* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
* http://www.gnu.org/copyleft/gpl.html
*
- * @addtogroup 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]+);
'zwj' => 8205,
'zwnj' => 8204 );
+/**
+ * Character entity aliases accepted by MediaWiki
+ */
+global $wgHtmlEntityAliases;
+$wgHtmlEntityAliases = array(
+ 'רלמ' => 'rlm',
+ 'رلم' => 'rlm',
+);
+
+
+/**
+ * XHTML sanitizer for MediaWiki
+ * @ingroup Parser
+ */
class Sanitizer {
/**
* Cleans up HTML, removes dangerous tags and attributes, and
* removes HTML comments
* @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
+ * @param $text String
+ * @param $processCallback Callback to do any variable or parameter replacements in HTML attribute values
+ * @param $args Array for the processing callback
+ * @param $extratags Array for any extra tags to include
+ * @param $removetags Array for any tags (default or extra) to exclude
* @return string
*/
- static function removeHTMLtags( $text, $processCallback = null, $args = array() ) {
+ static function removeHTMLtags( $text, $processCallback = null, $args = array(), $extratags = array(), $removetags = array() ) {
global $wgUseTidy;
- static $htmlpairs, $htmlsingle, $htmlsingleonly, $htmlnest, $tabletags,
- $htmllist, $listtags, $htmlsingleallowed, $htmlelements, $staticInitialised;
+ static $htmlpairsStatic, $htmlsingle, $htmlsingleonly, $htmlnest, $tabletags,
+ $htmllist, $listtags, $htmlsingleallowed, $htmlelementsStatic, $staticInitialised;
wfProfileIn( __METHOD__ );
if ( !$staticInitialised ) {
- $htmlpairs = array( # Tags that must be closed
+ $htmlpairsStatic = 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',
'li',
);
- $htmlsingleallowed = array_merge( $htmlsingle, $tabletags );
- $htmlelements = array_merge( $htmlsingle, $htmlpairs, $htmlnest );
+ $htmlsingleallowed = array_unique( array_merge( $htmlsingle, $tabletags ) );
+ $htmlelementsStatic = array_unique( array_merge( $htmlsingle, $htmlpairsStatic, $htmlnest ) );
# Convert them all to hashtables for faster lookup
- $vars = array( 'htmlpairs', 'htmlsingle', 'htmlsingleonly', 'htmlnest', 'tabletags',
- 'htmllist', 'listtags', 'htmlsingleallowed', 'htmlelements' );
+ $vars = array( 'htmlpairsStatic', 'htmlsingle', 'htmlsingleonly', 'htmlnest', 'tabletags',
+ 'htmllist', 'listtags', 'htmlsingleallowed', 'htmlelementsStatic' );
foreach ( $vars as $var ) {
$$var = array_flip( $$var );
}
$staticInitialised = true;
}
+ # Populate $htmlpairs and $htmlelements with the $extratags and $removetags arrays
+ $extratags = array_flip( $extratags );
+ $removetags = array_flip( $removetags );
+ $htmlpairs = array_merge( $extratags, $htmlpairsStatic );
+ $htmlelements = array_diff_key( array_merge( $extratags, $htmlelementsStatic ) , $removetags );
# Remove HTML comments
$text = Sanitizer::removeHTMLcomments( $text );
$optstack = array();
array_push ($optstack, $ot);
while ( ( ( $ot = @array_pop( $tagstack ) ) != $t ) &&
- isset( $htmlsingleallowed[$ot] ) )
+ isset( $htmlsingleallowed[$ot] ) )
{
array_push ($optstack, $ot);
}
* trailing spaces and one of the newlines.
*
* @private
- * @param string $text
+ * @param $text String
* @return string
*/
static function removeHTMLcomments( $text ) {
*
* - Discards attributes not on a whitelist for the given element
* - Unsafe style attributes are discarded
+ * - Invalid id attributes are reencoded
*
- * @param array $attribs
- * @param string $element
- * @return array
+ * @param $attribs Array
+ * @param $element String
+ * @return Array
*
* @todo Check for legal values where the DTD limits things.
* @todo Check for unique id attribute :P
*/
static function validateTagAttributes( $attribs, $element ) {
- $whitelist = array_flip( Sanitizer::attributeWhitelist( $element ) );
+ return Sanitizer::validateAttributes( $attribs,
+ Sanitizer::attributeWhitelist( $element ) );
+ }
+
+ /**
+ * 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 $attribs Array
+ * @param $whitelist Array: 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] ) ) {
}
}
- if ( $attribute === 'id' )
- $value = Sanitizer::escapeId( $value );
+ if ( $attribute === 'id' ) {
+ global $wgEnforceHtmlIds;
+ $value = Sanitizer::escapeId( $value,
+ $wgEnforceHtmlIds ? 'noninitial' : 'xml' );
+ }
// If this attribute was previously set, override it.
// Output should only have one attribute of each name.
}
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 $a Array
+ * @param $b Array
+ * @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
+ * @param $value String
+ * @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
# haxx0r
return false;
}
-
+
return $value;
}
* - Unsafe style attributes are discarded
* - Prepends space if there are attributes.
*
- * @param string $text
- * @param string $element
- * @return string
+ * @param $text String
+ * @param $element String
+ * @return String
*/
static function fixTagAttributes( $text, $element ) {
if( trim( $text ) == '' ) {
/**
* Encode an attribute value for HTML output.
- * @param $text
+ * @param $text String
* @return HTML-encoded text fragment
*/
static function encodeAttribute( $text ) {
- $encValue = htmlspecialchars( $text );
+ $encValue = htmlspecialchars( $text, ENT_QUOTES );
// Whitespace is normalized during attribute decoding,
// so if we've been passed non-spaces we must encode them
/**
* Encode an attribute value for HTML tags, with extra armoring
* against further wiki processing.
- * @param $text
+ * @param $text String
* @return HTML-encoded text fragment
*/
static function safeEncodeAttribute( $text ) {
* name attributes
* @see http://www.w3.org/TR/html401/struct/links.html#h-12.2.3 Anchors with the id attribute
*
- * @static
- *
- * @param string $id
- * @return string
+ * @param $id String: id to validate
+ * @param $options Mixed: string or array of strings (default is array()):
+ * 'noninitial': This is a non-initial fragment of an id, not a full id,
+ * so don't pay attention if the first character isn't valid at the
+ * beginning of an id.
+ * 'xml': Don't restrict the id to be HTML4-compatible. This option
+ * allows any alphabetic character to be used, per the XML standard.
+ * Therefore, it also completely changes the type of escaping: instead
+ * of weird dot-encoding, runs of invalid characters (mostly
+ * whitespace) are just compressed into a single underscore.
+ * @return String
*/
- static function escapeId( $id ) {
- static $replace = array(
- '%3A' => ':',
- '%' => '.'
- );
+ static function escapeId( $id, $options = array() ) {
+ $options = (array)$options;
+
+ if ( !in_array( 'xml', $options ) ) {
+ # HTML4-style escaping
+ static $replace = array(
+ '%3A' => ':',
+ '%' => '.'
+ );
+
+ $id = urlencode( Sanitizer::decodeCharReferences( strtr( $id, ' ', '_' ) ) );
+ $id = str_replace( array_keys( $replace ), array_values( $replace ), $id );
+
+ if ( !preg_match( '/^[a-zA-Z]/', $id )
+ && !in_array( 'noninitial', $options ) ) {
+ // Initial character must be a letter!
+ $id = "x$id";
+ }
+ return $id;
+ }
- $id = urlencode( Sanitizer::decodeCharReferences( strtr( $id, ' ', '_' ) ) );
+ # XML-style escaping. For the patterns used, see the XML 1.0 standard,
+ # 5th edition, NameStartChar and NameChar: <http://www.w3.org/TR/REC-xml/>
+ $nameStartChar = ':a-zA-Z_\xC0-\xD6\xD8-\xF6\xF8-\x{2FF}\x{370}-\x{37D}'
+ . '\x{37F}-\x{1FFF}\x{200C}-\x{200D}\x{2070}-\x{218F}\x{2C00}-\x{2FEF}'
+ . '\x{3001}-\x{D7FF}\x{F900}-\x{FDCF}\x{FDF0}-\x{FFFD}\x{10000}-\x{EFFFF}';
+ $nameChar = $nameStartChar . '.\-0-9\xB7\x{0300}-\x{036F}'
+ . '\x{203F}-\x{2040}';
+ # Replace _ as well so we don't get multiple consecutive underscores
+ $id = preg_replace( "/([^$nameChar]|_)+/u", '_', $id );
+ $id = trim( $id, '_' );
+
+ if ( !preg_match( "/^[$nameStartChar]/u", $id )
+ && !in_array( 'noninitial', $options ) ) {
+ $id = "_$id";
+ }
- return str_replace( array_keys( $replace ), array_values( $replace ), $id );
+ return $id;
}
/**
*
* @see http://www.w3.org/TR/CSS21/syndata.html Valid characters/format
*
- * @param string $class
- * @return string
+ * @param $class String
+ * @return String
*/
static function escapeClass( $class ) {
// Convert ugly stuff to underscores and kill underscores in ugly places
$class ), '_');
}
+ /**
+ * Given HTML input, escape with htmlspecialchars but un-escape entites.
+ * This allows (generally harmless) entities like to survive.
+ *
+ * @param $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
+ * @param $matches Array
* @return string
- * @private
*/
private static function armorLinksCallback( $matches ) {
return str_replace( ':', ':', $matches[1] );
* a partial tag string. Attribute names are forces to lowercase,
* character references are decoded to UTF-8 text.
*
- * @param string
- * @return array
+ * @param $text String
+ * @return Array
*/
- static function decodeTagAttributes( $text ) {
+ public static function decodeTagAttributes( $text ) {
$attribs = array();
if( trim( $text ) == '' ) {
* Pick the appropriate attribute value from a match set from the
* MW_ATTRIBS_REGEX matches.
*
- * @param array $set
- * @return string
- * @private
+ * @param $set Array
+ * @return String
*/
private static function getTagAttributeCallback( $set ) {
if( isset( $set[6] ) ) {
* but note that we're not returning the value, but are returning
* XML source fragments that will be slapped into output.
*
- * @param string $text
- * @return string
- * @private
+ * @param $text String
+ * @return String
*/
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
* c. use &#x, not &#X
* d. fix or reject non-valid attributes
*
- * @param string $text
- * @return string
+ * @param $text String
+ * @return String
* @private
*/
static function normalizeCharReferences( $text ) {
$text );
}
/**
- * @param string $matches
- * @return string
+ * @param $matches String
+ * @return String
*/
static function normalizeCharReferencesCallback( $matches ) {
$ret = null;
/**
* 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
+ * @param $name String
+ * @return String
*/
static function normalizeEntity( $name ) {
- global $wgHtmlEntities;
- if( isset( $wgHtmlEntities[$name] ) ) {
+ global $wgHtmlEntities, $wgHtmlEntityAliases;
+ if ( isset( $wgHtmlEntityAliases[$name] ) ) {
+ return "&{$wgHtmlEntityAliases[$name]};";
+ } elseif( isset( $wgHtmlEntities[$name] ) ) {
return "&$name;";
} else {
return "&$name;";
/**
* Returns true if a given Unicode codepoint is a valid character in XML.
- * @param int $codepoint
- * @return bool
+ * @param $codepoint Integer
+ * @return Boolean
*/
private static function validateCodepoint( $codepoint ) {
return ($codepoint == 0x09)
* Decode any character references, numeric or named entities,
* in the text and return a UTF-8 string.
*
- * @param string $text
- * @return string
- * @public
- * @static
+ * @param $text String
+ * @return String
*/
public static function decodeCharReferences( $text ) {
return preg_replace_callback(
}
/**
- * @param string $matches
- * @return string
+ * @param $matches String
+ * @return String
*/
static function decodeCharReferencesCallback( $matches ) {
if( $matches[1] != '' ) {
/**
* Return UTF-8 string for a codepoint if that is a valid
* character reference, otherwise U+FFFD REPLACEMENT CHARACTER.
- * @param int $codepoint
- * @return string
+ * @param $codepoint Integer
+ * @return String
* @private
*/
static function decodeChar( $codepoint ) {
* return the UTF-8 encoding of that character. Otherwise, returns
* pseudo-entity source (eg &foo;)
*
- * @param string $name
- * @return string
+ * @param $name Strings
+ * @return String
*/
static function decodeEntity( $name ) {
- global $wgHtmlEntities;
+ global $wgHtmlEntities, $wgHtmlEntityAliases;
+ if ( isset( $wgHtmlEntityAliases[$name] ) ) {
+ $name = $wgHtmlEntityAliases[$name];
+ }
if( isset( $wgHtmlEntities[$name] ) ) {
return codepointToUtf8( $wgHtmlEntities[$name] );
} else {
}
/**
- * Fetch the whitelist of acceptable attributes for a given
- * element name.
+ * Fetch the whitelist of acceptable attributes for a given element name.
*
- * @param string $element
- * @return array
+ * @param $element String
+ * @return Array
*/
static function attributeWhitelist( $element ) {
static $list;
}
/**
- * @todo Document it a bit
- * @return array
+ * Foreach array key (an allowed HTML element), return an array
+ * of allowed attributes
+ * @return Array
*/
static function setupAttributeWhitelist() {
$common = array( 'id', 'class', 'lang', 'dir', 'title', 'style' );
'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.
*
- * @param string $text HTML fragment
- * @return string
+ * Warning: this return value must be further escaped for literal
+ * inclusion in HTML output as of 1.10!
+ *
+ * @param $text String: HTML fragment
+ * @return String
*/
static function stripAllTags( $text ) {
# Actual <tags>
$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;
}
*
* Use for passing XHTML fragments to PHP's XML parsing functions
*
- * @return string
- * @static
+ * @return String
*/
static function hackDocType() {
global $wgHtmlEntities;
return $out;
}
- static function cleanUrl( $url, $hostname=true ) {
+ static function cleanUrl( $url ) {
# Normalize any HTML entities in input. They will be
# re-escaped by makeExternalLink().
$url = Sanitizer::decodeCharReferences( $url );
}
}
-
-?>
dépôts / lhc / web / wiklou.git / blobdiff