<?php
/**
- * Deal with importing all those nasssty globals and things
+ * Deal with importing all those nasty globals and things
*
* Copyright © 2003 Brion Vibber <brion@pobox.com>
- * http://www.mediawiki.org/
+ * https://www.mediawiki.org/
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* @file
*/
+use MediaWiki\MediaWikiServices;
+use MediaWiki\Session\Session;
+use MediaWiki\Session\SessionId;
+use MediaWiki\Session\SessionManager;
+
/**
* The WebRequest class encapsulates getting at data passed in the
- * URL or via a POSTed form, handling remove of "magic quotes" slashes,
- * stripping illegal input characters and normalizing Unicode sequences.
- *
- * Usually this is used via a global singleton, $wgRequest. You should
- * not create a second WebRequest object; make a FauxRequest object if
- * you want to pass arbitrary data to some function in place of the web
- * input.
+ * URL or via a POSTed form stripping illegal input characters and
+ * normalizing Unicode sequences.
*
* @ingroup HTTP
*/
class WebRequest {
- protected $data, $headers = array();
+ protected $data, $headers = [];
+
+ /**
+ * Flag to make WebRequest::getHeader return an array of values.
+ * @since 1.26
+ */
+ const GETHEADER_LIST = 1;
+
+ /**
+ * The unique request ID.
+ * @var string
+ */
+ private static $reqId;
/**
* Lazy-init response object
*/
private $response;
+ /**
+ * Cached client IP address
+ * @var string
+ */
+ private $ip;
+
+ /**
+ * The timestamp of the start of the request, with microsecond precision.
+ * @var float
+ */
+ protected $requestTime;
+
+ /**
+ * Cached URL protocol
+ * @var string
+ */
+ protected $protocol;
+
+ /**
+ * @var SessionId|null Session ID to use for this
+ * request. We can't save the session directly due to reference cycles not
+ * working too well (slow GC in Zend and never collected in HHVM).
+ */
+ protected $sessionId = null;
+
+ /** @var bool Whether this HTTP request is "safe" (even if it is an HTTP post) */
+ protected $markedAsSafe = false;
+
+ /**
+ * @codeCoverageIgnore
+ */
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
- /// demand; but currently we have no wrapper for $_SERVER etc.
- $this->checkMagicQuotes();
+ $this->requestTime = isset( $_SERVER['REQUEST_TIME_FLOAT'] )
+ ? $_SERVER['REQUEST_TIME_FLOAT'] : microtime( true );
// POST overrides GET data
// We don't use $_REQUEST here to avoid interference from cookies...
}
/**
- * Check for title, action, and/or variant data in the URL
- * and interpolate it into the GET variables.
- * This should only be run after $wgContLang is available,
- * as we may need the list of language variants to determine
- * available variant URLs.
+ * 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 string $want 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.
*/
- public function interpolateTitle() {
+ public static function getPathInfo( $want = 'all' ) {
global $wgUsePathInfo;
+ // PATH_INFO is mangled due to https://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 = [];
+ 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;
+ }
+ MediaWiki\suppressWarnings();
+ $a = parse_url( $url );
+ MediaWiki\restoreWarnings();
+ 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;
+ }
- // bug 16019: title interpolation on API queries is useless and possible harmful
- if ( defined( 'MW_API' ) ) {
- return;
- }
+ $router = new PathRouter;
+
+ // Raw PATH_INFO style
+ $router->add( "$wgScript/$1" );
- 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;
+ 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" );
}
- $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 $wgArticlePath;
+ if ( $wgArticlePath ) {
+ $router->add( $wgArticlePath );
+ }
- global $wgActionPaths;
- if( !$matches && $wgActionPaths ) {
- $matches = $this->extractTitle( $path, $wgActionPaths, 'action' );
- }
+ global $wgActionPaths;
+ if ( $wgActionPaths ) {
+ $router->add( $wgActionPaths, [ 'action' => '$key' ] );
+ }
- 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' );
- }
+ global $wgVariantArticlePath, $wgContLang;
+ if ( $wgVariantArticlePath ) {
+ $router->add( $wgVariantArticlePath,
+ [ 'variant' => '$2' ],
+ [ '$2' => $wgContLang->getVariants() ]
+ );
}
- } elseif ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
+
+ Hooks::run( 'WebRequestPathInfoRouter', [ $router ] );
+
+ $matches = $router->parse( $path );
+ }
+ } elseif ( $wgUsePathInfo ) {
+ if ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
// Mangled PATH_INFO
- // http://bugs.php.net/bug.php?id=31892
+ // https://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'] != '') ) {
+ } elseif ( isset( $_SERVER['PATH_INFO'] ) && $_SERVER['PATH_INFO'] != '' ) {
// Regular old PATH_INFO yay
$matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
}
- foreach( $matches as $key => $val) {
- $this->data[$key] = $_GET[$key] = $_REQUEST[$key] = $val;
- }
}
+
+ return $matches;
}
/**
- * Internal URL rewriting function; tries to extract page title and,
- * optionally, one other fixed parameter value from a URL path.
+ * Work out an appropriate URL prefix containing scheme and host, based on
+ * information detected from $_SERVER
*
- * @param $path string: the URL path given from the client
- * @param $bases array: one or more URLs, optionally with $1 at the end
- * @param $key string: if provided, the matching key in $bases will be
- * passed on as the value of this URL parameter
- * @return array of URL variables to interpolate; empty if no match
+ * @return string
*/
- private function extractTitle( $path, $bases, $key=false ) {
- foreach( (array)$bases as $keyValue => $base ) {
- // Find the part after $wgArticlePath
- $base = str_replace( '$1', '', $base );
- $baseLen = strlen( $base );
- if( substr( $path, 0, $baseLen ) == $base ) {
- $raw = substr( $path, $baseLen );
- if( $raw !== '' ) {
- $matches = array( 'title' => rawurldecode( $raw ) );
- if( $key ) {
- $matches[$key] = $keyValue;
- }
- return $matches;
- }
+ public static function detectServer() {
+ global $wgAssumeProxiesUseDefaultProtocolPorts;
+
+ $proto = self::detectProtocol();
+ $stdPort = $proto === 'https' ? 443 : 80;
+
+ $varNames = [ '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 ( $wgAssumeProxiesUseDefaultProtocolPorts && isset( $_SERVER['HTTP_X_FORWARDED_PROTO'] ) ) {
+ // T72021: Assume that upstream proxy is running on the default
+ // port based on the protocol. We have no reliable way to determine
+ // the actual port in use upstream.
+ $port = $stdPort;
+ } elseif ( $parts[1] === false ) {
+ if ( isset( $_SERVER['SERVER_PORT'] ) ) {
+ $port = $_SERVER['SERVER_PORT'];
+ } // else leave it as $stdPort
+ } else {
+ $port = $parts[1];
}
+ break;
}
- return array();
+
+ return $proto . '://' . IP::combineHostAndPort( $host, $port, $stdPort );
}
/**
- * Recursively strips slashes from the given array;
- * used for undoing the evil that is magic_quotes_gpc.
+ * Detect the protocol from $_SERVER.
+ * This is for use prior to Setup.php, when no WebRequest object is available.
+ * At other times, use the non-static function getProtocol().
*
- * @param $arr array: will be modified
- * @return array the original array
+ * @return string
*/
- private function &fix_magic_quotes( &$arr ) {
- foreach( $arr as $key => $val ) {
- if( is_array( $val ) ) {
- $this->fix_magic_quotes( $arr[$key] );
- } else {
- $arr[$key] = stripslashes( $val );
- }
+ public static function detectProtocol() {
+ if ( ( !empty( $_SERVER['HTTPS'] ) && $_SERVER['HTTPS'] !== 'off' ) ||
+ ( isset( $_SERVER['HTTP_X_FORWARDED_PROTO'] ) &&
+ $_SERVER['HTTP_X_FORWARDED_PROTO'] === 'https' ) ) {
+ return 'https';
+ } else {
+ return 'http';
}
- return $arr;
}
/**
- * If magic_quotes_gpc option is on, run the global arrays
- * through fix_magic_quotes to strip out the stupid slashes.
- * WARNING: This should only be done once! Running a second
- * time could damage the values.
- */
- private function checkMagicQuotes() {
- $mustFixQuotes = function_exists( 'get_magic_quotes_gpc' )
- && get_magic_quotes_gpc();
- if( $mustFixQuotes ) {
- $this->fix_magic_quotes( $_COOKIE );
- $this->fix_magic_quotes( $_ENV );
- $this->fix_magic_quotes( $_GET );
- $this->fix_magic_quotes( $_POST );
- $this->fix_magic_quotes( $_REQUEST );
- $this->fix_magic_quotes( $_SERVER );
+ * Get the number of seconds to have elapsed since request start,
+ * in fractional seconds, with microsecond resolution.
+ *
+ * @return float
+ * @since 1.25
+ */
+ public function getElapsedTime() {
+ return microtime( true ) - $this->requestTime;
+ }
+
+ /**
+ * Get the unique request ID.
+ * This is either the value of the UNIQUE_ID envvar (if present) or a
+ * randomly-generated 24-character string.
+ *
+ * @return string
+ * @since 1.27
+ */
+ public static function getRequestId() {
+ if ( !self::$reqId ) {
+ self::$reqId = isset( $_SERVER['UNIQUE_ID'] )
+ ? $_SERVER['UNIQUE_ID'] : wfRandomString( 24 );
}
+
+ return self::$reqId;
+ }
+
+ /**
+ * Override the unique request ID. This is for sub-requests, such as jobs,
+ * that wish to use the same id but are not part of the same execution context.
+ *
+ * @param string $id
+ * @since 1.27
+ */
+ public static function overrideRequestId( $id ) {
+ self::$reqId = $id;
+ }
+
+ /**
+ * Get the current URL protocol (http or https)
+ * @return string
+ */
+ public function getProtocol() {
+ if ( $this->protocol === null ) {
+ $this->protocol = self::detectProtocol();
+ }
+ return $this->protocol;
+ }
+
+ /**
+ * Check for title, action, and/or variant data in the URL
+ * and interpolate it into the GET variables.
+ * This should only be run after $wgContLang is available,
+ * as we may need the list of language variants to determine
+ * available variant URLs.
+ */
+ public function interpolateTitle() {
+ // T18019: title interpolation on API queries is useless and sometimes harmful
+ if ( defined( 'MW_API' ) ) {
+ return;
+ }
+
+ $matches = self::getPathInfo( 'title' );
+ foreach ( $matches as $key => $val ) {
+ $this->data[$key] = $_GET[$key] = $_REQUEST[$key] = $val;
+ }
+ }
+
+ /**
+ * URL rewriting function; tries to extract page title and,
+ * optionally, one other fixed parameter value from a URL path.
+ *
+ * @param string $path The URL path given from the client
+ * @param array $bases One or more URLs, optionally with $1 at the end
+ * @param string|bool $key If provided, the matching key in $bases will be
+ * passed on as the value of this URL parameter
+ * @return array Array of URL variables to interpolate; empty if no match
+ */
+ static function extractTitle( $path, $bases, $key = false ) {
+ foreach ( (array)$bases as $keyValue => $base ) {
+ // Find the part after $wgArticlePath
+ $base = str_replace( '$1', '', $base );
+ $baseLen = strlen( $base );
+ if ( substr( $path, 0, $baseLen ) == $base ) {
+ $raw = substr( $path, $baseLen );
+ if ( $raw !== '' ) {
+ $matches = [ 'title' => rawurldecode( $raw ) ];
+ if ( $key ) {
+ $matches[$key] = $keyValue;
+ }
+ return $matches;
+ }
+ }
+ }
+ return [];
}
/**
* Recursively normalizes UTF-8 strings in the given array.
*
- * @param $data string or array
- * @return cleaned-up version of the given
+ * @param string|array $data
+ * @return array|string Cleaned-up version of the given
* @private
*/
- function normalizeUnicode( $data ) {
- if( is_array( $data ) ) {
- foreach( $data as $key => $val ) {
+ public function normalizeUnicode( $data ) {
+ if ( is_array( $data ) ) {
+ foreach ( $data as $key => $val ) {
$data[$key] = $this->normalizeUnicode( $val );
}
} else {
global $wgContLang;
- $data = $wgContLang->normalize( $data );
+ $data = isset( $wgContLang ) ?
+ $wgContLang->normalize( $data ) :
+ UtfNormal\Validator::cleanUp( $data );
}
return $data;
}
/**
* Fetch a value from the given array or return $default if it's not set.
*
- * @param $arr Array
- * @param $name String
- * @param $default Mixed
+ * @param array $arr
+ * @param string $name
+ * @param mixed $default
* @return mixed
*/
private function getGPCVal( $arr, $name, $default ) {
# PHP is so nice to not touch input data, except sometimes:
- # http://us2.php.net/variables.external#language.variables.external.dot-in-names
+ # https://secure.php.net/variables.external#language.variables.external.dot-in-names
# Work around PHP *feature* to avoid *bugs* elsewhere.
$name = strtr( $name, '.', '_' );
- if( isset( $arr[$name] ) ) {
+ if ( isset( $arr[$name] ) ) {
global $wgContLang;
$data = $arr[$name];
- if( isset( $_GET[$name] ) && !is_array( $data ) ) {
+ if ( isset( $_GET[$name] ) && !is_array( $data ) ) {
# Check for alternate/legacy character encoding.
- if( isset( $wgContLang ) ) {
+ if ( isset( $wgContLang ) ) {
$data = $wgContLang->checkTitleEncoding( $data );
}
}
$data = $this->normalizeUnicode( $data );
return $data;
} else {
- taint( $default );
return $default;
}
}
+ /**
+ * Fetch a scalar from the input without normalization, or return $default
+ * if it's not set.
+ *
+ * Unlike self::getVal(), this does not perform any normalization on the
+ * input value.
+ *
+ * @since 1.28
+ * @param string $name
+ * @param string|null $default Optional default
+ * @return string|null
+ */
+ public function getRawVal( $name, $default = null ) {
+ $name = strtr( $name, '.', '_' ); // See comment in self::getGPCVal()
+ if ( isset( $this->data[$name] ) && !is_array( $this->data[$name] ) ) {
+ $val = $this->data[$name];
+ } else {
+ $val = $default;
+ }
+ if ( is_null( $val ) ) {
+ return $val;
+ } else {
+ return (string)$val;
+ }
+ }
+
/**
* Fetch a scalar from the input or return $default if it's not set.
* Returns a string. Arrays are discarded. Useful for
* non-freeform text inputs (e.g. predefined internal text keys
* selected by a drop-down menu). For freeform input, see getText().
*
- * @param $name String
- * @param $default String: optional default (or NULL)
- * @return String
+ * @param string $name
+ * @param string $default Optional default (or null)
+ * @return string|null
*/
public function getVal( $name, $default = null ) {
$val = $this->getGPCVal( $this->data, $name, $default );
- if( is_array( $val ) ) {
+ if ( is_array( $val ) ) {
$val = $default;
}
- if( is_null( $val ) ) {
+ if ( is_null( $val ) ) {
return $val;
} else {
return (string)$val;
}
/**
- * 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
- * @return Mixed: old value if one was present, null otherwise
+ * @param string $key Key name to use
+ * @param mixed $value Value to set
+ * @return mixed Old value if one was present, null otherwise
*/
public function setVal( $key, $value ) {
$ret = isset( $this->data[$key] ) ? $this->data[$key] : null;
return $ret;
}
+ /**
+ * Unset an arbitrary value from our get/post data.
+ *
+ * @param string $key Key name to use
+ * @return mixed Old value if one was present, null otherwise
+ */
+ public function unsetVal( $key ) {
+ if ( !isset( $this->data[$key] ) ) {
+ $ret = null;
+ } else {
+ $ret = $this->data[$key];
+ unset( $this->data[$key] );
+ }
+ return $ret;
+ }
+
/**
* Fetch an array from the input or return $default if it's not set.
* If source was scalar, will return an array with a single element.
- * If no source and no default, returns NULL.
+ * If no source and no default, returns null.
*
- * @param $name String
- * @param $default Array: optional default (or NULL)
- * @return Array
+ * @param string $name
+ * @param array $default Optional default (or null)
+ * @return array|null
*/
public function getArray( $name, $default = null ) {
$val = $this->getGPCVal( $this->data, $name, $default );
- if( is_null( $val ) ) {
+ if ( is_null( $val ) ) {
return null;
} else {
return (array)$val;
/**
* Fetch an array of integers, or return $default if it's not set.
* If source was scalar, will return an array with a single element.
- * If no source and no default, returns NULL.
+ * If no source and no default, returns null.
* If an array is returned, contents are guaranteed to be integers.
*
- * @param $name String
- * @param $default Array: option default (or NULL)
- * @return Array of ints
+ * @param string $name
+ * @param array $default Option default (or null)
+ * @return array Array of ints
*/
public function getIntArray( $name, $default = null ) {
$val = $this->getArray( $name, $default );
- if( is_array( $val ) ) {
+ if ( is_array( $val ) ) {
$val = array_map( 'intval', $val );
}
return $val;
* Guaranteed to return an integer; non-numeric input will typically
* return 0.
*
- * @param $name String
- * @param $default Integer
- * @return Integer
+ * @param string $name
+ * @param int $default
+ * @return int
*/
public function getInt( $name, $default = 0 ) {
- return intval( $this->getVal( $name, $default ) );
+ return intval( $this->getRawVal( $name, $default ) );
}
/**
* Guaranteed to return an integer or null; non-numeric input will
* typically return null.
*
- * @param $name String
- * @return Integer
+ * @param string $name
+ * @return int|null
*/
public function getIntOrNull( $name ) {
- $val = $this->getVal( $name );
+ $val = $this->getRawVal( $name );
return is_numeric( $val )
? intval( $val )
: null;
}
+ /**
+ * Fetch a floating point value from the input or return $default if not set.
+ * Guaranteed to return a float; non-numeric input will typically
+ * return 0.
+ *
+ * @since 1.23
+ * @param string $name
+ * @param float $default
+ * @return float
+ */
+ public function getFloat( $name, $default = 0.0 ) {
+ return floatval( $this->getRawVal( $name, $default ) );
+ }
+
/**
* Fetch a boolean value from the input or return $default if not set.
* Guaranteed to return true or false, with normal PHP semantics for
* boolean interpretation of strings.
*
- * @param $name String
- * @param $default Boolean
- * @return Boolean
+ * @param string $name
+ * @param bool $default
+ * @return bool
*/
public function getBool( $name, $default = false ) {
- return (bool)$this->getVal( $name, $default );
+ return (bool)$this->getRawVal( $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
* useful when interpreting information sent from JavaScript.
*
- * @param $name String
- * @param $default Boolean
- * @return Boolean
+ * @param string $name
+ * @param bool $default
+ * @return bool
*/
public function getFuzzyBool( $name, $default = false ) {
- return $this->getBool( $name, $default ) && strcasecmp( $this->getVal( $name ), 'false' ) !== 0;
+ return $this->getBool( $name, $default )
+ && strcasecmp( $this->getRawVal( $name ), 'false' ) !== 0;
}
/**
* value is (even "0"). Return false if the named value is not set.
* Example use is checking for the presence of check boxes in forms.
*
- * @param $name String
- * @return Boolean
+ * @param string $name
+ * @return bool
*/
public function getCheck( $name ) {
# Checkboxes and buttons are only present when clicked
- # Presence connotes truth, abscense false
- $val = $this->getVal( $name, null );
- return isset( $val );
+ # Presence connotes truth, absence false
+ return $this->getRawVal( $name, null ) !== null;
}
/**
* Fetch a text string from the given array or return $default if it's not
- * set. Carriage returns are stripped from the text, and with some language
- * modules there is an input transliteration applied. This should generally
- * be used for form <textarea> and <input> fields. Used for user-supplied
- * freeform text input (for which input transformations may be required - e.g.
- * Esperanto x-coding).
+ * set. Carriage returns are stripped from the text. This should generally
+ * be used for form "<textarea>" and "<input>" fields, and for
+ * user-supplied freeform text input.
*
- * @param $name String
- * @param $default String: optional
- * @return String
+ * @param string $name
+ * @param string $default Optional
+ * @return string
*/
public function getText( $name, $default = '' ) {
- global $wgContLang;
$val = $this->getVal( $name, $default );
- return str_replace( "\r\n", "\n",
- $wgContLang->recodeInput( $val ) );
+ return str_replace( "\r\n", "\n", $val );
}
/**
* 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();
$names = array_keys( $this->data );
}
- $retVal = array();
+ $retVal = [];
foreach ( $names as $name ) {
- $value = $this->getVal( $name );
+ $value = $this->getGPCVal( $this->data, $name, null );
if ( !is_null( $value ) ) {
$retVal[$name] = $value;
}
return $retVal;
}
+ /**
+ * Returns the names of all input values excluding those in $exclude.
+ *
+ * @param array $exclude
+ * @return array
+ */
+ public function getValueNames( $exclude = [] ) {
+ return array_diff( array_keys( $this->getValues() ), $exclude );
+ }
+
+ /**
+ * Get the values passed in the query string.
+ * No transformation is performed on the values.
+ *
+ * @codeCoverageIgnore
+ * @return array
+ */
+ public function getQueryValues() {
+ return $_GET;
+ }
+
+ /**
+ * Return the contents of the Query with no decoding. Use when you need to
+ * know exactly what was sent, e.g. for an OAuth signature over the elements.
+ *
+ * @codeCoverageIgnore
+ * @return string
+ */
+ public function getRawQueryString() {
+ return $_SERVER['QUERY_STRING'];
+ }
+
+ /**
+ * Return the contents of the POST with no decoding. Use when you need to
+ * know exactly what was sent, e.g. for an OAuth signature over the elements.
+ *
+ * @return string
+ */
+ public function getRawPostString() {
+ if ( !$this->wasPosted() ) {
+ return '';
+ }
+ return $this->getRawInput();
+ }
+
+ /**
+ * Return the raw request body, with no processing. Cached since some methods
+ * disallow reading the stream more than once. As stated in the php docs, this
+ * does not work with enctype="multipart/form-data".
+ *
+ * @return string
+ */
+ public function getRawInput() {
+ static $input = null;
+ if ( $input === null ) {
+ $input = file_get_contents( 'php://input' );
+ }
+ return $input;
+ }
+
+ /**
+ * Get the HTTP method used for this request.
+ *
+ * @return string
+ */
+ public function getMethod() {
+ return isset( $_SERVER['REQUEST_METHOD'] ) ? $_SERVER['REQUEST_METHOD'] : 'GET';
+ }
+
/**
* Returns true if the present request was reached by a POST operation,
* false otherwise (GET, HEAD, or command-line).
* Note that values retrieved by the object may come from the
* GET URL etc even on a POST request.
*
- * @return Boolean
+ * @return bool
*/
public function wasPosted() {
- return $_SERVER['REQUEST_METHOD'] == 'POST';
+ return $this->getMethod() == 'POST';
}
/**
- * Returns true if there is a session cookie set.
- * This does not necessarily mean that the user is logged in!
+ * Return the session for this request
*
- * If you want to check for an open session, use session_id()
- * instead; that will also tell you if the session was opened
- * during the current request (in which case the cookie will
- * be sent back to the client at the end of the script run).
+ * This might unpersist an existing session if it was invalid.
*
- * @return Boolean
+ * @since 1.27
+ * @note For performance, keep the session locally if you will be making
+ * much use of it instead of calling this method repeatedly.
+ * @return Session
+ */
+ public function getSession() {
+ if ( $this->sessionId !== null ) {
+ $session = SessionManager::singleton()->getSessionById( (string)$this->sessionId, true, $this );
+ if ( $session ) {
+ return $session;
+ }
+ }
+
+ $session = SessionManager::singleton()->getSessionForRequest( $this );
+ $this->sessionId = $session->getSessionId();
+ return $session;
+ }
+
+ /**
+ * Set the session for this request
+ * @since 1.27
+ * @private For use by MediaWiki\Session classes only
+ * @param SessionId $sessionId
+ */
+ public function setSessionId( SessionId $sessionId ) {
+ $this->sessionId = $sessionId;
+ }
+
+ /**
+ * Get the session id for this request, if any
+ * @since 1.27
+ * @private For use by MediaWiki\Session classes only
+ * @return SessionId|null
*/
- public function checkSessionCookie() {
- return isset( $_COOKIE[ session_name() ] );
+ public function getSessionId() {
+ return $this->sessionId;
}
/**
* Get a cookie from the $_COOKIE jar
*
- * @param $key String: the name of the cookie
- * @param $prefix String: a prefix to use for the cookie name, if not $wgCookiePrefix
- * @param $default Mixed: what to return if the value isn't found
- * @return Mixed: cookie value or $default if the cookie not set
+ * @param string $key The name of the cookie
+ * @param string $prefix A prefix to use for the cookie name, if not $wgCookiePrefix
+ * @param mixed $default What to return if the value isn't found
+ * @return mixed Cookie value or $default if the cookie not set
*/
public function getCookie( $key, $prefix = null, $default = null ) {
- if( $prefix === null ) {
+ if ( $prefix === null ) {
global $wgCookiePrefix;
$prefix = $wgCookiePrefix;
}
- return $this->getGPCVal( $_COOKIE, $prefix . $key , $default );
+ return $this->getGPCVal( $_COOKIE, $prefix . $key, $default );
}
/**
- * Return the path portion of the request URI.
+ * Return the path and query string portion of the main request URI.
+ * This will be suitable for use as a relative link in HTML output.
*
- * @return String
+ * @throws MWException
+ * @return string
*/
- public function getRequestURL() {
- if( isset( $_SERVER['REQUEST_URI']) && strlen($_SERVER['REQUEST_URI']) ) {
+ public static function getGlobalRequestURL() {
+ 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'] != '' ) {
+ if ( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
$base .= '?' . $_SERVER['QUERY_STRING'];
}
} 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 https://phabricator.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
// need to strip it or we get false-positive redirect loops
// or weird output URLs
$hash = strpos( $base, '#' );
- if( $hash !== false ) {
+ if ( $hash !== false ) {
$base = substr( $base, 0, $hash );
}
- if( $base{0} == '/' ) {
- return $base;
+
+ if ( $base[0] == '/' ) {
+ // More than one slash will look like it is protocol relative
+ return preg_replace( '!^/+!', '/', $base );
} else {
// We may get paths with a host prepended; strip it.
- return preg_replace( '!^[^:]+://[^/]+/!', '/', $base );
+ return preg_replace( '!^[^:]+://[^/]+/+!', '/', $base );
}
}
/**
- * Return the request URI with the canonical service and hostname.
+ * 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
+ * @throws MWException
+ * @return string
*/
- public function getFullRequestURL() {
- global $wgServer;
- return $wgServer . $this->getRequestURL();
+ public function getRequestURL() {
+ return self::getGlobalRequestURL();
}
/**
- * Take an arbitrary query and rewrite the present URL to include it
- * @param $query String: query string fragment; do not include initial '?'
+ * 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
+ * @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 );
+ public function getFullRequestURL() {
+ return wfExpandUrl( $this->getRequestURL(), PROTO_CURRENT );
}
/**
- * HTML-safe version of appendQuery().
- *
- * @param $query String: query string fragment; do not include initial '?'
- * @return String
+ * @param string $key
+ * @param string $value
+ * @return string
*/
- public function escapeAppendQuery( $query ) {
- return htmlspecialchars( $this->appendQuery( $query ) );
- }
-
- public function appendQueryValue( $key, $value, $onlyquery = false ) {
- return $this->appendQueryArray( array( $key => $value ), $onlyquery );
+ public function appendQueryValue( $key, $value ) {
+ return $this->appendQueryArray( [ $key => $value ] );
}
/**
* Appends or replaces value of query variables.
*
- * @param $array Array of values to replace/add to query
- * @param $onlyquery Bool: whether to only return the query string and not
- * the complete URL
- * @return String
+ * @param array $array Array of values to replace/add to query
+ * @return string
*/
- public function appendQueryArray( $array, $onlyquery = false ) {
- global $wgTitle;
- $newquery = $_GET;
+ public function appendQueryArray( $array ) {
+ $newquery = $this->getQueryValues();
unset( $newquery['title'] );
$newquery = array_merge( $newquery, $array );
- $query = wfArrayToCGI( $newquery );
- return $onlyquery ? $query : $wgTitle->getLocalURL( $query );
+
+ return wfArrayToCgi( $newquery );
}
/**
* defaults if not given. The limit must be positive and is capped at 5000.
* Offset must be positive but is not capped.
*
- * @param $deflimit Integer: limit to use if no input and the user hasn't set the option.
- * @param $optionname String: to specify an option other than rclimit to pull from.
- * @return array first element is limit, second is offset
+ * @param int $deflimit Limit to use if no input and the user hasn't set the option.
+ * @param string $optionname To specify an option other than rclimit to pull from.
+ * @return int[] First element is limit, second is offset
*/
public function getLimitOffset( $deflimit = 50, $optionname = 'rclimit' ) {
global $wgUser;
$limit = $this->getInt( 'limit', 0 );
- if( $limit < 0 ) {
+ if ( $limit < 0 ) {
$limit = 0;
}
- if( ( $limit == 0 ) && ( $optionname != '' ) ) {
- $limit = (int)$wgUser->getOption( $optionname );
+ if ( ( $limit == 0 ) && ( $optionname != '' ) ) {
+ $limit = $wgUser->getIntOption( $optionname );
}
- if( $limit <= 0 ) {
+ if ( $limit <= 0 ) {
$limit = $deflimit;
}
- if( $limit > 5000 ) {
+ if ( $limit > 5000 ) {
$limit = 5000; # We have *some* limits...
}
$offset = $this->getInt( 'offset', 0 );
- if( $offset < 0 ) {
+ if ( $offset < 0 ) {
$offset = 0;
}
- return array( $limit, $offset );
+ return [ $limit, $offset ];
}
/**
* Return the path to the temporary file where PHP has stored the upload.
*
- * @param $key String:
- * @return string or NULL if no such file.
+ * @param string $key
+ * @return string|null String or null if no such file.
*/
public function getFileTempname( $key ) {
$file = new WebRequestUpload( $this, $key );
return $file->getTempName();
}
- /**
- * Return the size of the upload, or 0.
- *
- * @deprecated
- * @param $key String:
- * @return integer
- */
- public function getFileSize( $key ) {
- $file = new WebRequestUpload( $this, $key );
- return $file->getSize();
- }
-
/**
* Return the upload error or 0
*
- * @param $key String:
- * @return integer
+ * @param string $key
+ * @return int
*/
public function getUploadError( $key ) {
$file = new WebRequestUpload( $this, $key );
*
* Other than this the name is not verified for being a safe filename.
*
- * @param $key String:
- * @return string or NULL if no such file.
+ * @param string $key
+ * @return string|null String or null if no such file.
*/
public function getFileName( $key ) {
$file = new WebRequestUpload( $this, $key );
/**
* Return a WebRequestUpload object corresponding to the key
*
- * @param @key string
+ * @param string $key
* @return WebRequestUpload
*/
public function getUpload( $key ) {
}
/**
- * Get a request header, or false if it isn't set
- * @param $name String: case-insensitive header name
+ * Initialise the header list
*/
- public function getHeader( $name ) {
- $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;
+ protected function initHeaders() {
+ if ( count( $this->headers ) ) {
+ return;
+ }
+
+ $apacheHeaders = function_exists( 'apache_request_headers' ) ? apache_request_headers() : false;
+ if ( $apacheHeaders ) {
+ foreach ( $apacheHeaders as $tempName => $tempValue ) {
+ $this->headers[strtoupper( $tempName )] = $tempValue;
}
} 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;
+ 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 data from $_SESSION
+ * Get an array containing all request headers
*
- * @param $key String: name of key in $_SESSION
- * @return Mixed
+ * @return array Mapping header name to its value
*/
- public function getSessionData( $key ) {
- if( !isset( $_SESSION[$key] ) ) {
- return null;
+ public function getAllHeaders() {
+ $this->initHeaders();
+ return $this->headers;
+ }
+
+ /**
+ * Get a request header, or false if it isn't set.
+ *
+ * @param string $name Case-insensitive header name
+ * @param int $flags Bitwise combination of:
+ * WebRequest::GETHEADER_LIST Treat the header as a comma-separated list
+ * of values, as described in RFC 2616 § 4.2.
+ * (since 1.26).
+ * @return string|array|bool False if header is unset; otherwise the
+ * header value(s) as either a string (the default) or an array, if
+ * WebRequest::GETHEADER_LIST flag was set.
+ */
+ public function getHeader( $name, $flags = 0 ) {
+ $this->initHeaders();
+ $name = strtoupper( $name );
+ if ( !isset( $this->headers[$name] ) ) {
+ return false;
+ }
+ $value = $this->headers[$name];
+ if ( $flags & self::GETHEADER_LIST ) {
+ $value = array_map( 'trim', explode( ',', $value ) );
}
- return $_SESSION[$key];
+ return $value;
+ }
+
+ /**
+ * Get data from the session
+ *
+ * @note Prefer $this->getSession() instead if making multiple calls.
+ * @param string $key Name of key in the session
+ * @return mixed
+ */
+ public function getSessionData( $key ) {
+ return $this->getSession()->get( $key );
}
/**
* Set session data
*
- * @param $key String: name of key in $_SESSION
- * @param $data Mixed
+ * @note Prefer $this->getSession() instead if making multiple calls.
+ * @param string $key Name of key in the session
+ * @param mixed $data
*/
public function setSessionData( $key, $data ) {
- $_SESSION[$key] = $data;
+ $this->getSession()->set( $key, $data );
}
/**
- * Check if Internet Explorer will detect an incorrect cache extension in
+ * 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 array $extWhitelist
+ * @throws HttpError
+ * @return bool
*/
- public function checkUrlExtension( $extWhitelist = array() ) {
- global $wgScriptExtension;
- $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
+ public function checkUrlExtension( $extWhitelist = [] ) {
+ $extWhitelist[] = 'php';
if ( IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist ) ) {
if ( !$this->wasPosted() ) {
$newUrl = IEUrlExtension::fixUrlForIE6(
return false;
}
}
- wfHttpError( 403, 'Forbidden',
+ throw new HttpError( 403,
'Invalid file extension found in the path info or query string.' );
-
- return false;
}
return true;
}
/**
- * Attempt to redirect to a URL with a QUERY_STRING that's not dangerous in
+ * 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 string $url
+ * @return bool
*/
protected function doSecurityRedirect( $url ) {
header( 'Location: ' . $url );
header( 'Content-Type: text/html' );
$encUrl = htmlspecialchars( $url );
echo <<<HTML
+<!DOCTYPE html>
<html>
<head>
<title>Security redirect</title>
<body>
<h1>Security redirect</h1>
<p>
-We can't serve non-HTML content from the URL you have requested, because
+We can't serve non-HTML content from the URL you have requested, because
Internet Explorer would interpret it as an incorrect and potentially dangerous
content type.</p>
-<p>Instead, please use <a href="$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>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>
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
- * is not HTML but may be detected as such.
- *
- * Various past attempts to use the URL to make this check have generally
- * run up against the fact that CGI does not provide a standard method to
- * determine the URL. PATH_INFO may be mangled (e.g. if cgi.fix_pathinfo=0),
- * 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).
- */
- public function isPathInfoBad() {
- global $wgScriptExtension;
- $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 then appearing time in the header in ascending order.
* May contain the "language" '*', which applies to languages other than those explicitly listed.
* This is aligned with rfc2616 section 14.4
+ * Preference for earlier languages appears in rfc3282 as an extension to HTTP/1.1.
*/
public function getAcceptLang() {
- // Modified version of code found at http://www.thefutureoftheweb.com/blog/use-accept-language-header
+ // Modified version of code found at
+ // http://www.thefutureoftheweb.com/blog/use-accept-language-header
$acceptLang = $this->getHeader( 'Accept-Language' );
if ( !$acceptLang ) {
- return array();
+ return [];
}
// Return the language codes in lower case
// 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]+)?)?)?/',
- $acceptLang, $lang_parse );
+ 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] ) ) {
- return array();
+ return [];
}
- // Create a list like "en" => 0.8
- $langs = array_combine( $lang_parse[1], $lang_parse[4] );
+ $langcodes = $lang_parse[1];
+ $qvalues = $lang_parse[4];
+ $indices = range( 0, count( $lang_parse[1] ) - 1 );
+
// Set default q factor to 1
- foreach ( $langs as $lang => $val ) {
- if ( $val === '' ) {
- $langs[$lang] = 1;
- } else if ( $val == 0 ) {
- unset($langs[$lang]);
+ foreach ( $indices as $index ) {
+ if ( $qvalues[$index] === '' ) {
+ $qvalues[$index] = 1;
+ } elseif ( $qvalues[$index] == 0 ) {
+ unset( $langcodes[$index], $qvalues[$index], $indices[$index] );
}
}
- // Sort list
- arsort( $langs, SORT_NUMERIC );
- return $langs;
- }
-}
+ // Sort list. First by $qvalues, then by order. Reorder $langcodes the same way
+ array_multisort( $qvalues, SORT_DESC, SORT_NUMERIC, $indices, $langcodes );
-/**
- * Object to access the $_FILES array
- */
-class WebRequestUpload {
- protected $request;
- protected $doesExist;
- protected $fileInfo;
+ // Create a list like "en" => 0.8
+ $langs = array_combine( $langcodes, $qvalues );
- /**
- * Constructor. Should only be called by WebRequest
- *
- * @param $request WebRequest The associated request
- * @param $key string Key in $_FILES array (name of form field)
- */
- public function __construct( $request, $key ) {
- $this->request = $request;
- $this->doesExist = isset( $_FILES[$key] );
- if ( $this->doesExist ) {
- $this->fileInfo = $_FILES[$key];
- }
+ return $langs;
}
/**
- * Return whether a file with this name was uploaded.
+ * Fetch the raw IP from the request
*
- * @return bool
- */
- public function exists() {
- return $this->doesExist;
- }
-
- /**
- * Return the original filename of the uploaded file
+ * @since 1.19
*
- * @return mixed Filename or null if non-existent
+ * @throws MWException
+ * @return string
*/
- public function getName() {
- if ( !$this->exists() ) {
+ protected function getRawIP() {
+ if ( !isset( $_SERVER['REMOTE_ADDR'] ) ) {
return null;
}
- global $wgContLang;
- $name = $this->fileInfo['name'];
+ if ( is_array( $_SERVER['REMOTE_ADDR'] ) || strpos( $_SERVER['REMOTE_ADDR'], ',' ) !== false ) {
+ throw new MWException( __METHOD__
+ . " : Could not determine the remote IP address due to multiple values." );
+ } else {
+ $ipchain = $_SERVER['REMOTE_ADDR'];
+ }
- # Safari sends filenames in HTML-encoded Unicode form D...
- # Horrid and evil! Let's try to make some kind of sense of it.
- $name = Sanitizer::decodeCharReferences( $name );
- $name = $wgContLang->normalize( $name );
- wfDebug( __METHOD__ . ": {$this->fileInfo['name']} normalized to '$name'\n" );
- return $name;
+ return IP::canonicalize( $ipchain );
}
/**
- * Return the file size of the uploaded file
+ * Work out the IP address based on various globals
+ * For trusted proxies, use the XFF client IP (first of the chain)
+ *
+ * @since 1.19
*
- * @return int File size or zero if non-existent
+ * @throws MWException
+ * @return string
*/
- public function getSize() {
- if ( !$this->exists() ) {
- return 0;
+ public function getIP() {
+ global $wgUsePrivateIPs;
+
+ # Return cached result
+ if ( $this->ip !== null ) {
+ return $this->ip;
}
- return $this->fileInfo['size'];
- }
+ # collect the originating ips
+ $ip = $this->getRawIP();
+ if ( !$ip ) {
+ throw new MWException( 'Unable to determine IP.' );
+ }
- /**
- * Return the path to the temporary file
- *
- * @return mixed Path or null if non-existent
- */
- public function getTempName() {
- if ( !$this->exists() ) {
- return null;
+ # Append XFF
+ $forwardedFor = $this->getHeader( 'X-Forwarded-For' );
+ if ( $forwardedFor !== false ) {
+ $proxyLookup = MediaWikiServices::getInstance()->getProxyLookup();
+ $isConfigured = $proxyLookup->isConfiguredProxy( $ip );
+ $ipchain = array_map( 'trim', explode( ',', $forwardedFor ) );
+ $ipchain = array_reverse( $ipchain );
+ 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). However, prefer private
+ # IP addresses over proxy servers controlled by this site (more sensible).
+ # Note that some XFF values might be "unknown" with Squid/Varnish.
+ foreach ( $ipchain as $i => $curIP ) {
+ $curIP = IP::sanitizeIP( IP::canonicalize( $curIP ) );
+ if ( !$curIP || !isset( $ipchain[$i + 1] ) || $ipchain[$i + 1] === 'unknown'
+ || !$proxyLookup->isTrustedProxy( $curIP )
+ ) {
+ break; // IP is not valid/trusted or does not point to anything
+ }
+ if (
+ IP::isPublic( $ipchain[$i + 1] ) ||
+ $wgUsePrivateIPs ||
+ $proxyLookup->isConfiguredProxy( $curIP ) // T50919; treat IP as sane
+ ) {
+ // Follow the next IP according to the proxy
+ $nextIP = IP::canonicalize( $ipchain[$i + 1] );
+ if ( !$nextIP && $isConfigured ) {
+ // We have not yet made it past CDN/proxy servers of this site,
+ // so either they are misconfigured or there is some IP spoofing.
+ throw new MWException( "Invalid IP given in XFF '$forwardedFor'." );
+ }
+ $ip = $nextIP;
+ // keep traversing the chain
+ continue;
+ }
+ break;
+ }
}
- return $this->fileInfo['tmp_name'];
+ # Allow extensions to improve our guess
+ Hooks::run( 'GetIP', [ &$ip ] );
+
+ if ( !$ip ) {
+ throw new MWException( "Unable to determine IP." );
+ }
+
+ wfDebug( "IP: $ip\n" );
+ $this->ip = $ip;
+ return $ip;
}
/**
- * Return the upload error. See link for explanation
- * http://www.php.net/manual/en/features.file-upload.errors.php
- *
- * @return int One of the UPLOAD_ constants, 0 if non-existent
+ * @param string $ip
+ * @return void
+ * @since 1.21
*/
- public function getError() {
- if ( !$this->exists() ) {
- return 0; # UPLOAD_ERR_OK
- }
-
- return $this->fileInfo['error'];
+ public function setIP( $ip ) {
+ $this->ip = $ip;
}
/**
- * Returns whether this upload failed because of overflow of a maximum set
- * in php.ini
+ * Check if this request uses a "safe" HTTP method
+ *
+ * Safe methods are verbs (e.g. GET/HEAD/OPTIONS) used for obtaining content. Such requests
+ * are not expected to mutate content, especially in ways attributable to the client. Verbs
+ * like POST and PUT are typical of non-safe requests which often change content.
*
* @return bool
+ * @see https://tools.ietf.org/html/rfc7231#section-4.2.1
+ * @see https://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html
+ * @since 1.28
*/
- public function isIniSizeOverflow() {
- if ( $this->getError() == UPLOAD_ERR_INI_SIZE ) {
- # PHP indicated that upload_max_filesize is exceeded
- return true;
- }
-
- $contentLength = $this->request->getHeader( 'CONTENT_LENGTH' );
- if ( $contentLength > wfShorthandToInteger( ini_get( 'post_max_size' ) ) ) {
- # post_max_size is exceeded
- return true;
+ public function hasSafeMethod() {
+ if ( !isset( $_SERVER['REQUEST_METHOD'] ) ) {
+ return false; // CLI mode
}
- return false;
+ return in_array( $_SERVER['REQUEST_METHOD'], [ 'GET', 'HEAD', 'OPTIONS', 'TRACE' ] );
}
-}
-
-/**
- * WebRequest clone which takes values from a provided array.
- *
- * @ingroup HTTP
- */
-class FauxRequest extends WebRequest {
- private $wasPosted = false;
- private $session = array();
/**
- * @param $data Array of *non*-urlencoded key => value pairs, the
- * fake GET/POST values
- * @param $wasPosted Bool: whether to treat the data as POST
- * @param $session Mixed: session array or null
+ * Whether this request should be identified as being "safe"
+ *
+ * This means that the client is not requesting any state changes and that database writes
+ * are not inherently required. Ideally, no visible updates would happen at all. If they
+ * must, then they should not be publically attributed to the end user.
+ *
+ * In more detail:
+ * - Cache populations and refreshes MAY occur.
+ * - Private user session updates and private server logging MAY occur.
+ * - Updates to private viewing activity data MAY occur via DeferredUpdates.
+ * - Other updates SHOULD NOT occur (e.g. modifying content assets).
+ *
+ * @return bool
+ * @see https://tools.ietf.org/html/rfc7231#section-4.2.1
+ * @see https://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html
+ * @since 1.28
*/
- public function __construct( $data, $wasPosted = false, $session = null ) {
- if( is_array( $data ) ) {
- $this->data = $data;
- } else {
- throw new MWException( "FauxRequest() got bogus data" );
+ public function isSafeRequest() {
+ if ( $this->markedAsSafe && $this->wasPosted() ) {
+ return true; // marked as a "safe" POST
}
- $this->wasPosted = $wasPosted;
- if( $session )
- $this->session = $session;
- }
-
- private function notImplemented( $method ) {
- throw new MWException( "{$method}() not implemented" );
- }
-
- public function getText( $name, $default = '' ) {
- # Override; don't recode since we're using internal data
- return (string)$this->getVal( $name, $default );
- }
-
- public function getValues() {
- return $this->data;
- }
-
- public function wasPosted() {
- return $this->wasPosted;
- }
- public function checkSessionCookie() {
- return false;
+ return $this->hasSafeMethod();
}
- public function getRequestURL() {
- $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 );
- }
-
- public function getHeader( $name ) {
- return isset( $this->headers[$name] ) ? $this->headers[$name] : false;
- }
-
- public function setHeader( $name, $val ) {
- $this->headers[$name] = $val;
- }
-
- public function getSessionData( $key ) {
- if( isset( $this->session[$key] ) )
- return $this->session[$key];
- }
-
- public function setSessionData( $key, $data ) {
- $this->session[$key] = $data;
- }
-
- public function isPathInfoBad( $extWhitelist = array() ) {
- return false;
- }
-
- public function checkUrlExtension( $extWhitelist = array() ) {
- return true;
+ /**
+ * Mark this request as identified as being nullipotent even if it is a POST request
+ *
+ * POST requests are often used due to the need for a client payload, even if the request
+ * is otherwise equivalent to a "safe method" request.
+ *
+ * @see https://tools.ietf.org/html/rfc7231#section-4.2.1
+ * @see https://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html
+ * @since 1.28
+ */
+ public function markAsSafeRequest() {
+ $this->markedAsSafe = true;
}
}
scripts.mit.edu / autoinstallsdev / mediawiki.git / blobdiff