* (bug 2067) Fixed crash on empty quoted HTML attribute
[lhc/web/wiklou.git] / includes / Sanitizer.php
1 <?php
2
3 /**
4 * (X)HTML sanitizer for MediaWiki
5 *
6 * Copyright (C) 2002-2005 Brion Vibber <brion@pobox.com> et al
7 * http://www.mediawiki.org/
8 *
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
13 *
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
18 *
19 * You should have received a copy of the GNU General Public License along
20 * with this program; if not, write to the Free Software Foundation, Inc.,
21 * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
22 * http://www.gnu.org/copyleft/gpl.html
23 *
24 * @package MediaWiki
25 * @subpackage Parser
26 */
27
28 class Sanitizer {
29 /**
30 * Cleans up HTML, removes dangerous tags and attributes, and
31 * removes HTML comments
32 * @access private
33 * @param string $text
34 * @return string
35 */
36 function removeHTMLtags( $text ) {
37 global $wgUseTidy, $wgUserHtml;
38 $fname = 'Parser::removeHTMLtags';
39 wfProfileIn( $fname );
40
41 if( $wgUserHtml ) {
42 $htmlpairs = array( # Tags that must be closed
43 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
44 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
45 'strike', 'strong', 'tt', 'var', 'div', 'center',
46 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
47 'ruby', 'rt' , 'rb' , 'rp', 'p', 'span'
48 );
49 $htmlsingle = array(
50 'br', 'hr', 'li', 'dt', 'dd'
51 );
52 $htmlnest = array( # Tags that can be nested--??
53 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
54 'dl', 'font', 'big', 'small', 'sub', 'sup', 'span'
55 );
56 $tabletags = array( # Can only appear inside table
57 'td', 'th', 'tr'
58 );
59 } else {
60 $htmlpairs = array();
61 $htmlsingle = array();
62 $htmlnest = array();
63 $tabletags = array();
64 }
65
66 $htmlsingle = array_merge( $tabletags, $htmlsingle );
67 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
68
69 # Remove HTML comments
70 $text = Sanitizer::removeHTMLcomments( $text );
71
72 $bits = explode( '<', $text );
73 $text = array_shift( $bits );
74 if(!$wgUseTidy) {
75 $tagstack = array(); $tablestack = array();
76 foreach ( $bits as $x ) {
77 $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
78 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
79 $x, $regs );
80 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
81 error_reporting( $prev );
82
83 $badtag = 0 ;
84 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
85 # Check our stack
86 if ( $slash ) {
87 # Closing a tag...
88 if ( ! in_array( $t, $htmlsingle ) &&
89 ( $ot = @array_pop( $tagstack ) ) != $t ) {
90 @array_push( $tagstack, $ot );
91 $badtag = 1;
92 } else {
93 if ( $t == 'table' ) {
94 $tagstack = array_pop( $tablestack );
95 }
96 $newparams = '';
97 }
98 } else {
99 # Keep track for later
100 if ( in_array( $t, $tabletags ) &&
101 ! in_array( 'table', $tagstack ) ) {
102 $badtag = 1;
103 } else if ( in_array( $t, $tagstack ) &&
104 ! in_array ( $t , $htmlnest ) ) {
105 $badtag = 1 ;
106 } else if ( ! in_array( $t, $htmlsingle ) ) {
107 if ( $t == 'table' ) {
108 array_push( $tablestack, $tagstack );
109 $tagstack = array();
110 }
111 array_push( $tagstack, $t );
112 }
113 # Strip non-approved attributes from the tag
114 $newparams = Sanitizer::fixTagAttributes( $params, $t );
115 }
116 if ( ! $badtag ) {
117 $rest = str_replace( '>', '&gt;', $rest );
118 $text .= "<$slash$t$newparams$brace$rest";
119 continue;
120 }
121 }
122 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
123 }
124 # Close off any remaining tags
125 while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
126 $text .= "</$t>\n";
127 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
128 }
129 } else {
130 # this might be possible using tidy itself
131 foreach ( $bits as $x ) {
132 preg_match( '/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/',
133 $x, $regs );
134 @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
135 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
136 $newparams = Sanitizer::fixTagAttributes( $params, $t );
137 $rest = str_replace( '>', '&gt;', $rest );
138 $text .= "<$slash$t$newparams$brace$rest";
139 } else {
140 $text .= '&lt;' . str_replace( '>', '&gt;', $x);
141 }
142 }
143 }
144 wfProfileOut( $fname );
145 return $text;
146 }
147
148 /**
149 * Remove '<!--', '-->', and everything between.
150 * To avoid leaving blank lines, when a comment is both preceded
151 * and followed by a newline (ignoring spaces), trim leading and
152 * trailing spaces and one of the newlines.
153 *
154 * @access private
155 * @param string $text
156 * @return string
157 */
158 function removeHTMLcomments( $text ) {
159 $fname='Parser::removeHTMLcomments';
160 wfProfileIn( $fname );
161 while (($start = strpos($text, '<!--')) !== false) {
162 $end = strpos($text, '-->', $start + 4);
163 if ($end === false) {
164 # Unterminated comment; bail out
165 break;
166 }
167
168 $end += 3;
169
170 # Trim space and newline if the comment is both
171 # preceded and followed by a newline
172 $spaceStart = max($start - 1, 0);
173 $spaceLen = $end - $spaceStart;
174 while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
175 $spaceStart--;
176 $spaceLen++;
177 }
178 while (substr($text, $spaceStart + $spaceLen, 1) === ' ')
179 $spaceLen++;
180 if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart + $spaceLen, 1) === "\n") {
181 # Remove the comment, leading and trailing
182 # spaces, and leave only one newline.
183 $text = substr_replace($text, "\n", $spaceStart, $spaceLen + 1);
184 }
185 else {
186 # Remove just the comment.
187 $text = substr_replace($text, '', $start, $end - $start);
188 }
189 }
190 wfProfileOut( $fname );
191 return $text;
192 }
193
194 /**
195 * Take a tag soup fragment listing an HTML element's attributes
196 * and normalize it to well-formed XML, discarding unwanted attributes.
197 *
198 * - Normalizes attribute names to lowercase
199 * - Discards attributes not on a whitelist for the given element
200 * - Turns broken or invalid entities into plaintext
201 * - Double-quotes all attribute values
202 * - Attributes without values are given the name as attribute
203 * - Double attributes are discarded
204 * - Unsafe style attributes are discarded
205 * - Prepends space if there are attributes.
206 *
207 * @param string $text
208 * @param string $element
209 * @return string
210 *
211 * @todo Check for legal values where the DTD limits things.
212 * @todo Check for unique id attribute :P
213 */
214 function fixTagAttributes( $text, $element ) {
215 if( trim( $text ) == '' ) {
216 return '';
217 }
218
219 # Unquoted attribute
220 # Since we quote this later, this can be anything distinguishable
221 # from the end of the attribute
222 $attrib = '[A-Za-z0-9]';
223 $space = '[\x09\x0a\x0d\x20]';
224 if( !preg_match_all(
225 "/(?:^|$space)($attrib+)
226 ($space*=$space*
227 (?:
228 # The attribute value: quoted or alone
229 \"([^<\"]*)\"
230 | '([^<']*)'
231 | ([a-zA-Z0-9!#$%&()*,\\-.\\/:;<>?@[\\]^_`{|}~]+)
232 | (\#[0-9a-fA-F]+) # Technically wrong, but lots of
233 # colors are specified like this.
234 # We'll be normalizing it.
235 )
236 )?(?=$space|\$)/sx",
237 $text,
238 $pairs,
239 PREG_SET_ORDER ) ) {
240 return '';
241 }
242
243 $whitelist = array_flip( Sanitizer::attributeWhitelist( $element ) );
244 $attribs = array();
245 foreach( $pairs as $set ) {
246 $attribute = strtolower( $set[1] );
247 if( !isset( $whitelist[$attribute] ) ) {
248 continue;
249 } elseif( isset( $set[6] ) ) {
250 # Illegal #XXXXXX color with no quotes.
251 $value = Sanitizer::normalizeAttributeValue( $set[6] );
252 } elseif( isset( $set[5] ) ) {
253 # No quotes.
254 $value = Sanitizer::normalizeAttributeValue( $set[5] );
255 } elseif( isset( $set[4] ) ) {
256 # Single-quoted
257 $value = str_replace( '"', '&quot;',
258 Sanitizer::normalizeAttributeValue( $set[4] ) );
259 } elseif( isset( $set[3] ) ) {
260 # Double-quoted
261 $value = Sanitizer::normalizeAttributeValue( $set[3] );
262 } elseif( !isset( $set[2] ) ) {
263 # In XHTML, attributes must have a value.
264 $value = $set[1];
265 } else {
266 wfDebugDieBacktrace( "Tag conditions not met. This should never happen and is a bug." );
267 }
268
269 # Strip javascript "expression" from stylesheets.
270 # http://msdn.microsoft.com/workshop/author/dhtml/overview/recalc.asp
271 if( $attribute == 'style' && preg_match(
272 '/(expression|tps*:\/\/|url\\s*\().*/is',
273 wfMungeToUtf8( $value ) ) ) {
274 # haxx0r
275 continue;
276 }
277
278 if( !isset( $attribs[$attribute] ) ) {
279 $attribs[$attribute] = "$attribute=\"$value\"";
280 }
281 }
282 if( empty( $attribs ) ) {
283 return '';
284 } else {
285 return ' ' . implode( ' ', $attribs );
286 }
287 }
288
289 /**
290 * Normalize whitespace and character references in an XML source-
291 * encoded text for an attribute value.
292 *
293 * See http://www.w3.org/TR/REC-xml/#AVNormalize for background,
294 * but note that we're not returning the value, but are returning
295 * XML source fragments that will be slapped into output.
296 *
297 * @param string $text
298 * @return string
299 * @access private
300 */
301 function normalizeAttributeValue( $text ) {
302 return preg_replace(
303 '/\r\n|[\x20\x0d\x0a\x09]/',
304 ' ',
305 Sanitizer::normalizeCharReferences( $text ) );
306 }
307
308 /**
309 * Ensure that any entities and character references are legal
310 * for XML and XHTML specifically. Any stray bits will be
311 * &amp;-escaped to result in a valid text fragment.
312 *
313 * a. any named char refs must be known in XHTML
314 * b. any numeric char refs must be legal chars, not invalid or forbidden
315 * c. use &#x, not &#X
316 * d. fix or reject non-valid attributes
317 *
318 * @param string $text
319 * @return string
320 * @access private
321 */
322 function normalizeCharReferences( $text ) {
323 return preg_replace_callback(
324 '/&([A-Za-z0-9]+);
325 |&\#([0-9]+);
326 |&\#x([0-9A-Za-z]+);
327 |&\#X([0-9A-Za-z]+);
328 |(&)/x',
329 array( 'Sanitizer', 'normalizeCharReferencesCallback' ),
330 $text );
331 }
332 /**
333 * @param string $matches
334 * @return string
335 */
336 function normalizeCharReferencesCallback( $matches ) {
337 $ret = null;
338 if( $matches[1] != '' ) {
339 $ret = Sanitizer::normalizeEntity( $matches[1] );
340 } elseif( $matches[2] != '' ) {
341 $ret = Sanitizer::decCharReference( $matches[2] );
342 } elseif( $matches[3] != '' ) {
343 $ret = Sanitizer::hexCharReference( $matches[3] );
344 } elseif( $matches[4] != '' ) {
345 $ret = Sanitizer::hexCharReference( $matches[4] );
346 }
347 if( is_null( $ret ) ) {
348 return htmlspecialchars( $matches[0] );
349 } else {
350 return $ret;
351 }
352 }
353
354 /**
355 * If the named entity is defined in the HTML 4.0/XHTML 1.0 DTD,
356 * return the named entity reference as is. Otherwise, returns
357 * HTML-escaped text of pseudo-entity source (eg &amp;foo;)
358 *
359 * @param string $name
360 * @return string
361 */
362 function normalizeEntity( $name ) {
363 # List of all named character entities defined in HTML 4.01
364 # http://www.w3.org/TR/html4/sgml/entities.html
365 static $htmlEntities = array(
366 'aacute' => true,
367 'Aacute' => true,
368 'acirc' => true,
369 'Acirc' => true,
370 'acute' => true,
371 'aelig' => true,
372 'AElig' => true,
373 'agrave' => true,
374 'Agrave' => true,
375 'alefsym' => true,
376 'alpha' => true,
377 'Alpha' => true,
378 'amp' => true,
379 'and' => true,
380 'ang' => true,
381 'apos' => true,
382 'aring' => true,
383 'Aring' => true,
384 'asymp' => true,
385 'atilde' => true,
386 'Atilde' => true,
387 'auml' => true,
388 'Auml' => true,
389 'bdquo' => true,
390 'beta' => true,
391 'Beta' => true,
392 'brvbar' => true,
393 'bull' => true,
394 'cap' => true,
395 'ccedil' => true,
396 'Ccedil' => true,
397 'cedil' => true,
398 'cent' => true,
399 'chi' => true,
400 'Chi' => true,
401 'circ' => true,
402 'clubs' => true,
403 'cong' => true,
404 'copy' => true,
405 'crarr' => true,
406 'cup' => true,
407 'curren' => true,
408 'dagger' => true,
409 'Dagger' => true,
410 'darr' => true,
411 'dArr' => true,
412 'deg' => true,
413 'delta' => true,
414 'Delta' => true,
415 'diams' => true,
416 'divide' => true,
417 'eacute' => true,
418 'Eacute' => true,
419 'ecirc' => true,
420 'Ecirc' => true,
421 'egrave' => true,
422 'Egrave' => true,
423 'empty' => true,
424 'emsp' => true,
425 'ensp' => true,
426 'epsilon' => true,
427 'Epsilon' => true,
428 'equiv' => true,
429 'eta' => true,
430 'Eta' => true,
431 'eth' => true,
432 'ETH' => true,
433 'euml' => true,
434 'Euml' => true,
435 'euro' => true,
436 'exist' => true,
437 'fnof' => true,
438 'forall' => true,
439 'frac12' => true,
440 'frac14' => true,
441 'frac34' => true,
442 'frasl' => true,
443 'gamma' => true,
444 'Gamma' => true,
445 'ge' => true,
446 'gt' => true,
447 'harr' => true,
448 'hArr' => true,
449 'hearts' => true,
450 'hellip' => true,
451 'iacute' => true,
452 'Iacute' => true,
453 'icirc' => true,
454 'Icirc' => true,
455 'iexcl' => true,
456 'igrave' => true,
457 'Igrave' => true,
458 'image' => true,
459 'infin' => true,
460 'int' => true,
461 'iota' => true,
462 'Iota' => true,
463 'iquest' => true,
464 'isin' => true,
465 'iuml' => true,
466 'Iuml' => true,
467 'kappa' => true,
468 'Kappa' => true,
469 'lambda' => true,
470 'Lambda' => true,
471 'lang' => true,
472 'laquo' => true,
473 'larr' => true,
474 'lArr' => true,
475 'lceil' => true,
476 'ldquo' => true,
477 'le' => true,
478 'lfloor' => true,
479 'lowast' => true,
480 'loz' => true,
481 'lrm' => true,
482 'lsaquo' => true,
483 'lsquo' => true,
484 'lt' => true,
485 'macr' => true,
486 'mdash' => true,
487 'micro' => true,
488 'middot' => true,
489 'minus' => true,
490 'mu' => true,
491 'Mu' => true,
492 'nabla' => true,
493 'nbsp' => true,
494 'ndash' => true,
495 'ne' => true,
496 'ni' => true,
497 'not' => true,
498 'notin' => true,
499 'nsub' => true,
500 'ntilde' => true,
501 'Ntilde' => true,
502 'nu' => true,
503 'Nu' => true,
504 'oacute' => true,
505 'Oacute' => true,
506 'ocirc' => true,
507 'Ocirc' => true,
508 'oelig' => true,
509 'OElig' => true,
510 'ograve' => true,
511 'Ograve' => true,
512 'oline' => true,
513 'omega' => true,
514 'Omega' => true,
515 'omicron' => true,
516 'Omicron' => true,
517 'oplus' => true,
518 'or' => true,
519 'ordf' => true,
520 'ordm' => true,
521 'oslash' => true,
522 'Oslash' => true,
523 'otilde' => true,
524 'Otilde' => true,
525 'otimes' => true,
526 'ouml' => true,
527 'Ouml' => true,
528 'para' => true,
529 'part' => true,
530 'permil' => true,
531 'perp' => true,
532 'phi' => true,
533 'Phi' => true,
534 'pi' => true,
535 'Pi' => true,
536 'piv' => true,
537 'plusmn' => true,
538 'pound' => true,
539 'prime' => true,
540 'Prime' => true,
541 'prod' => true,
542 'prop' => true,
543 'psi' => true,
544 'Psi' => true,
545 'quot' => true,
546 'radic' => true,
547 'rang' => true,
548 'raquo' => true,
549 'rarr' => true,
550 'rArr' => true,
551 'rceil' => true,
552 'rdquo' => true,
553 'real' => true,
554 'reg' => true,
555 'rfloor' => true,
556 'rho' => true,
557 'Rho' => true,
558 'rlm' => true,
559 'rsaquo' => true,
560 'rsquo' => true,
561 'sbquo' => true,
562 'scaron' => true,
563 'Scaron' => true,
564 'sdot' => true,
565 'sect' => true,
566 'shy' => true,
567 'sigma' => true,
568 'Sigma' => true,
569 'sigmaf' => true,
570 'sim' => true,
571 'spades' => true,
572 'sub' => true,
573 'sube' => true,
574 'sum' => true,
575 'sup' => true,
576 'sup1' => true,
577 'sup2' => true,
578 'sup3' => true,
579 'supe' => true,
580 'szlig' => true,
581 'tau' => true,
582 'Tau' => true,
583 'there4' => true,
584 'theta' => true,
585 'Theta' => true,
586 'thetasym' => true,
587 'thinsp' => true,
588 'thorn' => true,
589 'THORN' => true,
590 'tilde' => true,
591 'times' => true,
592 'trade' => true,
593 'uacute' => true,
594 'Uacute' => true,
595 'uarr' => true,
596 'uArr' => true,
597 'ucirc' => true,
598 'Ucirc' => true,
599 'ugrave' => true,
600 'Ugrave' => true,
601 'uml' => true,
602 'upsih' => true,
603 'upsilon' => true,
604 'Upsilon' => true,
605 'uuml' => true,
606 'Uuml' => true,
607 'weierp' => true,
608 'xi' => true,
609 'Xi' => true,
610 'yacute' => true,
611 'Yacute' => true,
612 'yen' => true,
613 'yuml' => true,
614 'Yuml' => true,
615 'zeta' => true,
616 'Zeta' => true,
617 'zwj' => true,
618 'zwnj' => true );
619 if( isset( $htmlEntities[$name] ) ) {
620 return "&$name;";
621 } else {
622 return "&amp;$name;";
623 }
624 }
625
626 function decCharReference( $codepoint ) {
627 $point = IntVal( $codepoint );
628 if( Sanitizer::validateCodepoint( $point ) ) {
629 return sprintf( '&#%d;', $point );
630 } else {
631 return null;
632 }
633 }
634
635 function hexCharReference( $codepoint ) {
636 $point = hexdec( $codepoint );
637 if( Sanitizer::validateCodepoint( $point ) ) {
638 return sprintf( '&#x%x;', $point );
639 } else {
640 return null;
641 }
642 }
643
644 /**
645 * Returns true if a given Unicode codepoint is a valid character in XML.
646 * @param int $codepoint
647 * @return bool
648 */
649 function validateCodepoint( $codepoint ) {
650 return ($codepoint == 0x09)
651 || ($codepoint == 0x0a)
652 || ($codepoint == 0x0d)
653 || ($codepoint >= 0x20 && $codepoint <= 0xd7ff)
654 || ($codepoint >= 0xe000 && $codepoint <= 0xfffd)
655 || ($codepoint >= 0x10000 && $codepoint <= 0x10ffff);
656 }
657
658 /**
659 * Fetch the whitelist of acceptable attributes for a given
660 * element name.
661 *
662 * @param string $element
663 * @return array
664 */
665 function attributeWhitelist( $element ) {
666 $list = Sanitizer::setupAttributeWhitelist();
667 return isset( $list[$element] )
668 ? $list[$element]
669 : array();
670 }
671
672 /**
673 * @return array
674 */
675 function setupAttributeWhitelist() {
676 $common = array( 'id', 'class', 'lang', 'dir', 'title', 'style' );
677 $block = array_merge( $common, array( 'align' ) );
678 $tablealign = array( 'align', 'char', 'charoff', 'valign' );
679 $tablecell = array( 'abbr',
680 'axis',
681 'headers',
682 'scope',
683 'rowspan',
684 'colspan',
685 'nowrap', # deprecated
686 'width', # deprecated
687 'height' # deprecated
688 );
689
690 # Numbers refer to sections in HTML 4.01 standard describing the element.
691 # See: http://www.w3.org/TR/html4/
692 $whitelist = array (
693 # 7.5.4
694 'div' => $block,
695 'center' => $common, # deprecated
696 'span' => $block, # ??
697
698 # 7.5.5
699 'h1' => $block,
700 'h2' => $block,
701 'h3' => $block,
702 'h4' => $block,
703 'h5' => $block,
704 'h6' => $block,
705
706 # 7.5.6
707 # address
708
709 # 8.2.4
710 # bdo
711
712 # 9.2.1
713 'em' => $common,
714 'strong' => $common,
715 'cite' => $common,
716 # dfn
717 'code' => $common,
718 # samp
719 # kbd
720 'var' => $common,
721 # abbr
722 # acronym
723
724 # 9.2.2
725 'blockquote' => array_merge( $common, array( 'cite' ) ),
726 # q
727
728 # 9.2.3
729 'sub' => $common,
730 'sup' => $common,
731
732 # 9.3.1
733 'p' => $block,
734
735 # 9.3.2
736 'br' => array( 'id', 'class', 'title', 'style', 'clear' ),
737
738 # 9.3.4
739 'pre' => array_merge( $common, array( 'width' ) ),
740
741 # 9.4
742 'ins' => array_merge( $common, array( 'cite', 'datetime' ) ),
743 'del' => array_merge( $common, array( 'cite', 'datetime' ) ),
744
745 # 10.2
746 'ul' => array_merge( $common, array( 'type' ) ),
747 'ol' => array_merge( $common, array( 'type', 'start' ) ),
748 'li' => array_merge( $common, array( 'type', 'value' ) ),
749
750 # 10.3
751 'dl' => $common,
752 'dd' => $common,
753 'dt' => $common,
754
755 # 11.2.1
756 'table' => array_merge( $common,
757 array( 'summary', 'width', 'border', 'frame',
758 'rules', 'cellspacing', 'cellpadding',
759 'align', 'bgcolor', 'frame', 'rules',
760 'border' ) ),
761
762 # 11.2.2
763 'caption' => array_merge( $common, array( 'align' ) ),
764
765 # 11.2.3
766 'thead' => array_merge( $common, $tablealign ),
767 'tfoot' => array_merge( $common, $tablealign ),
768 'tbody' => array_merge( $common, $tablealign ),
769
770 # 11.2.4
771 'colgroup' => array_merge( $common, array( 'span', 'width' ), $tablealign ),
772 'col' => array_merge( $common, array( 'span', 'width' ), $tablealign ),
773
774 # 11.2.5
775 'tr' => array_merge( $common, array( 'bgcolor' ), $tablealign ),
776
777 # 11.2.6
778 'td' => array_merge( $common, $tablecell, $tablealign ),
779 'th' => array_merge( $common, $tablecell, $tablealign ),
780
781 # 15.2.1
782 'tt' => $common,
783 'b' => $common,
784 'i' => $common,
785 'big' => $common,
786 'small' => $common,
787 'strike' => $common,
788 's' => $common,
789 'u' => $common,
790
791 # 15.2.2
792 'font' => array_merge( $common, array( 'size', 'color', 'face' ) ),
793 # basefont
794
795 # 15.3
796 'hr' => array_merge( $common, array( 'noshade', 'size', 'width' ) ),
797
798 # XHTML Ruby annotation text module, simple ruby only.
799 # http://www.w3c.org/TR/ruby/
800 'ruby' => $common,
801 # rbc
802 # rtc
803 'rb' => $common,
804 'rt' => $common, #array_merge( $common, array( 'rbspan' ) ),
805 'rp' => $common,
806 );
807 return $whitelist;
808 }
809
810 /**
811 * Take a fragment of (potentially invalid) HTML and return
812 * a version with any tags removed, encoded suitably for literal
813 * inclusion in an attribute value.
814 *
815 * @param string $text HTML fragment
816 * @return string
817 */
818 function stripAllTags( $text ) {
819 # Actual <tags>
820 $text = preg_replace( '/<[^>]*>/', '', $text );
821
822 # Normalize &entities and whitespace
823 $text = Sanitizer::normalizeAttributeValue( $text );
824
825 # Will be placed into "double-quoted" attributes,
826 # make sure remaining bits are safe.
827 $text = str_replace(
828 array('<', '>', '"'),
829 array('&lt;', '&gt;', '&quot;'),
830 $text );
831
832 return $text;
833 }
834
835 }
836
837 ?>