1614 } else {
1615 $text = $matches[0];
1616 }
1617 return $text;
1618 }
1619
1620 # Split template arguments
1621 function getTemplateArgs( $argsString ) {
1622 if ( $argsString === '' ) {
1623 return array();
1624 }
1625
1626 $args = explode( '|', substr( $argsString, 1 ) );
1627
1628 # If any of the arguments contains a '[[' but no ']]', it needs to be
1629 # merged with the next arg because the '|' character between belongs
1630 # to the link syntax and not the template parameter syntax.
1631 $argc = count($args);
1632 $i = 0;
1633 for ( $i = 0; $i < $argc-1; $i++ ) {
1634 if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1635 $args[$i] .= '|'.$args[$i+1];
1636 array_splice($args, $i+1, 1);
1637 $i--;
1638 $argc--;
1639 }
1640 }
1641
1642 return $args;
1643 }
1644
1645 /**
1646 * Return the text of a template, after recursively
1647 * replacing any variables or templates within the template.
1648 *
1649 * @param array $matches The parts of the template
1650 * $matches[1]: the title, i.e. the part before the |
1651 * $matches[2]: the parameters (including a leading |), if any
1652 * @return string the text of the template
1653 * @access private
1654 */
1655 function braceSubstitution( $matches ) {
1656 global $wgLinkCache, $wgLang;
1657 $fname = 'Parser::braceSubstitution';
1658 $found = false;
1659 $nowiki = false;
1660 $noparse = false;
1661 $itcamefromthedatabase = false;
1662
1663 $title = NULL;
1664
1665 # $part1 is the bit before the first |, and must contain only title characters
1666 # $args is a list of arguments, starting from index 0, not including $part1
1667
1668 $part1 = $matches[1];
1669 # If the second subpattern matched anything, it will start with |
1670
1671 $args = $this->getTemplateArgs($matches[2]);
1672 $argc = count( $args );
1673
1674 # {{{}}}
1675 if ( strpos( $matches[0], '{{{' ) !== false ) {
1676 $text = $matches[0];
1677 $found = true;
1678 $noparse = true;
1679 }
1680
1681 # SUBST
1682 if ( !$found ) {
1683 $mwSubst =& MagicWord::get( MAG_SUBST );
1684 if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1685 # One of two possibilities is true:
1686 # 1) Found SUBST but not in the PST phase
1687 # 2) Didn't find SUBST and in the PST phase
1688 # In either case, return without further processing
1689 $text = $matches[0];
1690 $found = true;
1691 $noparse = true;
1692 }
1693 }
1694
1695 # MSG, MSGNW and INT
1696 if ( !$found ) {
1697 # Check for MSGNW:
1698 $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1699 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1700 $nowiki = true;
1701 } else {
1702 # Remove obsolete MSG:
1703 $mwMsg =& MagicWord::get( MAG_MSG );
1704 $mwMsg->matchStartAndRemove( $part1 );
1705 }
1706
1707 # Check if it is an internal message
1708 $mwInt =& MagicWord::get( MAG_INT );
1709 if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1710 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1711 $text = wfMsgReal( $part1, $args, true );
1712 $found = true;
1713 }
1714 }
1715 }
1716
1717 # NS
1718 if ( !$found ) {
1719 # Check for NS: (namespace expansion)
1720 $mwNs = MagicWord::get( MAG_NS );
1721 if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1722 if ( intval( $part1 ) ) {
1723 $text = $wgLang->getNsText( intval( $part1 ) );
1724 $found = true;
1725 } else {
1726 $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1727 if ( !is_null( $index ) ) {
1728 $text = $wgLang->getNsText( $index );
1729 $found = true;
1730 }
1731 }
1732 }
1733 }
1734
1735 # LOCALURL and LOCALURLE
1736 if ( !$found ) {
1737 $mwLocal = MagicWord::get( MAG_LOCALURL );
1738 $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1739
1740 if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1741 $func = 'getLocalURL';
1742 } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1743 $func = 'escapeLocalURL';
1744 } else {
1745 $func = '';
1746 }
1747
1748 if ( $func !== '' ) {
1749 $title = Title::newFromText( $part1 );
1750 if ( !is_null( $title ) ) {
1751 if ( $argc > 0 ) {
1752 $text = $title->$func( $args[0] );
1753 } else {
1754 $text = $title->$func();
1755 }
1756 $found = true;
1757 }
1758 }
1759 }
1760
1761 # GRAMMAR
1762 if ( !$found && $argc == 1 ) {
1763 $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1764 if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1765 $text = $wgLang->convertGrammar( $args[0], $part1 );
1766 $found = true;
1767 }
1768 }
1769
1770 # Template table test
1771
1772 # Did we encounter this template already? If yes, it is in the cache
1773 # and we need to check for loops.
1774 if ( isset( $this->mTemplates[$part1] ) ) {
1775 # Infinite loop test
1776 if ( isset( $this->mTemplatePath[$part1] ) ) {
1777 $noparse = true;
1778 $found = true;
1779 }
1780 # set $text to cached message.
1781 $text = $this->mTemplates[$part1];
1782 $found = true;
1783 }
1784
1785 # Load from database
1786 if ( !$found ) {
1787 $title = Title::newFromText( $part1, NS_TEMPLATE );
1788 if ( !is_null( $title ) && !$title->isExternal() ) {
1789 # Check for excessive inclusion
1790 $dbk = $title->getPrefixedDBkey();
1791 if ( $this->incrementIncludeCount( $dbk ) ) {
1792 # This should never be reached.
1793 $article = new Article( $title );
1794 $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1795 if ( $articleContent !== false ) {
1796 $found = true;
1797 $text = $articleContent;
1798 $itcamefromthedatabase = true;
1799 }
1800 }
1801
1802 # If the title is valid but undisplayable, make a link to it
1803 if ( $this->mOutputType == OT_HTML && !$found ) {
1804 $text = '[['.$title->getPrefixedText().']]';
1805 $found = true;
1806 }
1807
1808 # Template cache array insertion
1809 $this->mTemplates[$part1] = $text;
1810 }
1811 }
1812
1813 # Recursive parsing, escaping and link table handling
1814 # Only for HTML output
1815 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1816 $text = wfEscapeWikiText( $text );
1817 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
1818 # Clean up argument array
1819 $assocArgs = array();
1820 $index = 1;
1821 foreach( $args as $arg ) {
1822 $eqpos = strpos( $arg, '=' );
1823 if ( $eqpos === false ) {
1824 $assocArgs[$index++] = $arg;
1825 } else {
1826 $name = trim( substr( $arg, 0, $eqpos ) );
1827 $value = trim( substr( $arg, $eqpos+1 ) );
1828 if ( $value === false ) {
1829 $value = '';
1830 }
1831 if ( $name !== false ) {
1832 $assocArgs[$name] = $value;
1833 }
1834 }
1835 }
1836
1837 # Add a new element to the templace recursion path
1838 $this->mTemplatePath[$part1] = 1;
1839
1840 $text = $this->strip( $text, $this->mStripState );
1841 $text = $this->removeHTMLtags( $text );
1842 $text = $this->replaceVariables( $text, $assocArgs );
1843
1844 # Resume the link cache and register the inclusion as a link
1845 if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
1846 $wgLinkCache->addLinkObj( $title );
1847 }
1848 }
1849
1850 # Empties the template path
1851 $this->mTemplatePath = array();
1852 if ( !$found ) {
1853 return $matches[0];
1854 } else {
1855 # replace ==section headers==
1856 # XXX this needs to go away once we have a better parser.
1857 if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
1858 if( !is_null( $title ) )
1859 $encodedname = base64_encode($title->getPrefixedDBkey());
1860 else
1861 $encodedname = base64_encode("");
1862 $matches = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
1863 PREG_SPLIT_DELIM_CAPTURE);
1864 $text = '';
1865 $nsec = 0;
1866 for( $i = 0; $i < count($matches); $i += 2 ) {
1867 $text .= $matches[$i];
1868 if (!isset($matches[$i + 1]) || $matches[$i + 1] == "") continue;
1869 $hl = $matches[$i + 1];
1870 if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
1871 $text .= $hl;
1872 continue;
1873 }
1874 preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
1875 $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
1876 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
1877
1878 $nsec++;
1879 }
1880 }
1881 return $text;
1882 }
1883 }
1884
1885 /**
1886 * Triple brace replacement -- used for template arguments
1887 * @access private
1888 */
1889 function argSubstitution( $matches ) {
1890 $arg = trim( $matches[1] );
1891 $text = $matches[0];
1892 $inputArgs = end( $this->mArgStack );
1893
1894 if ( array_key_exists( $arg, $inputArgs ) ) {
1895 $text = $inputArgs[$arg];
1896 }
1897
1898 return $text;
1899 }
1900
1901 /**
1902 * Returns true if the function is allowed to include this entity
1903 * @access private
1904 */
1905 function incrementIncludeCount( $dbk ) {
1906 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1907 $this->mIncludeCount[$dbk] = 0;
1908 }
1909 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1910 return true;
1911 } else {
1912 return false;
1913 }
1914 }
1915
1916
1917 /**
1918 * Cleans up HTML, removes dangerous tags and attributes, and
1919 * removes HTML comments
1920 * @access private
1921 */
1922 function removeHTMLtags( $text ) {
1923 global $wgUseTidy, $wgUserHtml;
1924 $fname = 'Parser::removeHTMLtags';
1925 wfProfileIn( $fname );
1926
1927 if( $wgUserHtml ) {
1928 $htmlpairs = array( # Tags that must be closed
1929 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1930 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1931 'strike', 'strong', 'tt', 'var', 'div', 'center',
1932 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1933 'ruby', 'rt' , 'rb' , 'rp', 'p'
1934 );
1935 $htmlsingle = array(
1936 'br', 'hr', 'li', 'dt', 'dd'
1937 );
1938 $htmlnest = array( # Tags that can be nested--??
1939 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1940 'dl', 'font', 'big', 'small', 'sub', 'sup'
1941 );
1942 $tabletags = array( # Can only appear inside table
1943 'td', 'th', 'tr'
1944 );
1945 } else {
1946 $htmlpairs = array();
1947 $htmlsingle = array();
1948 $htmlnest = array();
1949 $tabletags = array();
1950 }
1951
1952 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1953 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1954
1955 $htmlattrs = $this->getHTMLattrs () ;
1956
1957 # Remove HTML comments
1958 $text = $this->removeHTMLcomments( $text );
1959
1960 $bits = explode( '<', $text );
1961 $text = array_shift( $bits );
1962 if(!$wgUseTidy) {
1963 $tagstack = array(); $tablestack = array();
1964 foreach ( $bits as $x ) {
1965 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
1966 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
1967 $x, $regs );
1968 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1969 error_reporting( $prev );
1970
1971 $badtag = 0 ;
1972 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1973 # Check our stack
1974 if ( $slash ) {
1975 # Closing a tag...
1976 if ( ! in_array( $t, $htmlsingle ) &&
1977 ( $ot = @array_pop( $tagstack ) ) != $t ) {
1978 @array_push( $tagstack, $ot );
1979 $badtag = 1;
1980 } else {
1981 if ( $t == 'table' ) {
1982 $tagstack = array_pop( $tablestack );
1983 }
1984 $newparams = '';
1985 }
1986 } else {
1987 # Keep track for later
1988 if ( in_array( $t, $tabletags ) &&
1989 ! in_array( 'table', $tagstack ) ) {
1990 $badtag = 1;
1991 } else if ( in_array( $t, $tagstack ) &&
1992 ! in_array ( $t , $htmlnest ) ) {
1993 $badtag = 1 ;
1994 } else if ( ! in_array( $t, $htmlsingle ) ) {
1995 if ( $t == 'table' ) {
1996 array_push( $tablestack, $tagstack );
1997 $tagstack = array();
1998 }
1999 array_push( $tagstack, $t );
2000 }
2001 # Strip non-approved attributes from the tag
2002 $newparams = $this->fixTagAttributes($params);
2003
2004 }
2005 if ( ! $badtag ) {
2006 $rest = str_replace( '>', '&gt;', $rest );
2007 $text .= "<$slash$t $newparams$brace$rest";
2008 continue;
2009 }
2010 }
2011 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2012 }
2013 # Close off any remaining tags
2014 while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2015 $text .= "</$t>\n";
2016 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2017 }
2018 } else {
2019 # this might be possible using tidy itself
2020 foreach ( $bits as $x ) {
2021 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2022 $x, $regs );
2023 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2024 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2025 $newparams = $this->fixTagAttributes($params);
2026 $rest = str_replace( '>', '&gt;', $rest );
2027 $text .= "<$slash$t $newparams$brace$rest";
2028 } else {
2029 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2030 }
2031 }
2032 }
2033 wfProfileOut( $fname );
2034 return $text;
2035 }
2036
2037 /**
2038 * Remove '<!--', '-->', and everything between.
2039 * To avoid leaving blank lines, when a comment is both preceded
2040 * and followed by a newline (ignoring spaces), trim leading and
2041 * trailing spaces and one of the newlines.
2042 *
2043 * @access private
2044 */
2045 function removeHTMLcomments( $text ) {
2046 $fname='Parser::removeHTMLcomments';
2047 wfProfileIn( $fname );
2048 while (($start = strpos($text, '<!--')) !== false) {
2049 $end = strpos($text, '-->', $start + 4);
2050 if ($end === false) {
2051 # Unterminated comment; bail out
2052 break;
2053 }
2054
2055 $end += 3;
2056
2057 # Trim space and newline if the comment is both
2058 # preceded and followed by a newline
2059 $spaceStart = max($start - 1, 0);
2060 $spaceLen = $end - $spaceStart;
2061 while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2062 $spaceStart--;
2063 $spaceLen++;
2064 }
2065 while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2066 $spaceLen++;
2067 if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2068 # Remove the comment, leading and trailing
2069 # spaces, and leave only one newline.
2070 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2071 }
2072 else {
2073 # Remove just the comment.
2074 $text = substr_replace($text, '', $start, $end - $start);
2075 }
2076 }
2077 wfProfileOut( $fname );
2078 return $text;
2079 }
2080
2081 /**
2082 * This function accomplishes several tasks:
2083 * 1) Auto-number headings if that option is enabled
2084 * 2) Add an [edit] link to sections for logged in users who have enabled the option
2085 * 3) Add a Table of contents on the top for users who have enabled the option
2086 * 4) Auto-anchor headings
2087 *
2088 * It loops through all headlines, collects the necessary data, then splits up the
2089 * string and re-inserts the newly formatted headlines.
2090 * @access private
2091 */
2092 /* private */ function formatHeadings( $text, $isMain=true ) {
2093 global $wgInputEncoding, $wgMaxTocLevel, $wgLang, $wgLinkHolders;
2094
2095 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2096 $doShowToc = $this->mOptions->getShowToc();
2097 $forceTocHere = false;
2098 if( !$this->mTitle->userCanEdit() ) {
2099 $showEditLink = 0;
2100 $rightClickHack = 0;
2101 } else {
2102 $showEditLink = $this->mOptions->getEditSection();
2103 $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2104 }
2105
2106 # Inhibit editsection links if requested in the page
2107 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2108 if( $esw->matchAndRemove( $text ) ) {
2109 $showEditLink = 0;
2110 }
2111 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2112 # do not add TOC
2113 $mw =& MagicWord::get( MAG_NOTOC );
2114 if( $mw->matchAndRemove( $text ) ) {
2115 $doShowToc = 0;
2116 }
2117
2118 # never add the TOC to the Main Page. This is an entry page that should not
2119 # be more than 1-2 screens large anyway
2120 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2121 $doShowToc = 0;
2122 }
2123
2124 # Get all headlines for numbering them and adding funky stuff like [edit]
2125 # links - this is for later, but we need the number of headlines right now
2126 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
2127
2128 # if there are fewer than 4 headlines in the article, do not show TOC
2129 if( $numMatches < 4 ) {
2130 $doShowToc = 0;
2131 }
2132
2133 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2134 # override above conditions and always show TOC at that place
2135 $mw =& MagicWord::get( MAG_TOC );
2136 if ($mw->match( $text ) ) {
2137 $doShowToc = 1;
2138 $forceTocHere = true;
2139 } else {
2140 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2141 # override above conditions and always show TOC above first header
2142 $mw =& MagicWord::get( MAG_FORCETOC );
2143 if ($mw->matchAndRemove( $text ) ) {
2144 $doShowToc = 1;
2145 }
2146 }
2147
2148
2149
2150 # We need this to perform operations on the HTML
2151 $sk =& $this->mOptions->getSkin();
2152
2153 # headline counter
2154 $headlineCount = 0;
2155 $sectionCount = 0; # headlineCount excluding template sections
2156
2157 # Ugh .. the TOC should have neat indentation levels which can be
2158 # passed to the skin functions. These are determined here
2159 $toclevel = 0;
2160 $toc = '';
2161 $full = '';
2162 $head = array();
2163 $sublevelCount = array();
2164 $level = 0;
2165 $prevlevel = 0;
2166 foreach( $matches[3] as $headline ) {
2167 $istemplate = 0;
2168 $templatetitle = "";
2169 $templatesection = 0;
2170
2171 if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2172 $istemplate = 1;
2173 $templatetitle = base64_decode($mat[1]);
2174 $templatesection = 1 + (int)base64_decode($mat[2]);
2175 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2176 }
2177
2178 $numbering = '';
2179 if( $level ) {
2180 $prevlevel = $level;
2181 }
2182 $level = $matches[1][$headlineCount];
2183 if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2184 # reset when we enter a new level
2185 $sublevelCount[$level] = 0;
2186 $toc .= $sk->tocIndent( $level - $prevlevel );
2187 $toclevel += $level - $prevlevel;
2188 }
2189 if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2190 # reset when we step back a level
2191 $sublevelCount[$level+1]=0;
2192 $toc .= $sk->tocUnindent( $prevlevel - $level );
2193 $toclevel -= $prevlevel - $level;
2194 }
2195 # count number of headlines for each level
2196 @$sublevelCount[$level]++;
2197 if( $doNumberHeadings || $doShowToc ) {
2198 $dot = 0;
2199 for( $i = 1; $i <= $level; $i++ ) {
2200 if( !empty( $sublevelCount[$i] ) ) {
2201 if( $dot ) {
2202 $numbering .= '.';
2203 }
2204 $numbering .= $wgLang->formatNum( $sublevelCount[$i] );
2205 $dot = 1;
2206 }
2207 }
2208 }
2209
2210 # The canonized header is a version of the header text safe to use for links
2211 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2212 $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2213 $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2214
2215 # Remove link placeholders by the link text.
2216 # <!--LINK number-->
2217 # turns into
2218 # link text with suffix
2219 $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2220 "\$wgLinkHolders['texts'][\$1]",
2221 $canonized_headline );
2222
2223 # strip out HTML
2224 $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2225 $tocline = trim( $canonized_headline );
2226 $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2227 $replacearray = array(
2228 '%3A' => ':',
2229 '%' => '.'
2230 );
2231 $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2232 $refer[$headlineCount] = $canonized_headline;
2233
2234 # count how many in assoc. array so we can track dupes in anchors
2235 @$refers[$canonized_headline]++;
2236 $refcount[$headlineCount]=$refers[$canonized_headline];
2237
2238 # Prepend the number to the heading text
2239
2240 if( $doNumberHeadings || $doShowToc ) {
2241 $tocline = $numbering . ' ' . $tocline;
2242
2243 # Don't number the heading if it is the only one (looks silly)
2244 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2245 # the two are different if the line contains a link
2246 $headline=$numbering . ' ' . $headline;
2247 }
2248 }
2249
2250 # Create the anchor for linking from the TOC to the section
2251 $anchor = $canonized_headline;
2252 if($refcount[$headlineCount] > 1 ) {
2253 $anchor .= '_' . $refcount[$headlineCount];
2254 }
2255 if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2256 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2257 }
2258 if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2259 if ( empty( $head[$headlineCount] ) ) {
2260 $head[$headlineCount] = '';
2261 }
2262 if( $istemplate )
2263 $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2264 else
2265 $head[$headlineCount] .= $sk->editSectionLink($sectionCount+1);
2266 }
2267
2268 # Add the edit section span
2269 if( $rightClickHack ) {
2270 if( $istemplate )
2271 $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2272 else
2273 $headline = $sk->editSectionScript($sectionCount+1,$headline);
2274 }
2275
2276 # give headline the correct <h#> tag
2277 @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2278
2279 $headlineCount++;
2280 if( !$istemplate )
2281 $sectionCount++;
2282 }
2283
2284 if( $doShowToc ) {
2285 $toclines = $headlineCount;
2286 $toc .= $sk->tocUnindent( $toclevel );
2287 $toc = $sk->tocTable( $toc );
2288 }
2289
2290 # split up and insert constructed headlines
2291
2292 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2293 $i = 0;
2294
2295 foreach( $blocks as $block ) {
2296 if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2297 # This is the [edit] link that appears for the top block of text when
2298 # section editing is enabled
2299
2300 # Disabled because it broke block formatting
2301 # For example, a bullet point in the top line
2302 # $full .= $sk->editSectionLink(0);
2303 }
2304 $full .= $block;
2305 if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2306 # Top anchor now in skin
2307 $full = $full.$toc;
2308 }
2309
2310 if( !empty( $head[$i] ) ) {
2311 $full .= $head[$i];
2312 }
2313 $i++;
2314 }
2315 if($forceTocHere) {
2316 $mw =& MagicWord::get( MAG_TOC );
2317 return $mw->replace( $toc, $full );
2318 } else {
2319 return $full;
2320 }
2321 }
2322
2323 /**
2324 * Return an HTML link for the "ISBN 123456" text
2325 * @access private
2326 */
2327 function magicISBN( $text ) {
2328 global $wgLang;
2329 $fname = 'Parser::magicISBN';
2330 wfProfileIn( $fname );
2331
2332 $a = split( 'ISBN ', ' '.$text );
2333 if ( count ( $a ) < 2 ) {
2334 wfProfileOut( $fname );
2335 return $text;
2336 }
2337 $text = substr( array_shift( $a ), 1);
2338 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2339
2340 foreach ( $a as $x ) {
2341 $isbn = $blank = '' ;
2342 while ( ' ' == $x{0} ) {
2343 $blank .= ' ';
2344 $x = substr( $x, 1 );
2345 }
2346 if ( $x == '' ) { # blank isbn
2347 $text .= "ISBN $blank";
2348 continue;
2349 }
2350 while ( strstr( $valid, $x{0} ) != false ) {
2351 $isbn .= $x{0};
2352 $x = substr( $x, 1 );
2353 }
2354 $num = str_replace( '-', '', $isbn );
2355 $num = str_replace( ' ', '', $num );
2356
2357 if ( '' == $num ) {
2358 $text .= "ISBN $blank$x";
2359 } else {
2360 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2361 $text .= '<a href="' .
2362 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2363 "\" class=\"internal\">ISBN $isbn</a>";
2364 $text .= $x;
2365 }
2366 }
2367 wfProfileOut( $fname );
2368 return $text;
2369 }
2370
2371 /**
2372 * Return an HTML link for the "GEO ..." text
2373 * @access private
2374 */
2375 function magicGEO( $text ) {
2376 global $wgLang, $wgUseGeoMode;
2377 $fname = 'Parser::magicGEO';
2378 wfProfileIn( $fname );
2379
2380 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2381 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2382 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2383 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2384 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2385 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2386
2387 $a = split( 'GEO ', ' '.$text );
2388 if ( count ( $a ) < 2 ) {
2389 wfProfileOut( $fname );
2390 return $text;
2391 }
2392 $text = substr( array_shift( $a ), 1);
2393 $valid = '0123456789.+-:';
2394
2395 foreach ( $a as $x ) {
2396 $geo = $blank = '' ;
2397 while ( ' ' == $x{0} ) {
2398 $blank .= ' ';
2399 $x = substr( $x, 1 );
2400 }
2401 while ( strstr( $valid, $x{0} ) != false ) {
2402 $geo .= $x{0};
2403 $x = substr( $x, 1 );
2404 }
2405 $num = str_replace( '+', '', $geo );
2406 $num = str_replace( ' ', '', $num );
2407
2408 if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2409 $text .= "GEO $blank$x";
2410 } else {
2411 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2412 $text .= '<a href="' .
2413 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2414 "\" class=\"internal\">GEO $geo</a>";
2415 $text .= $x;
2416 }
2417 }
2418 wfProfileOut( $fname );
2419 return $text;
2420 }
2421
2422 /**
2423 * Return an HTML link for the "RFC 1234" text
2424 * @access private
2425 * @param string $text text to be processed
2426 */
2427 function magicRFC( $text ) {
2428 global $wgLang;
2429
2430 $valid = '0123456789';
2431 $internal = false;
2432
2433 $a = split( 'RFC ', ' '.$text );
2434 if ( count ( $a ) < 2 ) return $text;
2435 $text = substr( array_shift( $a ), 1);
2436
2437 /* Check if RFC keyword is preceed by [[.
2438 * This test is made here cause of the array_shift above
2439 * that prevent the test to be done in the foreach.
2440 */
2441 if(substr($text, -2) == '[[') { $internal = true; }
2442
2443 foreach ( $a as $x ) {
2444 /* token might be empty if we have RFC RFC 1234 */
2445 if($x=='') {
2446 $text.='RFC ';
2447 continue;
2448 }
2449
2450 $rfc = $blank = '' ;
2451
2452 /** remove and save whitespaces in $blank */
2453 while ( $x{0} == ' ' ) {
2454 $blank .= ' ';
2455 $x = substr( $x, 1 );
2456 }
2457
2458 /** remove and save the rfc number in $rfc */
2459 while ( strstr( $valid, $x{0} ) != false ) {
2460 $rfc .= $x{0};
2461 $x = substr( $x, 1 );
2462 }
2463
2464 if ( $rfc == '') {
2465 /* call back stripped spaces*/
2466 $text .= "RFC $blank$x";
2467 } elseif( $internal) {
2468 /* normal link */
2469 $text .= "RFC $rfc$x";
2470 } else {
2471 /* build the external link*/
2472 $url = wfmsg( 'rfcurl' );
2473 $url = str_replace( '$1', $rfc, $url);
2474 $sk =& $this->mOptions->getSkin();
2475 $la = $sk->getExternalLinkAttributes( $url, 'RFC '.$rfc );
2476 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2477 }
2478
2479 /* Check if the next RFC keyword is preceed by [[ */
2480 $internal = (substr($x,-2) == '[[');
2481 }
2482 return $text;
2483 }
2484
2485 /**
2486 * Transform wiki markup when saving a page by doing \r\n -> \n
2487 * conversion, substitting signatures, {{subst:}} templates, etc.
2488 *
2489 * @param string $text the text to transform
2490 * @param Title &$title the Title object for the current article
2491 * @param User &$user the User object describing the current user
2492 * @param ParserOptions $options parsing options
2493 * @param bool $clearState whether to clear the parser state first
2494 * @return string the altered wiki markup
2495 * @access public
2496 */
2497 function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2498 $this->mOptions = $options;
2499 $this->mTitle =& $title;
2500 $this->mOutputType = OT_WIKI;
2501
2502 if ( $clearState ) {
2503 $this->clearState();
2504 }
2505
2506 $stripState = false;
2507 $pairs = array(
2508 "\r\n" => "\n",
2509 );
2510 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2511 // now with regexes
2512 /*
2513 $pairs = array(
2514 "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2515 "/<br *?>/i" => "<br />",
2516 );
2517 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2518 */
2519 $text = $this->strip( $text, $stripState, false );
2520 $text = $this->pstPass2( $text, $user );
2521 $text = $this->unstrip( $text, $stripState );
2522 $text = $this->unstripNoWiki( $text, $stripState );
2523 return $text;
2524 }
2525
2526 /**
2527 * Pre-save transform helper function
2528 * @access private
2529 */
2530 function pstPass2( $text, &$user ) {
2531 global $wgLang, $wgLocaltimezone, $wgCurParser;
2532
2533 # Variable replacement
2534 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2535 $text = $this->replaceVariables( $text );
2536
2537 # Signatures
2538 #
2539 $n = $user->getName();
2540 $k = $user->getOption( 'nickname' );
2541 if ( '' == $k ) { $k = $n; }
2542 if(isset($wgLocaltimezone)) {
2543 $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2544 }
2545 /* Note: this is an ugly timezone hack for the European wikis */
2546 $d = $wgLang->timeanddate( date( 'YmdHis' ), false ) .
2547 ' (' . date( 'T' ) . ')';
2548 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2549
2550 $text = preg_replace( '/~~~~~/', $d, $text );
2551 $text = preg_replace( '/~~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2552 $text = preg_replace( '/~~~/', '[[' . $wgLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2553
2554 # Context links: [[|name]] and [[name (context)|]]
2555 #
2556 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2557 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2558 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2559 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2560
2561 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/"; # [[page (context)|]]
2562 $p2 = "/\[\[\\|({$tc}+)]]/"; # [[|page]]
2563 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/"; # [[namespace:page|]] and [[:namespace:page|]]
2564 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2565 $context = '';
2566 $t = $this->mTitle->getText();
2567 if ( preg_match( $conpat, $t, $m ) ) {
2568 $context = $m[2];
2569 }
2570 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2571 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2572 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2573
2574 if ( '' == $context ) {
2575 $text = preg_replace( $p2, '[[\\1]]', $text );
2576 } else {
2577 $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2578 }
2579
2580 # Trim trailing whitespace
2581 # MAG_END (__END__) tag allows for trailing
2582 # whitespace to be deliberately included
2583 $text = rtrim( $text );
2584 $mw =& MagicWord::get( MAG_END );
2585 $mw->matchAndRemove( $text );
2586
2587 return $text;
2588 }
2589
2590 /**
2591 * Set up some variables which are usually set up in parse()
2592 * so that an external function can call some class members with confidence
2593 * @access public
2594 */
2595 function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2596 $this->mTitle =& $title;
2597 $this->mOptions = $options;
2598 $this->mOutputType = $outputType;
2599 if ( $clearState ) {
2600 $this->clearState();
2601 }
2602 }
2603
2604 /**
2605 * Transform a MediaWiki message by replacing magic variables.
2606 *
2607 * @param string $text the text to transform
2608 * @param ParserOptions $options options
2609 * @return string the text with variables substituted
2610 * @access public
2611 */
2612 function transformMsg( $text, $options ) {
2613 global $wgTitle;
2614 static $executing = false;
2615
2616 # Guard against infinite recursion
2617 if ( $executing ) {
2618 return $text;
2619 }
2620 $executing = true;
2621
2622 $this->mTitle = $wgTitle;
2623 $this->mOptions = $options;
2624 $this->mOutputType = OT_MSG;
2625 $this->clearState();
2626 $text = $this->replaceVariables( $text );
2627
2628 $executing = false;
2629 return $text;
2630 }
2631
2632 /**
2633 * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2634 * Callback will be called with the text within
2635 * Transform and return the text within
2636 * @access public
2637 */
2638 function setHook( $tag, $callback ) {
2639 $oldVal = @$this->mTagHooks[$tag];
2640 $this->mTagHooks[$tag] = $callback;
2641 return $oldVal;
2642 }
2643 }
2644
2645 /**
2646 * @todo document
2647 * @package MediaWiki
2648 */
2649 class ParserOutput
2650 {
2651 var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2652 var $mCacheTime; # Used in ParserCache
2653
2654 function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2655 $containsOldMagic = false )
2656 {
2657 $this->mText = $text;
2658 $this->mLanguageLinks = $languageLinks;
2659 $this->mCategoryLinks = $categoryLinks;
2660 $this->mContainsOldMagic = $containsOldMagic;
2661 $this->mCacheTime = '';
2662 }
2663
2664 function getText() { return $this->mText; }
2665 function getLanguageLinks() { return $this->mLanguageLinks; }
2666 function getCategoryLinks() { return $this->mCategoryLinks; }
2667 function getCacheTime() { return $this->mCacheTime; }
2668 function containsOldMagic() { return $this->mContainsOldMagic; }
2669 function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2670 function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2671 function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2672 function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2673 function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2674
2675 function merge( $other ) {
2676 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2677 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2678 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2679 }
2680
2681 }
2682
2683 /**
2684 * Set options of the Parser
2685 * @todo document
2686 * @package MediaWiki
2687 */
2688 class ParserOptions
2689 {
2690 # All variables are private
2691 var $mUseTeX; # Use texvc to expand <math> tags
2692 var $mUseDynamicDates; # Use $wgDateFormatter to format dates
2693 var $mInterwikiMagic; # Interlanguage links are removed and returned in an array
2694 var $mAllowExternalImages; # Allow external images inline
2695 var $mSkin; # Reference to the preferred skin
2696 var $mDateFormat; # Date format index
2697 var $mEditSection; # Create "edit section" links
2698 var $mEditSectionOnRightClick; # Generate JavaScript to edit section on right click
2699 var $mNumberHeadings; # Automatically number headings
2700 var $mShowToc; # Show table of contents
2701
2702 function getUseTeX() { return $this->mUseTeX; }
2703 function getUseDynamicDates() { return $this->mUseDynamicDates; }
2704 function getInterwikiMagic() { return $this->mInterwikiMagic; }
2705 function getAllowExternalImages() { return $this->mAllowExternalImages; }
2706 function getSkin() { return $this->mSkin; }
2707 function getDateFormat() { return $this->mDateFormat; }
2708 function getEditSection() { return $this->mEditSection; }
2709 function getEditSectionOnRightClick() { return $this->mEditSectionOnRightClick; }
2710 function getNumberHeadings() { return $this->mNumberHeadings; }
2711 function getShowToc() { return $this->mShowToc; }
2712
2713 function setUseTeX( $x ) { return wfSetVar( $this->mUseTeX, $x ); }
2714 function setUseDynamicDates( $x ) { return wfSetVar( $this->mUseDynamicDates, $x ); }
2715 function setInterwikiMagic( $x ) { return wfSetVar( $this->mInterwikiMagic, $x ); }
2716 function setAllowExternalImages( $x ) { return wfSetVar( $this->mAllowExternalImages, $x ); }
2717 function setDateFormat( $x ) { return wfSetVar( $this->mDateFormat, $x ); }
2718 function setEditSection( $x ) { return wfSetVar( $this->mEditSection, $x ); }
2719 function setEditSectionOnRightClick( $x ) { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2720 function setNumberHeadings( $x ) { return wfSetVar( $this->mNumberHeadings, $x ); }
2721 function setShowToc( $x ) { return wfSetVar( $this->mShowToc, $x ); }
2722
2723 function setSkin( &$x ) { $this->mSkin =& $x; }
2724
2725 # Get parser options
2726 /* static */ function newFromUser( &$user ) {
2727 $popts = new ParserOptions;
2728 $popts->initialiseFromUser( $user );
2729 return $popts;
2730 }
2731
2732 # Get user options
2733 function initialiseFromUser( &$userInput ) {
2734 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2735
2736 $fname = 'ParserOptions::initialiseFromUser';
2737 wfProfileIn( $fname );
2738 if ( !$userInput ) {
2739 $user = new User;
2740 $user->setLoaded( true );
2741 } else {
2742 $user =& $userInput;
2743 }
2744
2745 $this->mUseTeX = $wgUseTeX;
2746 $this->mUseDynamicDates = $wgUseDynamicDates;
2747 $this->mInterwikiMagic = $wgInterwikiMagic;
2748 $this->mAllowExternalImages = $wgAllowExternalImages;
2749 wfProfileIn( $fname.'-skin' );
2750 $this->mSkin =& $user->getSkin();
2751 wfProfileOut( $fname.'-skin' );
2752 $this->mDateFormat = $user->getOption( 'date' );
2753 $this->mEditSection = $user->getOption( 'editsection' );
2754 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2755 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2756 $this->mShowToc = $user->getOption( 'showtoc' );
2757 wfProfileOut( $fname );
2758 }
2759
2760
2761 }
2762
2763 # Regex callbacks, used in Parser::replaceVariables
2764 function wfBraceSubstitution( $matches ) {
2765 global $wgCurParser;
2766 return $wgCurParser->braceSubstitution( $matches );
2767 }
2768
2769 function wfArgSubstitution( $matches ) {
2770 global $wgCurParser;
2771 return $wgCurParser->argSubstitution( $matches );
2772 }
2773
2774 function wfVariableSubstitution( $matches ) {
2775 global $wgCurParser;
2776 return $wgCurParser->variableSubstitution( $matches );
2777 }
2778
2779 /**
2780 * Return the total number of articles
2781 */
2782 function wfNumberOfArticles() {
2783 global $wgNumberOfArticles;
2784
2785 wfLoadSiteStats();
2786 return $wgNumberOfArticles;
2787 }
2788
2789 /**
2790 * Get various statistics from the database
2791 * @private
2792 */
2793 function wfLoadSiteStats() {
2794 global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
2795 $fname = 'wfLoadSiteStats';
2796
2797 if ( -1 != $wgNumberOfArticles ) return;
2798 $dbr =& wfGetDB( DB_SLAVE );
2799 $s = $dbr->getArray( 'site_stats',
2800 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
2801 array( 'ss_row_id' => 1 ), $fname
2802 );
2803
2804 if ( $s === false ) {
2805 return;
2806 } else {
2807 $wgTotalViews = $s->ss_total_views;
2808 $wgTotalEdits = $s->ss_total_edits;
2809 $wgNumberOfArticles = $s->ss_good_articles;
2810 }
2811 }
2812
2813 function wfEscapeHTMLTagsOnly( $in ) {
2814 return str_replace(
2815 array( '"', '>', '<' ),
2816 array( '&quot;', '&gt;', '&lt;' ),
2817 $in );
2818 }
2819
2820 ?>
MediaWiki Core