PHP Classes

File: includes/src/vendor/rmccue/requests/docs/why-requests.md

Recommend this page to a friend!
  Classes of Subin Siby   Lobby   includes/src/vendor/rmccue/requests/docs/why-requests.md   Download  
File: includes/src/vendor/rmccue/requests/docs/why-requests.md
Role: Auxiliary data
Content type: text/markdown
Description: Auxiliary data
Class: Lobby
Web OS to install and run Web applications
Author: By
Last change:
Date: 8 years ago
Size: 7,713 bytes
 

Contents

Class file image Download

Why Requests Instead of X?

This is a quick look at why you should use Requests instead of another solution. Keep in mind though that these are my point of view, and they may not be issues for you.

As always with software, you should choose what you think is best.

Why should I use Requests?

  1. Designed for maximum compatibility

    The realities of working with widely deployable software mean that awesome PHP features aren't always available. PHP 5.3, cURL, OpenSSL and more are not necessarily going to be available on every system. While you're welcome to require PHP 5.3, 5.4 or even 5.5, it's not our job to force you to use those.

    (The WordPress project estimates [about 60%][wpstats] of hosts are running PHP 5.2, so this is a serious issue for developers working on large deployable projects.)

    Don't worry though, Requests will automatically use better features where possible, giving you an extra speed boost with cURL.

  2. Simple API

    Requests' API is designed to be able to be learnt in 10 minutes. Everything from basic requests all the way up to advanced usage involving custom SSL certificates and stored cookies is handled by a simple API.

    Other HTTP libraries optimize for the library developer's time; Requests optimizes for your time.

  3. Thoroughly tested

    Requests is [continuously integrated with Travis][travis] and test coverage is [constantly monitored with Coveralls][coveralls] to give you confidence in the library. We aim for test coverage over 90% at all times, and new features require new tests to go along with them. This ensures that you can be confident in the quality of the code, as well as being able to update to the latest version of Requests without worrying about compatibility.

  4. Secure by default

    Unlike other HTTP libraries, Requests is secure by default. Requests is the first and currently only standalone HTTP library to [fully verify][requests_ssl] all HTTPS requests even without cURL. We also bundle the latest root certificate authorities to ensure that your secure requests are actually secure.

    (Of note is that WordPress as of version 3.7 also supports full checking of the certificates, thanks to [evangelism efforts on our behalf][wpssl]. Together, we are the only HTTP libraries in PHP to fully verify certificates to the same level as browsers.)

  5. Extensible from the core

    If you need low-level access to Requests' internals, simply plug your callbacks in via the built-in [hooking system][] and mess around as much as you want. Requests' simple hooking system is so powerful that both authentication handlers and cookie support is actually handled internally with hooks.

[coveralls]: https://coveralls.io/r/rmccue/Requests [hooking system]: hooks.md [requests_ssl]: https://github.com/rmccue/Requests/blob/master/library/Requests/SSL.php [travis]: https://travis-ci.org/rmccue/Requests [wpssl]: http://core.trac.wordpress.org/ticket/25007

Why shouldn't I use...

Requests isn't the first or only HTTP library in PHP and it's important to acknowledge the other solutions out there. Here's why you should use Requests instead of something else, in our opinion.

cURL

  1. Not every host has cURL installed

    cURL is far from being ubiquitous, so you can't rely on it always being available when distributing software. Anecdotal data collected from various projects indicates that cURL is available on roughly 90% of hosts, but that leaves 10% of hosts without it.

  2. cURL's interface sucks

    cURL's interface was designed for PHP 4, and hence uses resources with horrible functions such as `curl_setopt()`. Combined with that, it uses 229 global constants, polluting the global namespace horribly.

    Requests, on the other hand, exposes only a handful of classes to the global namespace, most of which are for internal use. You can learn to use the `Requests::request()` method and the `Requests_Response` object in the space of 10 minutes and you already know how to use Requests.

Guzzle

  1. Requires cURL and PHP 5.3+

    Guzzle is designed to be a client to fit a large number of installations, but as a result of optimizing for Guzzle developer time, it uses cURL as an underlying transport. As noted above, this is a majority of systems, but far from all.

    The same is true for PHP 5.3+. While we'd all love to rely on PHP's newer features, the fact is that a huge percentage of hosts are still running on PHP 5.2. (The WordPress project estimates [about 60%][wpstats] of hosts are running PHP 5.2.)

  2. Not just a HTTP client

    Guzzle is not intended to just be a HTTP client, but rather to be a full-featured REST client. Requests is just a HTTP client, intentionally. Our development strategy is to act as a low-level library that REST clients can easily be built on, not to provide the whole kitchen sink for you.

    If you want to rapidly develop a web service client using a framework, Guzzle will suit you perfectly. On the other hand, if you want a HTTP client without all the rest, Requests is the way to go.

[wpstats]: http://wordpress.org/about/stats/

Buzz

  1. Requires PHP 5.3+

    As with Guzzle, while PHP 5.3+ is awesome, you can't always rely on it being on a host. With widely distributable software, this is a huge problem.

  2. Not transport-transparent

    For making certain types of requests, such as multi-requests, you can't rely on a high-level abstraction and instead have to use the low-level transports. This really gains nothing (other than a fancy interface) over just using the methods directly and means that you can't rely on features to be available.

fsockopen

  1. Very low-level

    fsockopen is used for working with sockets directly, so it only knows about the transport layer (TCP in our case), not anything higher (i.e. HTTP on the application layer). To be able to use fsockopen as a HTTP client, you need to write all the HTTP code yourself, and once you're done, you'll end up with something that is almost exactly like Requests.

PEAR HTTP_Request2

  1. Requires PEAR

    PEAR is (in theory) a great distribution system (with a less than wonderful implementation), however it is not ubiquitous, as many hosts disable it to save on space that most people aren't going to use anyway.

    PEAR is also a pain for users. Users want to be able to download a zip of your project without needing to install anything else from PEAR.

    (If you really want though, Requests is available via PEAR. Check the README to see how to grab it.)

  2. Depends on other PEAR utilities

    HTTP\_Request2 requires Net_URL2 in order to function, locking you in to using PEAR for your project.

    Requests is entirely self-contained, and includes all the libraries it needs (for example, Requests\_IRI is based on ComplexPie\_IRI by Geoffrey Sneddon).

PECL HttpRequest

  1. Requires a PECL extension

    Similar to PEAR, users aren't big fans of installing extra libraries. Unlike PEAR though, PECL extensions require compiling, which end users will be unfamiliar with. In addition, on systems where users do not have full control over PHP, they will be unable to install custom extensions.

Zend Framework's Zend\_Http\_Client

  1. Requires other parts of the Zend Framework

    Similar to HTTP_Request2, Zend's client is not fully self-contained and requires other components from the framework.