# p3k-http
A simple HTTP client, used by https://p3k.io projects.
## Usage
### GET
```php
$http = new p3k\HTTP();
$headers = [
'Accept: text/html, */*'
];
$response = $http->get('http://example.com/', $headers);
```
### POST
```php
$http = new p3k\HTTP();
$headers = [
'Accept: application/json',
'Content-type: application/json'
];
$response = $http->post('http://example.com/', json_encode([
'foo' => 'bar'
], $headers);
```
### HEAD
```php
$http = new p3k\HTTP();
$headers = [
'Accept: text/html, */*'
];
$response = $http->head('http://example.com/', $headers);
```
### Safe mode
When the URLs you fetch come from other people (a user's website, a
discovered endpoint, a link in a post), turn on safe mode so they cannot
point your server at itself or at your private network:
```php
$http = new p3k\HTTP();
$http->set_safe_mode(true);
```
In safe mode:
* only `http` and `https` URLs are fetched (no `file`, `gopher`, `dict`, …)
* the host must resolve only to public addresses; loopback, private, link-local
and other reserved ranges are refused, as are hosts written as unusual IP
forms like `2130706433` or `0x7f.1`
* curl connects to exactly the addresses that were checked, so a DNS answer
that changes between the check and the request makes no difference
* redirects are followed one at a time and every hop is checked the same way;
`Authorization`, `Proxy-Authorization` and `Cookie` headers are dropped when a
redirect goes to a different origin
A refused request makes no connection and returns `code` 0 with `error` set to
`blocked_url` (or `dns_error` when the host does not resolve).
To reach a private server on purpose, such as a development site on your LAN,
allow it by hostname, address or CIDR range:
```php
$http->set_safe_mode(true, ['dev.example.com', '10.11.11.0/24']);
```
Custom transports get the same checks. A transport that implements
`p3k\HTTP\Pinnable` is also held to the checked addresses, as the curl transport is.
### Response
The get/post/head functions will return an array with the following properties:
* `code` - integer, the HTTP response code that was returned
* `headers` - array, the HTTP headers returned
* `rels` - array, the parsed HTTP rel values from any `Link` headers
* `body` - string, the body of the HTTP response, or false/omit for a HEAD request
* `error` - string, an error string. see below for the enumerated list.
* `error_description` - string,
* `url` - string, the final URL retrieved after following any redirects
* `debug` - string, the full HTTP response
#### `headers`
The `headers` key will be an array of all the header values returned. The values will be either a string or array depending on whether there were multiple values returned for a given header name.
```
[headers] => Array
(
[Server] => nginx/1.12.0
[Content-Type] => text/html; charset=UTF-8
[Transfer-Encoding] => chunked
[Connection] => keep-alive
[Cache-Control] => no-cache
[Link] => Array
(
[0] => ; rel="hub"
[1] => ; rel="authorization_endpoint"
[2] => ; rel="micropub"
[3] => ; rel="token_endpoint"
[4] => ; rel="self"
)
[Date] => Fri, 28 Apr 2017 18:40:42 GMT
[Strict-Transport-Security] => max-age=2592000
[X-No-Cache] => 0
[X-Cache] => EXPIRED
)
```
#### `rels`
The `rels` key will be the parsed version of any HTTP `Link` headers that contain a rel value. All values will be arrays even if there is only one value.
```
[rels] => Array
(
[hub] => Array
(
[0] => https://switchboard.p3k.io/
)
[authorization_endpoint] => Array
(
[0] => https://aaronparecki.com/auth
)
[micropub] => Array
(
[0] => https://aaronparecki.com/micropub
)
[token_endpoint] => Array
(
[0] => https://aaronparecki.com/auth/token
)
[self] => Array
(
[0] => https://aaronparecki.com/
)
)
```
### Options
There are a few options you can set when making HTTP requests, to configure how the request will behave.
* `$http->set_user_agent('String')` - A shortcut for setting the user agent header.
* `$http->set_max_redirects(2)` - The maximum number of redirects to follow. Defaults to 8.
* `$http->set_timeout(10)` - The timeout in seconds for waiting for a response. Defaults to 4.
#### User Agent
You can set the user agent when you first instantiate the HTTP object:
```php
$http = new p3k\HTTP('Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36 p3k-http/0.1.0');
$http->get('http://example.com/');
```
Alternately, you can change it before each request:
```php
$http = new p3k\HTTP();
$http->set_user_agent('Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/57.0.2987.133 Safari/537.36 p3k-http/0.1.0');
$http->get('http://example.com/');
```
## Transports
By default, the library will use the PHP curl functions to make the HTTP request.
You can optionally define your own transport to use to make HTTP requests instead. This allows you to use an existing HTTP request mechanism you may have rather than using curl or the PHP stream functions.
Define a new class that implements the `p3k\HTTP\Transport` interface. That interface documents the return values expected as well.
Then you can set the transport after you create the main HTTP object.
```php
$http = new p3k\HTTP();
$http->set_transport(new p3k\HTTP\Stream()); // or your custom class here
```
The library ships with two alternative transport mechanisms, `p3k\HTTP\Stream` and `p3k\HTTP\Test`. The Stream transport uses `file_get_contents` with all the necessary config options to make it work. This is useful when running in Google App Engine. The Test transport will read HTTP responses from files, so that you can write tests that simulate making HTTP calls. See https://github.com/aaronpk/XRay/tree/master/tests for an example of using this transport.
## License
Copyright 2017 by Aaron Parecki
Available under the MIT license.