Server and execution environment information

$_SERVER

(PHP 4 >= 4.1.0, PHP 5, PHP 7, PHP 8)

$_SERVER — Server and execution environment information

Description

$_SERVER is an array containing information such as headers, paths, and script locations. The entries in this array are created by the web server. There is no guarantee that every web server will provide any of these; servers may omit some, or provide others not listed here. That said, a large number of these variables are accounted for in the , so you should be able to expect those.

Indices

You may or may not find any of the following elements in $_SERVER. Note that few, if any, of these will be available (or indeed have any meaning) if running PHP on the .

'PHP_SELF'
The filename of the currently executing script, relative to the document root. For instance, $_SERVER['PHP_SELF'] in a script at the address http://example.com/foo/bar.php would be /foo/bar.php. The constant contains the full path and filename of the current (i.e. included) file. If PHP is running as a command-line processor this variable contains the script name.
''
Array of arguments passed to the script. When the script is run on the command line, this gives C-style access to the command line parameters. When called via the GET method, this will contain the query string.
''
Contains the number of command line parameters passed to the script (if run on the command line).
'GATEWAY_INTERFACE'
What revision of the CGI specification the server is using; e.g. 'CGI/1.1'.
'SERVER_ADDR'
The IP address of the server under which the current script is executing.
'SERVER_NAME'
The name of the server host under which the current script is executing. If the script is running on a virtual host, this will be the value defined for that virtual host.

Note: Under Apache 2, you must set UseCanonicalName = On and ServerName. Otherwise, this value reflects the hostname supplied by the client, which can be spoofed. It is not safe to rely on this value in security-dependent contexts.

'SERVER_SOFTWARE'
Server identification string, given in the headers when responding to requests.
'SERVER_PROTOCOL'
Name and revision of the information protocol via which the page was requested; e.g. 'HTTP/1.0';
'REQUEST_METHOD'
Which request method was used to access the page; e.g. 'GET', 'HEAD', 'POST', 'PUT'.

Note:

PHP script is terminated after sending headers (it means after producing any output without output buffering) if the request method was HEAD.

'REQUEST_TIME'
The timestamp of the start of the request.
'REQUEST_TIME_FLOAT'
The timestamp of the start of the request, with microsecond precision.
'QUERY_STRING'
The query string, if any, via which the page was accessed.
'DOCUMENT_ROOT'
The document root directory under which the current script is executing, as defined in the server's configuration file.
'HTTP_ACCEPT'
Contents of the Accept: header from the current request, if there is one.
'HTTP_ACCEPT_CHARSET'
Contents of the Accept-Charset: header from the current request, if there is one. Example: 'iso-8859-1,*,utf-8'.
'HTTP_ACCEPT_ENCODING'
Contents of the Accept-Encoding: header from the current request, if there is one. Example: 'gzip'.
'HTTP_ACCEPT_LANGUAGE'
Contents of the Accept-Language: header from the current request, if there is one. Example: 'en'.
'HTTP_CONNECTION'
Contents of the Connection: header from the current request, if there is one. Example: 'Keep-Alive'.
'HTTP_HOST'
Contents of the Host: header from the current request, if there is one.
'HTTP_REFERER'
The address of the page (if any) which referred the user agent to the current page. This is set by the user agent. Not all user agents will set this, and some provide the ability to modify HTTP_REFERER as a feature. In short, it cannot really be trusted.
'HTTP_USER_AGENT'
Contents of the User-Agent: header from the current request, if there is one. This is a string denoting the user agent being which is accessing the page. A typical example is: Mozilla/4.5 [en] (X11; U; Linux 2.2.9 i586). Among other things, you can use this value with to tailor your page's output to the capabilities of the user agent.
'HTTPS'
Set to a non-empty value if the script was queried through the HTTPS protocol.
'REMOTE_ADDR'
The IP address from which the user is viewing the current page.
'REMOTE_HOST'
The Host name from which the user is viewing the current page. The reverse dns lookup is based on the REMOTE_ADDR of the user.

Note: Your web server must be configured to create this variable. For example in Apache you'll need HostnameLookups On inside httpd.conf for it to exist. See also .

'REMOTE_PORT'
The port being used on the user's machine to communicate with the web server.
'REMOTE_USER'
The authenticated user.
'REDIRECT_REMOTE_USER'
The authenticated user if the request is internally redirected.
'SCRIPT_FILENAME'

The absolute pathname of the currently executing script.

Note:

If a script is executed with the CLI, as a relative path, such as file.php or ../file.php, $_SERVER['SCRIPT_FILENAME'] will contain the relative path specified by the user.

'SERVER_ADMIN'
The value given to the SERVER_ADMIN (for Apache) directive in the web server configuration file. If the script is running on a virtual host, this will be the value defined for that virtual host.
'SERVER_PORT'
The port on the server machine being used by the web server for communication. For default setups, this will be '80'; using SSL, for instance, will change this to whatever your defined secure HTTP port is.

Note: Under the Apache 2, you must set UseCanonicalName = On, as well as UseCanonicalPhysicalPort = On in order to get the physical (real) port, otherwise, this value can be spoofed and it may or may not return the physical port value. It is not safe to rely on this value in security-dependent contexts.

'SERVER_SIGNATURE'
String containing the server version and virtual host name which are added to server-generated pages, if enabled.
'PATH_TRANSLATED'
Filesystem- (not document root-) based path to the current script, after the server has done any virtual-to-real mapping.

Note: Apache 2 users may use AcceptPathInfo = On inside httpd.conf to define PATH_INFO.

'SCRIPT_NAME'
Contains the current script's path. This is useful for pages which need to point to themselves. The constant contains the full path and filename of the current (i.e. included) file.
'REQUEST_URI'
The URI which was given in order to access this page; for instance, '/index.html'.
'PHP_AUTH_DIGEST'
When doing Digest HTTP authentication this variable is set to the 'Authorization' header sent by the client (which you should then use to make the appropriate validation).
'PHP_AUTH_USER'
When doing HTTP authentication this variable is set to the username provided by the user.
'PHP_AUTH_PW'
When doing HTTP authentication this variable is set to the password provided by the user.
'AUTH_TYPE'
When doing HTTP authentication this variable is set to the authentication type.
'PATH_INFO'
Contains any client-provided pathname information trailing the actual script filename but preceding the query string, if available. For instance, if the current script was accessed via the URL http://www.example.com/php/path_info.php/some/stuff?foo=bar, then $_SERVER['PATH_INFO'] would contain /some/stuff.
'ORIG_PATH_INFO'
Original version of 'PATH_INFO' before processed by PHP.

Examples

Example #1 $_SERVER example <?php
echo $_SERVER['SERVER_NAME'];
?>

The above example will output something similar to: 

www.example.com

Notes

Note:

This is a 'superglobal', or automatic global, variable. This simply means that it is available in all scopes throughout a script. There is no need to do global $variable; to access it within functions or methods.

See Also

User Contributed Notes

DanielTahar To expand a bit on the price you could pay for relying on 'HTTP_REFERER': several large news sites I read often have paywalls, with cookies in place so you can only read X articles before you must subscribe; if using Incognito, they count the number of times you accessed via the same IP; everything to get you to subscribe. However, in order to be appealing, any visit where the 'HTTP_REFERER' is Google News will give you the entire article. I'm sure it's a dilemma their webmasters have, but for now any time someone sends you a story on one of them, all you have to do is search for the title and click the result from Google News. Bottom line: never count on it.

PS (1): ofcourse i'm talking about a friend. I pay for content.
PS (2): after some debate, the RFC decided to keep 'HTTP_REFERER', although it's misspelled.
Johan Winge It should probably be noted that the value of $_SERVER['SERVER_PROTOCOL'] will never contain the substring "HTTPS". Assuming this is a common source of bugs and confusion. Instead, see $_SERVER['HTTPS']. vcoletti at tiscali dot it To list all the $_SERVER parameters, simply do:

foreach ($_SERVER as $parm => $value)  echo "$parm = '$value'\n";

No need to list all possible keys of the array. lemonostif at gmail dot com PHP_SELF is a disgrace of a programmer's work. One of the most widespread PHP vulnerabilities since version 4 and the manual says nothing about the dangers. At least clarify that ITS VALUE CAN BE PROVIDED BY THE USER with capitals preferably if you want to make the internet a safer place... Mark Simon So near, and yet so far …

$_SERVER has nearly everything you need to know about the current web page environment. Something which would have been handy is easy access to the protocol and the actual web root.

For the protocol, you may or may not have $_SERVER['HTTPS'] and it may or may not be empty. For the web root, $_SERVER['DOCUMENT_ROOT'] depends on the server configuration, and doesn't work for virtual hosts.

For practical purposes, I normally include something like the following in my scripts:

<?php
    //    Web Root
    //    Usage: include("$root/includes/something.inc.php");
        $root = $_SERVER['WEB_ROOT'] = str_replace($_SERVER['SCRIPT_NAME'],'',$_SERVER['SCRIPT_FILENAME']);

    //    Host & Protocol
    //    Usage: $url = "$protocol://$host/images/something.jpg";
        $host = $_SERVER['HTTP_HOST'];
        $protocol=$_SERVER['PROTOCOL'] = isset($_SERVER['HTTPS']) && !empty($_SERVER['HTTPS']) ? 'https' : 'http';
?> Skylar A snippet to list the header values:

<?php

echo <<<END
<!DOCTYPE html>
<meta charset="UTF-8" />
<title>\$_SERVER</title>
<style>
    table {
        border-collapse: collapse;
    }
    td {
        border: 1px solid #999;
        padding: 3px;
    }
</style>
<table>
END;

foreach ($_SERVER as $k => $v) {
    $key = htmlentities($k);
    $value = htmlentities($v);
    echo "\n\t<tr>\n\t\t<td>$key\n\t\t<td>$value\n";
}
echo "</table>";
?> lilJoshu Remember,

Although $_SERVER["REQUEST_METHOD"] is initially built with GET, POST, PUT, HEAD in mind, a server can allow more.

This may be important if you're building a RESTful interfaces that will also use methods such as PATCH and DELETE.

Also important as a security risk as a possible point of injection. In the event of building something acting based on REQUEST_METHOD, it's recommended to put it in a switch statement.

<?php
switch ($_SERVER["REQUEST_METHOD"]){
   case "PUT":
       foo_replace_data();
       break;
   case "POST":
       foo_add_data();
       break;
   case "HEAD";
       foo_set_that_cookie();
       break;
   case "GET":
   default:
      foo_fetch_stuff()
      break;
}

?> razvan_bc at yahoo dot com In 2018 things can be pretty cool !!!
look around:

<?php
//save it index.php in a root folder

/*
    if($_SERVER["REQUEST_METHOD"]!='GET'){
        echo('it`s a POST');
    }else{   
        echo('it`s a GET');
    }
*/

$_REQ =& $_SERVER["REQUEST_METHOD"];
($_REQ=='POST')and($what='it`s a POST');
($_REQ=='GET')and($what='it`s a GET');
isset($what) and print($what);

?>
<form method="POST" action="">
<input type="text" name="postbaby" value="sadasdsa">
<input type="submit" value="POST THAT">
</form>
<br>
<form method="GET" action="">
<input type="text" name="getbaby" value="sadasdsa">
<input type="submit" value="GET THAT">
</form>
<hr><?php
phpinfo();
?>

what if you have a class with controllers?
i hope will not be depreciated cause STATIC CALLS I TEST ALREADY ARE VERY FAST:

<?php
//save it index.php in a root folder

/*
    if($_SERVER["REQUEST_METHOD"]!='GET'){
        echo('it`s a POST');
    }else{   
        echo('it`s a GET');
    }
*/
class forms {
    public static function type(){
        return $_SERVER["REQUEST_METHOD"];
    }
    public static function post($f){
        (self::type()=='POST')and $f();
    }
   
    public static function get($f){
        (self::type()=='GET')and $f();
    }
}

?>
<form method="POST" action="">
<input type="text" name="postbaby" value="sadasdsa">
<input type="submit" value="POST THAT">
</form>
<br>
<form method="GET" action="">
<input type="text" name="getbaby" value="sadasdsa">
<input type="submit" value="GET THAT">
</form>
<hr><?php
forms::post(function(){
    $what='it`s a POST';echo($what);
});

forms::get(function(){
    $what='it`s a GET';echo($what);
});
//phpinfo();
?> 2962051004 at qq dot com <?php
/*
Sometimes you will find that your website will not get the correct user IP after adding CDN, then this function will help you
*/
function real_ip()
{
   $ip = $_SERVER['REMOTE_ADDR'];
    if (isset($_SERVER['HTTP_X_FORWARDED_FOR']) && preg_match_all('#\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}#s', $_SERVER['HTTP_X_FORWARDED_FOR'], $matches)) {
        foreach ($matches[0] AS $xip) {
            if (!preg_match('#^(10|172\.16|192\.168)\.#', $xip)) {
                $ip = $xip;
                break;
            }
        }
    } elseif (isset($_SERVER['HTTP_CLIENT_IP']) && preg_match('/^([0-9]{1,3}\.){3}[0-9]{1,3}$/', $_SERVER['HTTP_CLIENT_IP'])) {
        $ip = $_SERVER['HTTP_CLIENT_IP'];
    } elseif (isset($_SERVER['HTTP_CF_CONNECTING_IP']) && preg_match('/^([0-9]{1,3}\.){3}[0-9]{1,3}$/', $_SERVER['HTTP_CF_CONNECTING_IP'])) {
        $ip = $_SERVER['HTTP_CF_CONNECTING_IP'];
    } elseif (isset($_SERVER['HTTP_X_REAL_IP']) && preg_match('/^([0-9]{1,3}\.){3}[0-9]{1,3}$/', $_SERVER['HTTP_X_REAL_IP'])) {
        $ip = $_SERVER['HTTP_X_REAL_IP'];
    }
    return $ip;

}
echo real_ip();

?> ywarnier at beeznest dot org Note that $_SERVER['REQUEST_URI'] might include the scheme and domain in certain cases.

This happens, for example, when calling the page through a call to stream_context_create() with a HTTP header of 'request_fulluri' set to 1.

For example:

$http = ['request_fulluri' => 1, /* other params here */];
$context = stream_context_create(array( 'http' => $http ));
$fp = fopen($some_url, 'rb', false, $context);

When outputting $_SERVER['REQUEST_URI'] on the server at $some_url, you will get
https://some_url/some_script.php

Remove the request_fulluri => 1 option, and $_SERVER['REQUEST_URI'] gets back to its "normal":
/some_script.php

Apparently, request_fulluri is useful when using some proxy servers.

In this case, there is no proper way to "detect" if this option was set or not, and you should probably use a combination of other $_SERVER[] elements (like REQUEST_SCHEME, SERVER_NAME and SERVER_PORT) to determine if this was the case.

One quick (and improvable) way to detect it would be to compare the start of the REQUEST_URI with REQUEST_SCHEME:

$scheme = $_SERVER['REQUEST_SCHEME'] . '://';
if (strcmp(substr($_SERVER['REQUEST_URI'], 0, strlen($scheme)), $scheme) === 0) {
    // request_fulluri was set
} centurianii at yahoo dot co dot uk If you apply redirection in ALL your requests using commands at the Apache virtual host file like:
RewriteEngine On
RewriteCond "%{REQUEST_URI}" "!=/index.php"
RewriteRule "^/(.*)$" "index.php?$1" [NC,NE,L,QSA]
you should expect some deviations in your $_SERVER global.

Say, you send a url of: [hostname here]/a/b?x=1&y=2
which makes Apache to modify to: /index.php?/a/b?x=1&y=2

Now your $_SERVER global contains among others:
'REQUEST_URI' => '/a/b?x=1&y=2', it retains the initial url after the host
 'QUERY_STRING' => 'a/b&x=1&y=2', notice how php replaces '?' with '&'
 'SCRIPT_NAME' => '/index.php', as it was intended to be.

To test your $_SERVER global:
function serverArray(){
   $arr = array();
   foreach($_SERVER as $key=>$value)
      $arr[] = '&nbsp;&nbsp;&nbsp;\'' . $key . '\' => \'' . (isset($value)? $value : '-') . '\'';
   return @\sort($arr)? '$_SERVER = array(<br />' . implode($arr, ',<br />') . '<br />);' : false;
}
echo serverArray(); pierstoval at example dot com As PHP $_SERVER var is populated with a lot of vars, I think it's important to say that it's also populated with environment vars.

For example, with a PHP script, we can have this:

    MY_ENV_VAR=Hello php -r 'echo $_SERVER["MY_ENV_VAR"];'
   
Will show "Hello".

But, internally, PHP makes sure that "internal" keys in $_SERVER are not overriden, so you wouldn't be able to do something like this:

    REQUEST_TIME=Hello php -r 'var_dump($_SERVER["REQUEST_TIME"]);'
   
Will show something like 1492897785

However, a lot of vars are still vulnerable from environment injection.

I created a gist here ( https://gist.github.com/Pierstoval/f287d3e61252e791a943dd73874ab5ee ) with my PHP configuration on windows with PHP7.0.15 on WSL with bash, the results are that the only "safe" vars are the following:

PHP_SELF
SCRIPT_NAME
SCRIPT_FILENAME
PATH_TRANSLATED
DOCUMENT_ROOT
REQUEST_TIME_FLOAT
REQUEST_TIME
argv
argc

All the rest can be overriden with environment vars, which is not very cool actually because it can break PHP applications sometimes...

(and I only tested on CLI, I had no patience to test with Apache mod_php or Nginx + PHP-FPM, but I can imagine that not a lot of $_SERVER properties are "that" secure...) chris at ocproducts dot com Guide to script parameters...

Data: $_GET
Data type: Array (map)
Purpose: Contains all GET parameters (i.e. a parsed URL query string).
Caveat: GET parameter names have to be compliant with PHP variable naming, e.g. dots are not allowed and get substituted.
Works on web mode: Yes
Works on CLI mode: No

Data: $_SERVER['QUERY_STRING']
Data type: String
Purpose: Gets an unparsed URL query string.
Caveat: Not set on all PHP environments, may need setting via http_build_query($_GET).
Works on web mode: Yes
Works on CLI mode: No

Data: $_SERVER['argv']
Data type: Array (list)
Purpose: Get CLI call parameters.
Works on web mode: Tenuous (just contains a single parameter, the query string)
Works on CLI mode: Yes chris at ocproducts dot com Guide to URL paths...

Data: $_SERVER['PHP_SELF']
Data type: String
Purpose: The URL path name of the current PHP file, including path-info (see $_SERVER['PATH_INFO']) and excluding URL query string. Includes leading slash.
Caveat: This is after URL rewrites (i.e. it's as seen by PHP, not necessarily the original call URL).
Works on web mode: Yes
Works on CLI mode: Tenuous (emulated to contain just the exact call path of the CLI script, with whatever exotic relative pathname you may call with, not made absolute and not normalised or pre-resolved)

Data: $_SERVER['SCRIPT_NAME']
Data type: String
Purpose: The URL path name of the current PHP file, excluding path-info and excluding URL query string. Includes leading slash.
Caveat: This is after URL rewrites (i.e. it's as seen by PHP, not necessarily the original call URL).
Caveat: Not set on all PHP environments, may need setting via preg_replace('#\.php/.*#', '.php', $_SERVER['PHP_SELF']).
Works on web mode: Yes
Works on CLI mode: Tenuous (emulated to contain just the exact call path of the CLI script, with whatever exotic relative pathname you may call with, not made absolute and not normalised or pre-resolved)

Data: $_SERVER['REDIRECT_URL']
Data type: String
Purpose: The URL path name of the current PHP file, path-info is N/A and excluding URL query string. Includes leading slash.
Caveat: This is before URL rewrites (i.e. it's as per the original call URL).
Caveat: Not set on all PHP environments, and definitely only ones with URL rewrites.
Works on web mode: Yes
Works on CLI mode: No

Data: $_SERVER['REQUEST_URI']
Data type: String
Purpose: The URL path name of the current PHP file, including path-info and including URL query string. Includes leading slash.
Caveat: This is before URL rewrites (i.e. it's as per the original call URL). *
*: I've seen at least one situation where this is not true (there was another $_SERVER variable to use instead supplied by the URL rewriter), but the author of the URL rewriter later fixed it so probably fair to dismiss this particular note.
Caveat: Not set on all PHP environments, may need setting via $_SERVER['REDIRECT_URL'] . '?' . http_build_query($_GET) [if $_SERVER['REDIRECT_URL'] is set, and imperfect as we don't know what GET parameters were originally passed vs which were injected in the URL rewrite] --otherwise-- $_SERVER['PHP_SELF'] . '?' . http_build_query($_GET).
Works on web mode: Yes
Works on CLI mode: No

Data: $_SERVER['PATH_INFO']
Data type: String
Purpose: Find the path-info, which is data after the .php filename in the URL call. It's a strange concept.
Caveat: Some environments may not support it, it is best avoided unless you have complete server control
Works on web mode: Yes
Works on CLI mode: No

Note that if something is not set it may be missing from $_SERVER, or it may be blank, so use PHP's 'empty' function for your test. chris at ocproducts dot com Guide to absolute paths...

Data: __FILE__
Data type: String
Purpose: The absolute pathname of the running PHP file, including the filename.
Caveat: This is not the file called by the PHP processor, it's what is running. So if you are inside an include, it's the include.
Caveat: Symbolic links are pre-resolved, so don't trust comparison of paths to be accurate.
Caveat: Don't assume all operating systems use '/' for the directory separator.
Works on web mode: Yes
Works on CLI mode: Yes

Data: __DIR__
Data type: String
Purpose: The absolute pathname to the running PHP file, excluding the filename
Caveat: This is not the file called by the PHP processor, it's what is running. So if you are inside an include, it's the include.
Caveat: Symbolic links are pre-resolved, so don't trust comparison of paths to be accurate.
Caveat: Don't assume all operating systems use '/' for the directory separator.
Works on web mode: Yes
Works on CLI mode: Yes

Data: $_SERVER['SCRIPT_FILENAME']
Data type: String
Purpose: The absolute pathname of the origin PHP file, including the filename
Caveat: Not set on all PHP environments, may need setting by copying from __FILE__ before other files are included.
Caveat: Symbolic links are not pre-resolved, use PHP's 'realpath' function if you need it resolved.
Caveat: Don't assume all operating systems use '/' for the directory separator.
Caveat: "Filename" makes you think it is just a filename, but it really is the full absolute pathname. Read the identifier as "Script's filesystem (path)name".
Works on web mode: Yes
Works on CLI mode: Yes

Data: $_SERVER['PATH_TRANSLATED']
Data type: String
Purpose: The absolute pathname of the origin PHP file, including the filename
Caveat: It's probably not set, best to just not use it. Just use realpath($_SERVER['SCRIPT_FILENAME']) (and be aware that itself may need to have been emulated).
Caveat: Symbolic links are pre-resolved, so don't trust comparison of paths to be accurate.
Caveat: Don't assume all operating systems use '/' for the directory separator.
Works on web mode: Yes
Works on CLI mode: No

Data: $_SERVER['DOCUMENT_ROOT']
Data type: String
Purpose: Get the absolute path to the web server's document root. No trailing slash.
Caveat: Don't trust this to be set, or set correctly, unless you control the server environment.
Caveat: May or may not have symbolic links pre-resolved, use PHP's 'realpath' function if you need it resolved.
Caveat: Don't assume all operating systems use '/' for the directory separator.
Works on web mode: Yes
Works on CLI mode: No

Note that if something is not set it may be missing from $_SERVER, or it may be blank, so use PHP's 'empty' function for your test.

Note that if you call "php --info" on the command line then naturally some of these settings are going to be blank, as no PHP file is involved. plugwash at p10link dot net Be aware that it's a bad idea to access x-forwarded-for and similar headers through this array. The header names are mangled when populating the array and this mangling can introduce spoofing vulnerabilities.

See jit_chavan at yahoo dot com searched $_SERVER["REDIRECT_URL"] for a while and noted that it is not mentioned in php documentation page itself. look like this is only generated by apache server(not others) and using   $_SERVER["REQUEST_URI"] will be useful in some cases as mine. wyattstorch42 at outlook dot com <?php
/*
 * I wrote this because I was including a file with classes in it. Let's say that
 * I have a contact page at mysite.com/contact/index.php and a Form class at
 * mysite.com/classes/Form.php. So in index.php, I have this statement:
 * require '../classes/Form.php';
 * The Form class includes a method to generate the HTML markup for a number of
 * form elements, including a CAPTCHA image and associated text field. To do so,
 * it must generate an <img /> element and give it a src of Form.php?captcha.
 * But I wanted it to automatically generate a src attribute without index.php
 * giving it a relative path. This script comes in handy by automatically
 * locating the directory that contains the included file (Form.php) and converting
 * it from an absolute path to a relative path that could be used for an img src,
 * an a href, a link href, etc.
 */
function relativeURL () {
    $dir = str_replace('\\', '/', __DIR__);
        // Resolves inconsistency with PATH_SEPARATOR on Windows vs. Linux
        // Use dirname(__FILE__) in place of __DIR__ for older PHP versions
    return substr($dir, strlen($_SERVER['DOCUMENT_ROOT']));
        // Clip off the part of the path outside of the document root
}

/*
 *contact/index.php
 */
require '../classes/Form.php';
new Form()->drawCaptchaField();
    // Writes: <img src="/classes/form.php?captcha" />

   
/*
 * classes/Form.php
 */
if (isset($_GET['captcha'])) {
    // generate/return CAPTCHA image
}

class Form {
    // ...
    public function drawCaptchaField () {
        echo '<img src="'.relativeURL().'?captcha" />';
    }
}
?> pomat at live dot it $_SERVER['DOCUMENT_ROOT'] may contain backslashes on windows systems, and of course it may or may not have a trailing slash (backslash).
I saw the following as an example of the proper way we're supposed to deal with this issue:

<?php
include(dirname($_SERVER['DOCUMENT_ROOT']) . DIRECTORY_SEPARATOR . 'file.php');
?>

Ok, the latter may be used to access a file inside the parent directory of the document root, but actually does not properly address the issue.
In the end, don't warry about. It should be safe to use forward slashes and append a trailing slash in all cases.
Let's say we have this:

<?php
$path = 'subdir/file.php';
$result = $_SERVER['DOCUMENT_ROOT'] . '/' . $path;
?>

On linux $result might be something like
1) "/var/www/subdir/file.php"
2) "/var/www//subdir/file.php"
String 2 is parsed the same as string 1 (have a try with command 'cd').

On windows $result might be something like
1) "C:/apache/htdocs/subdir/file.php"
2) "C:/apache/htdocs//subdir/file.php"
3) "C:\apache\htdocs/subdir/file.php"
4) "C:\apache\htdocs\/subdir/file.php"
All those strings are parsed as "C:\apache\htdocs\subdir\file.php" (have a try with 'cd'). zeufonlinux at gmail dot com Just a PHP file to put on your local server (as I don't have enough memory)

 <?php
 $indicesServer = array('PHP_SELF',
'argv',
'argc',
'GATEWAY_INTERFACE',
'SERVER_ADDR',
'SERVER_NAME',
'SERVER_SOFTWARE',
'SERVER_PROTOCOL',
'REQUEST_METHOD',
'REQUEST_TIME',
'REQUEST_TIME_FLOAT',
'QUERY_STRING',
'DOCUMENT_ROOT',
'HTTP_ACCEPT',
'HTTP_ACCEPT_CHARSET',
'HTTP_ACCEPT_ENCODING',
'HTTP_ACCEPT_LANGUAGE',
'HTTP_CONNECTION',
'HTTP_HOST',
'HTTP_REFERER',
'HTTP_USER_AGENT',
'HTTPS',
'REMOTE_ADDR',
'REMOTE_HOST',
'REMOTE_PORT',
'REMOTE_USER',
'REDIRECT_REMOTE_USER',
'SCRIPT_FILENAME',
'SERVER_ADMIN',
'SERVER_PORT',
'SERVER_SIGNATURE',
'PATH_TRANSLATED',
'SCRIPT_NAME',
'REQUEST_URI',
'PHP_AUTH_DIGEST',
'PHP_AUTH_USER',
'PHP_AUTH_PW',
'AUTH_TYPE',
'PATH_INFO',
'ORIG_PATH_INFO') ;

echo '<table >' ;
foreach ($indicesServer as $arg) {
    if (isset($_SERVER[$arg])) {
        echo '<tr><td>'.$arg.'</td><td>' . $_SERVER[$arg] . '</td></tr>' ;
    }
    else {
        echo '<tr><td>'.$arg.'</td><td>-</td></tr>' ;
    }
}
echo '</table>' ;

/*

That will give you the result of each variable like (if the file is server_indices.php at the root and Apache Web directory is in E:\web) :

PHP_SELF    /server_indices.php
argv    -
argc    -
GATEWAY_INTERFACE    CGI/1.1
SERVER_ADDR    127.0.0.1
SERVER_NAME    localhost
SERVER_SOFTWARE    Apache/2.2.22 (Win64) PHP/5.3.13
SERVER_PROTOCOL    HTTP/1.1
REQUEST_METHOD    GET
REQUEST_TIME    1361542579
REQUEST_TIME_FLOAT    -
QUERY_STRING   
DOCUMENT_ROOT    E:/web/
HTTP_ACCEPT    text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
HTTP_ACCEPT_CHARSET    ISO-8859-1,utf-8;q=0.7,*;q=0.3
HTTP_ACCEPT_ENCODING    gzip,deflate,sdch
HTTP_ACCEPT_LANGUAGE    fr-FR,fr;q=0.8,en-US;q=0.6,en;q=0.4
HTTP_CONNECTION    keep-alive
HTTP_HOST    localhost
HTTP_REFERER    http://localhost/
HTTP_USER_AGENT    Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.17 (KHTML, like Gecko) Chrome/24.0.1312.57 Safari/537.17
HTTPS    -
REMOTE_ADDR    127.0.0.1
REMOTE_HOST    -
REMOTE_PORT    65037
REMOTE_USER    -
REDIRECT_REMOTE_USER    -
SCRIPT_FILENAME    E:/web/server_indices.php
SERVER_ADMIN    myemail@personal.us
SERVER_PORT    80
SERVER_SIGNATURE   
PATH_TRANSLATED    -
SCRIPT_NAME    /server_indices.php
REQUEST_URI    /server_indices.php
PHP_AUTH_DIGEST    -
PHP_AUTH_USER    -
PHP_AUTH_PW    -
AUTH_TYPE    -
PATH_INFO    -
ORIG_PATH_INFO    -

*/
?> krinklemail at gmail dot com If requests to your PHP script send a header "Content-Type" or/ "Content-Length" it will, contrary to regular HTTP headers, not appear in $_SERVER as $_SERVER['HTTP_CONTENT_TYPE']. PHP removes these (per CGI/1.1 specification[1]) from the HTTP_ match group.

They are still accessible, but only if the request was a POST request. When it is, it'll be available as:
$_SERVER['CONTENT_LENGTH']
$_SERVER['CONTENT_TYPE']

[1] https://www.ietf.org/rfc/rfc3875 Tom Be warned that most contents of the Server-Array (even $_SERVER['SERVER_NAME']) are provided by the client and can be manipulated. They can also be used for injections and thus MUST be checked and treated like any other user input. picov at e-link dot it A simple function to detect if the current page address was rewritten by mod_rewrite:

<?php
public function urlWasRewritten() {
  $realScriptName=$_SERVER['SCRIPT_NAME'];
  $virtualScriptName=reset(explode("?", $_SERVER['REQUEST_URI']));
  return !($realScriptName==$virtualScriptName);
}
?> MarkAgius at markagius dot co dot uk You have missed 'REDIRECT_STATUS'

Very useful if you point all your error pages to the same file.

File; .htaccess
# .htaccess file.

ErrorDocument 404 /error-msg.php
ErrorDocument 500 /error-msg.php
ErrorDocument 400 /error-msg.php
ErrorDocument 401 /error-msg.php
ErrorDocument 403 /error-msg.php
# End of file.

File; error-msg.php
<?php
  $HttpStatus = $_SERVER["REDIRECT_STATUS"] ;
  if($HttpStatus==200) {print "Document has been processed and sent to you.";}
  if($HttpStatus==400) {print "Bad HTTP request ";}
  if($HttpStatus==401) {print "Unauthorized - Iinvalid password";}
  if($HttpStatus==403) {print "Forbidden";}
  if($HttpStatus==500) {print "Internal Server Error";}
  if($HttpStatus==418) {print "I'm a teapot! - This is a real value, defined in 1998";}

?> rulerof at gmail dot com I needed to get the full base directory of my script local to my webserver, IIS 7 on Windows 2008.

I ended up using this:

<?php
function GetBasePath() {
    return substr($_SERVER['SCRIPT_FILENAME'], 0, strlen($_SERVER['SCRIPT_FILENAME']) - strlen(strrchr($_SERVER['SCRIPT_FILENAME'], "\\")));
}
?>

And it returned C:\inetpub\wwwroot\<applicationfolder> as I had hoped. Stefano (info at sarchittu dot org) A way to get the absolute path of your page, independent from the site position (so works both on local machine and on server without setting anything) and from the server OS (works both on Unix systems and Windows systems).

The only parameter it requires is the folder in which you place this script
So, for istance, I'll place this into my SCRIPT folder, and I'll write SCRIPT word length in $conflen

<?php
$conflen=strlen('SCRIPT');
$B=substr(__FILE__,0,strrpos(__FILE__,'/'));
$A=substr($_SERVER['DOCUMENT_ROOT'], strrpos($_SERVER['DOCUMENT_ROOT'], $_SERVER['PHP_SELF']));
$C=substr($B,strlen($A));
$posconf=strlen($C)-$conflen-1;
$D=substr($C,1,$posconf);
$host=''.$_SERVER['SERVER_NAME'].'/'.$D;
?>

$host will finally contain the absolute path. dtomasiewicz at gmail dot com To get an associative array of HTTP request headers formatted similarly to get_headers(), this will do the trick:

<?php
/**
 * Transforms $_SERVER HTTP headers into a nice associative array. For example:
 *   array(
 *       'Referer' => 'example.com',
 *       'X-Requested-With' => 'XMLHttpRequest'
 *   )
 */
function get_request_headers() {
    $headers = array();
    foreach($_SERVER as $key => $value) {
        if(strpos($key, 'HTTP_') === 0) {
            $headers[str_replace(' ', '-', ucwords(str_replace('_', ' ', strtolower(substr($key, 5)))))] = $value;
        }
    }
    return $headers;
}
?> wbeaumo1 at gmail dot com Don't forget $_SERVER['HTTP_COOKIE']. It contains the raw value of the 'Cookie' header sent by the user agent. php at isnoop dot net Use the apache SetEnv directive to set arbitrary $_SERVER variables in your vhost or apache config.

SetEnv varname "variable value" mirko dot steiner at slashdevslashnull dot de <?php

// RFC 2616 compatible Accept Language Parser
//
// Hypertext Transfer Protocol -- HTTP/1.1

foreach (explode(',', $_SERVER['HTTP_ACCEPT_LANGUAGE']) as $lang) {
    $pattern = '/^(?P<primarytag>[a-zA-Z]{2,8})'.
    '(?:-(?P<subtag>[a-zA-Z]{2,8}))?(?:(?:;q=)'.
    '(?P<quantifier>\d\.\d))?$/';

    $splits = array();

    printf("Lang:,,%s''\n", $lang);
    if (preg_match($pattern, $lang, $splits)) {
        print_r($splits);
    } else {
        echo "\nno match\n";
    }
}

?>

example output:

Google Chrome 3.0.195.27 Windows xp

Lang:,,de-DE''
Array
(
    [0] => de-DE
    [primarytag] => de
    [1] => de
    [subtag] => DE
    [2] => DE
)
Lang:,,de;q=0.8''
Array
(
    [0] => de;q=0.8
    [primarytag] => de
    [1] => de
    [subtag] =>
    [2] =>
    [quantifier] => 0.8
    [3] => 0.8
)
Lang:,,en-US;q=0.6''
Array
(
    [0] => en-US;q=0.6
    [primarytag] => en
    [1] => en
    [subtag] => US
    [2] => US
    [quantifier] => 0.6
    [3] => 0.6
)
Lang:,,en;q=0.4''
Array
(
    [0] => en;q=0.4
    [primarytag] => en
    [1] => en
    [subtag] =>
    [2] =>
    [quantifier] => 0.4
    [3] => 0.4
) Lord Mac An even *more* improved version...

<?php
phpinfo(32);
?> steve at sc-fa dot com If you are serving from behind a proxy server, you will almost certainly save time by looking at what these $_SERVER variables do on your machine behind the proxy.  

$_SERVER['HTTP_X_FORWARDED_FOR'] in place of $_SERVER['REMOTE_ADDR']

$_SERVER['HTTP_X_FORWARDED_HOST'] and
$_SERVER['HTTP_X_FORWARDED_SERVER'] in place of (at least in our case,) $_SERVER['SERVER_NAME'] cupy at email dot cz Tech note:
$_SERVER['argc'] and $_SERVER['argv'][] has some funny behaviour,
used from linux (bash) commandline, when called like
"php ./script_name.php 0x020B"
there is everything correct, but
"./script_name.php 0x020B"
is not correct - "0" is passed instead of "0x020B" as $_SERVER['argv'][1] - see the script below.
Looks like the parameter is not passed well from bash to PHP.
(but, inspected on the level of bash, 0x020B is understood well as $1)

try this example:

------------->8------------------
cat ./script_name.php
#! /usr/bin/php

if( $_SERVER['argc'] == 2)
  {
    // funny... we have to do this trick to pass e.g. 0x020B from parameters
    // ignore this: "PHP Notice:  Undefined offset:  2 in ..."
    $EID = $_SERVER['argv'][1] + $_SERVER['argv'][2] + $_SERVER['argv'][3];
  }
 else
   {        // default
     $EID = 0x0210; // PPS failure
   } jarrod at squarecrow dot com $_SERVER['DOCUMENT_ROOT'] is incredibly useful especially when working in your development environment. If you're working on large projects you'll likely be including a large number of files into your pages. For example:

<?php
//Defines constants to use for "include" URLS - helps keep our paths clean

        define("REGISTRY_CLASSES",  $_SERVER['DOCUMENT_ROOT']."/SOAP/classes/");
        define("REGISTRY_CONTROLS", $_SERVER['DOCUMENT_ROOT']."/SOAP/controls/");

        define("STRING_BUILDER",     REGISTRY_CLASSES. "stringbuilder.php");
        define("SESSION_MANAGER",     REGISTRY_CLASSES. "sessionmanager.php");
        define("STANDARD_CONTROLS",    REGISTRY_CONTROLS."standardcontrols.php");
?>

In development environments, you're rarely working with your root folder, especially if you're running PHP locally on your box and using DOCUMENT_ROOT is a great way to maintain URL conformity. This will save you hours of work preparing your application for deployment from your box to a production server (not to mention save you the headache of include path failures). Richard York Not documented here is the fact that $_SERVER is populated with some pretty useful information when accessing PHP via the shell.

 ["_SERVER"]=>
  array(24) {
    ["MANPATH"]=>
    string(48) "/usr/share/man:/usr/local/share/man:/usr/X11/man"
    ["TERM"]=>
    string(11) "xterm-color"
    ["SHELL"]=>
    string(9) "/bin/bash"
    ["SSH_CLIENT"]=>
    string(20) "127.0.0.1 41242 22"
    ["OLDPWD"]=>
    string(60) "/Library/WebServer/Domains/www.example.com/private"
    ["SSH_TTY"]=>
    string(12) "/dev/ttys000"
    ["USER"]=>
    string(5) "username"
    ["MAIL"]=>
    string(15) "/var/mail/username"
    ["PATH"]=>
    string(57) "/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin"
    ["PWD"]=>
    string(56) "/Library/WebServer/Domains/www.example.com/www"
    ["SHLVL"]=>
    string(1) "1"
    ["HOME"]=>
    string(12) "/Users/username"
    ["LOGNAME"]=>
    string(5) "username"
    ["SSH_CONNECTION"]=>
    string(31) "127.0.0.1 41242 10.0.0.1 22"
    ["_"]=>
    string(12) "/usr/bin/php"
    ["__CF_USER_TEXT_ENCODING"]=>
    string(9) "0x1F5:0:0"
    ["PHP_SELF"]=>
    string(10) "Shell.php"
    ["SCRIPT_NAME"]=>
    string(10) "Shell.php"
    ["SCRIPT_FILENAME"]=>
    string(10) "Shell.php"
    ["PATH_TRANSLATED"]=>
    string(10) "Shell.php"
    ["DOCUMENT_ROOT"]=>
    string(0) ""
    ["REQUEST_TIME"]=>
    int(1247162183)
    ["argv"]=>
    array(1) {
      [0]=>
      string(10) "Shell.php"
    }
    ["argc"]=>
    int(1)
  } chris A table of everything in the $_SERVER array can be found near the bottom of the output of phpinfo(); pudding06 at gmail dot com Here's a simple, quick but effective way to block unwanted external visitors to your local server:

<?php
// only local requests
if ($_SERVER['REMOTE_ADDR'] !== '127.0.0.1') die(header("Location: /"));
?>

This will direct all external traffic to your home page. Of course you could send a 404 or other custom error. Best practice is not to stay on the page with a custom error message as you acknowledge that the page does exist. That's why I redirect unwanted calls to (for example) phpmyadmin. Vladimir Kornea 1. All elements of the $_SERVER array whose keys begin with 'HTTP_' come from HTTP request headers and are not to be trusted.

2. All HTTP headers sent to the script are made available through the $_SERVER array, with names prefixed by 'HTTP_'.

3. $_SERVER['PHP_SELF'] is dangerous if misused. If login.php/nearly_arbitrary_string is requested, $_SERVER['PHP_SELF'] will contain not just login.php, but the entire login.php/nearly_arbitrary_string. If you've printed $_SERVER['PHP_SELF'] as the value of the action attribute of your form tag without performing HTML encoding, an attacker can perform XSS attacks by offering users a link to your site such as this:

<a href='"><script type="text/javascript">...</script><span a="'>Example.com</a>

The javascript block would define an event handler function and bind it to the form's submit event. This event handler would load via an <img> tag an external file, with the submitted username and password as parameters.

Use $_SERVER['SCRIPT_NAME'] instead of $_SERVER['PHP_SELF']. HTML encode every string sent to the browser that should not be interpreted as HTML, unless you are absolutely certain that it cannot contain anything that the browser can interpret as HTML. info at mtprod dot com On Windows IIS 7 you must use $_SERVER['LOCAL_ADDR'] rather than $_SERVER['SERVER_ADDR'] to get the server's IP address. jonbarnett at gmail dot com It's worth noting that $_SERVER variables get created for any HTTP request headers, including those you might invent:

If the browser sends an HTTP request header of:
X-Debug-Custom: some string

Then:

<?php
$_SERVER['HTTP_X_DEBUG_CUSTOM']; // "some string"
?>

There are better ways to identify the HTTP request headers sent by the browser, but this is convenient if you know what to expect from, for example, an AJAX script with custom headers.

Works in PHP5 on Apache with mod_php.  Don't know if this is true from other environments. jette at nerdgirl dot dk Windows running IIS v6 does not include $_SERVER['SERVER_ADDR']

If you need to get the IP addresse, use this instead:

<?php
$ipAddress = gethostbyname($_SERVER['SERVER_NAME']);
?> Tonin When using the $_SERVER['SERVER_NAME'] variable in an apache virtual host setup with a ServerAlias directive, be sure to check the UseCanonicalName apache directive.  If it is On, this variable will always have the apache ServerName value.  If it is Off, it will have the value given by the headers sent by the browser.

Depending on what you want to do the content of this variable, put in On or Off. silverquick at gmail dot com I think the HTTPS element will only be present under Apache 2.x. It's not in the list of "special" variables here:

But it is here: