* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
* http://www.gnu.org/copyleft/gpl.html
*
- * @package MediaWiki
- * @subpackage Parser
+ * @file
+ * @ingroup Parser
*/
/**
* Sanitizer::normalizeCharReferences and Sanitizer::decodeCharReferences
*/
define( 'MW_CHAR_REFS_REGEX',
- '/&([A-Za-z0-9]+);
+ '/&([A-Za-z0-9\x80-\xff]+);
|&\#([0-9]+);
|&\#x([0-9A-Za-z]+);
|&\#X([0-9A-Za-z]+);
/**
* List of all named character entities defined in HTML 4.01
* http://www.w3.org/TR/html4/sgml/entities.html
- * @access private
+ * @private
*/
global $wgHtmlEntities;
$wgHtmlEntities = array(
'zwj' => 8205,
'zwnj' => 8204 );
-/** @package MediaWiki */
+/**
+ * Character entity aliases accepted by MediaWiki
+ */
+global $wgHtmlEntityAliases;
+$wgHtmlEntityAliases = array(
+ 'רלמ' => 'rlm',
+ 'رلم' => 'rlm',
+);
+
+
+/**
+ * XHTML sanitizer for MediaWiki
+ * @ingroup Parser
+ */
class Sanitizer {
+ const NONE = 0;
+ const INITIAL_NONLETTER = 1;
+
/**
* Cleans up HTML, removes dangerous tags and attributes, and
* removes HTML comments
- * @access private
+ * @private
* @param string $text
* @param callback $processCallback to do any variable or parameter replacements in HTML attribute values
* @param array $args for the processing callback
* @return string
*/
- function removeHTMLtags( $text, $processCallback = null, $args = array() ) {
- global $wgUseTidy, $wgUserHtml;
- $fname = 'Parser::removeHTMLtags';
- wfProfileIn( $fname );
+ static function removeHTMLtags( $text, $processCallback = null, $args = array(), $extratags = array() ) {
+ global $wgUseTidy;
+
+ static $htmlpairs, $htmlsingle, $htmlsingleonly, $htmlnest, $tabletags,
+ $htmllist, $listtags, $htmlsingleallowed, $htmlelements, $staticInitialised;
+
+ wfProfileIn( __METHOD__ );
- if( $wgUserHtml ) {
- $htmlpairs = array( # Tags that must be closed
+ if ( !$staticInitialised ) {
+
+ $htmlpairs = array_merge( $extratags, array( # Tags that must be closed
'b', 'del', 'i', 'ins', 'u', 'font', 'big', 'small', 'sub', 'sup', 'h1',
'h2', 'h3', 'h4', 'h5', 'h6', 'cite', 'code', 'em', 's',
'strike', 'strong', 'tt', 'var', 'div', 'center',
'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
- 'ruby', 'rt' , 'rb' , 'rp', 'p', 'span'
- );
+ 'ruby', 'rt' , 'rb' , 'rp', 'p', 'span', 'u'
+ ) );
$htmlsingle = array(
'br', 'hr', 'li', 'dt', 'dd'
);
'table', 'tr', 'td', 'th', 'div', 'blockquote', 'ol', 'ul',
'dl', 'font', 'big', 'small', 'sub', 'sup', 'span'
);
- $tabletags = array( # Can only appear inside table
- 'td', 'th', 'tr'
+ $tabletags = array( # Can only appear inside table, we will close them
+ 'td', 'th', 'tr',
);
- } else {
- $htmlpairs = array();
- $htmlsingle = array();
- $htmlnest = array();
- $tabletags = array();
- }
+ $htmllist = array( # Tags used by list
+ 'ul','ol',
+ );
+ $listtags = array( # Tags that can appear in a list
+ 'li',
+ );
+
+ $htmlsingleallowed = array_merge( $htmlsingle, $tabletags );
+ $htmlelements = array_merge( $htmlsingle, $htmlpairs, $htmlnest );
- $htmlsingle = array_merge( $tabletags, $htmlsingle );
- $htmlelements = array_merge( $htmlsingle, $htmlpairs );
+ # Convert them all to hashtables for faster lookup
+ $vars = array( 'htmlpairs', 'htmlsingle', 'htmlsingleonly', 'htmlnest', 'tabletags',
+ 'htmllist', 'listtags', 'htmlsingleallowed', 'htmlelements' );
+ foreach ( $vars as $var ) {
+ $$var = array_flip( $$var );
+ }
+ $staticInitialised = true;
+ }
# Remove HTML comments
$text = Sanitizer::removeHTMLcomments( $text );
-
$bits = explode( '<', $text );
- $text = array_shift( $bits );
+ $text = str_replace( '>', '>', array_shift( $bits ) );
if(!$wgUseTidy) {
- $tagstack = array(); $tablestack = array();
+ $tagstack = $tablestack = array();
foreach ( $bits as $x ) {
- $prev = error_reporting( E_ALL & ~( E_NOTICE | E_WARNING ) );
- preg_match( '/^(\\/?)(\\w+)([^>]*?)(\\/{0,1}>)([^<]*)$/',
- $x, $regs );
- list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
- error_reporting( $prev );
+ $regs = array();
+ if( preg_match( '!^(/?)(\\w+)([^>]*?)(/{0,1}>)([^<]*)$!', $x, $regs ) ) {
+ list( /* $qbar */, $slash, $t, $params, $brace, $rest ) = $regs;
+ } else {
+ $slash = $t = $params = $brace = $rest = null;
+ }
$badtag = 0 ;
- if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
+ if ( isset( $htmlelements[$t = strtolower( $t )] ) ) {
# Check our stack
if ( $slash ) {
# Closing a tag...
- if( in_array( $t, $htmlsingleonly ) ) {
+ if( isset( $htmlsingleonly[$t] ) ) {
$badtag = 1;
} elseif ( ( $ot = @array_pop( $tagstack ) ) != $t ) {
- @array_push( $tagstack, $ot );
- $badtag = 1;
+ if ( isset( $htmlsingleallowed[$ot] ) ) {
+ # Pop all elements with an optional close tag
+ # and see if we find a match below them
+ $optstack = array();
+ array_push ($optstack, $ot);
+ while ( ( ( $ot = @array_pop( $tagstack ) ) != $t ) &&
+ isset( $htmlsingleallowed[$ot] ) )
+ {
+ array_push ($optstack, $ot);
+ }
+ if ( $t != $ot ) {
+ # No match. Push the optinal elements back again
+ $badtag = 1;
+ while ( $ot = @array_pop( $optstack ) ) {
+ array_push( $tagstack, $ot );
+ }
+ }
+ } else {
+ @array_push( $tagstack, $ot );
+ # <li> can be nested in <ul> or <ol>, skip those cases:
+ if(!(isset( $htmllist[$ot] ) && isset( $listtags[$t] ) )) {
+ $badtag = 1;
+ }
+ }
} else {
if ( $t == 'table' ) {
$tagstack = array_pop( $tablestack );
}
- $newparams = '';
}
+ $newparams = '';
} else {
# Keep track for later
- if ( in_array( $t, $tabletags ) &&
+ if ( isset( $tabletags[$t] ) &&
! in_array( 'table', $tagstack ) ) {
$badtag = 1;
} else if ( in_array( $t, $tagstack ) &&
- ! in_array ( $t , $htmlnest ) ) {
+ ! isset( $htmlnest [$t ] ) ) {
$badtag = 1 ;
- } elseif( in_array( $t, $htmlsingleonly ) ) {
+ # Is it a self closed htmlpair ? (bug 5487)
+ } else if( $brace == '/>' &&
+ isset( $htmlpairs[$t] ) ) {
+ $badtag = 1;
+ } elseif( isset( $htmlsingleonly[$t] ) ) {
# Hack to force empty tag for uncloseable elements
$brace = '/>';
+ } else if( isset( $htmlsingle[$t] ) ) {
+ # Hack to not close $htmlsingle tags
+ $brace = NULL;
+ } else if( isset( $tabletags[$t] )
+ && in_array($t ,$tagstack) ) {
+ // New table tag but forgot to close the previous one
+ $text .= "</$t>";
} else {
if ( $t == 'table' ) {
array_push( $tablestack, $tagstack );
}
if ( ! $badtag ) {
$rest = str_replace( '>', '>', $rest );
- $close = ( $brace == '/>' ) ? ' /' : '';
+ $close = ( $brace == '/>' && !$slash ) ? ' /' : '';
$text .= "<$slash$t$newparams$close>$rest";
continue;
}
foreach ( $bits as $x ) {
preg_match( '/^(\\/?)(\\w+)([^>]*?)(\\/{0,1}>)([^<]*)$/',
$x, $regs );
- @list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
- if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
+ @list( /* $qbar */, $slash, $t, $params, $brace, $rest ) = $regs;
+ if ( isset( $htmlelements[$t = strtolower( $t )] ) ) {
if( is_callable( $processCallback ) ) {
call_user_func_array( $processCallback, array( &$params, $args ) );
}
}
}
}
- wfProfileOut( $fname );
+ wfProfileOut( __METHOD__ );
return $text;
}
* and followed by a newline (ignoring spaces), trim leading and
* trailing spaces and one of the newlines.
*
- * @access private
+ * @private
* @param string $text
* @return string
*/
- function removeHTMLcomments( $text ) {
- $fname='Parser::removeHTMLcomments';
- wfProfileIn( $fname );
+ static function removeHTMLcomments( $text ) {
+ wfProfileIn( __METHOD__ );
while (($start = strpos($text, '<!--')) !== false) {
$end = strpos($text, '-->', $start + 4);
if ($end === false) {
$text = substr_replace($text, '', $start, $end - $start);
}
}
- wfProfileOut( $fname );
+ wfProfileOut( __METHOD__ );
return $text;
}
/**
- * Take a tag soup fragment listing an HTML element's attributes
- * and normalize it to well-formed XML, discarding unwanted attributes.
+ * Take an array of attribute names and values and normalize or discard
+ * illegal values for the given element type.
*
- * - Normalizes attribute names to lowercase
* - Discards attributes not on a whitelist for the given element
- * - Turns broken or invalid entities into plaintext
- * - Double-quotes all attribute values
- * - Attributes without values are given the name as attribute
- * - Double attributes are discarded
* - Unsafe style attributes are discarded
- * - Prepends space if there are attributes.
+ * - Invalid id attributes are reencoded
*
- * @param string $text
+ * @param array $attribs
* @param string $element
- * @return string
+ * @return array
*
* @todo Check for legal values where the DTD limits things.
* @todo Check for unique id attribute :P
*/
- function fixTagAttributes( $text, $element ) {
- if( trim( $text ) == '' ) {
- return '';
- }
-
- # Unquoted attribute
- # Since we quote this later, this can be anything distinguishable
- # from the end of the attribute
- $pairs = array();
- if( !preg_match_all(
- MW_ATTRIBS_REGEX,
- $text,
- $pairs,
- PREG_SET_ORDER ) ) {
- return '';
- }
+ static function validateTagAttributes( $attribs, $element ) {
+ return Sanitizer::validateAttributes( $attribs,
+ Sanitizer::attributeWhitelist( $element ) );
+ }
- $whitelist = array_flip( Sanitizer::attributeWhitelist( $element ) );
- $attribs = array();
- foreach( $pairs as $set ) {
- $attribute = strtolower( $set[1] );
+ /**
+ * Take an array of attribute names and values and normalize or discard
+ * illegal values for the given whitelist.
+ *
+ * - Discards attributes not the given whitelist
+ * - Unsafe style attributes are discarded
+ * - Invalid id attributes are reencoded
+ *
+ * @param array $attribs
+ * @param array $whitelist list of allowed attribute names
+ * @return array
+ *
+ * @todo Check for legal values where the DTD limits things.
+ * @todo Check for unique id attribute :P
+ */
+ static function validateAttributes( $attribs, $whitelist ) {
+ $whitelist = array_flip( $whitelist );
+ $out = array();
+ foreach( $attribs as $attribute => $value ) {
if( !isset( $whitelist[$attribute] ) ) {
continue;
}
-
- $raw = Sanitizer::getTagAttributeCallback( $set );
- $value = Sanitizer::normalizeAttributeValue( $raw );
-
# Strip javascript "expression" from stylesheets.
# http://msdn.microsoft.com/workshop/author/dhtml/overview/recalc.asp
if( $attribute == 'style' ) {
- $stripped = Sanitizer::decodeCharReferences( $value );
-
- // Remove any comments; IE gets token splitting wrong
- $stripped = preg_replace( '!/\\*.*?\\*/!S', ' ', $stripped );
- $value = htmlspecialchars( $stripped );
-
- // ... and continue checks
- $stripped = preg_replace( '!\\\\([0-9A-Fa-f]{1,6})[ \\n\\r\\t\\f]?!e',
- 'codepointToUtf8(hexdec("$1"))', $stripped );
- $stripped = str_replace( '\\', '', $stripped );
- if( preg_match( '/(expression|tps*:\/\/|url\\s*\().*/is',
- $stripped ) ) {
+ $value = Sanitizer::checkCss( $value );
+ if( $value === false ) {
# haxx0r
continue;
}
if ( $attribute === 'id' )
$value = Sanitizer::escapeId( $value );
- # Templates and links may be expanded in later parsing,
- # creating invalid or dangerous output. Suppress this.
- $value = strtr( $value, array(
- '<' => '<', // This should never happen,
- '>' => '>', // we've received invalid input
- '"' => '"', // which should have been escaped.
- '{' => '{',
- '[' => '[',
- "''" => '''',
- 'ISBN' => 'ISBN',
- 'RFC' => 'RFC',
- 'PMID' => 'PMID',
- ) );
-
- # Stupid hack
- $value = preg_replace_callback(
- '/(' . wfUrlProtocols() . ')/',
- array( 'Sanitizer', 'armorLinksCallback' ),
- $value );
-
// If this attribute was previously set, override it.
// Output should only have one attribute of each name.
- $attribs[$attribute] = "$attribute=\"$value\"";
+ $out[$attribute] = $value;
+ }
+ return $out;
+ }
+
+ /**
+ * Merge two sets of HTML attributes. Conflicting items in the second set
+ * will override those in the first, except for 'class' attributes which
+ * will be combined (if they're both strings).
+ *
+ * @todo implement merging for other attributes such as style
+ * @param array $a
+ * @param array $b
+ * @return array
+ */
+ static function mergeAttributes( $a, $b ) {
+ $out = array_merge( $a, $b );
+ if( isset( $a['class'] ) && isset( $b['class'] )
+ && is_string( $a['class'] ) && is_string( $b['class'] )
+ && $a['class'] !== $b['class'] ) {
+ $classes = preg_split( '/\s+/', "{$a['class']} {$b['class']}",
+ -1, PREG_SPLIT_NO_EMPTY );
+ $out['class'] = implode( ' ', array_unique( $classes ) );
+ }
+ return $out;
+ }
+
+ /**
+ * Pick apart some CSS and check it for forbidden or unsafe structures.
+ * Returns a sanitized string, or false if it was just too evil.
+ *
+ * Currently URL references, 'expression', 'tps' are forbidden.
+ *
+ * @param string $value
+ * @return mixed
+ */
+ static function checkCss( $value ) {
+ $stripped = Sanitizer::decodeCharReferences( $value );
+
+ // Remove any comments; IE gets token splitting wrong
+ $stripped = StringUtils::delimiterReplace( '/*', '*/', ' ', $stripped );
+
+ $value = $stripped;
+
+ // ... and continue checks
+ $stripped = preg_replace( '!\\\\([0-9A-Fa-f]{1,6})[ \\n\\r\\t\\f]?!e',
+ 'codepointToUtf8(hexdec("$1"))', $stripped );
+ $stripped = str_replace( '\\', '', $stripped );
+ if( preg_match( '/(?:expression|tps*:\/\/|url\\s*\().*/is',
+ $stripped ) ) {
+ # haxx0r
+ return false;
+ }
+
+ return $value;
+ }
+
+ /**
+ * Take a tag soup fragment listing an HTML element's attributes
+ * and normalize it to well-formed XML, discarding unwanted attributes.
+ * Output is safe for further wikitext processing, with escaping of
+ * values that could trigger problems.
+ *
+ * - Normalizes attribute names to lowercase
+ * - Discards attributes not on a whitelist for the given element
+ * - Turns broken or invalid entities into plaintext
+ * - Double-quotes all attribute values
+ * - Attributes without values are given the name as attribute
+ * - Double attributes are discarded
+ * - Unsafe style attributes are discarded
+ * - Prepends space if there are attributes.
+ *
+ * @param string $text
+ * @param string $element
+ * @return string
+ */
+ static function fixTagAttributes( $text, $element ) {
+ if( trim( $text ) == '' ) {
+ return '';
}
+ $stripped = Sanitizer::validateTagAttributes(
+ Sanitizer::decodeTagAttributes( $text ), $element );
+
+ $attribs = array();
+ foreach( $stripped as $attribute => $value ) {
+ $encAttribute = htmlspecialchars( $attribute );
+ $encValue = Sanitizer::safeEncodeAttribute( $value );
+
+ $attribs[] = "$encAttribute=\"$encValue\"";
+ }
return count( $attribs ) ? ' ' . implode( ' ', $attribs ) : '';
}
+ /**
+ * Encode an attribute value for HTML output.
+ * @param $text
+ * @return HTML-encoded text fragment
+ */
+ static function encodeAttribute( $text ) {
+ $encValue = htmlspecialchars( $text, ENT_QUOTES );
+
+ // Whitespace is normalized during attribute decoding,
+ // so if we've been passed non-spaces we must encode them
+ // ahead of time or they won't be preserved.
+ $encValue = strtr( $encValue, array(
+ "\n" => ' ',
+ "\r" => ' ',
+ "\t" => '	',
+ ) );
+
+ return $encValue;
+ }
+
+ /**
+ * Encode an attribute value for HTML tags, with extra armoring
+ * against further wiki processing.
+ * @param $text
+ * @return HTML-encoded text fragment
+ */
+ static function safeEncodeAttribute( $text ) {
+ $encValue = Sanitizer::encodeAttribute( $text );
+
+ # Templates and links may be expanded in later parsing,
+ # creating invalid or dangerous output. Suppress this.
+ $encValue = strtr( $encValue, array(
+ '<' => '<', // This should never happen,
+ '>' => '>', // we've received invalid input
+ '"' => '"', // which should have been escaped.
+ '{' => '{',
+ '[' => '[',
+ "''" => '''',
+ 'ISBN' => 'ISBN',
+ 'RFC' => 'RFC',
+ 'PMID' => 'PMID',
+ '|' => '|',
+ '__' => '__',
+ ) );
+
+ # Stupid hack
+ $encValue = preg_replace_callback(
+ '/(' . wfUrlProtocols() . ')/',
+ array( 'Sanitizer', 'armorLinksCallback' ),
+ $encValue );
+ return $encValue;
+ }
+
/**
* Given a value escape it so that it can be used in an id attribute and
* return it, this does not validate the value however (see first link)
*
- * @link http://www.w3.org/TR/html401/types.html#type-name Valid characters
+ * @see http://www.w3.org/TR/html401/types.html#type-name Valid characters
* in the id and
* name attributes
- * @link http://www.w3.org/TR/html401/struct/links.html#h-12.2.3 Anchors with the id attribute
- *
- * @bug 4461
+ * @see http://www.w3.org/TR/html401/struct/links.html#h-12.2.3 Anchors with the id attribute
*
- * @static
- *
- * @param string $id
+ * @param string $id Id to validate
+ * @param int $flags Currently only two values: Sanitizer::INITIAL_NONLETTER
+ * (default) permits initial non-letter characters,
+ * such as if you're adding a prefix to them.
+ * Sanitizer::NONE will prepend an 'x' if the id
+ * would otherwise start with a nonletter.
* @return string
*/
- function escapeId( $id ) {
+ static function escapeId( $id, $flags = Sanitizer::INITIAL_NONLETTER ) {
static $replace = array(
'%3A' => ':',
'%' => '.'
);
$id = urlencode( Sanitizer::decodeCharReferences( strtr( $id, ' ', '_' ) ) );
+ $id = str_replace( array_keys( $replace ), array_values( $replace ), $id );
+
+ if( ~$flags & Sanitizer::INITIAL_NONLETTER
+ && !preg_match( '/[a-zA-Z]/', $id[0] ) ) {
+ // Initial character must be a letter!
+ $id = "x$id";
+ }
+ return $id;
+ }
+
+ /**
+ * Given a value, escape it so that it can be used as a CSS class and
+ * return it.
+ *
+ * @todo For extra validity, input should be validated UTF-8.
+ *
+ * @see http://www.w3.org/TR/CSS21/syndata.html Valid characters/format
+ *
+ * @param string $class
+ * @return string
+ */
+ static function escapeClass( $class ) {
+ // Convert ugly stuff to underscores and kill underscores in ugly places
+ return rtrim(preg_replace(
+ array('/(^[0-9\\-])|[\\x00-\\x20!"#$%&\'()*+,.\\/:;<=>?@[\\]^`{|}~]|\\xC2\\xA0/','/_+/'),
+ '_',
+ $class ), '_');
+ }
- return str_replace( array_keys( $replace ), array_values( $replace ), $id );
+ /**
+ * Given HTML input, escape with htmlspecialchars but un-escape entites.
+ * This allows (generally harmless) entities like to survive.
+ *
+ * @param string $html String to escape
+ * @return string Escaped input
+ */
+ static function escapeHtmlAllowEntities( $html ) {
+ # It seems wise to escape ' as well as ", as a matter of course. Can't
+ # hurt.
+ $html = htmlspecialchars( $html, ENT_QUOTES );
+ $html = str_replace( '&', '&', $html );
+ $html = Sanitizer::normalizeCharReferences( $html );
+ return $html;
}
/**
* Regex replace callback for armoring links against further processing.
* @param array $matches
* @return string
- * @access private
+ * @private
*/
- function armorLinksCallback( $matches ) {
+ private static function armorLinksCallback( $matches ) {
return str_replace( ':', ':', $matches[1] );
}
* @param string
* @return array
*/
- function decodeTagAttributes( $text ) {
+ public static function decodeTagAttributes( $text ) {
$attribs = array();
if( trim( $text ) == '' ) {
foreach( $pairs as $set ) {
$attribute = strtolower( $set[1] );
$value = Sanitizer::getTagAttributeCallback( $set );
+
+ // Normalize whitespace
+ $value = preg_replace( '/[\t\r\n ]+/', ' ', $value );
+ $value = trim( $value );
+
+ // Decode character references
$attribs[$attribute] = Sanitizer::decodeCharReferences( $value );
}
return $attribs;
*
* @param array $set
* @return string
- * @access private
+ * @private
*/
- function getTagAttributeCallback( $set ) {
+ private static function getTagAttributeCallback( $set ) {
if( isset( $set[6] ) ) {
# Illegal #XXXXXX color with no quotes.
return $set[6];
# For 'reduced' form, return explicitly the attribute name here.
return $set[1];
} else {
- wfDebugDieBacktrace( "Tag conditions not met. This should never happen and is a bug." );
+ throw new MWException( "Tag conditions not met. This should never happen and is a bug." );
}
}
*
* @param string $text
* @return string
- * @access private
+ * @private
*/
- function normalizeAttributeValue( $text ) {
+ private static function normalizeAttributeValue( $text ) {
return str_replace( '"', '"',
- preg_replace(
- '/\r\n|[\x20\x0d\x0a\x09]/',
- ' ',
+ self::normalizeWhitespace(
Sanitizer::normalizeCharReferences( $text ) ) );
}
+ private static function normalizeWhitespace( $text ) {
+ return preg_replace(
+ '/\r\n|[\x20\x0d\x0a\x09]/',
+ ' ',
+ $text );
+ }
+
/**
* Ensure that any entities and character references are legal
* for XML and XHTML specifically. Any stray bits will be
*
* @param string $text
* @return string
- * @access private
+ * @private
*/
- function normalizeCharReferences( $text ) {
+ static function normalizeCharReferences( $text ) {
return preg_replace_callback(
MW_CHAR_REFS_REGEX,
array( 'Sanitizer', 'normalizeCharReferencesCallback' ),
* @param string $matches
* @return string
*/
- function normalizeCharReferencesCallback( $matches ) {
+ static function normalizeCharReferencesCallback( $matches ) {
$ret = null;
if( $matches[1] != '' ) {
$ret = Sanitizer::normalizeEntity( $matches[1] );
/**
* If the named entity is defined in the HTML 4.0/XHTML 1.0 DTD,
- * return the named entity reference as is. Otherwise, returns
- * HTML-escaped text of pseudo-entity source (eg &foo;)
+ * return the named entity reference as is. If the entity is a
+ * MediaWiki-specific alias, returns the HTML equivalent. Otherwise,
+ * returns HTML-escaped text of pseudo-entity source (eg &foo;)
*
* @param string $name
* @return string
+ * @static
*/
- function normalizeEntity( $name ) {
- global $wgHtmlEntities;
- if( isset( $wgHtmlEntities[$name] ) ) {
+ static function normalizeEntity( $name ) {
+ global $wgHtmlEntities, $wgHtmlEntityAliases;
+ if ( isset( $wgHtmlEntityAliases[$name] ) ) {
+ return "&{$wgHtmlEntityAliases[$name]};";
+ } elseif( isset( $wgHtmlEntities[$name] ) ) {
return "&$name;";
} else {
return "&$name;";
}
}
- function decCharReference( $codepoint ) {
+ static function decCharReference( $codepoint ) {
$point = intval( $codepoint );
if( Sanitizer::validateCodepoint( $point ) ) {
return sprintf( '&#%d;', $point );
}
}
- function hexCharReference( $codepoint ) {
+ static function hexCharReference( $codepoint ) {
$point = hexdec( $codepoint );
if( Sanitizer::validateCodepoint( $point ) ) {
return sprintf( '&#x%x;', $point );
* @param int $codepoint
* @return bool
*/
- function validateCodepoint( $codepoint ) {
+ private static function validateCodepoint( $codepoint ) {
return ($codepoint == 0x09)
|| ($codepoint == 0x0a)
|| ($codepoint == 0x0d)
*
* @param string $text
* @return string
- * @access public
+ * @public
+ * @static
*/
- function decodeCharReferences( $text ) {
+ public static function decodeCharReferences( $text ) {
return preg_replace_callback(
MW_CHAR_REFS_REGEX,
array( 'Sanitizer', 'decodeCharReferencesCallback' ),
* @param string $matches
* @return string
*/
- function decodeCharReferencesCallback( $matches ) {
+ static function decodeCharReferencesCallback( $matches ) {
if( $matches[1] != '' ) {
return Sanitizer::decodeEntity( $matches[1] );
} elseif( $matches[2] != '' ) {
* character reference, otherwise U+FFFD REPLACEMENT CHARACTER.
* @param int $codepoint
* @return string
- * @access private
+ * @private
*/
- function decodeChar( $codepoint ) {
+ static function decodeChar( $codepoint ) {
if( Sanitizer::validateCodepoint( $codepoint ) ) {
return codepointToUtf8( $codepoint );
} else {
* @param string $name
* @return string
*/
- function decodeEntity( $name ) {
- global $wgHtmlEntities;
+ static function decodeEntity( $name ) {
+ global $wgHtmlEntities, $wgHtmlEntityAliases;
+ if ( isset( $wgHtmlEntityAliases[$name] ) ) {
+ $name = $wgHtmlEntityAliases[$name];
+ }
if( isset( $wgHtmlEntities[$name] ) ) {
return codepointToUtf8( $wgHtmlEntities[$name] );
} else {
* @param string $element
* @return array
*/
- function attributeWhitelist( $element ) {
+ static function attributeWhitelist( $element ) {
static $list;
if( !isset( $list ) ) {
$list = Sanitizer::setupAttributeWhitelist();
}
/**
+ * Foreach array key (an allowed HTML element), return an array
+ * of allowed attributes
* @return array
*/
- function setupAttributeWhitelist() {
+ static function setupAttributeWhitelist() {
$common = array( 'id', 'class', 'lang', 'dir', 'title', 'style' );
$block = array_merge( $common, array( 'align' ) );
$tablealign = array( 'align', 'char', 'charoff', 'valign' );
# 11.2.1
'table' => array_merge( $common,
array( 'summary', 'width', 'border', 'frame',
- 'rules', 'cellspacing', 'cellpadding',
- 'align', 'bgcolor', 'frame', 'rules',
- 'border' ) ),
+ 'rules', 'cellspacing', 'cellpadding',
+ 'align', 'bgcolor',
+ ) ),
# 11.2.2
'caption' => array_merge( $common, array( 'align' ) ),
'td' => array_merge( $common, $tablecell, $tablealign ),
'th' => array_merge( $common, $tablecell, $tablealign ),
+ # 13.2
+ # Not usually allowed, but may be used for extension-style hooks
+ # such as <math> when it is rasterized
+ 'img' => array_merge( $common, array( 'alt' ) ),
+
# 15.2.1
'tt' => $common,
'b' => $common,
'rb' => $common,
'rt' => $common, #array_merge( $common, array( 'rbspan' ) ),
'rp' => $common,
+
+ # MathML root element, where used for extensions
+ # 'title' may not be 100% valid here; it's XHTML
+ # http://www.w3.org/TR/REC-MathML/
+ 'math' => array( 'class', 'style', 'id', 'title' ),
);
return $whitelist;
}
/**
* Take a fragment of (potentially invalid) HTML and return
- * a version with any tags removed, encoded suitably for literal
- * inclusion in an attribute value.
+ * a version with any tags removed, encoded as plain text.
+ *
+ * Warning: this return value must be further escaped for literal
+ * inclusion in HTML output as of 1.10!
*
* @param string $text HTML fragment
* @return string
*/
- function stripAllTags( $text ) {
+ static function stripAllTags( $text ) {
# Actual <tags>
- $text = preg_replace( '/ < .*? > /x', '', $text );
+ $text = StringUtils::delimiterReplace( '<', '>', '', $text );
# Normalize &entities and whitespace
- $text = Sanitizer::normalizeAttributeValue( $text );
-
- # Will be placed into "double-quoted" attributes,
- # make sure remaining bits are safe.
- $text = str_replace(
- array('<', '>', '"'),
- array('<', '>', '"'),
- $text );
+ $text = self::decodeCharReferences( $text );
+ $text = self::normalizeWhitespace( $text );
return $text;
}
* @return string
* @static
*/
- function hackDocType() {
+ static function hackDocType() {
global $wgHtmlEntities;
$out = "<!DOCTYPE html [\n";
foreach( $wgHtmlEntities as $entity => $codepoint ) {
return $out;
}
-}
+ static function cleanUrl( $url ) {
+ # Normalize any HTML entities in input. They will be
+ # re-escaped by makeExternalLink().
+ $url = Sanitizer::decodeCharReferences( $url );
+
+ # Escape any control characters introduced by the above step
+ $url = preg_replace( '/[\][<>"\\x00-\\x20\\x7F]/e', "urlencode('\\0')", $url );
+
+ # Validate hostname portion
+ $matches = array();
+ if( preg_match( '!^([^:]+:)(//[^/]+)?(.*)$!iD', $url, $matches ) ) {
+ list( /* $whole */, $protocol, $host, $rest ) = $matches;
+
+ // Characters that will be ignored in IDNs.
+ // http://tools.ietf.org/html/3454#section-3.1
+ // Strip them before further processing so blacklists and such work.
+ $strip = "/
+ \\s| # general whitespace
+ \xc2\xad| # 00ad SOFT HYPHEN
+ \xe1\xa0\x86| # 1806 MONGOLIAN TODO SOFT HYPHEN
+ \xe2\x80\x8b| # 200b ZERO WIDTH SPACE
+ \xe2\x81\xa0| # 2060 WORD JOINER
+ \xef\xbb\xbf| # feff ZERO WIDTH NO-BREAK SPACE
+ \xcd\x8f| # 034f COMBINING GRAPHEME JOINER
+ \xe1\xa0\x8b| # 180b MONGOLIAN FREE VARIATION SELECTOR ONE
+ \xe1\xa0\x8c| # 180c MONGOLIAN FREE VARIATION SELECTOR TWO
+ \xe1\xa0\x8d| # 180d MONGOLIAN FREE VARIATION SELECTOR THREE
+ \xe2\x80\x8c| # 200c ZERO WIDTH NON-JOINER
+ \xe2\x80\x8d| # 200d ZERO WIDTH JOINER
+ [\xef\xb8\x80-\xef\xb8\x8f] # fe00-fe00f VARIATION SELECTOR-1-16
+ /xuD";
+
+ $host = preg_replace( $strip, '', $host );
+
+ // @fixme: validate hostnames here
+
+ return $protocol . $host . $rest;
+ } else {
+ return $url;
+ }
+ }
-?>
+}