Return current web path in PHP

Question

Hi All

Currently developing a PHP framework and have ran into my first problem. I need to be able to drop the framework into any folder on a server, no matter how many folders deep, and need to find that directory to use as a base URL.

For example, it currently works if I put the framework in the root of the server (http://cms.dev/), but if I were to put it in http://cms.dev/folder/ it does not work.

Please advise, any comments welcome.

BenTheDesigner

Answer 1

Check out PHP's dirname() function.

Answer 2

To get the current path of the file you must use:

$path=getcwd();

This will return you if in windows for example C:\blah\blah\blah with no file name.

Answer 3

__FILE__ is a magic constant that returns the entire path of the current script. Combine with dirname and add ".." appropriately. It's more reliable than getcwd, since it cannot change during execution.

You can then strip off the web root, to get the relative path to your script (should map to URL). There are many $_SERVER variables that have this information. I suggest using the file system to determine:

1. If your script is publicly accessible?
2. At which depth / URL prefix?

Then combine with your base URL. If your script's path == /home/public/foo_1/script.php ... and your $_SERVER['DOCUMENT_ROOT'] == /home/public

Then you can rewrite your URL as /foo_1/script.php. You don't need the fully qualified URL, unless you want it. This technique works best if you execute it from a central location, like an autoloader.

Answer 4

Theres a bunch of useful stuff available for things like this in the $_SERVER array. Do a var_dump($_SERVER); to see which element(s) of the array you need.

Answer 5

Sounds like you are looking for the relative path.

$_SERVER['SCRIPT_FILENAME']

should do the trick. The php.net site has good documentation on what is available to that http://php.net/manual/en/reserved.variables.server.php

Also, if you are ever curious about what else is there, and what the values are

<?php print_r($_SERVER);  ?>  

will tell you more that you thought you could know.

Answer 6

dirname(__FILE__);
basename(__FILE__);
print_r($_SERVER);
pathinfo('www/htdocs/index.html');
realpath('../../dir1/');
parse_url('http://username:password@hostname/path?arg=value#anchor');

Answer 7

In order to make urls work check the base tag:

<base href="http://cms.dev/folder/" />

If the PHP file paths are the issue go with Pestilance's advice:

dirname(__FILE__) // returns the directory of current file

Answer 8

There are four existing answers, but they all seem to deal with file paths, and you're asking about a base URL for web requests.

Given any web request, you get a bunch of keys in $_SERVER that may be helpful. For example, in your mock example, you might have the following:

• http://cms.dev/folder/ — $_SERVER['REQUEST_URI'] == /folder/
• http://cms.dev/folder/index.php — $_SERVER['REQUEST_URI'] == /folder/index.php
• http://cms.dev/folder/index.php/some/pathinfo — $_SERVER['REQUEST_URI'] == /folder/index.php/some/pathinfo
• http://cms.dev/folder/some/modrewrite — $_SERVER['REQUEST_URI'] == /folder/some/modrewrite

Thinking critically, how would you pull out the base URL for any given subrequest? In certain cases you can look at $_SERVER['REQUEST_URI'] and strip off trailing elements if you know how deep in your hierarchy the request is. (For example, if your script is two folders deep, strip off the last two path elements.) When PATH_INFO or mod_rewrite are in use, things become less clear: as longer and longer URLs are provided, there is no clear indication where the paths end and the dynamic URL begins.

This is why WordPress, MediaWiki, phpBB, phpMyAdmin, and every application I've ever written has the user manually specify a base URL as part of the application configuration.