dépôts / lhc / web / wiklou.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
history | raw | HEAD
Fix bug 561: {{/Subpage}} acts like {{Template:/Subpage}}, by
[lhc/web/wiklou.git] / includes / Parser.php
1 <?php
2
3 /**
4 * File for Parser and related classes
5 *
6 * @package MediaWiki
7 * @version $Id$
8 */
9
10 /**
11 * Variable substitution O(N^2) attack
12 *
13 * Without countermeasures, it would be possible to attack the parser by saving
14 * a page filled with a large number of inclusions of large pages. The size of
15 * the generated page would be proportional to the square of the input size.
16 * Hence, we limit the number of inclusions of any given page, thus bringing any
17 * attack back to O(N).
18 */
19 define( 'MAX_INCLUDE_REPEAT', 100 );
20 define( 'MAX_INCLUDE_SIZE', 1000000 ); // 1 Million
21
22 # Allowed values for $mOutputType
23 define( 'OT_HTML', 1 );
24 define( 'OT_WIKI', 2 );
25 define( 'OT_MSG' , 3 );
26
27 # string parameter for extractTags which will cause it
28 # to strip HTML comments in addition to regular
29 # <XML>-style tags. This should not be anything we
30 # may want to use in wikisyntax
31 define( 'STRIP_COMMENTS', 'HTMLCommentStrip' );
32
33 # prefix for escaping, used in two functions at least
34 define( 'UNIQ_PREFIX', 'NaodW29');
35
36 # Constants needed for external link processing
37 define( 'URL_PROTOCOLS', 'http|https|ftp|irc|gopher|news|mailto' );
38 define( 'HTTP_PROTOCOLS', 'http|https' );
39 # Everything except bracket, space, or control characters
40 define( 'EXT_LINK_URL_CLASS', '[^]\\x00-\\x20\\x7F]' );
41 define( 'INVERSE_EXT_LINK_URL_CLASS', '[\]\\x00-\\x20\\x7F]' );
42 # Including space
43 define( 'EXT_LINK_TEXT_CLASS', '[^\]\\x00-\\x1F\\x7F]' );
44 define( 'EXT_IMAGE_FNAME_CLASS', '[A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]' );
45 define( 'EXT_IMAGE_EXTENSIONS', 'gif|png|jpg|jpeg' );
46 define( 'EXT_LINK_BRACKETED', '/\[(('.URL_PROTOCOLS.'):'.EXT_LINK_URL_CLASS.'+) *('.EXT_LINK_TEXT_CLASS.'*?)\]/S' );
47 define( 'EXT_IMAGE_REGEX',
48 '/^('.HTTP_PROTOCOLS.':)'. # Protocol
49 '('.EXT_LINK_URL_CLASS.'+)\\/'. # Hostname and path
50 '('.EXT_IMAGE_FNAME_CLASS.'+)\\.((?i)'.EXT_IMAGE_EXTENSIONS.')$/S' # Filename
51 );
52
53 /**
54 * PHP Parser
55 *
56 * Processes wiki markup
57 *
58 * <pre>
59 * There are three main entry points into the Parser class:
60 * parse()
61 * produces HTML output
62 * preSaveTransform().
63 * produces altered wiki markup.
64 * transformMsg()
65 * performs brace substitution on MediaWiki messages
66 *
67 * Globals used:
68 * objects: $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
69 *
70 * NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
71 *
72 * settings:
73 * $wgUseTex*, $wgUseDynamicDates*, $wgInterwikiMagic*,
74 * $wgNamespacesWithSubpages, $wgAllowExternalImages*,
75 * $wgLocaltimezone
76 *
77 * * only within ParserOptions
78 * </pre>
79 *
80 * @package MediaWiki
81 */
82 class Parser
83 {
84 /**#@+
85 * @access private
86 */
87 # Persistent:
88 var $mTagHooks;
89
90 # Cleared with clearState():
91 var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
92 var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
93
94 # Temporary:
95 var $mOptions, $mTitle, $mOutputType,
96 $mTemplates, // cache of already loaded templates, avoids
97 // multiple SQL queries for the same string
98 $mTemplatePath; // stores an unsorted hash of all the templates already loaded
99 // in this path. Used for loop detection.
100
101 /**#@-*/
102
103 /**
104 * Constructor
105 *
106 * @access public
107 */
108 function Parser() {
109 $this->mTemplates = array();
110 $this->mTemplatePath = array();
111 $this->mTagHooks = array();
112 $this->clearState();
113 }
114
115 /**
116 * Clear Parser state
117 *
118 * @access private
119 */
120 function clearState() {
121 $this->mOutput = new ParserOutput;
122 $this->mAutonumber = 0;
123 $this->mLastSection = "";
124 $this->mDTopen = false;
125 $this->mVariables = false;
126 $this->mIncludeCount = array();
127 $this->mStripState = array();
128 $this->mArgStack = array();
129 $this->mInPre = false;
130 }
131
132 /**
133 * First pass--just handle <nowiki> sections, pass the rest off
134 * to internalParse() which does all the real work.
135 *
136 * @access private
137 * @return ParserOutput a ParserOutput
138 */
139 function parse( $text, &$title, $options, $linestart = true, $clearState = true ) {
140 global $wgUseTidy;
141 $fname = 'Parser::parse';
142 wfProfileIn( $fname );
143
144 if ( $clearState ) {
145 $this->clearState();
146 }
147
148 $this->mOptions = $options;
149 $this->mTitle =& $title;
150 $this->mOutputType = OT_HTML;
151
152 $stripState = NULL;
153 $text = $this->strip( $text, $this->mStripState );
154 $text = $this->internalParse( $text, $linestart );
155 $text = $this->unstrip( $text, $this->mStripState );
156 # Clean up special characters, only run once, next-to-last before doBlockLevels
157 if(!$wgUseTidy) {
158 $fixtags = array(
159 # french spaces, last one Guillemet-left
160 # only if there is something before the space
161 '/(.) (?=\\?|:|;|!|\\302\\273)/i' => '\\1&nbsp;\\2',
162 # french spaces, Guillemet-right
163 "/(\\302\\253) /i"=>"\\1&nbsp;",
164 '/<hr *>/i' => '<hr />',
165 '/<br *>/i' => '<br />',
166 '/<center *>/i' => '<div class="center">',
167 '/<\\/center *>/i' => '</div>',
168 # Clean up spare ampersands; note that we probably ought to be
169 # more careful about named entities.
170 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&amp;'
171 );
172 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
173 } else {
174 $fixtags = array(
175 # french spaces, last one Guillemet-left
176 '/ (\\?|:|;|!|\\302\\273)/i' => '&nbsp;\\1',
177 # french spaces, Guillemet-right
178 '/(\\302\\253) /i' => '\\1&nbsp;',
179 '/<center *>/i' => '<div class="center">',
180 '/<\\/center *>/i' => '</div>'
181 );
182 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
183 }
184 # only once and last
185 $text = $this->doBlockLevels( $text, $linestart );
186 $text = $this->unstripNoWiki( $text, $this->mStripState );
187 $this->mOutput->setText( $text );
188 wfProfileOut( $fname );
189 return $this->mOutput;
190 }
191
192 /**
193 * Get a random string
194 *
195 * @access private
196 * @static
197 */
198 function getRandomString() {
199 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
200 }
201
202 /**
203 * Replaces all occurrences of <$tag>content</$tag> in the text
204 * with a random marker and returns the new text. the output parameter
205 * $content will be an associative array filled with data on the form
206 * $unique_marker => content.
207 *
208 * If $content is already set, the additional entries will be appended
209 * If $tag is set to STRIP_COMMENTS, the function will extract
210 * <!-- HTML comments -->
211 *
212 * @access private
213 * @static
214 */
215 function extractTags($tag, $text, &$content, $uniq_prefix = ''){
216 $rnd = $uniq_prefix . '-' . $tag . Parser::getRandomString();
217 if ( !$content ) {
218 $content = array( );
219 }
220 $n = 1;
221 $stripped = '';
222
223 while ( '' != $text ) {
224 if($tag==STRIP_COMMENTS) {
225 $p = preg_split( '/<!--/i', $text, 2 );
226 } else {
227 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
228 }
229 $stripped .= $p[0];
230 if ( ( count( $p ) < 2 ) || ( '' == $p[1] ) ) {
231 $text = '';
232 } else {
233 if($tag==STRIP_COMMENTS) {
234 $q = preg_split( '/-->/i', $p[1], 2 );
235 } else {
236 $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
237 }
238 $marker = $rnd . sprintf('%08X', $n++);
239 $content[$marker] = $q[0];
240 $stripped .= $marker;
241 $text = $q[1];
242 }
243 }
244 return $stripped;
245 }
246
247 /**
248 * Strips and renders nowiki, pre, math, hiero
249 * If $render is set, performs necessary rendering operations on plugins
250 * Returns the text, and fills an array with data needed in unstrip()
251 * If the $state is already a valid strip state, it adds to the state
252 *
253 * @param bool $stripcomments when set, HTML comments <!-- like this -->
254 * will be stripped in addition to other tags. This is important
255 * for section editing, where these comments cause confusion when
256 * counting the sections in the wikisource
257 *
258 * @access private
259 */
260 function strip( $text, &$state, $stripcomments = false ) {
261 $render = ($this->mOutputType == OT_HTML);
262 $html_content = array();
263 $nowiki_content = array();
264 $math_content = array();
265 $pre_content = array();
266 $comment_content = array();
267 $ext_content = array();
268
269 # Replace any instances of the placeholders
270 $uniq_prefix = UNIQ_PREFIX;
271 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
272
273 # html
274 global $wgRawHtml, $wgWhitelistEdit;
275 if( $wgRawHtml && $wgWhitelistEdit ) {
276 $text = Parser::extractTags('html', $text, $html_content, $uniq_prefix);
277 foreach( $html_content as $marker => $content ) {
278 if ($render ) {
279 # Raw and unchecked for validity.
280 $html_content[$marker] = $content;
281 } else {
282 $html_content[$marker] = '<html>'.$content.'</html>';
283 }
284 }
285 }
286
287 # nowiki
288 $text = Parser::extractTags('nowiki', $text, $nowiki_content, $uniq_prefix);
289 foreach( $nowiki_content as $marker => $content ) {
290 if( $render ){
291 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
292 } else {
293 $nowiki_content[$marker] = '<nowiki>'.$content.'</nowiki>';
294 }
295 }
296
297 # math
298 $text = Parser::extractTags('math', $text, $math_content, $uniq_prefix);
299 foreach( $math_content as $marker => $content ){
300 if( $render ) {
301 if( $this->mOptions->getUseTeX() ) {
302 $math_content[$marker] = renderMath( $content );
303 } else {
304 $math_content[$marker] = '&lt;math&gt;'.$content.'&lt;math&gt;';
305 }
306 } else {
307 $math_content[$marker] = '<math>'.$content.'</math>';
308 }
309 }
310
311 # pre
312 $text = Parser::extractTags('pre', $text, $pre_content, $uniq_prefix);
313 foreach( $pre_content as $marker => $content ){
314 if( $render ){
315 $pre_content[$marker] = '<pre>' . wfEscapeHTMLTagsOnly( $content ) . '</pre>';
316 } else {
317 $pre_content[$marker] = '<pre>'.$content.'</pre>';
318 }
319 }
320
321 # Comments
322 if($stripcomments) {
323 $text = Parser::extractTags(STRIP_COMMENTS, $text, $comment_content, $uniq_prefix);
324 foreach( $comment_content as $marker => $content ){
325 $comment_content[$marker] = '<!--'.$content.'-->';
326 }
327 }
328
329 # Extensions
330 foreach ( $this->mTagHooks as $tag => $callback ) {
331 $ext_contents[$tag] = array();
332 $text = Parser::extractTags( $tag, $text, $ext_content[$tag], $uniq_prefix );
333 foreach( $ext_content[$tag] as $marker => $content ) {
334 if ( $render ) {
335 $ext_content[$tag][$marker] = $callback( $content );
336 } else {
337 $ext_content[$tag][$marker] = "<$tag>$content</$tag>";
338 }
339 }
340 }
341
342 # Merge state with the pre-existing state, if there is one
343 if ( $state ) {
344 $state['html'] = $state['html'] + $html_content;
345 $state['nowiki'] = $state['nowiki'] + $nowiki_content;
346 $state['math'] = $state['math'] + $math_content;
347 $state['pre'] = $state['pre'] + $pre_content;
348 $state['comment'] = $state['comment'] + $comment_content;
349
350 foreach( $ext_content as $tag => $array ) {
351 if ( array_key_exists( $tag, $state ) ) {
352 $state[$tag] = $state[$tag] + $array;
353 }
354 }
355 } else {
356 $state = array(
357 'html' => $html_content,
358 'nowiki' => $nowiki_content,
359 'math' => $math_content,
360 'pre' => $pre_content,
361 'comment' => $comment_content,
362 ) + $ext_content;
363 }
364 return $text;
365 }
366
367 /**
368 * restores pre, math, and heiro removed by strip()
369 *
370 * always call unstripNoWiki() after this one
371 * @access private
372 */
373 function unstrip( $text, &$state ) {
374 # Must expand in reverse order, otherwise nested tags will be corrupted
375 $contentDict = end( $state );
376 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
377 if( key($state) != 'nowiki' && key($state) != 'html') {
378 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
379 $text = str_replace( key( $contentDict ), $content, $text );
380 }
381 }
382 }
383
384 return $text;
385 }
386
387 /**
388 * always call this after unstrip() to preserve the order
389 *
390 * @access private
391 */
392 function unstripNoWiki( $text, &$state ) {
393 # Must expand in reverse order, otherwise nested tags will be corrupted
394 for ( $content = end($state['nowiki']); $content !== false; $content = prev( $state['nowiki'] ) ) {
395 $text = str_replace( key( $state['nowiki'] ), $content, $text );
396 }
397
398 global $wgRawHtml;
399 if ($wgRawHtml) {
400 for ( $content = end($state['html']); $content !== false; $content = prev( $state['html'] ) ) {
401 $text = str_replace( key( $state['html'] ), $content, $text );
402 }
403 }
404
405 return $text;
406 }
407
408 /**
409 * Add an item to the strip state
410 * Returns the unique tag which must be inserted into the stripped text
411 * The tag will be replaced with the original text in unstrip()
412 *
413 * @access private
414 */
415 function insertStripItem( $text, &$state ) {
416 $rnd = UNIQ_PREFIX . '-item' . Parser::getRandomString();
417 if ( !$state ) {
418 $state = array(
419 'html' => array(),
420 'nowiki' => array(),
421 'math' => array(),
422 'pre' => array()
423 );
424 }
425 $state['item'][$rnd] = $text;
426 return $rnd;
427 }
428
429 /**
430 * Return allowed HTML attributes
431 *
432 * @access private
433 */
434 function getHTMLattrs () {
435 $htmlattrs = array( # Allowed attributes--no scripting, etc.
436 'title', 'align', 'lang', 'dir', 'width', 'height',
437 'bgcolor', 'clear', /* BR */ 'noshade', /* HR */
438 'cite', /* BLOCKQUOTE, Q */ 'size', 'face', 'color',
439 /* FONT */ 'type', 'start', 'value', 'compact',
440 /* For various lists, mostly deprecated but safe */
441 'summary', 'width', 'border', 'frame', 'rules',
442 'cellspacing', 'cellpadding', 'valign', 'char',
443 'charoff', 'colgroup', 'col', 'span', 'abbr', 'axis',
444 'headers', 'scope', 'rowspan', 'colspan', /* Tables */
445 'id', 'class', 'name', 'style' /* For CSS */
446 );
447 return $htmlattrs ;
448 }
449
450 /**
451 * Remove non approved attributes and javascript in css
452 *
453 * @access private
454 */
455 function fixTagAttributes ( $t ) {
456 if ( trim ( $t ) == '' ) return '' ; # Saves runtime ;-)
457 $htmlattrs = $this->getHTMLattrs() ;
458
459 # Strip non-approved attributes from the tag
460 $t = preg_replace(
461 '/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e',
462 "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
463 $t);
464
465 $t = str_replace ( '<></>' , '' , $t ) ; # This should fix bug 980557
466
467 # Strip javascript "expression" from stylesheets. Brute force approach:
468 # If anythin offensive is found, all attributes of the HTML tag are dropped
469
470 if( preg_match(
471 '/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is',
472 wfMungeToUtf8( $t ) ) )
473 {
474 $t='';
475 }
476
477 return trim ( $t ) ;
478 }
479
480 /**
481 * interface with html tidy, used if $wgUseTidy = true
482 *
483 * @access public
484 * @static
485 */
486 function tidy ( $text ) {
487 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
488 global $wgInputEncoding, $wgOutputEncoding;
489 $fname = 'Parser::tidy';
490 wfProfileIn( $fname );
491
492 $cleansource = '';
493 $opts = '';
494 switch(strtoupper($wgOutputEncoding)) {
495 case 'ISO-8859-1':
496 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -latin1':' -raw';
497 break;
498 case 'UTF-8':
499 $opts .= ($wgInputEncoding == $wgOutputEncoding)? ' -utf8':' -raw';
500 break;
501 default:
502 $opts .= ' -raw';
503 }
504
505 $wrappedtext = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
506 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
507 '<head><title>test</title></head><body>'.$text.'</body></html>';
508 $descriptorspec = array(
509 0 => array('pipe', 'r'),
510 1 => array('pipe', 'w'),
511 2 => array('file', '/dev/null', 'a')
512 );
513 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts$opts", $descriptorspec, $pipes);
514 if (is_resource($process)) {
515 fwrite($pipes[0], $wrappedtext);
516 fclose($pipes[0]);
517 while (!feof($pipes[1])) {
518 $cleansource .= fgets($pipes[1], 1024);
519 }
520 fclose($pipes[1]);
521 $return_value = proc_close($process);
522 }
523
524 wfProfileOut( $fname );
525
526 if( $cleansource == '' && $text != '') {
527 wfDebug( "Tidy error detected!\n" );
528 return $text . "\n<!-- Tidy found serious XHTML errors -->\n";
529 } else {
530 return $cleansource;
531 }
532 }
533
534 /**
535 * parse the wiki syntax used to render tables
536 *
537 * @access private
538 */
539 function doTableStuff ( $t ) {
540 $fname = 'Parser::doTableStuff';
541 wfProfileIn( $fname );
542
543 $t = explode ( "\n" , $t ) ;
544 $td = array () ; # Is currently a td tag open?
545 $ltd = array () ; # Was it TD or TH?
546 $tr = array () ; # Is currently a tr tag open?
547 $ltr = array () ; # tr attributes
548 $indent_level = 0; # indent level of the table
549 foreach ( $t AS $k => $x )
550 {
551 $x = trim ( $x ) ;
552 $fc = substr ( $x , 0 , 1 ) ;
553 if ( preg_match( '/^(:*)\{\|(.*)$/', $x, $matches ) ) {
554 $indent_level = strlen( $matches[1] );
555 $t[$k] = "\n" .
556 str_repeat( '<dl><dd>', $indent_level ) .
557 '<table ' . $this->fixTagAttributes ( $matches[2] ) . '>' ;
558 array_push ( $td , false ) ;
559 array_push ( $ltd , '' ) ;
560 array_push ( $tr , false ) ;
561 array_push ( $ltr , '' ) ;
562 }
563 else if ( count ( $td ) == 0 ) { } # Don't do any of the following
564 else if ( '|}' == substr ( $x , 0 , 2 ) ) {
565 $z = "</table>\n" ;
566 $l = array_pop ( $ltd ) ;
567 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
568 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
569 array_pop ( $ltr ) ;
570 $t[$k] = $z . str_repeat( '</dd></dl>', $indent_level );
571 }
572 else if ( '|-' == substr ( $x , 0 , 2 ) ) { # Allows for |---------------
573 $x = substr ( $x , 1 ) ;
574 while ( $x != '' && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
575 $z = '' ;
576 $l = array_pop ( $ltd ) ;
577 if ( array_pop ( $tr ) ) $z = '</tr>' . $z ;
578 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
579 array_pop ( $ltr ) ;
580 $t[$k] = $z ;
581 array_push ( $tr , false ) ;
582 array_push ( $td , false ) ;
583 array_push ( $ltd , '' ) ;
584 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
585 }
586 else if ( '|' == $fc || '!' == $fc || '|+' == substr ( $x , 0 , 2 ) ) { # Caption
587 if ( '|+' == substr ( $x , 0 , 2 ) ) {
588 $fc = '+' ;
589 $x = substr ( $x , 1 ) ;
590 }
591 $after = substr ( $x , 1 ) ;
592 if ( $fc == '!' ) $after = str_replace ( '!!' , '||' , $after ) ;
593 $after = explode ( '||' , $after ) ;
594 $t[$k] = '' ;
595 foreach ( $after AS $theline )
596 {
597 $z = '' ;
598 if ( $fc != '+' )
599 {
600 $tra = array_pop ( $ltr ) ;
601 if ( !array_pop ( $tr ) ) $z = '<tr '.$tra.">\n" ;
602 array_push ( $tr , true ) ;
603 array_push ( $ltr , '' ) ;
604 }
605
606 $l = array_pop ( $ltd ) ;
607 if ( array_pop ( $td ) ) $z = '</'.$l.'>' . $z ;
608 if ( $fc == '|' ) $l = 'td' ;
609 else if ( $fc == '!' ) $l = 'th' ;
610 else if ( $fc == '+' ) $l = 'caption' ;
611 else $l = '' ;
612 array_push ( $ltd , $l ) ;
613 $y = explode ( '|' , $theline , 2 ) ;
614 if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
615 else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
616 $t[$k] .= $y ;
617 array_push ( $td , true ) ;
618 }
619 }
620 }
621
622 # Closing open td, tr && table
623 while ( count ( $td ) > 0 )
624 {
625 if ( array_pop ( $td ) ) $t[] = '</td>' ;
626 if ( array_pop ( $tr ) ) $t[] = '</tr>' ;
627 $t[] = '</table>' ;
628 }
629
630 $t = implode ( "\n" , $t ) ;
631 # $t = $this->removeHTMLtags( $t );
632 wfProfileOut( $fname );
633 return $t ;
634 }
635
636 /**
637 * Helper function for parse() that transforms wiki markup into
638 * HTML. Only called for $mOutputType == OT_HTML.
639 *
640 * @access private
641 */
642 function internalParse( $text, $linestart, $args = array(), $isMain=true ) {
643 global $wgContLang;
644
645 $fname = 'Parser::internalParse';
646 wfProfileIn( $fname );
647
648 $text = $this->removeHTMLtags( $text );
649 $text = $this->replaceVariables( $text, $args );
650
651 $text = $wgContLang->convert($text);
652
653 $text = preg_replace( '/(^|\n)-----*/', '\\1<hr />', $text );
654
655 $text = $this->doHeadings( $text );
656 if($this->mOptions->getUseDynamicDates()) {
657 global $wgDateFormatter;
658 $text = $wgDateFormatter->reformat( $this->mOptions->getDateFormat(), $text );
659 }
660 $text = $this->doAllQuotes( $text );
661 $text = $this->doMagicLinks( $text );
662 $text = $this->replaceInternalLinks ( $text );
663 # Another call to replace links and images inside captions of images
664 $text = $this->replaceInternalLinks ( $text );
665 $text = $this->replaceExternalLinks( $text );
666 $text = $this->doTableStuff( $text );
667 $text = $this->formatHeadings( $text, $isMain );
668 $sk =& $this->mOptions->getSkin();
669 $text = $sk->transformContent( $text );
670
671 wfProfileOut( $fname );
672 return $text;
673 }
674
675 /**
676 * Replace special strings like "ISBN xxx" and "RFC xxx" with
677 * magic external links.
678 *
679 * @access private
680 */
681 function &doMagicLinks( &$text ) {
682 global $wgUseGeoMode;
683 $text = $this->magicISBN( $text );
684 if ( isset( $wgUseGeoMode ) && $wgUseGeoMode ) {
685 $text = $this->magicGEO( $text );
686 }
687 $text = $this->magicRFC( $text );
688 return $text;
689 }
690
691 /**
692 * Parse ^^ tokens and return html
693 *
694 * @access private
695 */
696 function doExponent ( $text ) {
697 $fname = 'Parser::doExponent';
698 wfProfileIn( $fname);
699 $text = preg_replace('/\^\^(.*)\^\^/','<small><sup>\\1</sup></small>', $text);
700 wfProfileOut( $fname);
701 return $text;
702 }
703
704 /**
705 * Parse headers and return html
706 *
707 * @access private
708 */
709 function doHeadings( $text ) {
710 $fname = 'Parser::doHeadings';
711 wfProfileIn( $fname );
712 for ( $i = 6; $i >= 1; --$i ) {
713 $h = substr( '======', 0, $i );
714 $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
715 "<h{$i}>\\1</h{$i}>\\2", $text );
716 }
717 wfProfileOut( $fname );
718 return $text;
719 }
720
721 /**
722 * Replace single quotes with HTML markup
723 * @access private
724 * @return string the altered text
725 */
726 function doAllQuotes( $text ) {
727 $fname = 'Parser::doAllQuotes';
728 wfProfileIn( $fname );
729 $outtext = '';
730 $lines = explode( "\n", $text );
731 foreach ( $lines as $line ) {
732 $outtext .= $this->doQuotes ( $line ) . "\n";
733 }
734 $outtext = substr($outtext, 0,-1);
735 wfProfileOut( $fname );
736 return $outtext;
737 }
738
739 /**
740 * Helper function for doAllQuotes()
741 * @access private
742 */
743 function doQuotes( $text ) {
744 $arr = preg_split ("/(''+)/", $text, -1, PREG_SPLIT_DELIM_CAPTURE);
745 if (count ($arr) == 1)
746 return $text;
747 else
748 {
749 # First, do some preliminary work. This may shift some apostrophes from
750 # being mark-up to being text. It also counts the number of occurrences
751 # of bold and italics mark-ups.
752 $i = 0;
753 $numbold = 0;
754 $numitalics = 0;
755 foreach ($arr as $r)
756 {
757 if (($i % 2) == 1)
758 {
759 # If there are ever four apostrophes, assume the first is supposed to
760 # be text, and the remaining three constitute mark-up for bold text.
761 if (strlen ($arr[$i]) == 4)
762 {
763 $arr[$i-1] .= "'";
764 $arr[$i] = "'''";
765 }
766 # If there are more than 5 apostrophes in a row, assume they're all
767 # text except for the last 5.
768 else if (strlen ($arr[$i]) > 5)
769 {
770 $arr[$i-1] .= str_repeat ("'", strlen ($arr[$i]) - 5);
771 $arr[$i] = "'''''";
772 }
773 # Count the number of occurrences of bold and italics mark-ups.
774 # We are not counting sequences of five apostrophes.
775 if (strlen ($arr[$i]) == 2) $numitalics++; else
776 if (strlen ($arr[$i]) == 3) $numbold++; else
777 if (strlen ($arr[$i]) == 5) { $numitalics++; $numbold++; }
778 }
779 $i++;
780 }
781
782 # If there is an odd number of both bold and italics, it is likely
783 # that one of the bold ones was meant to be an apostrophe followed
784 # by italics. Which one we cannot know for certain, but it is more
785 # likely to be one that has a single-letter word before it.
786 if (($numbold % 2 == 1) && ($numitalics % 2 == 1))
787 {
788 $i = 0;
789 $firstsingleletterword = -1;
790 $firstmultiletterword = -1;
791 $firstspace = -1;
792 foreach ($arr as $r)
793 {
794 if (($i % 2 == 1) and (strlen ($r) == 3))
795 {
796 $x1 = substr ($arr[$i-1], -1);
797 $x2 = substr ($arr[$i-1], -2, 1);
798 if ($x1 == ' ') {
799 if ($firstspace == -1) $firstspace = $i;
800 } else if ($x2 == ' ') {
801 if ($firstsingleletterword == -1) $firstsingleletterword = $i;
802 } else {
803 if ($firstmultiletterword == -1) $firstmultiletterword = $i;
804 }
805 }
806 $i++;
807 }
808
809 # If there is a single-letter word, use it!
810 if ($firstsingleletterword > -1)
811 {
812 $arr [ $firstsingleletterword ] = "''";
813 $arr [ $firstsingleletterword-1 ] .= "'";
814 }
815 # If not, but there's a multi-letter word, use that one.
816 else if ($firstmultiletterword > -1)
817 {
818 $arr [ $firstmultiletterword ] = "''";
819 $arr [ $firstmultiletterword-1 ] .= "'";
820 }
821 # ... otherwise use the first one that has neither.
822 # (notice that it is possible for all three to be -1 if, for example,
823 # there is only one pentuple-apostrophe in the line)
824 else if ($firstspace > -1)
825 {
826 $arr [ $firstspace ] = "''";
827 $arr [ $firstspace-1 ] .= "'";
828 }
829 }
830
831 # Now let's actually convert our apostrophic mush to HTML!
832 $output = '';
833 $buffer = '';
834 $state = '';
835 $i = 0;
836 foreach ($arr as $r)
837 {
838 if (($i % 2) == 0)
839 {
840 if ($state == 'both')
841 $buffer .= $r;
842 else
843 $output .= $r;
844 }
845 else
846 {
847 if (strlen ($r) == 2)
848 {
849 if ($state == 'i')
850 { $output .= '</i>'; $state = ''; }
851 else if ($state == 'bi')
852 { $output .= '</i>'; $state = 'b'; }
853 else if ($state == 'ib')
854 { $output .= '</b></i><b>'; $state = 'b'; }
855 else if ($state == 'both')
856 { $output .= '<b><i>'.$buffer.'</i>'; $state = 'b'; }
857 else # $state can be 'b' or ''
858 { $output .= '<i>'; $state .= 'i'; }
859 }
860 else if (strlen ($r) == 3)
861 {
862 if ($state == 'b')
863 { $output .= '</b>'; $state = ''; }
864 else if ($state == 'bi')
865 { $output .= '</i></b><i>'; $state = 'i'; }
866 else if ($state == 'ib')
867 { $output .= '</b>'; $state = 'i'; }
868 else if ($state == 'both')
869 { $output .= '<i><b>'.$buffer.'</b>'; $state = 'i'; }
870 else # $state can be 'i' or ''
871 { $output .= '<b>'; $state .= 'b'; }
872 }
873 else if (strlen ($r) == 5)
874 {
875 if ($state == 'b')
876 { $output .= '</b><i>'; $state = 'i'; }
877 else if ($state == 'i')
878 { $output .= '</i><b>'; $state = 'b'; }
879 else if ($state == 'bi')
880 { $output .= '</i></b>'; $state = ''; }
881 else if ($state == 'ib')
882 { $output .= '</b></i>'; $state = ''; }
883 else if ($state == 'both')
884 { $output .= '<i><b>'.$buffer.'</b></i>'; $state = ''; }
885 else # ($state == '')
886 { $buffer = ''; $state = 'both'; }
887 }
888 }
889 $i++;
890 }
891 # Now close all remaining tags. Notice that the order is important.
892 if ($state == 'b' || $state == 'ib')
893 $output .= '</b>';
894 if ($state == 'i' || $state == 'bi' || $state == 'ib')
895 $output .= '</i>';
896 if ($state == 'bi')
897 $output .= '</b>';
898 if ($state == 'both')
899 $output .= '<b><i>'.$buffer.'</i></b>';
900 return $output;
901 }
902 }
903
904 /**
905 * Replace external links
906 *
907 * Note: we have to do external links before the internal ones,
908 * and otherwise take great care in the order of things here, so
909 * that we don't end up interpreting some URLs twice.
910 *
911 * @access private
912 */
913 function replaceExternalLinks( $text ) {
914 $fname = 'Parser::replaceExternalLinks';
915 wfProfileIn( $fname );
916
917 $sk =& $this->mOptions->getSkin();
918 $linktrail = wfMsgForContent('linktrail');
919 $bits = preg_split( EXT_LINK_BRACKETED, $text, -1, PREG_SPLIT_DELIM_CAPTURE );
920
921 $s = $this->replaceFreeExternalLinks( array_shift( $bits ) );
922
923 $i = 0;
924 while ( $i<count( $bits ) ) {
925 $url = $bits[$i++];
926 $protocol = $bits[$i++];
927 $text = $bits[$i++];
928 $trail = $bits[$i++];
929
930 # If the link text is an image URL, replace it with an <img> tag
931 # This happened by accident in the original parser, but some people used it extensively
932 $img = $this->maybeMakeImageLink( $text );
933 if ( $img !== false ) {
934 $text = $img;
935 }
936
937 $dtrail = '';
938
939 # No link text, e.g. [http://domain.tld/some.link]
940 if ( $text == '' ) {
941 # Autonumber if allowed
942 if ( strpos( HTTP_PROTOCOLS, $protocol ) !== false ) {
943 $text = '[' . ++$this->mAutonumber . ']';
944 } else {
945 # Otherwise just use the URL
946 $text = htmlspecialchars( $url );
947 }
948 } else {
949 # Have link text, e.g. [http://domain.tld/some.link text]s
950 # Check for trail
951 if ( preg_match( $linktrail, $trail, $m2 ) ) {
952 $dtrail = $m2[1];
953 $trail = $m2[2];
954 }
955 }
956
957 $encUrl = htmlspecialchars( $url );
958 # Bit in parentheses showing the URL for the printable version
959 if( $url == $text || preg_match( "!$protocol://" . preg_quote( $text, '/' ) . "/?$!", $url ) ) {
960 $paren = '';
961 } else {
962 # Expand the URL for printable version
963 if ( ! $sk->suppressUrlExpansion() ) {
964 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $encUrl ) . "</i>)</span>";
965 } else {
966 $paren = '';
967 }
968 }
969
970 # Process the trail (i.e. everything after this link up until start of the next link),
971 # replacing any non-bracketed links
972 $trail = $this->replaceFreeExternalLinks( $trail );
973
974 $la = $sk->getExternalLinkAttributes( $url, $text );
975
976 # Use the encoded URL
977 # This means that users can paste URLs directly into the text
978 # Funny characters like &ouml; aren't valid in URLs anyway
979 # This was changed in August 2004
980 $s .= "<a href=\"{$url}\"{$la}>{$text}</a>{$dtrail}{$paren}{$trail}";
981 }
982
983 wfProfileOut( $fname );
984 return $s;
985 }
986
987 /**
988 * Replace anything that looks like a URL with a link
989 * @access private
990 */
991 function replaceFreeExternalLinks( $text ) {
992 $bits = preg_split( '/((?:'.URL_PROTOCOLS.'):)/', $text, -1, PREG_SPLIT_DELIM_CAPTURE );
993 $s = array_shift( $bits );
994 $i = 0;
995
996 $sk =& $this->mOptions->getSkin();
997
998 while ( $i < count( $bits ) ){
999 $protocol = $bits[$i++];
1000 $remainder = $bits[$i++];
1001
1002 if ( preg_match( '/^('.EXT_LINK_URL_CLASS.'+)(.*)$/s', $remainder, $m ) ) {
1003 # Found some characters after the protocol that look promising
1004 $url = $protocol . $m[1];
1005 $trail = $m[2];
1006
1007 # Move trailing punctuation to $trail
1008 $sep = ',;\.:!?';
1009 # If there is no left bracket, then consider right brackets fair game too
1010 if ( strpos( $url, '(' ) === false ) {
1011 $sep .= ')';
1012 }
1013
1014 $numSepChars = strspn( strrev( $url ), $sep );
1015 if ( $numSepChars ) {
1016 $trail = substr( $url, -$numSepChars ) . $trail;
1017 $url = substr( $url, 0, -$numSepChars );
1018 }
1019
1020 # Replace &amp; from obsolete syntax with &
1021 $url = str_replace( '&amp;', '&', $url );
1022
1023 # Is this an external image?
1024 $text = $this->maybeMakeImageLink( $url );
1025 if ( $text === false ) {
1026 # Not an image, make a link
1027 $text = $sk->makeExternalLink( $url, $url );
1028 }
1029 $s .= $text . $trail;
1030 } else {
1031 $s .= $protocol . $remainder;
1032 }
1033 }
1034 return $s;
1035 }
1036
1037 /**
1038 * make an image if it's allowed
1039 * @access private
1040 */
1041 function maybeMakeImageLink( $url ) {
1042 $sk =& $this->mOptions->getSkin();
1043 $text = false;
1044 if ( $this->mOptions->getAllowExternalImages() ) {
1045 if ( preg_match( EXT_IMAGE_REGEX, $url ) ) {
1046 # Image found
1047 $text = $sk->makeImage( htmlspecialchars( $url ) );
1048 }
1049 }
1050 return $text;
1051 }
1052
1053 /**
1054 * Process [[ ]] wikilinks
1055 *
1056 * @access private
1057 */
1058 function replaceInternalLinks( $s ) {
1059 global $wgLang, $wgContLang, $wgLinkCache;
1060 static $fname = 'Parser::replaceInternalLinks' ;
1061 wfProfileIn( $fname );
1062
1063 wfProfileIn( $fname.'-setup' );
1064 static $tc = FALSE;
1065 # the % is needed to support urlencoded titles as well
1066 if ( !$tc ) { $tc = Title::legalChars() . '#%'; }
1067 $sk =& $this->mOptions->getSkin();
1068
1069 $redirect = MagicWord::get ( MAG_REDIRECT ) ;
1070
1071 $a = explode( '[[', ' ' . $s );
1072 $s = array_shift( $a );
1073 $s = substr( $s, 1 );
1074
1075 # Match a link having the form [[namespace:link|alternate]]trail
1076 static $e1 = FALSE;
1077 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
1078 # Match the end of a line for a word that's not followed by whitespace,
1079 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
1080 static $e2 = '/^(.*?)([a-zA-Z\x80-\xff]+)$/sD';
1081
1082 $useLinkPrefixExtension = $wgContLang->linkPrefixExtension();
1083 # Special and Media are pseudo-namespaces; no pages actually exist in them
1084
1085 $nottalk = !Namespace::isTalk( $this->mTitle->getNamespace() );
1086
1087 if ( $useLinkPrefixExtension ) {
1088 if ( preg_match( $e2, $s, $m ) ) {
1089 $first_prefix = $m[2];
1090 $s = $m[1];
1091 } else {
1092 $first_prefix = false;
1093 }
1094 } else {
1095 $prefix = '';
1096 }
1097
1098 wfProfileOut( $fname.'-setup' );
1099
1100 # start procedeeding each line
1101 foreach ( $a as $line ) {
1102 wfProfileIn( $fname.'-prefixhandling' );
1103 if ( $useLinkPrefixExtension ) {
1104 if ( preg_match( $e2, $s, $m ) ) {
1105 $prefix = $m[2];
1106 $s = $m[1];
1107 } else {
1108 $prefix='';
1109 }
1110 # first link
1111 if($first_prefix) {
1112 $prefix = $first_prefix;
1113 $first_prefix = false;
1114 }
1115 }
1116 wfProfileOut( $fname.'-prefixhandling' );
1117
1118 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
1119 $text = $m[2];
1120 # fix up urlencoded title texts
1121 if(preg_match('/%/', $m[1] )) $m[1] = urldecode($m[1]);
1122 $trail = $m[3];
1123 } else { # Invalid form; output directly
1124 $s .= $prefix . '[[' . $line ;
1125 continue;
1126 }
1127
1128 # Don't allow internal links to pages containing
1129 # PROTO: where PROTO is a valid URL protocol; these
1130 # should be external links.
1131 if (preg_match('/((?:'.URL_PROTOCOLS.'):)/', $m[1])) {
1132 $s .= $prefix . '[[' . $line ;
1133 continue;
1134 }
1135
1136 # Make subpage if necessary, and check for leading ':'
1137 $noforce = true;
1138 $link = $this->maybeDoSubpageLink( $m[1], $noforce, $text );
1139
1140 $wasblank = ( '' == $text );
1141 if( $wasblank ) $text = $link;
1142
1143 $nt = Title::newFromText( $link );
1144 if( !$nt ) {
1145 $s .= $prefix . '[[' . $line;
1146 continue;
1147 }
1148
1149 $ns = $nt->getNamespace();
1150 $iw = $nt->getInterWiki();
1151
1152 # Link not escaped by : , create the various objects
1153 if( $noforce ) {
1154
1155 # Interwikis
1156 if( $iw && $this->mOptions->getInterwikiMagic() && $nottalk && $wgContLang->getLanguageName( $iw ) ) {
1157 array_push( $this->mOutput->mLanguageLinks, $nt->getFullText() );
1158 $tmp = $prefix . $trail ;
1159 $s .= (trim($tmp) == '')? '': $tmp;
1160 continue;
1161 }
1162
1163 if ( $ns == NS_IMAGE ) {
1164 $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
1165 $wgLinkCache->addImageLinkObj( $nt );
1166 continue;
1167 }
1168
1169 if ( $ns == NS_CATEGORY ) {
1170 $t = $nt->getText() ;
1171 $nnt = Title::newFromText ( Namespace::getCanonicalName(NS_CATEGORY).':'.$t ) ;
1172
1173 $wgLinkCache->suspend(); # Don't save in links/brokenlinks
1174 $pPLC=$sk->postParseLinkColour();
1175 $sk->postParseLinkColour( false );
1176 $t = $sk->makeLinkObj( $nnt, $t, '', '' , $prefix );
1177 $sk->postParseLinkColour( $pPLC );
1178 $wgLinkCache->resume();
1179
1180 if ( $wasblank ) {
1181 if ( $this->mTitle->getNamespace() == NS_CATEGORY ) {
1182 $sortkey = $this->mTitle->getText();
1183 } else {
1184 $sortkey = $this->mTitle->getPrefixedText();
1185 }
1186 } else {
1187 $sortkey = $text;
1188 }
1189 $wgLinkCache->addCategoryLinkObj( $nt, $sortkey );
1190 $this->mOutput->mCategoryLinks[] = $t ;
1191 $s .= $prefix . $trail ;
1192 continue;
1193 }
1194 }
1195
1196 if( ( $nt->getPrefixedText() === $this->mTitle->getPrefixedText() ) &&
1197 ( strpos( $link, '#' ) === FALSE ) ) {
1198 # Self-links are handled specially; generally de-link and change to bold.
1199 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, '', $trail );
1200 continue;
1201 }
1202
1203 if( $ns == NS_MEDIA ) {
1204 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1205 $wgLinkCache->addImageLinkObj( $nt );
1206 continue;
1207 } elseif( $ns == NS_SPECIAL ) {
1208 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, '', $trail );
1209 continue;
1210 }
1211 $s .= $sk->makeLinkObj( $nt, $text, '', $trail, $prefix );
1212 }
1213 wfProfileOut( $fname );
1214 return $s;
1215 }
1216
1217 /**
1218 * Handle link to subpage if necessary
1219 * @param $target string the source of the link
1220 * @param $target set to true unless target began with ':'
1221 * @param &$text the link text, modified as necessary
1222 * @return string the full name of the link
1223 * @access private
1224 */
1225 function maybeDoSubpageLink($target, &$noforce, &$text) {
1226 # Valid link forms:
1227 # Foobar -- normal
1228 # :Foobar -- override special treatment of prefix (images, language links)
1229 # /Foobar -- convert to CurrentPage/Foobar
1230 # /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
1231 global $wgNamespacesWithSubpages;
1232
1233 # Look at the first character
1234 $c = $target{0};
1235 $noforce = ($c != ':');
1236
1237 # subpage
1238 if( $c == '/' ) {
1239 # / at end means we don't want the slash to be shown
1240 if(substr($target,-1,1)=='/') {
1241 $target=substr($target,1,-1);
1242 $noslash=$target;
1243 } else {
1244 $noslash=substr($target,1);
1245 }
1246
1247 # Some namespaces don't allow subpages
1248 if(!empty($wgNamespacesWithSubpages[$this->mTitle->getNamespace()])) {
1249 # subpages allowed here
1250 $ret = $this->mTitle->getPrefixedText(). '/' . trim($noslash);
1251 if( '' === $text ) {
1252 $text = $target;
1253 } # this might be changed for ugliness reasons
1254 } else {
1255 # no subpage allowed, use standard link
1256 $ret = $target;
1257 }
1258 } elseif( $noforce ) {
1259 # no subpage
1260 $ret = $target;
1261 } else {
1262 # We don't want to keep the first character
1263 $ret = substr( $target, 1 );
1264 }
1265
1266 return $ret;
1267 }
1268
1269 /**#@+
1270 * Used by doBlockLevels()
1271 * @access private
1272 */
1273 /* private */ function closeParagraph() {
1274 $result = '';
1275 if ( '' != $this->mLastSection ) {
1276 $result = '</' . $this->mLastSection . ">\n";
1277 }
1278 $this->mInPre = false;
1279 $this->mLastSection = '';
1280 return $result;
1281 }
1282 # getCommon() returns the length of the longest common substring
1283 # of both arguments, starting at the beginning of both.
1284 #
1285 /* private */ function getCommon( $st1, $st2 ) {
1286 $fl = strlen( $st1 );
1287 $shorter = strlen( $st2 );
1288 if ( $fl < $shorter ) { $shorter = $fl; }
1289
1290 for ( $i = 0; $i < $shorter; ++$i ) {
1291 if ( $st1{$i} != $st2{$i} ) { break; }
1292 }
1293 return $i;
1294 }
1295 # These next three functions open, continue, and close the list
1296 # element appropriate to the prefix character passed into them.
1297 #
1298 /* private */ function openList( $char ) {
1299 $result = $this->closeParagraph();
1300
1301 if ( '*' == $char ) { $result .= '<ul><li>'; }
1302 else if ( '#' == $char ) { $result .= '<ol><li>'; }
1303 else if ( ':' == $char ) { $result .= '<dl><dd>'; }
1304 else if ( ';' == $char ) {
1305 $result .= '<dl><dt>';
1306 $this->mDTopen = true;
1307 }
1308 else { $result = '<!-- ERR 1 -->'; }
1309
1310 return $result;
1311 }
1312
1313 /* private */ function nextItem( $char ) {
1314 if ( '*' == $char || '#' == $char ) { return '</li><li>'; }
1315 else if ( ':' == $char || ';' == $char ) {
1316 $close = '</dd>';
1317 if ( $this->mDTopen ) { $close = '</dt>'; }
1318 if ( ';' == $char ) {
1319 $this->mDTopen = true;
1320 return $close . '<dt>';
1321 } else {
1322 $this->mDTopen = false;
1323 return $close . '<dd>';
1324 }
1325 }
1326 return '<!-- ERR 2 -->';
1327 }
1328
1329 /* private */ function closeList( $char ) {
1330 if ( '*' == $char ) { $text = '</li></ul>'; }
1331 else if ( '#' == $char ) { $text = '</li></ol>'; }
1332 else if ( ':' == $char ) {
1333 if ( $this->mDTopen ) {
1334 $this->mDTopen = false;
1335 $text = '</dt></dl>';
1336 } else {
1337 $text = '</dd></dl>';
1338 }
1339 }
1340 else { return '<!-- ERR 3 -->'; }
1341 return $text."\n";
1342 }
1343 /**#@-*/
1344
1345 /**
1346 * Make lists from lines starting with ':', '*', '#', etc.
1347 *
1348 * @access private
1349 * @return string the lists rendered as HTML
1350 */
1351 function doBlockLevels( $text, $linestart ) {
1352 $fname = 'Parser::doBlockLevels';
1353 wfProfileIn( $fname );
1354
1355 # Parsing through the text line by line. The main thing
1356 # happening here is handling of block-level elements p, pre,
1357 # and making lists from lines starting with * # : etc.
1358 #
1359 $textLines = explode( "\n", $text );
1360
1361 $lastPrefix = $output = $lastLine = '';
1362 $this->mDTopen = $inBlockElem = false;
1363 $prefixLength = 0;
1364 $paragraphStack = false;
1365
1366 if ( !$linestart ) {
1367 $output .= array_shift( $textLines );
1368 }
1369 foreach ( $textLines as $oLine ) {
1370 $lastPrefixLength = strlen( $lastPrefix );
1371 $preCloseMatch = preg_match('/<\\/pre/i', $oLine );
1372 $preOpenMatch = preg_match('/<pre/i', $oLine );
1373 if ( !$this->mInPre ) {
1374 # Multiple prefixes may abut each other for nested lists.
1375 $prefixLength = strspn( $oLine, '*#:;' );
1376 $pref = substr( $oLine, 0, $prefixLength );
1377
1378 # eh?
1379 $pref2 = str_replace( ';', ':', $pref );
1380 $t = substr( $oLine, $prefixLength );
1381 $this->mInPre = !empty($preOpenMatch);
1382 } else {
1383 # Don't interpret any other prefixes in preformatted text
1384 $prefixLength = 0;
1385 $pref = $pref2 = '';
1386 $t = $oLine;
1387 }
1388
1389 # List generation
1390 if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1391 # Same as the last item, so no need to deal with nesting or opening stuff
1392 $output .= $this->nextItem( substr( $pref, -1 ) );
1393 $paragraphStack = false;
1394
1395 if ( substr( $pref, -1 ) == ';') {
1396 # The one nasty exception: definition lists work like this:
1397 # ; title : definition text
1398 # So we check for : in the remainder text to split up the
1399 # title and definition, without b0rking links.
1400 # FIXME: This is not foolproof. Something better in Tokenizer might help.
1401 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1402 $term = $match[1];
1403 $output .= $term . $this->nextItem( ':' );
1404 $t = $match[2];
1405 }
1406 }
1407 } elseif( $prefixLength || $lastPrefixLength ) {
1408 # Either open or close a level...
1409 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1410 $paragraphStack = false;
1411
1412 while( $commonPrefixLength < $lastPrefixLength ) {
1413 $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1414 --$lastPrefixLength;
1415 }
1416 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1417 $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1418 }
1419 while ( $prefixLength > $commonPrefixLength ) {
1420 $char = substr( $pref, $commonPrefixLength, 1 );
1421 $output .= $this->openList( $char );
1422
1423 if ( ';' == $char ) {
1424 # FIXME: This is dupe of code above
1425 if( preg_match( '/^(.*?(?:\s|&nbsp;)):(.*)$/', $t, $match ) ) {
1426 $term = $match[1];
1427 $output .= $term . $this->nextItem( ':' );
1428 $t = $match[2];
1429 }
1430 }
1431 ++$commonPrefixLength;
1432 }
1433 $lastPrefix = $pref2;
1434 }
1435 if( 0 == $prefixLength ) {
1436 # No prefix (not in list)--go to paragraph mode
1437 $uniq_prefix = UNIQ_PREFIX;
1438 // XXX: use a stack for nestable elements like span, table and div
1439 $openmatch = preg_match('/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|<p|<ul|<li|<\\/tr|<\\/td|<\\/th)/i', $t );
1440 $closematch = preg_match(
1441 '/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'.
1442 '<td|<th|<div|<\\/div|<hr|<\\/pre|<\\/p|'.$uniq_prefix.'-pre|<\\/li|<\\/ul)/i', $t );
1443 if ( $openmatch or $closematch ) {
1444 $paragraphStack = false;
1445 $output .= $this->closeParagraph();
1446 if($preOpenMatch and !$preCloseMatch) {
1447 $this->mInPre = true;
1448 }
1449 if ( $closematch ) {
1450 $inBlockElem = false;
1451 } else {
1452 $inBlockElem = true;
1453 }
1454 } else if ( !$inBlockElem && !$this->mInPre ) {
1455 if ( ' ' == $t{0} and ( $this->mLastSection == 'pre' or trim($t) != '' ) ) {
1456 // pre
1457 if ($this->mLastSection != 'pre') {
1458 $paragraphStack = false;
1459 $output .= $this->closeParagraph().'<pre>';
1460 $this->mLastSection = 'pre';
1461 }
1462 $t = substr( $t, 1 );
1463 } else {
1464 // paragraph
1465 if ( '' == trim($t) ) {
1466 if ( $paragraphStack ) {
1467 $output .= $paragraphStack.'<br />';
1468 $paragraphStack = false;
1469 $this->mLastSection = 'p';
1470 } else {
1471 if ($this->mLastSection != 'p' ) {
1472 $output .= $this->closeParagraph();
1473 $this->mLastSection = '';
1474 $paragraphStack = '<p>';
1475 } else {
1476 $paragraphStack = '</p><p>';
1477 }
1478 }
1479 } else {
1480 if ( $paragraphStack ) {
1481 $output .= $paragraphStack;
1482 $paragraphStack = false;
1483 $this->mLastSection = 'p';
1484 } else if ($this->mLastSection != 'p') {
1485 $output .= $this->closeParagraph().'<p>';
1486 $this->mLastSection = 'p';
1487 }
1488 }
1489 }
1490 }
1491 }
1492 if ($paragraphStack === false) {
1493 $output .= $t."\n";
1494 }
1495 }
1496 while ( $prefixLength ) {
1497 $output .= $this->closeList( $pref2{$prefixLength-1} );
1498 --$prefixLength;
1499 }
1500 if ( '' != $this->mLastSection ) {
1501 $output .= '</' . $this->mLastSection . '>';
1502 $this->mLastSection = '';
1503 }
1504
1505 wfProfileOut( $fname );
1506 return $output;
1507 }
1508
1509 /**
1510 * Return value of a magic variable (like PAGENAME)
1511 *
1512 * @access private
1513 */
1514 function getVariableValue( $index ) {
1515 global $wgContLang, $wgSitename, $wgServer;
1516
1517 switch ( $index ) {
1518 case MAG_CURRENTMONTH:
1519 return $wgContLang->formatNum( date( 'm' ) );
1520 case MAG_CURRENTMONTHNAME:
1521 return $wgContLang->getMonthName( date('n') );
1522 case MAG_CURRENTMONTHNAMEGEN:
1523 return $wgContLang->getMonthNameGen( date('n') );
1524 case MAG_CURRENTDAY:
1525 return $wgContLang->formatNum( date('j') );
1526 case MAG_PAGENAME:
1527 return $this->mTitle->getText();
1528 case MAG_PAGENAMEE:
1529 return $this->mTitle->getPartialURL();
1530 case MAG_NAMESPACE:
1531 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1532 return $wgContLang->getNsText($this->mTitle->getNamespace()); # Patch by Dori
1533 case MAG_CURRENTDAYNAME:
1534 return $wgContLang->getWeekdayName( date('w')+1 );
1535 case MAG_CURRENTYEAR:
1536 return $wgContLang->formatNum( date( 'Y' ) );
1537 case MAG_CURRENTTIME:
1538 return $wgContLang->time( wfTimestampNow(), false );
1539 case MAG_NUMBEROFARTICLES:
1540 return $wgContLang->formatNum( wfNumberOfArticles() );
1541 case MAG_SITENAME:
1542 return $wgSitename;
1543 case MAG_SERVER:
1544 return $wgServer;
1545 default:
1546 return NULL;
1547 }
1548 }
1549
1550 /**
1551 * initialise the magic variables (like CURRENTMONTHNAME)
1552 *
1553 * @access private
1554 */
1555 function initialiseVariables() {
1556 global $wgVariableIDs;
1557 $this->mVariables = array();
1558 foreach ( $wgVariableIDs as $id ) {
1559 $mw =& MagicWord::get( $id );
1560 $mw->addToArray( $this->mVariables, $this->getVariableValue( $id ) );
1561 }
1562 }
1563
1564 /**
1565 * Replace magic variables, templates, and template arguments
1566 * with the appropriate text. Templates are substituted recursively,
1567 * taking care to avoid infinite loops.
1568 *
1569 * Note that the substitution depends on value of $mOutputType:
1570 * OT_WIKI: only {{subst:}} templates
1571 * OT_MSG: only magic variables
1572 * OT_HTML: all templates and magic variables
1573 *
1574 * @param string $tex The text to transform
1575 * @param array $args Key-value pairs representing template parameters to substitute
1576 * @access private
1577 */
1578 function replaceVariables( $text, $args = array() ) {
1579 global $wgLang, $wgScript, $wgArticlePath;
1580
1581 # Prevent too big inclusions
1582 if(strlen($text)> MAX_INCLUDE_SIZE)
1583 return $text;
1584
1585 $fname = 'Parser::replaceVariables';
1586 wfProfileIn( $fname );
1587
1588 $titleChars = Title::legalChars();
1589
1590 # This function is called recursively. To keep track of arguments we need a stack:
1591 array_push( $this->mArgStack, $args );
1592
1593 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1594 $GLOBALS['wgCurParser'] =& $this;
1595
1596 # Variable substitution
1597 $text = preg_replace_callback( "/{{([$titleChars]*?)}}/", 'wfVariableSubstitution', $text );
1598
1599 if ( $this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI ) {
1600 # Argument substitution
1601 $text = preg_replace_callback( "/{{{([$titleChars]*?)}}}/", 'wfArgSubstitution', $text );
1602 }
1603 # Template substitution
1604 $regex = '/(\\n|{)?{{(['.$titleChars.']*)(\\|.*?|)}}/s';
1605 $text = preg_replace_callback( $regex, 'wfBraceSubstitution', $text );
1606
1607 array_pop( $this->mArgStack );
1608
1609 wfProfileOut( $fname );
1610 return $text;
1611 }
1612
1613 /**
1614 * Replace magic variables
1615 * @access private
1616 */
1617 function variableSubstitution( $matches ) {
1618 if ( !$this->mVariables ) {
1619 $this->initialiseVariables();
1620 }
1621 $skip = false;
1622 if ( $this->mOutputType == OT_WIKI ) {
1623 # Do only magic variables prefixed by SUBST
1624 $mwSubst =& MagicWord::get( MAG_SUBST );
1625 if (!$mwSubst->matchStartAndRemove( $matches[1] ))
1626 $skip = true;
1627 # Note that if we don't substitute the variable below,
1628 # we don't remove the {{subst:}} magic word, in case
1629 # it is a template rather than a magic variable.
1630 }
1631 if ( !$skip && array_key_exists( $matches[1], $this->mVariables ) ) {
1632 $text = $this->mVariables[$matches[1]];
1633 $this->mOutput->mContainsOldMagic = true;
1634 } else {
1635 $text = $matches[0];
1636 }
1637 return $text;
1638 }
1639
1640 # Split template arguments
1641 function getTemplateArgs( $argsString ) {
1642 if ( $argsString === '' ) {
1643 return array();
1644 }
1645
1646 $args = explode( '|', substr( $argsString, 1 ) );
1647
1648 # If any of the arguments contains a '[[' but no ']]', it needs to be
1649 # merged with the next arg because the '|' character between belongs
1650 # to the link syntax and not the template parameter syntax.
1651 $argc = count($args);
1652 $i = 0;
1653 for ( $i = 0; $i < $argc-1; $i++ ) {
1654 if ( substr_count ( $args[$i], '[[' ) != substr_count ( $args[$i], ']]' ) ) {
1655 $args[$i] .= '|'.$args[$i+1];
1656 array_splice($args, $i+1, 1);
1657 $i--;
1658 $argc--;
1659 }
1660 }
1661
1662 return $args;
1663 }
1664
1665 /**
1666 * Return the text of a template, after recursively
1667 * replacing any variables or templates within the template.
1668 *
1669 * @param array $matches The parts of the template
1670 * $matches[1]: the title, i.e. the part before the |
1671 * $matches[2]: the parameters (including a leading |), if any
1672 * @return string the text of the template
1673 * @access private
1674 */
1675 function braceSubstitution( $matches ) {
1676 global $wgLinkCache, $wgContLang;
1677 $fname = 'Parser::braceSubstitution';
1678 $found = false;
1679 $nowiki = false;
1680 $noparse = false;
1681
1682 $title = NULL;
1683
1684 # Need to know if the template comes at the start of a line,
1685 # to treat the beginning of the template like the beginning
1686 # of a line for tables and block-level elements.
1687 $linestart = $matches[1];
1688
1689 # $part1 is the bit before the first |, and must contain only title characters
1690 # $args is a list of arguments, starting from index 0, not including $part1
1691
1692 $part1 = $matches[2];
1693 # If the third subpattern matched anything, it will start with |
1694
1695 $args = $this->getTemplateArgs($matches[3]);
1696 $argc = count( $args );
1697
1698 # Don't parse {{{}}} because that's only for template arguments
1699 if ( $linestart === '{' ) {
1700 $text = $matches[0];
1701 $found = true;
1702 $noparse = true;
1703 }
1704
1705 # SUBST
1706 if ( !$found ) {
1707 $mwSubst =& MagicWord::get( MAG_SUBST );
1708 if ( $mwSubst->matchStartAndRemove( $part1 ) xor ($this->mOutputType == OT_WIKI) ) {
1709 # One of two possibilities is true:
1710 # 1) Found SUBST but not in the PST phase
1711 # 2) Didn't find SUBST and in the PST phase
1712 # In either case, return without further processing
1713 $text = $matches[0];
1714 $found = true;
1715 $noparse = true;
1716 }
1717 }
1718
1719 # MSG, MSGNW and INT
1720 if ( !$found ) {
1721 # Check for MSGNW:
1722 $mwMsgnw =& MagicWord::get( MAG_MSGNW );
1723 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1724 $nowiki = true;
1725 } else {
1726 # Remove obsolete MSG:
1727 $mwMsg =& MagicWord::get( MAG_MSG );
1728 $mwMsg->matchStartAndRemove( $part1 );
1729 }
1730
1731 # Check if it is an internal message
1732 $mwInt =& MagicWord::get( MAG_INT );
1733 if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1734 if ( $this->incrementIncludeCount( 'int:'.$part1 ) ) {
1735 $text = $linestart . wfMsgReal( $part1, $args, true );
1736 $found = true;
1737 }
1738 }
1739 }
1740
1741 # NS
1742 if ( !$found ) {
1743 # Check for NS: (namespace expansion)
1744 $mwNs = MagicWord::get( MAG_NS );
1745 if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1746 if ( intval( $part1 ) ) {
1747 $text = $linestart . $wgContLang->getNsText( intval( $part1 ) );
1748 $found = true;
1749 } else {
1750 $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1751 if ( !is_null( $index ) ) {
1752 $text = $linestart . $wgContLang->getNsText( $index );
1753 $found = true;
1754 }
1755 }
1756 }
1757 }
1758
1759 # LOCALURL and LOCALURLE
1760 if ( !$found ) {
1761 $mwLocal = MagicWord::get( MAG_LOCALURL );
1762 $mwLocalE = MagicWord::get( MAG_LOCALURLE );
1763
1764 if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1765 $func = 'getLocalURL';
1766 } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1767 $func = 'escapeLocalURL';
1768 } else {
1769 $func = '';
1770 }
1771
1772 if ( $func !== '' ) {
1773 $title = Title::newFromText( $part1 );
1774 if ( !is_null( $title ) ) {
1775 if ( $argc > 0 ) {
1776 $text = $linestart . $title->$func( $args[0] );
1777 } else {
1778 $text = $linestart . $title->$func();
1779 }
1780 $found = true;
1781 }
1782 }
1783 }
1784
1785 # GRAMMAR
1786 if ( !$found && $argc == 1 ) {
1787 $mwGrammar =& MagicWord::get( MAG_GRAMMAR );
1788 if ( $mwGrammar->matchStartAndRemove( $part1 ) ) {
1789 $text = $linestart . $wgContLang->convertGrammar( $args[0], $part1 );
1790 $found = true;
1791 }
1792 }
1793
1794 # Template table test
1795
1796 # Did we encounter this template already? If yes, it is in the cache
1797 # and we need to check for loops.
1798 if ( !$found && isset( $this->mTemplates[$part1] ) ) {
1799 # set $text to cached message.
1800 $text = $linestart . $this->mTemplates[$part1];
1801 $found = true;
1802
1803 # Infinite loop test
1804 if ( isset( $this->mTemplatePath[$part1] ) ) {
1805 $noparse = true;
1806 $found = true;
1807 $text .= '<!-- WARNING: template loop detected -->';
1808 }
1809 }
1810
1811 # Load from database
1812 $itcamefromthedatabase = false;
1813 if ( !$found ) {
1814 $ns = NS_TEMPLATE;
1815 $part1 = $this->maybeDoSubpageLink( $part1, $noforce, $subpage='' );
1816 if ($subpage !== '') {
1817 $ns = $this->mTitle->getNamespace();
1818 }
1819 $title = Title::newFromText( $part1, $ns );
1820 if ( !is_null( $title ) && !$title->isExternal() ) {
1821 # Check for excessive inclusion
1822 $dbk = $title->getPrefixedDBkey();
1823 if ( $this->incrementIncludeCount( $dbk ) ) {
1824 # This should never be reached.
1825 $article = new Article( $title );
1826 $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1827 if ( $articleContent !== false ) {
1828 $found = true;
1829 $text = $linestart . $articleContent;
1830 $itcamefromthedatabase = true;
1831 }
1832 }
1833
1834 # If the title is valid but undisplayable, make a link to it
1835 if ( $this->mOutputType == OT_HTML && !$found ) {
1836 $text = $linestart . '[['.$title->getPrefixedText().']]';
1837 $found = true;
1838 }
1839
1840 # Template cache array insertion
1841 $this->mTemplates[$part1] = $text;
1842 }
1843 }
1844
1845 # Recursive parsing, escaping and link table handling
1846 # Only for HTML output
1847 if ( $nowiki && $found && $this->mOutputType == OT_HTML ) {
1848 $text = wfEscapeWikiText( $text );
1849 } elseif ( ($this->mOutputType == OT_HTML || $this->mOutputType == OT_WIKI) && $found && !$noparse) {
1850 # Clean up argument array
1851 $assocArgs = array();
1852 $index = 1;
1853 foreach( $args as $arg ) {
1854 $eqpos = strpos( $arg, '=' );
1855 if ( $eqpos === false ) {
1856 $assocArgs[$index++] = $arg;
1857 } else {
1858 $name = trim( substr( $arg, 0, $eqpos ) );
1859 $value = trim( substr( $arg, $eqpos+1 ) );
1860 if ( $value === false ) {
1861 $value = '';
1862 }
1863 if ( $name !== false ) {
1864 $assocArgs[$name] = $value;
1865 }
1866 }
1867 }
1868
1869 # Add a new element to the templace recursion path
1870 $this->mTemplatePath[$part1] = 1;
1871
1872 $text = $this->strip( $text, $this->mStripState );
1873 $text = $this->removeHTMLtags( $text );
1874 $text = $this->replaceVariables( $text, $assocArgs );
1875
1876 # Resume the link cache and register the inclusion as a link
1877 if ( $this->mOutputType == OT_HTML && !is_null( $title ) ) {
1878 $wgLinkCache->addLinkObj( $title );
1879 }
1880
1881 # If the template begins with a table or block-level
1882 # element, it should be treated as beginning a new line.
1883 if ($linestart !== '\n' && preg_match('/^({\\||:|;|#|\*)/', $text)) {
1884 $text = "\n" . $text;
1885 }
1886 }
1887
1888 # Empties the template path
1889 $this->mTemplatePath = array();
1890 if ( !$found ) {
1891 return $matches[0];
1892 } else {
1893 # replace ==section headers==
1894 # XXX this needs to go away once we have a better parser.
1895 if ( $this->mOutputType != OT_WIKI && $itcamefromthedatabase ) {
1896 if( !is_null( $title ) )
1897 $encodedname = base64_encode($title->getPrefixedDBkey());
1898 else
1899 $encodedname = base64_encode("");
1900 $m = preg_split('/(^={1,6}.*?={1,6}\s*?$)/m', $text, -1,
1901 PREG_SPLIT_DELIM_CAPTURE);
1902 $text = '';
1903 $nsec = 0;
1904 for( $i = 0; $i < count($m); $i += 2 ) {
1905 $text .= $m[$i];
1906 if (!isset($m[$i + 1]) || $m[$i + 1] == "") continue;
1907 $hl = $m[$i + 1];
1908 if( strstr($hl, "<!--MWTEMPLATESECTION") ) {
1909 $text .= $hl;
1910 continue;
1911 }
1912 preg_match('/^(={1,6})(.*?)(={1,6})\s*?$/m', $hl, $m2);
1913 $text .= $m2[1] . $m2[2] . "<!--MWTEMPLATESECTION="
1914 . $encodedname . "&" . base64_encode("$nsec") . "-->" . $m2[3];
1915
1916 $nsec++;
1917 }
1918 }
1919 }
1920
1921 # Empties the template path
1922 $this->mTemplatePath = array();
1923 if ( !$found ) {
1924 return $matches[0];
1925 } else {
1926 return $text;
1927 }
1928 }
1929
1930 /**
1931 * Triple brace replacement -- used for template arguments
1932 * @access private
1933 */
1934 function argSubstitution( $matches ) {
1935 $arg = trim( $matches[1] );
1936 $text = $matches[0];
1937 $inputArgs = end( $this->mArgStack );
1938
1939 if ( array_key_exists( $arg, $inputArgs ) ) {
1940 $text = $inputArgs[$arg];
1941 }
1942
1943 return $text;
1944 }
1945
1946 /**
1947 * Returns true if the function is allowed to include this entity
1948 * @access private
1949 */
1950 function incrementIncludeCount( $dbk ) {
1951 if ( !array_key_exists( $dbk, $this->mIncludeCount ) ) {
1952 $this->mIncludeCount[$dbk] = 0;
1953 }
1954 if ( ++$this->mIncludeCount[$dbk] <= MAX_INCLUDE_REPEAT ) {
1955 return true;
1956 } else {
1957 return false;
1958 }
1959 }
1960
1961
1962 /**
1963 * Cleans up HTML, removes dangerous tags and attributes, and
1964 * removes HTML comments
1965 * @access private
1966 */
1967 function removeHTMLtags( $text ) {
1968 global $wgUseTidy, $wgUserHtml;
1969 $fname = 'Parser::removeHTMLtags';
1970 wfProfileIn( $fname );
1971
1972 if( $wgUserHtml ) {
1973 $htmlpairs = array( # Tags that must be closed
1974 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
1975 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
1976 'strike', 'strong', 'tt', 'var', 'div', 'center',
1977 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
1978 'ruby', 'rt' , 'rb' , 'rp', 'p'
1979 );
1980 $htmlsingle = array(
1981 'br', 'hr', 'li', 'dt', 'dd'
1982 );
1983 $htmlnest = array( # Tags that can be nested--??
1984 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
1985 'dl', 'font', 'big', 'small', 'sub', 'sup'
1986 );
1987 $tabletags = array( # Can only appear inside table
1988 'td', 'th', 'tr'
1989 );
1990 } else {
1991 $htmlpairs = array();
1992 $htmlsingle = array();
1993 $htmlnest = array();
1994 $tabletags = array();
1995 }
1996
1997 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1998 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1999
2000 $htmlattrs = $this->getHTMLattrs () ;
2001
2002 # Remove HTML comments
2003 $text = $this->removeHTMLcomments( $text );
2004
2005 $bits = explode( '<', $text );
2006 $text = array_shift( $bits );
2007 if(!$wgUseTidy) {
2008 $tagstack = array(); $tablestack = array();
2009 foreach ( $bits as $x ) {
2010 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
2011 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2012 $x, $regs );
2013 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2014 error_reporting( $prev );
2015
2016 $badtag = 0 ;
2017 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2018 # Check our stack
2019 if ( $slash ) {
2020 # Closing a tag...
2021 if ( ! in_array( $t, $htmlsingle ) &&
2022 ( $ot = @array_pop( $tagstack ) ) != $t ) {
2023 @array_push( $tagstack, $ot );
2024 $badtag = 1;
2025 } else {
2026 if ( $t == 'table' ) {
2027 $tagstack = array_pop( $tablestack );
2028 }
2029 $newparams = '';
2030 }
2031 } else {
2032 # Keep track for later
2033 if ( in_array( $t, $tabletags ) &&
2034 ! in_array( 'table', $tagstack ) ) {
2035 $badtag = 1;
2036 } else if ( in_array( $t, $tagstack ) &&
2037 ! in_array ( $t , $htmlnest ) ) {
2038 $badtag = 1 ;
2039 } else if ( ! in_array( $t, $htmlsingle ) ) {
2040 if ( $t == 'table' ) {
2041 array_push( $tablestack, $tagstack );
2042 $tagstack = array();
2043 }
2044 array_push( $tagstack, $t );
2045 }
2046 # Strip non-approved attributes from the tag
2047 $newparams = $this->fixTagAttributes($params);
2048
2049 }
2050 if ( ! $badtag ) {
2051 $rest = str_replace( '>', '&gt;', $rest );
2052 $text .= "<$slash$t $newparams$brace$rest";
2053 continue;
2054 }
2055 }
2056 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2057 }
2058 # Close off any remaining tags
2059 while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
2060 $text .= "</$t>\n";
2061 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
2062 }
2063 } else {
2064 # this might be possible using tidy itself
2065 foreach ( $bits as $x ) {
2066 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
2067 $x, $regs );
2068 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
2069 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
2070 $newparams = $this->fixTagAttributes($params);
2071 $rest = str_replace( '>', '&gt;', $rest );
2072 $text .= "<$slash$t $newparams$brace$rest";
2073 } else {
2074 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
2075 }
2076 }
2077 }
2078 wfProfileOut( $fname );
2079 return $text;
2080 }
2081
2082 /**
2083 * Remove '<!--', '-->', and everything between.
2084 * To avoid leaving blank lines, when a comment is both preceded
2085 * and followed by a newline (ignoring spaces), trim leading and
2086 * trailing spaces and one of the newlines.
2087 *
2088 * @access private
2089 */
2090 function removeHTMLcomments( $text ) {
2091 $fname='Parser::removeHTMLcomments';
2092 wfProfileIn( $fname );
2093 while (($start = strpos($text, '<!--')) !== false) {
2094 $end = strpos($text, '-->', $start + 4);
2095 if ($end === false) {
2096 # Unterminated comment; bail out
2097 break;
2098 }
2099
2100 $end += 3;
2101
2102 # Trim space and newline if the comment is both
2103 # preceded and followed by a newline
2104 $spaceStart = max($start - 1, 0);
2105 $spaceLen = $end - $spaceStart;
2106 while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
2107 $spaceStart--;
2108 $spaceLen++;
2109 }
2110 while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
2111 $spaceLen++;
2112 if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
2113 # Remove the comment, leading and trailing
2114 # spaces, and leave only one newline.
2115 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
2116 }
2117 else {
2118 # Remove just the comment.
2119 $text = substr_replace($text, '', $start, $end - $start);
2120 }
2121 }
2122 wfProfileOut( $fname );
2123 return $text;
2124 }
2125
2126 /**
2127 * This function accomplishes several tasks:
2128 * 1) Auto-number headings if that option is enabled
2129 * 2) Add an [edit] link to sections for logged in users who have enabled the option
2130 * 3) Add a Table of contents on the top for users who have enabled the option
2131 * 4) Auto-anchor headings
2132 *
2133 * It loops through all headlines, collects the necessary data, then splits up the
2134 * string and re-inserts the newly formatted headlines.
2135 * @access private
2136 */
2137 /* private */ function formatHeadings( $text, $isMain=true ) {
2138 global $wgInputEncoding, $wgMaxTocLevel, $wgContLang, $wgLinkHolders;
2139
2140 $doNumberHeadings = $this->mOptions->getNumberHeadings();
2141 $doShowToc = $this->mOptions->getShowToc();
2142 $forceTocHere = false;
2143 if( !$this->mTitle->userCanEdit() ) {
2144 $showEditLink = 0;
2145 $rightClickHack = 0;
2146 } else {
2147 $showEditLink = $this->mOptions->getEditSection();
2148 $rightClickHack = $this->mOptions->getEditSectionOnRightClick();
2149 }
2150
2151 # Inhibit editsection links if requested in the page
2152 $esw =& MagicWord::get( MAG_NOEDITSECTION );
2153 if( $esw->matchAndRemove( $text ) ) {
2154 $showEditLink = 0;
2155 }
2156 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
2157 # do not add TOC
2158 $mw =& MagicWord::get( MAG_NOTOC );
2159 if( $mw->matchAndRemove( $text ) ) {
2160 $doShowToc = 0;
2161 }
2162
2163 # never add the TOC to the Main Page. This is an entry page that should not
2164 # be more than 1-2 screens large anyway
2165 if( $this->mTitle->getPrefixedText() == wfMsg('mainpage') ) {
2166 $doShowToc = 0;
2167 }
2168
2169 # Get all headlines for numbering them and adding funky stuff like [edit]
2170 # links - this is for later, but we need the number of headlines right now
2171 $numMatches = preg_match_all( '/<H([1-6])(.*?' . '>)(.*?)<\/H[1-6]>/i', $text, $matches );
2172
2173 # if there are fewer than 4 headlines in the article, do not show TOC
2174 if( $numMatches < 4 ) {
2175 $doShowToc = 0;
2176 }
2177
2178 # if the string __TOC__ (not case-sensitive) occurs in the HTML,
2179 # override above conditions and always show TOC at that place
2180 $mw =& MagicWord::get( MAG_TOC );
2181 if ($mw->match( $text ) ) {
2182 $doShowToc = 1;
2183 $forceTocHere = true;
2184 } else {
2185 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
2186 # override above conditions and always show TOC above first header
2187 $mw =& MagicWord::get( MAG_FORCETOC );
2188 if ($mw->matchAndRemove( $text ) ) {
2189 $doShowToc = 1;
2190 }
2191 }
2192
2193
2194
2195 # We need this to perform operations on the HTML
2196 $sk =& $this->mOptions->getSkin();
2197
2198 # headline counter
2199 $headlineCount = 0;
2200 $sectionCount = 0; # headlineCount excluding template sections
2201
2202 # Ugh .. the TOC should have neat indentation levels which can be
2203 # passed to the skin functions. These are determined here
2204 $toclevel = 0;
2205 $toc = '';
2206 $full = '';
2207 $head = array();
2208 $sublevelCount = array();
2209 $level = 0;
2210 $prevlevel = 0;
2211 foreach( $matches[3] as $headline ) {
2212 $istemplate = 0;
2213 $templatetitle = "";
2214 $templatesection = 0;
2215
2216 if (preg_match("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", $headline, $mat)) {
2217 $istemplate = 1;
2218 $templatetitle = base64_decode($mat[1]);
2219 $templatesection = 1 + (int)base64_decode($mat[2]);
2220 $headline = preg_replace("/<!--MWTEMPLATESECTION=([^&]+)&([^_]+)-->/", "", $headline);
2221 }
2222
2223 $numbering = '';
2224 if( $level ) {
2225 $prevlevel = $level;
2226 }
2227 $level = $matches[1][$headlineCount];
2228 if( ( $doNumberHeadings || $doShowToc ) && $prevlevel && $level > $prevlevel ) {
2229 # reset when we enter a new level
2230 $sublevelCount[$level] = 0;
2231 $toc .= $sk->tocIndent( $level - $prevlevel );
2232 $toclevel += $level - $prevlevel;
2233 }
2234 if( ( $doNumberHeadings || $doShowToc ) && $level < $prevlevel ) {
2235 # reset when we step back a level
2236 $sublevelCount[$level+1]=0;
2237 $toc .= $sk->tocUnindent( $prevlevel - $level );
2238 $toclevel -= $prevlevel - $level;
2239 }
2240 # count number of headlines for each level
2241 @$sublevelCount[$level]++;
2242 if( $doNumberHeadings || $doShowToc ) {
2243 $dot = 0;
2244 for( $i = 1; $i <= $level; $i++ ) {
2245 if( !empty( $sublevelCount[$i] ) ) {
2246 if( $dot ) {
2247 $numbering .= '.';
2248 }
2249 $numbering .= $wgContLang->formatNum( $sublevelCount[$i] );
2250 $dot = 1;
2251 }
2252 }
2253 }
2254
2255 # The canonized header is a version of the header text safe to use for links
2256 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
2257 $canonized_headline = $this->unstrip( $headline, $this->mStripState );
2258 $canonized_headline = $this->unstripNoWiki( $headline, $this->mStripState );
2259
2260 # Remove link placeholders by the link text.
2261 # <!--LINK number-->
2262 # turns into
2263 # link text with suffix
2264 $canonized_headline = preg_replace( '/<!--LINK ([0-9]*)-->/e',
2265 "\$wgLinkHolders['texts'][\$1]",
2266 $canonized_headline );
2267
2268 # strip out HTML
2269 $canonized_headline = preg_replace( '/<.*?' . '>/','',$canonized_headline );
2270 $tocline = trim( $canonized_headline );
2271 $canonized_headline = urlencode( do_html_entity_decode( str_replace(' ', '_', $tocline), ENT_COMPAT, $wgInputEncoding ) );
2272 $replacearray = array(
2273 '%3A' => ':',
2274 '%' => '.'
2275 );
2276 $canonized_headline = str_replace(array_keys($replacearray),array_values($replacearray),$canonized_headline);
2277 $refer[$headlineCount] = $canonized_headline;
2278
2279 # count how many in assoc. array so we can track dupes in anchors
2280 @$refers[$canonized_headline]++;
2281 $refcount[$headlineCount]=$refers[$canonized_headline];
2282
2283 # Prepend the number to the heading text
2284
2285 if( $doNumberHeadings || $doShowToc ) {
2286 $tocline = $numbering . ' ' . $tocline;
2287
2288 # Don't number the heading if it is the only one (looks silly)
2289 if( $doNumberHeadings && count( $matches[3] ) > 1) {
2290 # the two are different if the line contains a link
2291 $headline=$numbering . ' ' . $headline;
2292 }
2293 }
2294
2295 # Create the anchor for linking from the TOC to the section
2296 $anchor = $canonized_headline;
2297 if($refcount[$headlineCount] > 1 ) {
2298 $anchor .= '_' . $refcount[$headlineCount];
2299 }
2300 if( $doShowToc && ( !isset($wgMaxTocLevel) || $toclevel<$wgMaxTocLevel ) ) {
2301 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
2302 }
2303 if( $showEditLink && ( !$istemplate || $templatetitle !== "" ) ) {
2304 if ( empty( $head[$headlineCount] ) ) {
2305 $head[$headlineCount] = '';
2306 }
2307 if( $istemplate )
2308 $head[$headlineCount] .= $sk->editSectionLinkForOther($templatetitle, $templatesection);
2309 else
2310 $head[$headlineCount] .= $sk->editSectionLink($sectionCount+1);
2311 }
2312
2313 # Add the edit section span
2314 if( $rightClickHack ) {
2315 if( $istemplate )
2316 $headline = $sk->editSectionScriptForOther($templatetitle, $templatesection, $headline);
2317 else
2318 $headline = $sk->editSectionScript($sectionCount+1,$headline);
2319 }
2320
2321 # give headline the correct <h#> tag
2322 @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline.'</h'.$level.'>';
2323
2324 $headlineCount++;
2325 if( !$istemplate )
2326 $sectionCount++;
2327 }
2328
2329 if( $doShowToc ) {
2330 $toclines = $headlineCount;
2331 $toc .= $sk->tocUnindent( $toclevel );
2332 $toc = $sk->tocTable( $toc );
2333 }
2334
2335 # split up and insert constructed headlines
2336
2337 $blocks = preg_split( '/<H[1-6].*?' . '>.*?<\/H[1-6]>/i', $text );
2338 $i = 0;
2339
2340 foreach( $blocks as $block ) {
2341 if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
2342 # This is the [edit] link that appears for the top block of text when
2343 # section editing is enabled
2344
2345 # Disabled because it broke block formatting
2346 # For example, a bullet point in the top line
2347 # $full .= $sk->editSectionLink(0);
2348 }
2349 $full .= $block;
2350 if( $doShowToc && !$i && $isMain && !$forceTocHere) {
2351 # Top anchor now in skin
2352 $full = $full.$toc;
2353 }
2354
2355 if( !empty( $head[$i] ) ) {
2356 $full .= $head[$i];
2357 }
2358 $i++;
2359 }
2360 if($forceTocHere) {
2361 $mw =& MagicWord::get( MAG_TOC );
2362 return $mw->replace( $toc, $full );
2363 } else {
2364 return $full;
2365 }
2366 }
2367
2368 /**
2369 * Return an HTML link for the "ISBN 123456" text
2370 * @access private
2371 */
2372 function magicISBN( $text ) {
2373 global $wgLang;
2374 $fname = 'Parser::magicISBN';
2375 wfProfileIn( $fname );
2376
2377 $a = split( 'ISBN ', ' '.$text );
2378 if ( count ( $a ) < 2 ) {
2379 wfProfileOut( $fname );
2380 return $text;
2381 }
2382 $text = substr( array_shift( $a ), 1);
2383 $valid = '0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ';
2384
2385 foreach ( $a as $x ) {
2386 $isbn = $blank = '' ;
2387 while ( ' ' == $x{0} ) {
2388 $blank .= ' ';
2389 $x = substr( $x, 1 );
2390 }
2391 if ( $x == '' ) { # blank isbn
2392 $text .= "ISBN $blank";
2393 continue;
2394 }
2395 while ( strstr( $valid, $x{0} ) != false ) {
2396 $isbn .= $x{0};
2397 $x = substr( $x, 1 );
2398 }
2399 $num = str_replace( '-', '', $isbn );
2400 $num = str_replace( ' ', '', $num );
2401
2402 if ( '' == $num ) {
2403 $text .= "ISBN $blank$x";
2404 } else {
2405 $titleObj = Title::makeTitle( NS_SPECIAL, 'Booksources' );
2406 $text .= '<a href="' .
2407 $titleObj->escapeLocalUrl( 'isbn='.$num ) .
2408 "\" class=\"internal\">ISBN $isbn</a>";
2409 $text .= $x;
2410 }
2411 }
2412 wfProfileOut( $fname );
2413 return $text;
2414 }
2415
2416 /**
2417 * Return an HTML link for the "GEO ..." text
2418 * @access private
2419 */
2420 function magicGEO( $text ) {
2421 global $wgLang, $wgUseGeoMode;
2422 $fname = 'Parser::magicGEO';
2423 wfProfileIn( $fname );
2424
2425 # These next five lines are only for the ~35000 U.S. Census Rambot pages...
2426 $directions = array ( 'N' => 'North' , 'S' => 'South' , 'E' => 'East' , 'W' => 'West' ) ;
2427 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2428 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['N']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2429 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['W']}/" , "(GEO +\$1.\$2.\$3:-\$4.\$5.\$6)" , $text ) ;
2430 $text = preg_replace ( "/(\d+)&deg;(\d+)'(\d+)\" {$directions['S']}, (\d+)&deg;(\d+)'(\d+)\" {$directions['E']}/" , "(GEO +\$1.\$2.\$3:+\$4.\$5.\$6)" , $text ) ;
2431
2432 $a = split( 'GEO ', ' '.$text );
2433 if ( count ( $a ) < 2 ) {
2434 wfProfileOut( $fname );
2435 return $text;
2436 }
2437 $text = substr( array_shift( $a ), 1);
2438 $valid = '0123456789.+-:';
2439
2440 foreach ( $a as $x ) {
2441 $geo = $blank = '' ;
2442 while ( ' ' == $x{0} ) {
2443 $blank .= ' ';
2444 $x = substr( $x, 1 );
2445 }
2446 while ( strstr( $valid, $x{0} ) != false ) {
2447 $geo .= $x{0};
2448 $x = substr( $x, 1 );
2449 }
2450 $num = str_replace( '+', '', $geo );
2451 $num = str_replace( ' ', '', $num );
2452
2453 if ( '' == $num || count ( explode ( ':' , $num , 3 ) ) < 2 ) {
2454 $text .= "GEO $blank$x";
2455 } else {
2456 $titleObj = Title::makeTitle( NS_SPECIAL, 'Geo' );
2457 $text .= '<a href="' .
2458 $titleObj->escapeLocalUrl( 'coordinates='.$num ) .
2459 "\" class=\"internal\">GEO $geo</a>";
2460 $text .= $x;
2461 }
2462 }
2463 wfProfileOut( $fname );
2464 return $text;
2465 }
2466
2467 /**
2468 * Return an HTML link for the "RFC 1234" text
2469 * @access private
2470 * @param string $text text to be processed
2471 */
2472 function magicRFC( $text ) {
2473 global $wgLang;
2474
2475 $valid = '0123456789';
2476 $internal = false;
2477
2478 $a = split( 'RFC ', ' '.$text );
2479 if ( count ( $a ) < 2 ) return $text;
2480 $text = substr( array_shift( $a ), 1);
2481
2482 /* Check if RFC keyword is preceed by [[.
2483 * This test is made here cause of the array_shift above
2484 * that prevent the test to be done in the foreach.
2485 */
2486 if(substr($text, -2) == '[[') { $internal = true; }
2487
2488 foreach ( $a as $x ) {
2489 /* token might be empty if we have RFC RFC 1234 */
2490 if($x=='') {
2491 $text.='RFC ';
2492 continue;
2493 }
2494
2495 $rfc = $blank = '' ;
2496
2497 /** remove and save whitespaces in $blank */
2498 while ( $x{0} == ' ' ) {
2499 $blank .= ' ';
2500 $x = substr( $x, 1 );
2501 }
2502
2503 /** remove and save the rfc number in $rfc */
2504 while ( strstr( $valid, $x{0} ) != false ) {
2505 $rfc .= $x{0};
2506 $x = substr( $x, 1 );
2507 }
2508
2509 if ( $rfc == '') {
2510 /* call back stripped spaces*/
2511 $text .= "RFC $blank$x";
2512 } elseif( $internal) {
2513 /* normal link */
2514 $text .= "RFC $rfc$x";
2515 } else {
2516 /* build the external link*/
2517 $url = wfmsg( 'rfcurl' );
2518 $url = str_replace( '$1', $rfc, $url);
2519 $sk =& $this->mOptions->getSkin();
2520 $la = $sk->getExternalLinkAttributes( $url, 'RFC '.$rfc );
2521 $text .= "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
2522 }
2523
2524 /* Check if the next RFC keyword is preceed by [[ */
2525 $internal = (substr($x,-2) == '[[');
2526 }
2527 return $text;
2528 }
2529
2530 /**
2531 * Transform wiki markup when saving a page by doing \r\n -> \n
2532 * conversion, substitting signatures, {{subst:}} templates, etc.
2533 *
2534 * @param string $text the text to transform
2535 * @param Title &$title the Title object for the current article
2536 * @param User &$user the User object describing the current user
2537 * @param ParserOptions $options parsing options
2538 * @param bool $clearState whether to clear the parser state first
2539 * @return string the altered wiki markup
2540 * @access public
2541 */
2542 function preSaveTransform( $text, &$title, &$user, $options, $clearState = true ) {
2543 $this->mOptions = $options;
2544 $this->mTitle =& $title;
2545 $this->mOutputType = OT_WIKI;
2546
2547 if ( $clearState ) {
2548 $this->clearState();
2549 }
2550
2551 $stripState = false;
2552 $pairs = array(
2553 "\r\n" => "\n",
2554 );
2555 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
2556 // now with regexes
2557 /*
2558 $pairs = array(
2559 "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
2560 "/<br *?>/i" => "<br />",
2561 );
2562 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
2563 */
2564 $text = $this->strip( $text, $stripState, false );
2565 $text = $this->pstPass2( $text, $user );
2566 $text = $this->unstrip( $text, $stripState );
2567 $text = $this->unstripNoWiki( $text, $stripState );
2568 return $text;
2569 }
2570
2571 /**
2572 * Pre-save transform helper function
2573 * @access private
2574 */
2575 function pstPass2( $text, &$user ) {
2576 global $wgLang, $wgContLang, $wgLocaltimezone, $wgCurParser;
2577
2578 # Variable replacement
2579 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
2580 $text = $this->replaceVariables( $text );
2581
2582 # Signatures
2583 #
2584 $n = $user->getName();
2585 $k = $user->getOption( 'nickname' );
2586 if ( '' == $k ) { $k = $n; }
2587 if(isset($wgLocaltimezone)) {
2588 $oldtz = getenv('TZ'); putenv('TZ='.$wgLocaltimezone);
2589 }
2590 /* Note: this is an ugly timezone hack for the European wikis */
2591 $d = $wgContLang->timeanddate( date( 'YmdHis' ), false ) .
2592 ' (' . date( 'T' ) . ')';
2593 if(isset($wgLocaltimezone)) putenv('TZ='.$oldtzs);
2594
2595 $text = preg_replace( '/~~~~~/', $d, $text );
2596 $text = preg_replace( '/~~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]] $d", $text );
2597 $text = preg_replace( '/~~~/', '[[' . $wgContLang->getNsText( NS_USER ) . ":$n|$k]]", $text );
2598
2599 # Context links: [[|name]] and [[name (context)|]]
2600 #
2601 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2602 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2603 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2604 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2605
2606 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/"; # [[page (context)|]]
2607 $p2 = "/\[\[\\|({$tc}+)]]/"; # [[|page]]
2608 $p3 = "/\[\[(:*$namespacechar+):({$np}+)\\|]]/"; # [[namespace:page|]] and [[:namespace:page|]]
2609 $p4 = "/\[\[(:*$namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/"; # [[ns:page (cont)|]] and [[:ns:page (cont)|]]
2610 $context = '';
2611 $t = $this->mTitle->getText();
2612 if ( preg_match( $conpat, $t, $m ) ) {
2613 $context = $m[2];
2614 }
2615 $text = preg_replace( $p4, '[[\\1:\\2 (\\3)|\\2]]', $text );
2616 $text = preg_replace( $p1, '[[\\1 (\\2)|\\1]]', $text );
2617 $text = preg_replace( $p3, '[[\\1:\\2|\\2]]', $text );
2618
2619 if ( '' == $context ) {
2620 $text = preg_replace( $p2, '[[\\1]]', $text );
2621 } else {
2622 $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2623 }
2624
2625 # Trim trailing whitespace
2626 # MAG_END (__END__) tag allows for trailing
2627 # whitespace to be deliberately included
2628 $text = rtrim( $text );
2629 $mw =& MagicWord::get( MAG_END );
2630 $mw->matchAndRemove( $text );
2631
2632 return $text;
2633 }
2634
2635 /**
2636 * Set up some variables which are usually set up in parse()
2637 * so that an external function can call some class members with confidence
2638 * @access public
2639 */
2640 function startExternalParse( &$title, $options, $outputType, $clearState = true ) {
2641 $this->mTitle =& $title;
2642 $this->mOptions = $options;
2643 $this->mOutputType = $outputType;
2644 if ( $clearState ) {
2645 $this->clearState();
2646 }
2647 }
2648
2649 /**
2650 * Transform a MediaWiki message by replacing magic variables.
2651 *
2652 * @param string $text the text to transform
2653 * @param ParserOptions $options options
2654 * @return string the text with variables substituted
2655 * @access public
2656 */
2657 function transformMsg( $text, $options ) {
2658 global $wgTitle;
2659 static $executing = false;
2660
2661 # Guard against infinite recursion
2662 if ( $executing ) {
2663 return $text;
2664 }
2665 $executing = true;
2666
2667 $this->mTitle = $wgTitle;
2668 $this->mOptions = $options;
2669 $this->mOutputType = OT_MSG;
2670 $this->clearState();
2671 $text = $this->replaceVariables( $text );
2672
2673 $executing = false;
2674 return $text;
2675 }
2676
2677 /**
2678 * Create an HTML-style tag, e.g. <yourtag>special text</yourtag>
2679 * Callback will be called with the text within
2680 * Transform and return the text within
2681 * @access public
2682 */
2683 function setHook( $tag, $callback ) {
2684 $oldVal = @$this->mTagHooks[$tag];
2685 $this->mTagHooks[$tag] = $callback;
2686 return $oldVal;
2687 }
2688 }
2689
2690 /**
2691 * @todo document
2692 * @package MediaWiki
2693 */
2694 class ParserOutput
2695 {
2696 var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2697 var $mCacheTime; # Used in ParserCache
2698
2699 function ParserOutput( $text = '', $languageLinks = array(), $categoryLinks = array(),
2700 $containsOldMagic = false )
2701 {
2702 $this->mText = $text;
2703 $this->mLanguageLinks = $languageLinks;
2704 $this->mCategoryLinks = $categoryLinks;
2705 $this->mContainsOldMagic = $containsOldMagic;
2706 $this->mCacheTime = '';
2707 }
2708
2709 function getText() { return $this->mText; }
2710 function getLanguageLinks() { return $this->mLanguageLinks; }
2711 function getCategoryLinks() { return $this->mCategoryLinks; }
2712 function getCacheTime() { return $this->mCacheTime; }
2713 function containsOldMagic() { return $this->mContainsOldMagic; }
2714 function setText( $text ) { return wfSetVar( $this->mText, $text ); }
2715 function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks, $ll ); }
2716 function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks, $cl ); }
2717 function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic, $com ); }
2718 function setCacheTime( $t ) { return wfSetVar( $this->mCacheTime, $t ); }
2719
2720 function merge( $other ) {
2721 $this->mLanguageLinks = array_merge( $this->mLanguageLinks, $other->mLanguageLinks );
2722 $this->mCategoryLinks = array_merge( $this->mCategoryLinks, $this->mLanguageLinks );
2723 $this->mContainsOldMagic = $this->mContainsOldMagic || $other->mContainsOldMagic;
2724 }
2725
2726 }
2727
2728 /**
2729 * Set options of the Parser
2730 * @todo document
2731 * @package MediaWiki
2732 */
2733 class ParserOptions
2734 {
2735 # All variables are private
2736 var $mUseTeX; # Use texvc to expand <math> tags
2737 var $mUseDynamicDates; # Use $wgDateFormatter to format dates
2738 var $mInterwikiMagic; # Interlanguage links are removed and returned in an array
2739 var $mAllowExternalImages; # Allow external images inline
2740 var $mSkin; # Reference to the preferred skin
2741 var $mDateFormat; # Date format index
2742 var $mEditSection; # Create "edit section" links
2743 var $mEditSectionOnRightClick; # Generate JavaScript to edit section on right click
2744 var $mNumberHeadings; # Automatically number headings
2745 var $mShowToc; # Show table of contents
2746
2747 function getUseTeX() { return $this->mUseTeX; }
2748 function getUseDynamicDates() { return $this->mUseDynamicDates; }
2749 function getInterwikiMagic() { return $this->mInterwikiMagic; }
2750 function getAllowExternalImages() { return $this->mAllowExternalImages; }
2751 function getSkin() { return $this->mSkin; }
2752 function getDateFormat() { return $this->mDateFormat; }
2753 function getEditSection() { return $this->mEditSection; }
2754 function getEditSectionOnRightClick() { return $this->mEditSectionOnRightClick; }
2755 function getNumberHeadings() { return $this->mNumberHeadings; }
2756 function getShowToc() { return $this->mShowToc; }
2757
2758 function setUseTeX( $x ) { return wfSetVar( $this->mUseTeX, $x ); }
2759 function setUseDynamicDates( $x ) { return wfSetVar( $this->mUseDynamicDates, $x ); }
2760 function setInterwikiMagic( $x ) { return wfSetVar( $this->mInterwikiMagic, $x ); }
2761 function setAllowExternalImages( $x ) { return wfSetVar( $this->mAllowExternalImages, $x ); }
2762 function setDateFormat( $x ) { return wfSetVar( $this->mDateFormat, $x ); }
2763 function setEditSection( $x ) { return wfSetVar( $this->mEditSection, $x ); }
2764 function setEditSectionOnRightClick( $x ) { return wfSetVar( $this->mEditSectionOnRightClick, $x ); }
2765 function setNumberHeadings( $x ) { return wfSetVar( $this->mNumberHeadings, $x ); }
2766 function setShowToc( $x ) { return wfSetVar( $this->mShowToc, $x ); }
2767
2768 function setSkin( &$x ) { $this->mSkin =& $x; }
2769
2770 # Get parser options
2771 /* static */ function newFromUser( &$user ) {
2772 $popts = new ParserOptions;
2773 $popts->initialiseFromUser( $user );
2774 return $popts;
2775 }
2776
2777 # Get user options
2778 function initialiseFromUser( &$userInput ) {
2779 global $wgUseTeX, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2780
2781 $fname = 'ParserOptions::initialiseFromUser';
2782 wfProfileIn( $fname );
2783 if ( !$userInput ) {
2784 $user = new User;
2785 $user->setLoaded( true );
2786 } else {
2787 $user =& $userInput;
2788 }
2789
2790 $this->mUseTeX = $wgUseTeX;
2791 $this->mUseDynamicDates = $wgUseDynamicDates;
2792 $this->mInterwikiMagic = $wgInterwikiMagic;
2793 $this->mAllowExternalImages = $wgAllowExternalImages;
2794 wfProfileIn( $fname.'-skin' );
2795 $this->mSkin =& $user->getSkin();
2796 wfProfileOut( $fname.'-skin' );
2797 $this->mDateFormat = $user->getOption( 'date' );
2798 $this->mEditSection = $user->getOption( 'editsection' );
2799 $this->mEditSectionOnRightClick = $user->getOption( 'editsectiononrightclick' );
2800 $this->mNumberHeadings = $user->getOption( 'numberheadings' );
2801 $this->mShowToc = $user->getOption( 'showtoc' );
2802 wfProfileOut( $fname );
2803 }
2804
2805
2806 }
2807
2808 # Regex callbacks, used in Parser::replaceVariables
2809 function wfBraceSubstitution( $matches ) {
2810 global $wgCurParser;
2811 return $wgCurParser->braceSubstitution( $matches );
2812 }
2813
2814 function wfArgSubstitution( $matches ) {
2815 global $wgCurParser;
2816 return $wgCurParser->argSubstitution( $matches );
2817 }
2818
2819 function wfVariableSubstitution( $matches ) {
2820 global $wgCurParser;
2821 return $wgCurParser->variableSubstitution( $matches );
2822 }
2823
2824 /**
2825 * Return the total number of articles
2826 */
2827 function wfNumberOfArticles() {
2828 global $wgNumberOfArticles;
2829
2830 wfLoadSiteStats();
2831 return $wgNumberOfArticles;
2832 }
2833
2834 /**
2835 * Get various statistics from the database
2836 * @private
2837 */
2838 function wfLoadSiteStats() {
2839 global $wgNumberOfArticles, $wgTotalViews, $wgTotalEdits;
2840 $fname = 'wfLoadSiteStats';
2841
2842 if ( -1 != $wgNumberOfArticles ) return;
2843 $dbr =& wfGetDB( DB_SLAVE );
2844 $s = $dbr->getArray( 'site_stats',
2845 array( 'ss_total_views', 'ss_total_edits', 'ss_good_articles' ),
2846 array( 'ss_row_id' => 1 ), $fname
2847 );
2848
2849 if ( $s === false ) {
2850 return;
2851 } else {
2852 $wgTotalViews = $s->ss_total_views;
2853 $wgTotalEdits = $s->ss_total_edits;
2854 $wgNumberOfArticles = $s->ss_good_articles;
2855 }
2856 }
2857
2858 function wfEscapeHTMLTagsOnly( $in ) {
2859 return str_replace(
2860 array( '"', '>', '<' ),
2861 array( '&quot;', '&gt;', '&lt;' ),
2862 $in );
2863 }
2864
2865 ?>
Wiklou, le Wiki du Wiklou