3 * XHTML sanitizer for MediaWiki
5 * Copyright (C) 2002-2005 Brion Vibber <brion@pobox.com> et al
6 * http://www.mediawiki.org/
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License along
19 * with this program; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
21 * http://www.gnu.org/copyleft/gpl.html
28 * Regular expression to match various types of character references in
29 * Sanitizer::normalizeCharReferences and Sanitizer::decodeCharReferences
31 define( 'MW_CHAR_REFS_REGEX',
39 * Regular expression to match HTML/XML attribute pairs within a tag.
40 * Allows some... latitude.
41 * Used in Sanitizer::fixTagAttributes and Sanitizer::decodeTagAttributes
43 $attrib = '[A-Za-z0-9]';
44 $space = '[\x09\x0a\x0d\x20]';
45 define( 'MW_ATTRIBS_REGEX',
46 "/(?:^|$space)($attrib+)
49 # The attribute value: quoted or alone
52 | ([a-zA-Z0-9!#$%&()*,\\-.\\/:;<>?@[\\]^_`{|}~]+)
53 | (\#[0-9a-fA-F]+) # Technically wrong, but lots of
54 # colors are specified like this.
55 # We'll be normalizing it.
57 )?(?=$space|\$)/sx" );
60 * List of all named character entities defined in HTML 4.01
61 * http://www.w3.org/TR/html4/sgml/entities.html
64 global $wgHtmlEntities;
65 $wgHtmlEntities = array(
319 /** @package MediaWiki */
322 * Cleans up HTML, removes dangerous tags and attributes, and
323 * removes HTML comments
325 * @param string $text
326 * @param callback $processCallback to do any variable or parameter replacements in HTML attribute values
327 * @param array $args for the processing callback
330 static function removeHTMLtags( $text, $processCallback = null, $args = array() ) {
331 global $wgUseTidy, $wgUserHtml;
333 static $htmlpairs, $htmlsingle, $htmlsingleonly, $htmlnest, $tabletags,
334 $htmllist, $listtags, $htmlsingleallowed, $htmlelements, $staticInitialised;
336 wfProfileIn( __METHOD__
);
338 if ( !$staticInitialised ) {
340 $htmlpairs = array( # Tags that must be closed
341 'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
342 'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
343 'strike', 'strong', 'tt', 'var', 'div', 'center',
344 'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
345 'ruby', 'rt' , 'rb' , 'rp', 'p', 'span', 'u'
348 'br', 'hr', 'li', 'dt', 'dd'
350 $htmlsingleonly = array( # Elements that cannot have close tags
353 $htmlnest = array( # Tags that can be nested--??
354 'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
355 'dl', 'font', 'big', 'small', 'sub', 'sup', 'span'
357 $tabletags = array( # Can only appear inside table
360 $htmllist = array( # Tags used by list
363 $listtags = array( # Tags that can appear in a list
368 $htmlpairs = array();
369 $htmlsingle = array();
371 $tabletags = array();
374 $htmlsingleallowed = array_merge( $htmlsingle, $tabletags );
375 $htmlelements = array_merge( $htmlsingle, $htmlpairs, $htmlnest );
377 # Convert them all to hashtables for faster lookup
378 $vars = array( 'htmlpairs', 'htmlsingle', 'htmlsingleonly', 'htmlnest', 'tabletags',
379 'htmllist', 'listtags', 'htmlsingleallowed', 'htmlelements' );
380 foreach ( $vars as $var ) {
381 $
$var = array_flip( $
$var );
383 $staticInitialised = true;
386 # Remove HTML comments
387 $text = Sanitizer
::removeHTMLcomments( $text );
388 $bits = explode( '<', $text );
389 $text = array_shift( $bits );
391 $tagstack = $tablestack = array();
392 foreach ( $bits as $x ) {
393 $prev = error_reporting( E_ALL
& ~
( E_NOTICE | E_WARNING
) );
395 preg_match( '!^(/?)(\\w+)([^>]*?)(/{0,1}>)([^<]*)$!', $x, $regs );
396 list( /* $qbar */, $slash, $t, $params, $brace, $rest ) = $regs;
397 error_reporting( $prev );
400 if ( isset( $htmlelements[$t = strtolower( $t )] ) ) {
404 if( isset( $htmlsingleonly[$t] ) ) {
406 } elseif ( ( $ot = @array_pop
( $tagstack ) ) != $t ) {
407 if ( isset( $htmlsingleallowed[$ot] ) ) {
408 # Pop all elements with an optional close tag
409 # and see if we find a match below them
411 array_push ($optstack, $ot);
412 while ( ( ( $ot = @array_pop
( $tagstack ) ) != $t ) &&
413 isset( $htmlsingleallowed[$ot] ) )
415 array_push ($optstack, $ot);
418 # No match. Push the optinal elements back again
420 while ( $ot = @array_pop
( $optstack ) ) {
421 array_push( $tagstack, $ot );
425 @array_push
( $tagstack, $ot );
426 # <li> can be nested in <ul> or <ol>, skip those cases:
427 if(!(isset( $htmllist[$ot] ) && isset( $listtags[$t] ) )) {
432 if ( $t == 'table' ) {
433 $tagstack = array_pop( $tablestack );
438 # Keep track for later
439 if ( isset( $tabletags[$t] ) &&
440 ! in_array( 'table', $tagstack ) ) {
442 } else if ( in_array( $t, $tagstack ) &&
443 ! isset( $htmlnest [$t ] ) ) {
445 #Â Is it a self closed htmlpair ? (bug 5487)
446 } else if( $brace == '/>' &&
447 isset( $htmlpairs[$t] ) ) {
449 } elseif( isset( $htmlsingleonly[$t] ) ) {
450 # Hack to force empty tag for uncloseable elements
452 } else if( isset( $htmlsingle[$t] ) ) {
453 # Hack to not close $htmlsingle tags
456 if ( $t == 'table' ) {
457 array_push( $tablestack, $tagstack );
460 array_push( $tagstack, $t );
463 # Replace any variables or template parameters with
465 if( is_callable( $processCallback ) ) {
466 call_user_func_array( $processCallback, array( &$params, $args ) );
469 # Strip non-approved attributes from the tag
470 $newparams = Sanitizer
::fixTagAttributes( $params, $t );
473 $rest = str_replace( '>', '>', $rest );
474 $close = ( $brace == '/>' ) ?
' /' : '';
475 $text .= "<$slash$t$newparams$close>$rest";
479 $text .= '<' . str_replace( '>', '>', $x);
481 # Close off any remaining tags
482 while ( is_array( $tagstack ) && ($t = array_pop( $tagstack )) ) {
484 if ( $t == 'table' ) { $tagstack = array_pop( $tablestack ); }
487 # this might be possible using tidy itself
488 foreach ( $bits as $x ) {
489 preg_match( '/^(\\/?)(\\w+)([^>]*?)(\\/{0,1}>)([^<]*)$/',
491 @list
( /* $qbar */, $slash, $t, $params, $brace, $rest ) = $regs;
492 if ( isset( $htmlelements[$t = strtolower( $t )] ) ) {
493 if( is_callable( $processCallback ) ) {
494 call_user_func_array( $processCallback, array( &$params, $args ) );
496 $newparams = Sanitizer
::fixTagAttributes( $params, $t );
497 $rest = str_replace( '>', '>', $rest );
498 $text .= "<$slash$t$newparams$brace$rest";
500 $text .= '<' . str_replace( '>', '>', $x);
504 wfProfileOut( __METHOD__
);
509 * Remove '<!--', '-->', and everything between.
510 * To avoid leaving blank lines, when a comment is both preceded
511 * and followed by a newline (ignoring spaces), trim leading and
512 * trailing spaces and one of the newlines.
515 * @param string $text
518 static function removeHTMLcomments( $text ) {
519 wfProfileIn( __METHOD__
);
520 while (($start = strpos($text, '<!--')) !== false) {
521 $end = strpos($text, '-->', $start +
4);
522 if ($end === false) {
523 # Unterminated comment; bail out
529 # Trim space and newline if the comment is both
530 # preceded and followed by a newline
531 $spaceStart = max($start - 1, 0);
532 $spaceLen = $end - $spaceStart;
533 while (substr($text, $spaceStart, 1) === ' ' && $spaceStart > 0) {
537 while (substr($text, $spaceStart +
$spaceLen, 1) === ' ')
539 if (substr($text, $spaceStart, 1) === "\n" and substr($text, $spaceStart +
$spaceLen, 1) === "\n") {
540 # Remove the comment, leading and trailing
541 # spaces, and leave only one newline.
542 $text = substr_replace($text, "\n", $spaceStart, $spaceLen +
1);
545 # Remove just the comment.
546 $text = substr_replace($text, '', $start, $end - $start);
549 wfProfileOut( __METHOD__
);
554 * Take an array of attribute names and values and normalize or discard
555 * illegal values for the given element type.
557 * - Discards attributes not on a whitelist for the given element
558 * - Unsafe style attributes are discarded
560 * @param array $attribs
561 * @param string $element
564 * @todo Check for legal values where the DTD limits things.
565 * @todo Check for unique id attribute :P
567 static function validateTagAttributes( $attribs, $element ) {
568 $whitelist = array_flip( Sanitizer
::attributeWhitelist( $element ) );
570 foreach( $attribs as $attribute => $value ) {
571 if( !isset( $whitelist[$attribute] ) ) {
574 # Strip javascript "expression" from stylesheets.
575 # http://msdn.microsoft.com/workshop/author/dhtml/overview/recalc.asp
576 if( $attribute == 'style' ) {
577 $value = Sanitizer
::checkCss( $value );
578 if( $value === false ) {
584 if ( $attribute === 'id' )
585 $value = Sanitizer
::escapeId( $value );
587 // If this attribute was previously set, override it.
588 // Output should only have one attribute of each name.
589 $out[$attribute] = $value;
595 * Pick apart some CSS and check it for forbidden or unsafe structures.
596 * Returns a sanitized string, or false if it was just too evil.
598 * Currently URL references, 'expression', 'tps' are forbidden.
600 * @param string $value
603 static function checkCss( $value ) {
604 $stripped = Sanitizer
::decodeCharReferences( $value );
606 // Remove any comments; IE gets token splitting wrong
607 $stripped = StringUtils
::delimiterReplace( '/*', '*/', ' ', $stripped );
611 // ... and continue checks
612 $stripped = preg_replace( '!\\\\([0-9A-Fa-f]{1,6})[ \\n\\r\\t\\f]?!e',
613 'codepointToUtf8(hexdec("$1"))', $stripped );
614 $stripped = str_replace( '\\', '', $stripped );
615 if( preg_match( '/(expression|tps*:\/\/|url\\s*\().*/is',
625 * Take a tag soup fragment listing an HTML element's attributes
626 * and normalize it to well-formed XML, discarding unwanted attributes.
627 * Output is safe for further wikitext processing, with escaping of
628 * values that could trigger problems.
630 * - Normalizes attribute names to lowercase
631 * - Discards attributes not on a whitelist for the given element
632 * - Turns broken or invalid entities into plaintext
633 * - Double-quotes all attribute values
634 * - Attributes without values are given the name as attribute
635 * - Double attributes are discarded
636 * - Unsafe style attributes are discarded
637 * - Prepends space if there are attributes.
639 * @param string $text
640 * @param string $element
643 static function fixTagAttributes( $text, $element ) {
644 if( trim( $text ) == '' ) {
648 $stripped = Sanitizer
::validateTagAttributes(
649 Sanitizer
::decodeTagAttributes( $text ), $element );
652 foreach( $stripped as $attribute => $value ) {
653 $encAttribute = htmlspecialchars( $attribute );
654 $encValue = Sanitizer
::safeEncodeAttribute( $value );
656 $attribs[] = "$encAttribute=\"$encValue\"";
658 return count( $attribs ) ?
' ' . implode( ' ', $attribs ) : '';
662 * Encode an attribute value for HTML output.
664 * @return HTML-encoded text fragment
666 static function encodeAttribute( $text ) {
667 $encValue = htmlspecialchars( $text );
669 // Whitespace is normalized during attribute decoding,
670 // so if we've been passed non-spaces we must encode them
671 // ahead of time or they won't be preserved.
672 $encValue = strtr( $encValue, array(
682 * Encode an attribute value for HTML tags, with extra armoring
683 * against further wiki processing.
685 * @return HTML-encoded text fragment
687 static function safeEncodeAttribute( $text ) {
688 $encValue = Sanitizer
::encodeAttribute( $text );
690 # Templates and links may be expanded in later parsing,
691 # creating invalid or dangerous output. Suppress this.
692 $encValue = strtr( $encValue, array(
693 '<' => '<', // This should never happen,
694 '>' => '>', // we've received invalid input
695 '"' => '"', // which should have been escaped.
698 "''" => '''',
699 'ISBN' => 'ISBN',
701 'PMID' => 'PMID',
707 $encValue = preg_replace_callback(
708 '/(' . wfUrlProtocols() . ')/',
709 array( 'Sanitizer', 'armorLinksCallback' ),
715 * Given a value escape it so that it can be used in an id attribute and
716 * return it, this does not validate the value however (see first link)
718 * @link http://www.w3.org/TR/html401/types.html#type-name Valid characters
721 * @link http://www.w3.org/TR/html401/struct/links.html#h-12.2.3 Anchors with the id attribute
730 static function escapeId( $id ) {
731 static $replace = array(
736 $id = urlencode( Sanitizer
::decodeCharReferences( strtr( $id, ' ', '_' ) ) );
738 return str_replace( array_keys( $replace ), array_values( $replace ), $id );
742 * Given a value, escape it so that it can be used as a CSS class and
745 * TODO: For extra validity, input should be validated UTF-8.
747 * @link http://www.w3.org/TR/CSS21/syndata.html Valid characters/format
749 * @param string $class
752 static function escapeClass( $class ) {
753 // Convert ugly stuff to underscores and kill underscores in ugly places
754 return rtrim(preg_replace(
755 array('/(^[0-9\\-])|[\\x00-\\x20!"#$%&\'()*+,.\\/:;<=>?@[\\]^`{|}~]|\\xC2\\xA0/','/_+/'),
761 * Regex replace callback for armoring links against further processing.
762 * @param array $matches
766 private static function armorLinksCallback( $matches ) {
767 return str_replace( ':', ':', $matches[1] );
771 * Return an associative array of attribute names and values from
772 * a partial tag string. Attribute names are forces to lowercase,
773 * character references are decoded to UTF-8 text.
778 static function decodeTagAttributes( $text ) {
781 if( trim( $text ) == '' ) {
794 foreach( $pairs as $set ) {
795 $attribute = strtolower( $set[1] );
796 $value = Sanitizer
::getTagAttributeCallback( $set );
798 // Normalize whitespace
799 $value = preg_replace( '/[\t\r\n ]+/', ' ', $value );
800 $value = trim( $value );
802 // Decode character references
803 $attribs[$attribute] = Sanitizer
::decodeCharReferences( $value );
809 * Pick the appropriate attribute value from a match set from the
810 * MW_ATTRIBS_REGEX matches.
816 private static function getTagAttributeCallback( $set ) {
817 if( isset( $set[6] ) ) {
818 # Illegal #XXXXXX color with no quotes.
820 } elseif( isset( $set[5] ) ) {
823 } elseif( isset( $set[4] ) ) {
826 } elseif( isset( $set[3] ) ) {
829 } elseif( !isset( $set[2] ) ) {
830 # In XHTML, attributes must have a value.
831 # For 'reduced' form, return explicitly the attribute name here.
834 throw new MWException( "Tag conditions not met. This should never happen and is a bug." );
839 * Normalize whitespace and character references in an XML source-
840 * encoded text for an attribute value.
842 * See http://www.w3.org/TR/REC-xml/#AVNormalize for background,
843 * but note that we're not returning the value, but are returning
844 * XML source fragments that will be slapped into output.
846 * @param string $text
850 private static function normalizeAttributeValue( $text ) {
851 return str_replace( '"', '"',
853 '/\r\n|[\x20\x0d\x0a\x09]/',
855 Sanitizer
::normalizeCharReferences( $text ) ) );
859 * Ensure that any entities and character references are legal
860 * for XML and XHTML specifically. Any stray bits will be
861 * &-escaped to result in a valid text fragment.
863 * a. any named char refs must be known in XHTML
864 * b. any numeric char refs must be legal chars, not invalid or forbidden
865 * c. use &#x, not &#X
866 * d. fix or reject non-valid attributes
868 * @param string $text
872 static function normalizeCharReferences( $text ) {
873 return preg_replace_callback(
875 array( 'Sanitizer', 'normalizeCharReferencesCallback' ),
879 * @param string $matches
882 static function normalizeCharReferencesCallback( $matches ) {
884 if( $matches[1] != '' ) {
885 $ret = Sanitizer
::normalizeEntity( $matches[1] );
886 } elseif( $matches[2] != '' ) {
887 $ret = Sanitizer
::decCharReference( $matches[2] );
888 } elseif( $matches[3] != '' ) {
889 $ret = Sanitizer
::hexCharReference( $matches[3] );
890 } elseif( $matches[4] != '' ) {
891 $ret = Sanitizer
::hexCharReference( $matches[4] );
893 if( is_null( $ret ) ) {
894 return htmlspecialchars( $matches[0] );
901 * If the named entity is defined in the HTML 4.0/XHTML 1.0 DTD,
902 * return the named entity reference as is. Otherwise, returns
903 * HTML-escaped text of pseudo-entity source (eg &foo;)
905 * @param string $name
909 static function normalizeEntity( $name ) {
910 global $wgHtmlEntities;
911 if( isset( $wgHtmlEntities[$name] ) ) {
914 return "&$name;";
918 static function decCharReference( $codepoint ) {
919 $point = intval( $codepoint );
920 if( Sanitizer
::validateCodepoint( $point ) ) {
921 return sprintf( '&#%d;', $point );
927 static function hexCharReference( $codepoint ) {
928 $point = hexdec( $codepoint );
929 if( Sanitizer
::validateCodepoint( $point ) ) {
930 return sprintf( '&#x%x;', $point );
937 * Returns true if a given Unicode codepoint is a valid character in XML.
938 * @param int $codepoint
941 private static function validateCodepoint( $codepoint ) {
942 return ($codepoint == 0x09)
943 ||
($codepoint == 0x0a)
944 ||
($codepoint == 0x0d)
945 ||
($codepoint >= 0x20 && $codepoint <= 0xd7ff)
946 ||
($codepoint >= 0xe000 && $codepoint <= 0xfffd)
947 ||
($codepoint >= 0x10000 && $codepoint <= 0x10ffff);
951 * Decode any character references, numeric or named entities,
952 * in the text and return a UTF-8 string.
954 * @param string $text
959 public static function decodeCharReferences( $text ) {
960 return preg_replace_callback(
962 array( 'Sanitizer', 'decodeCharReferencesCallback' ),
967 * @param string $matches
970 static function decodeCharReferencesCallback( $matches ) {
971 if( $matches[1] != '' ) {
972 return Sanitizer
::decodeEntity( $matches[1] );
973 } elseif( $matches[2] != '' ) {
974 return Sanitizer
::decodeChar( intval( $matches[2] ) );
975 } elseif( $matches[3] != '' ) {
976 return Sanitizer
::decodeChar( hexdec( $matches[3] ) );
977 } elseif( $matches[4] != '' ) {
978 return Sanitizer
::decodeChar( hexdec( $matches[4] ) );
980 # Last case should be an ampersand by itself
985 * Return UTF-8 string for a codepoint if that is a valid
986 * character reference, otherwise U+FFFD REPLACEMENT CHARACTER.
987 * @param int $codepoint
991 static function decodeChar( $codepoint ) {
992 if( Sanitizer
::validateCodepoint( $codepoint ) ) {
993 return codepointToUtf8( $codepoint );
995 return UTF8_REPLACEMENT
;
1000 * If the named entity is defined in the HTML 4.0/XHTML 1.0 DTD,
1001 * return the UTF-8 encoding of that character. Otherwise, returns
1002 * pseudo-entity source (eg &foo;)
1004 * @param string $name
1007 static function decodeEntity( $name ) {
1008 global $wgHtmlEntities;
1009 if( isset( $wgHtmlEntities[$name] ) ) {
1010 return codepointToUtf8( $wgHtmlEntities[$name] );
1017 * Fetch the whitelist of acceptable attributes for a given
1020 * @param string $element
1023 static function attributeWhitelist( $element ) {
1025 if( !isset( $list ) ) {
1026 $list = Sanitizer
::setupAttributeWhitelist();
1028 return isset( $list[$element] )
1034 * @todo Document it a bit
1037 static function setupAttributeWhitelist() {
1038 $common = array( 'id', 'class', 'lang', 'dir', 'title', 'style' );
1039 $block = array_merge( $common, array( 'align' ) );
1040 $tablealign = array( 'align', 'char', 'charoff', 'valign' );
1041 $tablecell = array( 'abbr',
1047 'nowrap', # deprecated
1048 'width', # deprecated
1049 'height', # deprecated
1050 'bgcolor' # deprecated
1053 # Numbers refer to sections in HTML 4.01 standard describing the element.
1054 # See: http://www.w3.org/TR/html4/
1055 $whitelist = array (
1058 'center' => $common, # deprecated
1059 'span' => $block, # ??
1077 'strong' => $common,
1088 'blockquote' => array_merge( $common, array( 'cite' ) ),
1099 'br' => array( 'id', 'class', 'title', 'style', 'clear' ),
1102 'pre' => array_merge( $common, array( 'width' ) ),
1105 'ins' => array_merge( $common, array( 'cite', 'datetime' ) ),
1106 'del' => array_merge( $common, array( 'cite', 'datetime' ) ),
1109 'ul' => array_merge( $common, array( 'type' ) ),
1110 'ol' => array_merge( $common, array( 'type', 'start' ) ),
1111 'li' => array_merge( $common, array( 'type', 'value' ) ),
1119 'table' => array_merge( $common,
1120 array( 'summary', 'width', 'border', 'frame',
1121 'rules', 'cellspacing', 'cellpadding',
1126 'caption' => array_merge( $common, array( 'align' ) ),
1129 'thead' => array_merge( $common, $tablealign ),
1130 'tfoot' => array_merge( $common, $tablealign ),
1131 'tbody' => array_merge( $common, $tablealign ),
1134 'colgroup' => array_merge( $common, array( 'span', 'width' ), $tablealign ),
1135 'col' => array_merge( $common, array( 'span', 'width' ), $tablealign ),
1138 'tr' => array_merge( $common, array( 'bgcolor' ), $tablealign ),
1141 'td' => array_merge( $common, $tablecell, $tablealign ),
1142 'th' => array_merge( $common, $tablecell, $tablealign ),
1150 'strike' => $common,
1155 'font' => array_merge( $common, array( 'size', 'color', 'face' ) ),
1159 'hr' => array_merge( $common, array( 'noshade', 'size', 'width' ) ),
1161 # XHTML Ruby annotation text module, simple ruby only.
1162 # http://www.w3c.org/TR/ruby/
1167 'rt' => $common, #array_merge( $common, array( 'rbspan' ) ),
1174 * Take a fragment of (potentially invalid) HTML and return
1175 * a version with any tags removed, encoded suitably for literal
1176 * inclusion in an attribute value.
1178 * @param string $text HTML fragment
1181 static function stripAllTags( $text ) {
1183 $text = StringUtils
::delimiterReplace( '<', '>', '', $text );
1185 # Normalize &entities and whitespace
1186 $text = Sanitizer
::normalizeAttributeValue( $text );
1188 # Will be placed into "double-quoted" attributes,
1189 # make sure remaining bits are safe.
1190 $text = str_replace(
1191 array('<', '>', '"'),
1192 array('<', '>', '"'),
1199 * Hack up a private DOCTYPE with HTML's standard entity declarations.
1200 * PHP 4 seemed to know these if you gave it an HTML doctype, but
1203 * Use for passing XHTML fragments to PHP's XML parsing functions
1208 static function hackDocType() {
1209 global $wgHtmlEntities;
1210 $out = "<!DOCTYPE html [\n";
1211 foreach( $wgHtmlEntities as $entity => $codepoint ) {
1212 $out .= "<!ENTITY $entity \"&#$codepoint;\">";
1218 static function cleanUrl( $url, $hostname=true ) {
1219 # Normalize any HTML entities in input. They will be
1220 # re-escaped by makeExternalLink().
1221 $url = Sanitizer
::decodeCharReferences( $url );
1223 # Escape any control characters introduced by the above step
1224 $url = preg_replace( '/[\][<>"\\x00-\\x20\\x7F]/e', "urlencode('\\0')", $url );
1226 # Validate hostname portion
1228 if( preg_match( '!^([^:]+:)(//[^/]+)?(.*)$!iD', $url, $matches ) ) {
1229 list( /* $whole */, $protocol, $host, $rest ) = $matches;
1231 // Characters that will be ignored in IDNs.
1232 // http://tools.ietf.org/html/3454#section-3.1
1233 // Strip them before further processing so blacklists and such work.
1235 \\s| # general whitespace
1236 \xc2\xad| # 00ad SOFT HYPHEN
1237 \xe1\xa0\x86| # 1806 MONGOLIAN TODO SOFT HYPHEN
1238 \xe2\x80\x8b| # 200b ZERO WIDTH SPACE
1239 \xe2\x81\xa0| # 2060 WORD JOINER
1240 \xef\xbb\xbf| # feff ZERO WIDTH NO-BREAK SPACE
1241 \xcd\x8f| # 034f COMBINING GRAPHEME JOINER
1242 \xe1\xa0\x8b| # 180b MONGOLIAN FREE VARIATION SELECTOR ONE
1243 \xe1\xa0\x8c| # 180c MONGOLIAN FREE VARIATION SELECTOR TWO
1244 \xe1\xa0\x8d| # 180d MONGOLIAN FREE VARIATION SELECTOR THREE
1245 \xe2\x80\x8c| # 200c ZERO WIDTH NON-JOINER
1246 \xe2\x80\x8d| # 200d ZERO WIDTH JOINER
1247 [\xef\xb8\x80-\xef\xb8\x8f] # fe00-fe00f VARIATION SELECTOR-1-16
1250 $host = preg_replace( $strip, '', $host );
1252 // @fixme: validate hostnames here
1254 return $protocol . $host . $rest;