*/
private $response;
+ /**
+ * Cached client IP address
+ * @var String
+ */
+ private $ip;
+
public function __construct() {
/// @todo FIXME: This preemptive de-quoting can interfere with other web libraries
/// and increases our memory footprint. It would be cleaner to do on
}
$matches = self::extractTitle( $path, $variantPaths, 'variant' );
}
+
+ wfRunHooks( 'WebRequestGetPathInfoRequestURI', array( $path, &$matches ) );
}
} elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
// Mangled PATH_INFO
return $matches;
}
+ /**
+ * Work out an appropriate URL prefix containing scheme and host, based on
+ * information detected from $_SERVER
+ *
+ * @return string
+ */
+ public static function detectServer() {
+ list( $proto, $stdPort ) = self::detectProtocolAndStdPort();
+
+ $varNames = array( 'HTTP_HOST', 'SERVER_NAME', 'HOSTNAME', 'SERVER_ADDR' );
+ $host = 'localhost';
+ $port = $stdPort;
+ foreach ( $varNames as $varName ) {
+ if ( !isset( $_SERVER[$varName] ) ) {
+ continue;
+ }
+ $parts = IP::splitHostAndPort( $_SERVER[$varName] );
+ if ( !$parts ) {
+ // Invalid, do not use
+ continue;
+ }
+ $host = $parts[0];
+ if ( $parts[1] === false ) {
+ if ( isset( $_SERVER['SERVER_PORT'] ) ) {
+ $port = $_SERVER['SERVER_PORT'];
+ } // else leave it as $stdPort
+ } else {
+ $port = $parts[1];
+ }
+ break;
+ }
+
+ return $proto . '://' . IP::combineHostAndPort( $host, $port, $stdPort );
+ }
+
+ /**
+ * @return array
+ */
+ public static function detectProtocolAndStdPort() {
+ return ( isset( $_SERVER['HTTPS'] ) && $_SERVER['HTTPS'] == 'on' ) ? array( 'https', 443 ) : array( 'http', 80 );
+ }
+
+ /**
+ * @return string
+ */
+ public static function detectProtocol() {
+ list( $proto, $stdPort ) = self::detectProtocolAndStdPort();
+ return $proto;
+ }
+
/**
* Check for title, action, and/or variant data in the URL
* and interpolate it into the GET variables.
}
/**
- * Internal URL rewriting function; tries to extract page title and,
+ * URL rewriting function; tries to extract page title and,
* optionally, one other fixed parameter value from a URL path.
*
* @param $path string: the URL path given from the client
* passed on as the value of this URL parameter
* @return array of URL variables to interpolate; empty if no match
*/
- private static function extractTitle( $path, $bases, $key=false ) {
+ static function extractTitle( $path, $bases, $key = false ) {
foreach( (array)$bases as $keyValue => $base ) {
// Find the part after $wgArticlePath
$base = str_replace( '$1', '', $base );
* used for undoing the evil that is magic_quotes_gpc.
*
* @param $arr array: will be modified
+ * @param $recursion bool Used to modify behaviour based on recursion
* @return array the original array
*/
- private function &fix_magic_quotes( &$arr ) {
+ private function &fix_magic_quotes( &$arr, $recursion = false ) {
+ $clean = array();
foreach( $arr as $key => $val ) {
if( is_array( $val ) ) {
- $this->fix_magic_quotes( $arr[$key] );
+ $cleanKey = !$recursion ? stripslashes( $key ) : $key;
+ $clean[$cleanKey] = $this->fix_magic_quotes( $arr[$key], true );
} else {
- $arr[$key] = stripslashes( $val );
+ $cleanKey = stripslashes( $key );
+ $clean[$cleanKey] = stripslashes( $val );
}
}
+ $arr = $clean;
return $arr;
}
}
} else {
global $wgContLang;
- $data = $wgContLang->normalize( $data );
+ $data = isset( $wgContLang ) ? $wgContLang->normalize( $data ) : UtfNormal::cleanUp( $data );
}
return $data;
}
}
/**
- * Set an aribtrary value into our get/post data.
+ * Set an arbitrary value into our get/post data.
*
* @param $key String: key name to use
* @param $value Mixed: value to set
* Extracts the given named values into an array.
* If no arguments are given, returns all input values.
* No transformation is performed on the values.
+ *
+ * @return array
*/
public function getValues() {
$names = func_get_args();
return $retVal;
}
+ /**
+ * Returns the names of all input values excluding those in $exclude.
+ *
+ * @param $exclude Array
+ * @return array
+ */
+ public function getValueNames( $exclude = array() ) {
+ return array_diff( array_keys( $this->getValues() ), $exclude );
+ }
+
/**
* Get the values passed in the query string.
* No transformation is performed on the values.
* @return Boolean
*/
public function wasPosted() {
- return $_SERVER['REQUEST_METHOD'] == 'POST';
+ return isset( $_SERVER['REQUEST_METHOD'] ) && $_SERVER['REQUEST_METHOD'] == 'POST';
}
/**
}
/**
- * Return the path portion of the request URI.
+ * Return the path and query string portion of the request URI.
+ * This will be suitable for use as a relative link in HTML output.
*
* @return String
*/
}
/**
- * Return the request URI with the canonical service and hostname.
+ * Return the request URI with the canonical service and hostname, path,
+ * and query string. This will be suitable for use as an absolute link
+ * in HTML or other output.
+ *
+ * If $wgServer is protocol-relative, this will return a fully
+ * qualified URL with the protocol that was used for this request.
*
* @return String
*/
public function getFullRequestURL() {
- global $wgServer;
- return $wgServer . $this->getRequestURL();
+ return wfExpandUrl( $this->getRequestURL(), PROTO_CURRENT );
}
/**
return htmlspecialchars( $this->appendQuery( $query ) );
}
+ /**
+ * @param $key
+ * @param $value
+ * @param $onlyquery bool
+ * @return String
+ */
public function appendQueryValue( $key, $value, $onlyquery = false ) {
return $this->appendQueryArray( array( $key => $value ), $onlyquery );
}
/**
* Return a WebRequestUpload object corresponding to the key
*
- * @param @key string
+ * @param $key string
* @return WebRequestUpload
*/
public function getUpload( $key ) {
$this->headers[ strtoupper( $tempName ) ] = $tempValue;
}
} else {
- $headers = $_SERVER;
foreach ( $_SERVER as $name => $value ) {
if ( substr( $name, 0, 5 ) === 'HTTP_' ) {
$name = str_replace( '_', '-', substr( $name, 5 ) );
/**
* Get a request header, or false if it isn't set
* @param $name String: case-insensitive header name
+ *
+ * @return string|false
*/
public function getHeader( $name ) {
$this->initHeaders();
$_SESSION[$key] = $data;
}
+ /**
+ * Check if Internet Explorer will detect an incorrect cache extension in
+ * PATH_INFO or QUERY_STRING. If the request can't be allowed, show an error
+ * message or redirect to a safer URL. Returns true if the URL is OK, and
+ * false if an error message has been shown and the request should be aborted.
+ *
+ * @param $extWhitelist array
+ * @return bool
+ */
+ public function checkUrlExtension( $extWhitelist = array() ) {
+ global $wgScriptExtension;
+ $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
+ if ( IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist ) ) {
+ if ( !$this->wasPosted() ) {
+ $newUrl = IEUrlExtension::fixUrlForIE6(
+ $this->getFullRequestURL(), $extWhitelist );
+ if ( $newUrl !== false ) {
+ $this->doSecurityRedirect( $newUrl );
+ return false;
+ }
+ }
+ throw new HttpError( 403,
+ 'Invalid file extension found in the path info or query string.' );
+ }
+ return true;
+ }
+
+ /**
+ * Attempt to redirect to a URL with a QUERY_STRING that's not dangerous in
+ * IE 6. Returns true if it was successful, false otherwise.
+ *
+ * @param $url string
+ * @return bool
+ */
+ protected function doSecurityRedirect( $url ) {
+ header( 'Location: ' . $url );
+ header( 'Content-Type: text/html' );
+ $encUrl = htmlspecialchars( $url );
+ echo <<<HTML
+<html>
+<head>
+<title>Security redirect</title>
+</head>
+<body>
+<h1>Security redirect</h1>
+<p>
+We can't serve non-HTML content from the URL you have requested, because
+Internet Explorer would interpret it as an incorrect and potentially dangerous
+content type.</p>
+<p>Instead, please use <a href="$encUrl">this URL</a>, which is the same as the URL you have requested, except that
+"&*" is appended. This prevents Internet Explorer from seeing a bogus file
+extension.
+</p>
+</body>
+</html>
+HTML;
+ echo "\n";
+ return true;
+ }
+
/**
* Returns true if the PATH_INFO ends with an extension other than a script
* extension. This could confuse IE for scripts that send arbitrary data which
* Also checks for anything that looks like a file extension at the end of
* QUERY_STRING, since IE 6 and earlier will use this to get the file type
* if there was no dot before the question mark (bug 28235).
- */
- public function isPathInfoBad() {
- global $wgScriptExtension;
-
- if ( $this->isQueryStringBad() ) {
- return true;
- }
-
- if ( !isset( $_SERVER['PATH_INFO'] ) ) {
- return false;
- }
- $pi = $_SERVER['PATH_INFO'];
- $dotPos = strrpos( $pi, '.' );
- if ( $dotPos === false ) {
- return false;
- }
- $ext = substr( $pi, $dotPos );
- return !in_array( $ext, array( $wgScriptExtension, '.php', '.php5' ) );
- }
-
- /**
- * Determine what extension IE6 will infer from a certain query string.
- * If the URL has an extension before the question mark, IE6 will use
- * that and ignore the query string, but per the comment at
- * isPathInfoBad() we don't have a reliable way to determine the URL,
- * so isPathInfoBad() just passes in the query string for $url.
- * All entry points have safe extensions (php, php5) anyway, so
- * checking the query string is possibly overly paranoid but never
- * insecure.
*
- * The criteria for finding an extension are as follows:
- * - a possible extension is a dot followed by one or more characters not
- * in <>\"/:|?.#
- * - if we find a possible extension followed by the end of the string or
- * a #, that's our extension
- * - if we find a possible extension followed by a ?, that's our extension
- * - UNLESS it's exe, dll or cgi, in which case we ignore it and continue
- * searching for another possible extension
- * - if we find a possible extension followed by a dot or another illegal
- * character, we ignore it and continue searching
- *
- * @param $url string URL
- * @return mixed Detected extension (string), or false if none found
- */
- public static function findIE6Extension( $url ) {
- $pos = 0;
- $hashPos = strpos( $url, '#' );
- if ( $hashPos !== false ) {
- $urlLength = $hashPos;
- } else {
- $urlLength = strlen( $url );
- }
- $remainingLength = $urlLength;
- while ( $remainingLength > 0 ) {
- // Skip ahead to the next dot
- $pos += strcspn( $url, '.', $pos, $remainingLength );
- if ( $pos >= $urlLength ) {
- // End of string, we're done
- return false;
- }
-
- // We found a dot. Skip past it
- $pos++;
- $remainingLength = $urlLength - $pos;
-
- // Check for illegal characters in our prospective extension,
- // or for another dot
- $nextPos = $pos + strcspn( $url, "<>\\\"/:|?*.", $pos, $remainingLength );
- if ( $nextPos >= $urlLength ) {
- // No illegal character or next dot
- // We have our extension
- return substr( $url, $pos, $urlLength - $pos );
- }
- if ( $url[$nextPos] === '?' ) {
- // We've found a legal extension followed by a question mark
- // If the extension is NOT exe, dll or cgi, return it
- $extension = substr( $url, $pos, $nextPos - $pos );
- if ( strcasecmp( $extension, 'exe' ) && strcasecmp( $extension, 'dll' ) &&
- strcasecmp( $extension, 'cgi' ) )
- {
- return $extension;
- }
- // Else continue looking
- }
- // We found an illegal character or another dot
- // Skip to that character and continue the loop
- $pos = $nextPos + 1;
- $remainingLength = $urlLength - $pos;
- }
- return false;
- }
-
- /**
- * Check for a bad query string, which IE 6 will use as a potentially
- * insecure cache file extension. See bug 28235. Returns true if the
- * request should be disallowed.
- *
- * @return Boolean
+ * @deprecated Use checkUrlExtension().
+ *
+ * @param $extWhitelist array
+ *
+ * @return bool
*/
- public function isQueryStringBad() {
- if ( !isset( $_SERVER['QUERY_STRING'] ) ) {
- return false;
- }
-
- $extension = self::findIE6Extension( $_SERVER['QUERY_STRING'] );
- if ( strval( $extension ) === '' ) {
- /* No extension or empty extension (false/'') */
- return false;
- }
-
- /* Only consider the extension understood by IE to be potentially
- * dangerous if it is made of normal characters (so it is more
- * likely to be registered with an application)
- * Compromise with api.php convenience. Considers for instance
- * that no sane application will register a dangerous file type
- * in an extension containing an ampersand.
- */
- return (bool)preg_match( '/^[a-zA-Z0-9_-]+$/', $extension );
+ public function isPathInfoBad( $extWhitelist = array() ) {
+ global $wgScriptExtension;
+ $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
+ return IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist );
}
/**
foreach ( $langs as $lang => $val ) {
if ( $val === '' ) {
$langs[$lang] = 1;
- } else if ( $val == 0 ) {
+ } elseif ( $val == 0 ) {
unset($langs[$lang]);
}
}
arsort( $langs, SORT_NUMERIC );
return $langs;
}
+
+ /**
+ * Fetch the raw IP from the request
+ *
+ * @return String
+ */
+ protected function getRawIP() {
+ if ( isset( $_SERVER['REMOTE_ADDR'] ) ) {
+ return IP::canonicalize( $_SERVER['REMOTE_ADDR'] );
+ } else {
+ return null;
+ }
+ }
+
+ /**
+ * Work out the IP address based on various globals
+ * For trusted proxies, use the XFF client IP (first of the chain)
+ * @return string
+ */
+ public function getIP() {
+ global $wgUsePrivateIPs;
+
+ # Return cached result
+ if ( $this->ip !== null ) {
+ return $this->ip;
+ }
+
+ # collect the originating ips
+ $ip = $this->getRawIP();
+
+ # Append XFF
+ $forwardedFor = $this->getHeader( 'X-Forwarded-For' );
+ if ( $forwardedFor !== false ) {
+ $ipchain = array_map( 'trim', explode( ',', $forwardedFor ) );
+ $ipchain = array_reverse( $ipchain );
+ if ( $ip ) {
+ array_unshift( $ipchain, $ip );
+ }
+
+ # Step through XFF list and find the last address in the list which is a trusted server
+ # Set $ip to the IP address given by that trusted server, unless the address is not sensible (e.g. private)
+ foreach ( $ipchain as $i => $curIP ) {
+ $curIP = IP::canonicalize( $curIP );
+ if ( wfIsTrustedProxy( $curIP ) ) {
+ if ( isset( $ipchain[$i + 1] ) ) {
+ if ( $wgUsePrivateIPs || IP::isPublic( $ipchain[$i + 1 ] ) ) {
+ $ip = $ipchain[$i + 1];
+ }
+ }
+ } else {
+ break;
+ }
+ }
+ }
+
+ # Allow extensions to improve our guess
+ wfRunHooks( 'GetIP', array( &$ip ) );
+
+ if ( !$ip ) {
+ throw new MWException( "Unable to determine IP" );
+ }
+
+ wfDebug( "IP: $ip\n" );
+ $this->ip = $ip;
+ return $ip;
+ }
}
/**
throw new MWException( "{$method}() not implemented" );
}
+ /**
+ * @param $name string
+ * @param $default string
+ * @return string
+ */
public function getText( $name, $default = '' ) {
# Override; don't recode since we're using internal data
return (string)$this->getVal( $name, $default );
}
+ /**
+ * @return Array
+ */
public function getValues() {
return $this->data;
}
+ /**
+ * @return array
+ */
public function getQueryValues() {
if ( $this->wasPosted ) {
return array();
}
}
+ /**
+ * @return bool
+ */
public function wasPosted() {
return $this->wasPosted;
}
$this->notImplemented( __METHOD__ );
}
+ /**
+ * @param $name
+ * @return bool|string
+ */
public function getHeader( $name ) {
return isset( $this->headers[$name] ) ? $this->headers[$name] : false;
}
+ /**
+ * @param $name string
+ * @param $val string
+ */
public function setHeader( $name, $val ) {
$this->headers[$name] = $val;
}
return $this->session;
}
- public function isPathInfoBad() {
+ /**
+ * @param array $extWhitelist
+ * @return bool
+ */
+ public function isPathInfoBad( $extWhitelist = array() ) {
return false;
}
+
+ /**
+ * @param array $extWhitelist
+ * @return bool
+ */
+ public function checkUrlExtension( $extWhitelist = array() ) {
+ return true;
+ }
+
+ /**
+ * @return string
+ */
+ protected function getRawIP() {
+ return '127.0.0.1';
+ }
}