CI框架源码阅读---------Input.php

<?php  if ( ! defined('BASEPATH')) exit('No direct script access allowed');
/**
 * CodeIgniter
 *
 * An open source application development framework for PHP 5.1.6 or newer
 *
 * @package		CodeIgniter
 * @author		ExpressionEngine Dev Team
 * @copyright	Copyright (c) 2008 - 2011, EllisLab, Inc.
 * @license		http://codeigniter.com/user_guide/license.html
 * @link		http://codeigniter.com
 * @since		Version 1.0
 * @filesource
 */

// ------------------------------------------------------------------------

/**
 * Input Class
 * 
 * Pre-processes global input data for security
 *
 * @package		CodeIgniter
 * @subpackage	Libraries
 * @category	Input
 * @author		ExpressionEngine Dev Team
 * @link		http://codeigniter.com/user_guide/libraries/input.html
 */
class CI_Input {

	/**
	 * IP address of the current user
	 * 当前用户的ip地址
	 * @var string
	 */
	var $ip_address				= FALSE;
	/**
	 * user agent (web browser) being used by the current user
	 * 当前用户（web浏览器）代理
	 * @var string
	 */
	var $user_agent				= FALSE;
	/**
	 * If FALSE, then $_GET will be set to an empty array
	 * 如果是FALSE ， $_GET将被设置为空数组
	 * @var bool
	 */
	var $_allow_get_array		= TRUE;
	/**
	 * If TRUE, then newlines are standardized
	 * 如果为TRUR，新行将被标准化
	 *
	 * @var bool
	 */
	var $_standardize_newlines	= TRUE;
	/**
	 * Determines whether the XSS filter is always active when GET, POST or COOKIE data is encountered
	 * Set automatically based on config setting
	 * 决定是否总是在GET ,POST , COOKIE数据中进行XSS过滤
     * 在配置选项里面配置是否自动开启
	 * 
	 * @var bool
	 */
	var $_enable_xss			= FALSE;
	/**
	 * Enables a CSRF cookie token to be set.
	 * Set automatically based on config setting
	 * 允许CSRF cookie令牌
	 *
	 * @var bool
	 */
	var $_enable_csrf			= FALSE;
	/**
	 * List of all HTTP request headers
	 * HTTP请求头部的列表
	 * @var array
	 */
	protected $headers			= array();

	/**
	 * Constructor
	 * 设置是否全局允许XSS处理和是否允许使用$_GET数组
	 * Sets whether to globally enable the XSS processing
	 * and whether to allow the $_GET array
	 *
	 * @return	void
	 */
	public function __construct()
	{
		log_message('debug', "Input Class Initialized");

		// 从配置文件中获取是否进行全局允许使用$_GET XSS过滤和csrf保护
		$this->_allow_get_array	= (config_item('allow_get_array') === TRUE);
		$this->_enable_xss		= (config_item('global_xss_filtering') === TRUE);
		$this->_enable_csrf		= (config_item('csrf_protection') === TRUE);
		
		// 清除globals变量，在开启了globals_register的情况下，相当于关闭了此配置。
		// 开启一道 安全防护
		
		global $SEC;
		$this->security =& $SEC;

		// Do we need the UTF-8 class?
		if (UTF8_ENABLED === TRUE)
		{
			global $UNI;
			$this->uni =& $UNI;
		}

		// Sanitize global arrays
		$this->_sanitize_globals();
	}

	// --------------------------------------------------------------------

	/**
	 * Fetch from array
	 * 从$array获取值，如果设置了xss_clean 那么进行过滤 
	 * This is a helper function to retrieve 检索 values from global arrays
	 * 这是一个帮助函数用来从全局数组中检索
	 *
	 * @access	private
	 * @param	array
	 * @param	string
	 * @param	bool
	 * @return	string
	 */
	function _fetch_from_array(&$array, $index = '', $xss_clean = FALSE)
	{
		if ( ! isset($array[$index]))
		{
			return FALSE;
		}

		if ($xss_clean === TRUE)
		{
			return $this->security->xss_clean($array[$index]);
		}

		return $array[$index];
	}

	// --------------------------------------------------------------------

	/**
	* Fetch an item from the GET array
	* 获取过滤后的GET数组
	* @access	public
	* @param	string
	* @param	bool
	* @return	string
	*/
	function get($index = NULL, $xss_clean = FALSE)
	{
		// Check if a field has been provided
		// 检查是否一个字段已经被提供
		if ($index === NULL AND ! empty($_GET))
		{
			$get = array();

			// loop through the full _GET array
			// 遍历_GET数组
			foreach (array_keys($_GET) as $key)
			{
				$get[$key] = $this->_fetch_from_array($_GET, $key, $xss_clean);
			}
			return $get;
		}

		return $this->_fetch_from_array($_GET, $index, $xss_clean);
	}

	// --------------------------------------------------------------------

	/**
	* Fetch an item from the POST array
	* 获取过滤后的$_POST值
	* @access	public
	* @param	string
	* @param	bool
	* @return	string
	*/
	function post($index = NULL, $xss_clean = FALSE)
	{
		// Check if a field has been provided
		if ($index === NULL AND ! empty($_POST))
		{
			$post = array();

			// Loop through the full _POST array and return it
			foreach (array_keys($_POST) as $key)
			{
				$post[$key] = $this->_fetch_from_array($_POST, $key, $xss_clean);
			}
			return $post;
		}

		return $this->_fetch_from_array($_POST, $index, $xss_clean);
	}


	// --------------------------------------------------------------------

	/**
	* Fetch an item from either the GET array or the POST
	* 从get和post中获取值， post优先
	* @access	public
	* @param	string	The index key
	* @param	bool	XSS cleaning
	* @return	string
	*/
	function get_post($index = '', $xss_clean = FALSE)
	{
		if ( ! isset($_POST[$index]) )
		{
			return $this->get($index, $xss_clean);
		}
		else
		{
			return $this->post($index, $xss_clean);
		}
	}

	// --------------------------------------------------------------------

	/**
	* Fetch an item from the COOKIE array
	* 返回过滤后的COOKIE值
	* @access	public
	* @param	string
	* @param	bool
	* @return	string
	*/
	function cookie($index = '', $xss_clean = FALSE)
	{
		return $this->_fetch_from_array($_COOKIE, $index, $xss_clean);
	}

	// ------------------------------------------------------------------------

	/**
	* Set cookie
	*
	* Accepts six parameter, or you can submit an associative
	* array in the first parameter containing all the values.
	* 接收6个参数或者接收一个关联数组里面包含所有的值
	* @access	public
	* @param	mixed
	* @param	string	the value of the cookie
	* @param	string	the number of seconds until expiration
	* @param	string	the cookie domain.  Usually:  .yourdomain.com
	* @param	string	the cookie path
	* @param	string	the cookie prefix
	* @param	bool	true makes the cookie secure
	* @return	void
	*/
	function set_cookie($name = '', $value = '', $expire = '', $domain = '', $path = '/', $prefix = '', $secure = FALSE)
	{
		// 如果第一个值是数组 将数组中的值分别赋值给留个参数
		if (is_array($name))
		{
			// always leave 'name' in last place, as the loop will break otherwise, due to $$item
			foreach (array('value', 'expire', 'domain', 'path', 'prefix', 'secure', 'name') as $item)
			{
				if (isset($name[$item]))
				{
					$$item = $name[$item];
				}
			}
		}

		// 如果某个参数为默认值但是config.php中的配置不是默认值
		// 则使用config.php中的配置值
		if ($prefix == '' AND config_item('cookie_prefix') != '')
		{
			$prefix = config_item('cookie_prefix');
		}
		if ($domain == '' AND config_item('cookie_domain') != '')
		{
			$domain = config_item('cookie_domain');
		}
		if ($path == '/' AND config_item('cookie_path') != '/')
		{
			$path = config_item('cookie_path');
		}
		if ($secure == FALSE AND config_item('cookie_secure') != FALSE)
		{
			$secure = config_item('cookie_secure');
		}

		if ( ! is_numeric($expire))
		{
			$expire = time() - 86500;
		}
		else
		{
			$expire = ($expire > 0) ? time() + $expire : 0;
		}

		setcookie($prefix.$name, $value, $expire, $path, $domain, $secure);
	}

	// --------------------------------------------------------------------

	/**
	* Fetch an item from the SERVER array
	* 返回过滤后的$_SERVER值
	* @access	public
	* @param	string
	* @param	bool
	* @return	string
	*/
	function server($index = '', $xss_clean = FALSE)
	{
		return $this->_fetch_from_array($_SERVER, $index, $xss_clean);
	}

	// --------------------------------------------------------------------

	/**
	* Fetch the IP Address
	* 返回当前用户的IP。如果IP地址无效，返回0.0.0.0的IP：
	* @return	string
	*/
	public function ip_address()
	{
		// 如果已经有了ip_address 则返回
		if ($this->ip_address !== FALSE)
		{
			return $this->ip_address;
		}

		$proxy_ips = config_item('proxy_ips');
		if ( ! empty($proxy_ips))
		{
			$proxy_ips = explode(',', str_replace(' ', '', $proxy_ips));
			foreach (array('HTTP_X_FORWARDED_FOR', 'HTTP_CLIENT_IP', 'HTTP_X_CLIENT_IP', 'HTTP_X_CLUSTER_CLIENT_IP') as $header)
			{
				if (($spoof = $this->server($header)) !== FALSE)
				{
					// Some proxies typically list the whole chain of IP
					// addresses through which the client has reached us.
					// e.g. client_ip, proxy_ip1, proxy_ip2, etc.
					if (strpos($spoof, ',') !== FALSE)
					{
						$spoof = explode(',', $spoof, 2);
						$spoof = $spoof[0];
					}

					if ( ! $this->valid_ip($spoof))
					{
						$spoof = FALSE;
					}
					else
					{
						break;
					}
				}
			}

			$this->ip_address = ($spoof !== FALSE && in_array($_SERVER['REMOTE_ADDR'], $proxy_ips, TRUE))
				? $spoof : $_SERVER['REMOTE_ADDR'];
		}
		else
		{
			$this->ip_address = $_SERVER['REMOTE_ADDR'];
		}

		if ( ! $this->valid_ip($this->ip_address))
		{
			$this->ip_address = '0.0.0.0';
		}

		return $this->ip_address;
	}

	// --------------------------------------------------------------------

	/**
	* Validate IP Address
	* 测试输入的IP地址是不是有效，返回布尔值TRUE或者FALSE。 
	* 注意：$this->input->ip_address()自动测试输入的IP地址本身格式是不是有效。
	* @access	public
	* @param	string
	* @param	string	ipv4 or ipv6
	* @return	bool
	*/
	public function valid_ip($ip, $which = '')
	{
		$which = strtolower($which);

		// First check if filter_var is available
		if (is_callable('filter_var'))
		{
			switch ($which) {
				case 'ipv4':
					$flag = FILTER_FLAG_IPV4;
					break;
				case 'ipv6':
					$flag = FILTER_FLAG_IPV6;
					break;
				default:
					$flag = '';
					break;
			}

			return (bool) filter_var($ip, FILTER_VALIDATE_IP, $flag);
		}

		if ($which !== 'ipv6' && $which !== 'ipv4')
		{
			if (strpos($ip, ':') !== FALSE)
			{
				$which = 'ipv6';
			}
			elseif (strpos($ip, '.') !== FALSE)
			{
				$which = 'ipv4';
			}
			else
			{
				return FALSE;
			}
		}

		$func = '_valid_'.$which;
		return $this->$func($ip);
	}

	// --------------------------------------------------------------------

	/**
	* Validate IPv4 Address
	* 验证ipv4地址
	* Updated version suggested by Geert De Deckere
	*
	* @access	protected
	* @param	string
	* @return	bool
	*/
	protected function _valid_ipv4($ip)
	{
		$ip_segments = explode('.', $ip);

		// Always 4 segments needed
		if (count($ip_segments) !== 4)
		{
			return FALSE;
		}
		// IP can not start with 0
		if ($ip_segments[0][0] == '0')
		{
			return FALSE;
		}

		// Check each segment
		foreach ($ip_segments as $segment)
		{
			// IP segments must be digits and can not be
			// longer than 3 digits or greater then 255
			if ($segment == '' OR preg_match("/[^0-9]/", $segment) OR $segment > 255 OR strlen($segment) > 3)
			{
				return FALSE;
			}
		}

		return TRUE;
	}

	// --------------------------------------------------------------------

	/**
	* Validate IPv6 Address
	* 验证ipv6地址
	* @access	protected
	* @param	string
	* @return	bool
	*/
	protected function _valid_ipv6($str)
	{
		// 8 groups, separated by :
		// 0-ffff per group
		// one set of consecutive 0 groups can be collapsed to ::

		$groups = 8;
		$collapsed = FALSE;

		$chunks = array_filter(
			preg_split('/(:{1,2})/', $str, NULL, PREG_SPLIT_DELIM_CAPTURE)
		);

		// Rule out easy nonsense
		if (current($chunks) == ':' OR end($chunks) == ':')
		{
			return FALSE;
		}

		// PHP supports IPv4-mapped IPv6 addresses, so we'll expect those as well
		if (strpos(end($chunks), '.') !== FALSE)
		{
			$ipv4 = array_pop($chunks);

			if ( ! $this->_valid_ipv4($ipv4))
			{
				return FALSE;
			}

			$groups--;
		}

		while ($seg = array_pop($chunks))
		{
			if ($seg[0] == ':')
			{
				if (--$groups == 0)
				{
					return FALSE;	// too many groups
				}

				if (strlen($seg) > 2)
				{
					return FALSE;	// long separator
				}

				if ($seg == '::')
				{
					if ($collapsed)
					{
						return FALSE;	// multiple collapsed
					}

					$collapsed = TRUE;
				}
			}
			elseif (preg_match("/[^0-9a-f]/i", $seg) OR strlen($seg) > 4)
			{
				return FALSE; // invalid segment
			}
		}

		return $collapsed OR $groups == 1;
	}

	// --------------------------------------------------------------------

	/**
	* User Agent
	* 返回当前用户正在使用的浏览器的user agent信息。 如果不能得到数据，返回FALSE。
	* 一般user_agent为空的时候被认定为手机访问，或者curl的抓取，或则会蜘蛛抓取
	* @access	public
	* @return	string
	*/
	function user_agent()
	{
		if ($this->user_agent !== FALSE)
		{
			return $this->user_agent;
		}

		$this->user_agent = ( ! isset($_SERVER['HTTP_USER_AGENT'])) ? FALSE : $_SERVER['HTTP_USER_AGENT'];

		return $this->user_agent;
	}

	// --------------------------------------------------------------------

	/**
	* Sanitize Globals
	* 清理全局数组
	* This function does the following:
	* 这个函数做了下面的操作：
	* Unsets $_GET data (if query strings are not enabled)
	* 销毁$_GET （如果query strings 没有开启）
	* Unsets all globals if register_globals is enabled
	* 销毁所有全局数组如果register_globals开启
	*
	* Standardizes newline characters to \n
	* 标准化换行符\n
	* @access	private
	* @return	void
	*/
	function _sanitize_globals()
	{
		// It would be "wrong" to unset any of these GLOBALS.
		// 销毁下面的全局数组将是错误的。
		$protected = array('_SERVER', '_GET', '_POST', '_FILES', '_REQUEST',
							'_SESSION', '_ENV', 'GLOBALS', 'HTTP_RAW_POST_DATA',
							'system_folder', 'application_folder', 'BM', 'EXT',
							'CFG', 'URI', 'RTR', 'OUT', 'IN');

		// Unset globals for securiy.为了安全销毁除了上面之外的全局数组
		// This is effectively the same as register_globals = off
		// 这样的效果和register_globals是相同的
        // 经过下面处理后，所有的非保护的全局变量将被删除掉
		
		foreach (array($_GET, $_POST, $_COOKIE) as $global)
		{
			if ( ! is_array($global))
			{
				if ( ! in_array($global, $protected))
				{
					global $$global;
					$$global = NULL;
				}
			}
			else
			{
				foreach ($global as $key => $val)
				{
					if ( ! in_array($key, $protected))
					{
						global $$key;
						$$key = NULL;
					}
				}
			}
		}

		// Is $_GET data allowed? If not we'll set the $_GET to an empty array
		// 是否允许$_GET数据？ 如果不允许的话，设置$_GET为空数组
		if ($this->_allow_get_array == FALSE)
		{
			$_GET = array();
		}
		else
		{
			if (is_array($_GET) AND count($_GET) > 0)
			{
				foreach ($_GET as $key => $val)
				{
					$_GET[$this->_clean_input_keys($key)] = $this->_clean_input_data($val);
				}
			}
		}

		// Clean $_POST Data
		// 过滤$_POST 数组
		if (is_array($_POST) AND count($_POST) > 0)
		{
			foreach ($_POST as $key => $val)
			{
				$_POST[$this->_clean_input_keys($key)] = $this->_clean_input_data($val);
			}
		}

		// Clean $_COOKIE Data
		// 过滤$_COOKIE数组
		if (is_array($_COOKIE) AND count($_COOKIE) > 0)
		{
			// Also get rid of specially treated cookies that might be set by a server
			// or silly application, that are of no use to a CI application anyway
			// but that when present will trip our 'Disallowed Key Characters' alarm
			// http://www.ietf.org/rfc/rfc2109.txt
			// note that the key names below are single quoted strings, and are not PHP variables
			unset($_COOKIE['$Version']);
			unset($_COOKIE['$Path']);
			unset($_COOKIE['$Domain']);

			foreach ($_COOKIE as $key => $val)
			{
				$_COOKIE[$this->_clean_input_keys($key)] = $this->_clean_input_data($val);
			}
		}

		// Sanitize PHP_SELF
		$_SERVER['PHP_SELF'] = strip_tags($_SERVER['PHP_SELF']);


		// CSRF Protection check on HTTP requests
		// ＣＳＲＦ保护检测http请求
		if ($this->_enable_csrf == TRUE && ! $this->is_cli_request())
		{
			$this->security->csrf_verify();
		}

		log_message('debug', "Global POST and COOKIE data sanitized");
	}

	// --------------------------------------------------------------------

	/**
	* Clean Input Data
	* 过滤input数据
	* This is a helper function. It escapes data and
	* standardizes newline characters to \n
	*
	* @access	private
	* @param	string
	* @return	string
	*/
	function _clean_input_data($str)
	{
		if (is_array($str))
		{
			$new_array = array();
			foreach ($str as $key => $val)
			{
				$new_array[$this->_clean_input_keys($key)] = $this->_clean_input_data($val);
			}
			return $new_array;
		}

		/* We strip slashes if magic quotes is on to keep things consistent
			如果小于PHP5.4版本，并且get_magic_quotes_gpc开启了，则去掉斜线。	
		   NOTE: In PHP 5.4 get_magic_quotes_gpc() will always return 0 and
		   it will probably not exist in future versions at all。
		   注意：在PHP5.4及之后版本，get_magic_quotes_gpc()将总是返回0，
		   在后续版本中可能会移除该特性
		*/
		if ( ! is_php('5.4') && get_magic_quotes_gpc())
		{
			$str = stripslashes($str);
		}

		// Clean UTF-8 if supported 如果支持清理utf8
		if (UTF8_ENABLED === TRUE)
		{
			$str = $this->uni->clean_string($str);
		}

		// Remove control characters 
		$str = remove_invisible_characters($str);

		// Should we filter the input data?
		if ($this->_enable_xss === TRUE)
		{
			$str = $this->security->xss_clean($str);
		}

		// Standardize newlines if needed
		if ($this->_standardize_newlines == TRUE)
		{
			if (strpos($str, "\r") !== FALSE)
			{
				$str = str_replace(array("\r\n", "\r", "\r\n\n"), PHP_EOL, $str);
			}
		}

		return $str;
	}

	// --------------------------------------------------------------------

	/**
	* Clean Keys
	* 过滤键值 
	* This is a helper function. To prevent malicious users
	* from trying to exploit keys we make sure that keys are
	* only named with alpha-numeric text and a few other items.
	*
	* @access	private
	* @param	string
	* @return	string
	*/
	function _clean_input_keys($str)
	{
		if ( ! preg_match("/^[a-z0-9:_\/-]+$/i", $str))
		{
			exit('Disallowed Key Characters.');
		}

		// Clean UTF-8 if supported
		if (UTF8_ENABLED === TRUE)
		{
			$str = $this->uni->clean_string($str);
		}

		return $str;
	}

	// --------------------------------------------------------------------

	/**
	 * Request Headers
	 * 返回请求头（header）数组。
	 * In Apache, you can simply call apache_request_headers(), however for
	 * people running other webservers the function is undefined.
	 *
	 * @param	bool XSS cleaning
	 *
	 * @return array
	 */
	public function request_headers($xss_clean = FALSE)
	{
		// Look at Apache go!
		if (function_exists('apache_request_headers'))
		{
			$headers = apache_request_headers();
		}
		else
		{
			$headers['Content-Type'] = (isset($_SERVER['CONTENT_TYPE'])) ? $_SERVER['CONTENT_TYPE'] : @getenv('CONTENT_TYPE');

			foreach ($_SERVER as $key => $val)
			{
				if (strncmp($key, 'HTTP_', 5) === 0)
				{
					$headers[substr($key, 5)] = $this->_fetch_from_array($_SERVER, $key, $xss_clean);
				}
			}
		}

		// take SOME_HEADER and turn it into Some-Header
		foreach ($headers as $key => $val)
		{
			$key = str_replace('_', ' ', strtolower($key));
			$key = str_replace(' ', '-', ucwords($key));

			$this->headers[$key] = $val;
		}

		return $this->headers;
	}

	// --------------------------------------------------------------------

	/**
	 * Get Request Header
	 * 返回请求头（request header）数组中某一个元素的值
	 * Returns the value of a single member of the headers class member
	 *
	 * @param 	string		array key for $this->headers
	 * @param	boolean		XSS Clean or not
	 * @return 	mixed		FALSE on failure, string on success
	 */
	public function get_request_header($index, $xss_clean = FALSE)
	{
		if (empty($this->headers))
		{
			$this->request_headers();
		}

		if ( ! isset($this->headers[$index]))
		{
			return FALSE;
		}

		if ($xss_clean === TRUE)
		{
			return $this->security->xss_clean($this->headers[$index]);
		}

		return $this->headers[$index];
	}

	// --------------------------------------------------------------------

	/**
	 * Is ajax Request? 
	 * 判断是否为ajax请求
	 * Test to see if a request contains the HTTP_X_REQUESTED_WITH header
	 *
	 * @return 	boolean
	 */
	public function is_ajax_request()
	{
		return ($this->server('HTTP_X_REQUESTED_WITH') === 'XMLHttpRequest');
	}

	// --------------------------------------------------------------------

	/**
	 * Is cli Request?
	 * 判断是否来自cli请求
	 * Test to see if a request was made from the command line
	 *
	 * @return 	bool
	 */
	public function is_cli_request()
	{
		return (php_sapi_name() === 'cli' OR defined('STDIN'));
	}

}

/* End of file Input.php */
/* Location: ./system/core/Input.php */