w/s
[lhc/web/wiklou.git] / includes / WebRequest.php
index 1c2e22e..39f9cb8 100644 (file)
@@ -44,8 +44,14 @@ class WebRequest {
         */
        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
+               /// @todo FIXME: This preemptive de-quoting can interfere with other web libraries
                ///        and increases our memory footprint. It would be cleaner to do on
                ///        demand; but currently we have no wrapper for $_SERVER etc.
                $this->checkMagicQuotes();
@@ -55,6 +61,143 @@ class WebRequest {
                $this->data = $_POST + $_GET;
        }
 
+       /**
+        * Extract relevant query arguments from the http request uri's path
+        * to be merged with the normal php provided query arguments.
+        * Tries to use the REQUEST_URI data if available and parses it
+        * according to the wiki's configuration looking for any known pattern.
+        *
+        * If the REQUEST_URI is not provided we'll fall back on the PATH_INFO
+        * provided by the server if any and use that to set a 'title' parameter.
+        *
+        * @param $want string: If this is not 'all', then the function
+        * will return an empty array if it determines that the URL is
+        * inside a rewrite path.
+        *
+        * @return Array: Any query arguments found in path matches.
+        */
+       static public function getPathInfo( $want = 'all' ) {
+               // PATH_INFO is mangled due to http://bugs.php.net/bug.php?id=31892
+               // And also by Apache 2.x, double slashes are converted to single slashes.
+               // So we will use REQUEST_URI if possible.
+               $matches = array();
+               if ( !empty( $_SERVER['REQUEST_URI'] ) ) {
+                       // Slurp out the path portion to examine...
+                       $url = $_SERVER['REQUEST_URI'];
+                       if ( !preg_match( '!^https?://!', $url ) ) {
+                               $url = 'http://unused' . $url;
+                       }
+                       $a = parse_url( $url );
+                       if( $a ) {
+                               $path = isset( $a['path'] ) ? $a['path'] : '';
+
+                               global $wgScript;
+                               if( $path == $wgScript && $want !== 'all' ) {
+                                       // Script inside a rewrite path?
+                                       // Abort to keep from breaking...
+                                       return $matches;
+                               }
+
+                               $router = new PathRouter;
+
+                               // Raw PATH_INFO style
+                               $router->add( "$wgScript/$1" );
+
+                               if( isset( $_SERVER['SCRIPT_NAME'] )
+                                       && preg_match( '/\.php5?/', $_SERVER['SCRIPT_NAME'] ) )
+                               {
+                                       # Check for SCRIPT_NAME, we handle index.php explicitly
+                                       # But we do have some other .php files such as img_auth.php
+                                       # Don't let root article paths clober the parsing for them
+                                       $router->add( $_SERVER['SCRIPT_NAME'] . "/$1" );
+                               }
+
+                               global $wgArticlePath;
+                               if( $wgArticlePath ) {
+                                       $router->add( $wgArticlePath );
+                               }
+
+                               global $wgActionPaths;
+                               if( $wgActionPaths ) {
+                                       $router->add( $wgActionPaths, array( 'action' => '$key' ) );
+                               }
+
+                               global $wgVariantArticlePath, $wgContLang;
+                               if( $wgVariantArticlePath ) {
+                                       $router->add( $wgVariantArticlePath,
+                                               array( 'variant' => '$2'),
+                                               array( '$2' => $wgContLang->getVariants() )
+                                       );
+                               }
+
+                               wfRunHooks( 'WebRequestPathInfoRouter', array( $router ) );
+
+                               $matches = $router->parse( $path );
+                       }
+               } elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
+                       // Mangled PATH_INFO
+                       // http://bugs.php.net/bug.php?id=31892
+                       // Also reported when ini_get('cgi.fix_pathinfo')==false
+                       $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
+
+               } elseif ( isset( $_SERVER['PATH_INFO'] ) && ($_SERVER['PATH_INFO'] != '') ) {
+                       // Regular old PATH_INFO yay
+                       $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
+               }
+
+               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.
@@ -71,60 +214,7 @@ class WebRequest {
                }
 
                if ( $wgUsePathInfo ) {
-                       // PATH_INFO is mangled due to http://bugs.php.net/bug.php?id=31892
-                       // And also by Apache 2.x, double slashes are converted to single slashes.
-                       // So we will use REQUEST_URI if possible.
-                       $matches = array();
-
-                       if ( !empty( $_SERVER['REQUEST_URI'] ) ) {
-                               // Slurp out the path portion to examine...
-                               $url = $_SERVER['REQUEST_URI'];
-                               if ( !preg_match( '!^https?://!', $url ) ) {
-                                       $url = 'http://unused' . $url;
-                               }
-                               $a = parse_url( $url );
-                               if( $a ) {
-                                       $path = isset( $a['path'] ) ? $a['path'] : '';
-
-                                       global $wgScript;
-                                       if( $path == $wgScript ) {
-                                               // Script inside a rewrite path?
-                                               // Abort to keep from breaking...
-                                               return;
-                                       }
-                                       // Raw PATH_INFO style
-                                       $matches = $this->extractTitle( $path, "$wgScript/$1" );
-
-                                       global $wgArticlePath;
-                                       if( !$matches && $wgArticlePath ) {
-                                               $matches = $this->extractTitle( $path, $wgArticlePath );
-                                       }
-
-                                       global $wgActionPaths;
-                                       if( !$matches && $wgActionPaths ) {
-                                               $matches = $this->extractTitle( $path, $wgActionPaths, 'action' );
-                                       }
-
-                                       global $wgVariantArticlePath, $wgContLang;
-                                       if( !$matches && $wgVariantArticlePath ) {
-                                               $variantPaths = array();
-                                               foreach( $wgContLang->getVariants() as $variant ) {
-                                                       $variantPaths[$variant] =
-                                                               str_replace( '$2', $variant, $wgVariantArticlePath );
-                                               }
-                                               $matches = $this->extractTitle( $path, $variantPaths, 'variant' );
-                                       }
-                               }
-                       } elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
-                               // Mangled PATH_INFO
-                               // http://bugs.php.net/bug.php?id=31892
-                               // Also reported when ini_get('cgi.fix_pathinfo')==false
-                               $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
-
-                       } elseif ( isset( $_SERVER['PATH_INFO'] ) && ($_SERVER['PATH_INFO'] != '') ) {
-                               // Regular old PATH_INFO yay
-                               $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
-                       }
+                       $matches = self::getPathInfo( 'title' );
                        foreach( $matches as $key => $val) {
                                $this->data[$key] = $_GET[$key] = $_REQUEST[$key] = $val;
                        }
@@ -132,7 +222,7 @@ class WebRequest {
        }
 
        /**
-        * 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
@@ -141,7 +231,7 @@ class WebRequest {
         *             passed on as the value of this URL parameter
         * @return array of URL variables to interpolate; empty if no match
         */
-       private 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 );
@@ -165,16 +255,24 @@ class WebRequest {
         * used for undoing the evil that is magic_quotes_gpc.
         *
         * @param $arr array: will be modified
+        * @param $topLevel bool Specifies if the array passed is from the top
+        * level of the source. In PHP5 magic_quotes only escapes the first level
+        * of keys that belong to an array.
         * @return array the original array
+        * @see http://www.php.net/manual/en/function.get-magic-quotes-gpc.php#49612
         */
-       private function &fix_magic_quotes( &$arr ) {
+       private function &fix_magic_quotes( &$arr, $topLevel = true ) {
+               $clean = array();
                foreach( $arr as $key => $val ) {
                        if( is_array( $val ) ) {
-                               $this->fix_magic_quotes( $arr[$key] );
+                               $cleanKey = $topLevel ? stripslashes( $key ) : $key;
+                               $clean[$cleanKey] = $this->fix_magic_quotes( $arr[$key], false );
                        } else {
-                               $arr[$key] = stripslashes( $val );
+                               $cleanKey = stripslashes( $key );
+                               $clean[$cleanKey] = stripslashes( $val );
                        }
                }
+               $arr = $clean;
                return $arr;
        }
 
@@ -211,7 +309,7 @@ class WebRequest {
                        }
                } else {
                        global $wgContLang;
-                       $data = $wgContLang->normalize( $data );
+                       $data = isset( $wgContLang ) ? $wgContLang->normalize( $data ) : UtfNormal::cleanUp( $data );
                }
                return $data;
        }
@@ -269,7 +367,7 @@ class WebRequest {
        }
 
        /**
-        * 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
@@ -357,7 +455,7 @@ class WebRequest {
        public function getBool( $name, $default = false ) {
                return (bool)$this->getVal( $name, $default );
        }
-       
+
        /**
         * Fetch a boolean value from the input or return $default if not set.
         * Unlike getBool, the string "false" will result in boolean false, which is
@@ -409,6 +507,8 @@ class WebRequest {
         * 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();
@@ -426,6 +526,26 @@ class WebRequest {
                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 Array
+        */
+        public function getQueryValues() {
+               return $_GET;
+        }
+
        /**
         * Returns true if the present request was reached by a POST operation,
         * false otherwise (GET, HEAD, or command-line).
@@ -436,7 +556,7 @@ class WebRequest {
         * @return Boolean
         */
        public function wasPosted() {
-               return $_SERVER['REQUEST_METHOD'] == 'POST';
+               return isset( $_SERVER['REQUEST_METHOD'] ) && $_SERVER['REQUEST_METHOD'] == 'POST';
        }
 
        /**
@@ -471,15 +591,18 @@ class WebRequest {
        }
 
        /**
-        * 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
         */
        public function getRequestURL() {
-               if( isset( $_SERVER['REQUEST_URI']) && strlen($_SERVER['REQUEST_URI']) ) {
+               if( isset( $_SERVER['REQUEST_URI'] ) && strlen( $_SERVER['REQUEST_URI'] ) ) {
                        $base = $_SERVER['REQUEST_URI'];
-               } elseif( isset( $_SERVER['SCRIPT_NAME'] ) ) {
+               } elseif ( isset( $_SERVER['HTTP_X_ORIGINAL_URL'] ) && strlen( $_SERVER['HTTP_X_ORIGINAL_URL'] ) ) {
                        // Probably IIS; doesn't set REQUEST_URI
+                       $base = $_SERVER['HTTP_X_ORIGINAL_URL'];
+               } elseif( isset( $_SERVER['SCRIPT_NAME'] ) ) {
                        $base = $_SERVER['SCRIPT_NAME'];
                        if( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
                                $base .= '?' . $_SERVER['QUERY_STRING'];
@@ -487,8 +610,8 @@ class WebRequest {
                } else {
                        // This shouldn't happen!
                        throw new MWException( "Web server doesn't provide either " .
-                               "REQUEST_URI or SCRIPT_NAME. Report details of your " .
-                               "web server configuration to http://bugzilla.wikimedia.org/" );
+                               "REQUEST_URI, HTTP_X_ORIGINAL_URL or SCRIPT_NAME. Report details " .
+                               "of your web server configuration to http://bugzilla.wikimedia.org/" );
                }
                // User-agents should not send a fragment with the URI, but
                // if they do, and the web server passes it on to us, we
@@ -498,7 +621,7 @@ class WebRequest {
                if( $hash !== false ) {
                        $base = substr( $base, 0, $hash );
                }
-               if( $base{0} == '/' ) {
+               if( $base[0] == '/' ) {
                        return $base;
                } else {
                        // We may get paths with a host prepended; strip it.
@@ -507,13 +630,17 @@ class WebRequest {
        }
 
        /**
-        * 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 );
        }
 
        /**
@@ -523,25 +650,7 @@ class WebRequest {
         * @return String
         */
        public function appendQuery( $query ) {
-               global $wgTitle;
-               $basequery = '';
-               foreach( $_GET as $var => $val ) {
-                       if ( $var == 'title' ) {
-                               continue;
-                       }
-                       if ( is_array( $val ) ) {
-                               /* This will happen given a request like
-                                * http://en.wikipedia.org/w/index.php?title[]=Special:Userlogin&returnto[]=Main_Page
-                                */
-                               continue;
-                       }
-                       $basequery .= '&' . urlencode( $var ) . '=' . urlencode( $val );
-               }
-               $basequery .= '&' . $query;
-
-               # Trim the extra &
-               $basequery = substr( $basequery, 1 );
-               return $wgTitle->getLocalURL( $basequery );
+               return $this->appendQueryArray( wfCgiToArray( $query ) );
        }
 
        /**
@@ -554,6 +663,12 @@ class WebRequest {
                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 );
        }
@@ -568,7 +683,7 @@ class WebRequest {
         */
        public function appendQueryArray( $array, $onlyquery = false ) {
                global $wgTitle;
-               $newquery = $_GET;
+               $newquery = $this->getQueryValues();
                unset( $newquery['title'] );
                $newquery = array_merge( $newquery, $array );
                $query = wfArrayToCGI( $newquery );
@@ -623,11 +738,12 @@ class WebRequest {
        /**
         * Return the size of the upload, or 0.
         *
-        * @deprecated
+        * @deprecated since 1.17
         * @param $key String:
         * @return integer
         */
        public function getFileSize( $key ) {
+               wfDeprecated( __METHOD__, '1.17' );
                $file = new WebRequestUpload( $this, $key );
                return $file->getSize();
        }
@@ -662,7 +778,7 @@ class WebRequest {
        /**
         * Return a WebRequestUpload object corresponding to the key
         *
-        * @param @key string
+        * @param $key string
         * @return WebRequestUpload
         */
        public function getUpload( $key ) {
@@ -684,33 +800,53 @@ class WebRequest {
                return $this->response;
        }
 
+       /**
+        * Initialise the header list
+        */
+       private function initHeaders() {
+               if ( count( $this->headers ) ) {
+                       return;
+               }
+
+               if ( function_exists( 'apache_request_headers' ) ) {
+                       foreach ( apache_request_headers() as $tempName => $tempValue ) {
+                               $this->headers[ strtoupper( $tempName ) ] = $tempValue;
+                       }
+               } else {
+                       foreach ( $_SERVER as $name => $value ) {
+                               if ( substr( $name, 0, 5 ) === 'HTTP_' ) {
+                                       $name = str_replace( '_', '-',  substr( $name, 5 ) );
+                                       $this->headers[$name] = $value;
+                               } elseif ( $name === 'CONTENT_LENGTH' ) {
+                                       $this->headers['CONTENT-LENGTH'] = $value;
+                               }
+                       }
+               }
+       }
+
+       /**
+        * Get an array containing all request headers
+        *
+        * @return Array mapping header name to its value
+        */
+       public function getAllHeaders() {
+               $this->initHeaders();
+               return $this->headers;
+       }
+
        /**
         * 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();
                $name = strtoupper( $name );
-               if ( function_exists( 'apache_request_headers' ) ) {
-                       if ( !$this->headers ) {
-                               foreach ( apache_request_headers() as $tempName => $tempValue ) {
-                                       $this->headers[ strtoupper( $tempName ) ] = $tempValue;
-                               }
-                       }
-                       if ( isset( $this->headers[$name] ) ) {
-                               return $this->headers[$name];
-                       } else {
-                               return false;
-                       }
+               if ( isset( $this->headers[$name] ) ) {
+                       return $this->headers[$name];
                } else {
-                       $name = 'HTTP_' . str_replace( '-', '_', $name );
-                       if ( $name === 'HTTP_CONTENT_LENGTH' && !isset( $_SERVER[$name] ) ) {
-                               $name = 'CONTENT_LENGTH';
-                       }
-                       if ( isset( $_SERVER[$name] ) ) {
-                               return $_SERVER[$name];
-                       } else {
-                               return false;
-                       }
+                       return false;
                }
        }
 
@@ -737,6 +873,66 @@ class WebRequest {
                $_SESSION[$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
+"&amp;*" 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
@@ -748,25 +944,27 @@ class WebRequest {
         * but only by prefixing it with the script name and maybe some other stuff,
         * the extension is not mangled. So this should be a reasonably portable
         * way to perform this security check.
+        *
+        * 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).
+        *
+        * @deprecated Use checkUrlExtension().
+        *
+        * @param $extWhitelist array
+        *
+        * @return bool
         */
-       public function isPathInfoBad() {
+       public function isPathInfoBad( $extWhitelist = array() ) {
+               wfDeprecated( __METHOD__, '1.17' );
                global $wgScriptExtension;
-
-               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' ) );
+               $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
+               return IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist );
        }
 
        /**
         * Parse the Accept-Language header sent by the client into an array
-        * @return array( languageCode => q-value ) sorted by q-value in descending order
+        * @return array array( languageCode => q-value ) sorted by q-value in descending order
         * May contain the "language" '*', which applies to languages other than those explicitly listed.
         * This is aligned with rfc2616 section 14.4
         */
@@ -782,7 +980,7 @@ class WebRequest {
 
                // Break up string into pieces (languages and q factors)
                $lang_parse = null;
-               preg_match_all( '/([a-z]{1,8}(-[a-z]{1,8})?|\*)\s*(;\s*q\s*=\s*(1|0(\.[0-9]+)?)?)?/',
+               preg_match_all( '/([a-z]{1,8}(-[a-z]{1,8})*|\*)\s*(;\s*q\s*=\s*(1(\.0{0,3})?|0(\.[0-9]{0,3})?)?)?/',
                        $acceptLang, $lang_parse );
 
                if ( !count( $lang_parse[1] ) ) {
@@ -795,7 +993,7 @@ class WebRequest {
                foreach ( $langs as $lang => $val ) {
                        if ( $val === '' ) {
                                $langs[$lang] = 1;
-                       } else if ( $val == 0 ) {
+                       } elseif ( $val == 0 ) {
                                unset($langs[$lang]);
                        }
                }
@@ -804,6 +1002,77 @@ class WebRequest {
                arsort( $langs, SORT_NUMERIC );
                return $langs;
        }
+
+       /**
+        * Fetch the raw IP from the request
+        *
+        * @since 1.19
+        *
+        * @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)
+        * 
+        * @since 1.19
+        *
+        * @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;
+       }
 }
 
 /**
@@ -935,7 +1204,7 @@ class FauxRequest extends WebRequest {
         * @param $wasPosted Bool: whether to treat the data as POST
         * @param $session Mixed: session array or null
         */
-       public function __construct( $data, $wasPosted = false, $session = null ) {
+       public function __construct( $data = array(), $wasPosted = false, $session = null ) {
                if( is_array( $data ) ) {
                        $this->data = $data;
                } else {
@@ -946,19 +1215,45 @@ class FauxRequest extends WebRequest {
                        $this->session = $session;
        }
 
+       /**
+        * @param $method string
+        * @throws MWException
+        */
        private function notImplemented( $method ) {
                throw new MWException( "{$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();
+               } else {
+                       return $this->data;
+               }
+       }
+
+       /**
+        * @return bool
+        */
        public function wasPosted() {
                return $this->wasPosted;
        }
@@ -971,46 +1266,114 @@ class FauxRequest extends WebRequest {
                $this->notImplemented( __METHOD__ );
        }
 
-       public function appendQuery( $query ) {
-               global $wgTitle;
-               $basequery = '';
-               foreach( $this->data as $var => $val ) {
-                       if ( $var == 'title' ) {
-                               continue;
-                       }
-                       if ( is_array( $val ) ) {
-                               /* This will happen given a request like
-                                * http://en.wikipedia.org/w/index.php?title[]=Special:Userlogin&returnto[]=Main_Page
-                                */
-                               continue;
-                       }
-                       $basequery .= '&' . urlencode( $var ) . '=' . urlencode( $val );
-               }
-               $basequery .= '&' . $query;
-
-               # Trim the extra &
-               $basequery = substr( $basequery, 1 );
-               return $wgTitle->getLocalURL( $basequery );
-       }
-
+       /**
+        * @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;
        }
 
+       /**
+        * @param $key
+        * @return mixed
+        */
        public function getSessionData( $key ) {
                if( isset( $this->session[$key] ) )
                        return $this->session[$key];
        }
 
+       /**
+        * @param $key
+        * @param $data
+        */
        public function setSessionData( $key, $data ) {
                $this->session[$key] = $data;
        }
 
-       public function isPathInfoBad() {
+       /**
+        * @return array|Mixed|null
+        */
+       public function getSessionArray() {
+               return $this->session;
+       }
+
+       /**
+        * @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';
+       }
+}
+
+/**
+ * Similar to FauxRequest, but only fakes URL parameters and method
+ * (POST or GET) and use the base request for the remaining stuff
+ * (cookies, session and headers).
+ *
+ * @ingroup HTTP
+ */
+class DerivativeRequest extends FauxRequest {
+       private $base;
+
+       public function __construct( WebRequest $base, $data, $wasPosted = false ) {
+               $this->base = $base;
+               parent::__construct( $data, $wasPosted );
+       }
+
+       public function getCookie( $key, $prefix = null, $default = null ) {
+               return $this->base->getCookie( $key, $prefix, $default );
+       }
+
+       public function checkSessionCookie() {
+               return $this->base->checkSessionCookie();
+       }
+
+       public function getHeader( $name ) {
+               return $this->base->getHeader( $name );
+       }
+
+       public function getAllHeaders() {
+               return $this->base->getAllHeaders();
+       }
+
+       public function getSessionData( $key ) {
+               return $this->base->getSessionData( $key );
+       }
+
+       public function setSessionData( $key, $data ) {
+               return $this->base->setSessionData( $key, $data );
+       }
+
+       public function getAcceptLang() {
+               return $this->base->getAcceptLang();
+       }
+
+       public function getIP() {
+               return $this->base->getIP();
+       }
 }