/bower

A package manager for the web

Primary LanguageJavaScriptMIT LicenseMIT

BOWER Build Status

Introduction

Bower is a package manager for the web. Bower lets you easily install assets such as images, CSS and JavaScript, and manages dependencies for you.

For example, to install a package, run:

bower install jquery

This will download jQuery to ./components/jquery. That's it. The idea is that Bower does package management and package management only.

Installing Bower

Bower is installed using Node and npm (oh my, how meta).

npm install bower -g

Usage

Your best friend at this stage is probably bower --help.

To install a package:

bower install jquery
bower install git://github.com/components/jquery.git
bower install components/jquery (same as above)
bower install http://foo.com/jquery.awesome-plugin.js
bower install ./repos/jquery

As you can see, packages can be installed by name, Git endpoint, GitHub shorthand, URL or local path. If you install from a URL that points to a zip or tar file, bower will automatically extract its contents. When tags are available in the endpoint, you can specify a semver tag to fetch concrete versions:

bower install jquery#1.8.1
bower install git://github.com/components/jquery.git#~1.8.1
bower install components/jquery#1.8.x

Bower also works with private Git repositories. Simply reference them by their SSH endpoint:

bower install git@github.com:user/private-package.git

View all packages available through Bower's registry.

During install you can have Bower add an entry to your component.json as well:

bower install --save jquery

To update a package, reference it by name:

bower update jquery-ui

To list installed packages:

bower list

To search for packages:

bower search [name]

To list all the available packages, just call bower search without specifying a name.

To clean the cache:

bower cache-clean [name]

Several packages can be cleaned at the same time. To clean the entire cache, just call bower cache-clean without any names. Also, both the install and update commands have a --force flag that tells bower to bypass the cache and always fetch remote sources.

You can disable colors by using the --no-color flag.

Bower Configuration

Bower can be configured by creating a .bowerrc file in your home folder (usually ~/.bowerrc) with one or all of the following configuration parameters. You can also configure Bower on a per-project basis by creating a .bowerrc file in the project directory, Bower will merge this configuration with the configuration found in your home directory. This allows you to version your project specific Bower configuration with the rest of your code base.

{
  "directory" : "components",
  "json"      : "component.json",
  "endpoint"  : "https://bower.herokuapp.com"
}

To run your own Bower Endpoint for custom components/packages that are behind a firewall you can use a simple implementation of bower server at https://github.com/twitter/bower-server.

The searchpath array provides additional URLs of read-only Bower registries that should be consulted to look up components. This is most typically used if your business wishes to house some components internally while still taking advantage of public Bower registries. For example, you might configure the following:

{
  "directory"  : "components",
  "json"       : "component.json",
  "endpoint"   : "http://bower.mycompany.com",
  "searchpath" : ["https://bower.herokuapp.com"]
}

Bower will first look to http://bower.mycompany.com while trying to find your components. If not found, the main registry at https://bower.herokuapp.com will be consulted to see if a copy of the resource can be retrieved.

Defining a package

You can create a component.json file in your project's root, specifying all of its dependencies. This is similar to Node's package.json, or Ruby's Gemfile, and is useful for locking down a project's dependencies.

{
  "name": "myProject",
  "version": "1.0.0",
  "main": "./path/to/main.css",
  "dependencies": {
    "jquery": "~1.7.2"
  }
}

Put this under your project's root, listing all of your dependencies. When you run bower install, Bower will read this component.json file, resolve all the relevant dependencies and install them.

For now, name, version, main, dependencies, and ignore are the only properties that are used by Bower. If you have several files you're distributing as part of your package, pass an array to main like this:

{
  "name": "myProject",
  "version": "1.0.0",
  "main": ["./path/to/app.css", "./path/to/app.js", "./path/to/sprite.img"],
  "dependencies": {
    "jquery": "~1.7.2"
  }
}

Bower only recognizes versions that follow the semver specification. There should only be at most one file per file type in the main list. So only one .js or .css.

You can also point to packages by adding their URL or file path in the dependency's property.

{
  "dependencies": {
    "eventEmitter": "Wolfy87/EventEmitter", // GitHub short URL
    "eventEmitter": "Wolfy87/EventEmitter#>=3", // with version
    "eventEmitter": "git://github.com/Wolfy87/EventEmitter",
    "eventEmitter": "git@github.com:Wolfy87/EventEmitter.git"
  }
}

Chances are you have a bunch of extra stuff in the repo that are not needed in production. List these non-necessary file paths in ignore.

{
  "ignore": [
    "tests/",
    "**/*.txt"
  ]
}

Installing dependencies

Dependencies are installed locally via the bower install command. First they’re resolved to find conflicts. Then they’re downloaded and unpacked in a local subdirectory called ./components, for example:

/component.json
/components/jquery/index.js
/components/jquery/component.json

You can also install packages one at a time bower install git://my/git/thing

There are no system wide dependencies, no dependencies are shared between different apps, and the dependency tree is flat.

Deploying

The easiest approach is to use Bower statically, just reference the packages manually from a script tag:

<script src="components/jquery/index.js"></script>

For more complex projects, you'll probably want to concatenate your scripts. Bower is just a package manager, but there are lots of awesome libraries out there to help you do this, such as Sprockets and RequireJS.

For example, to use Sprockets:

environment = Sprockets::Environment.new
environment.append_path 'components'
environment.append_path 'public'
run environment

Package Consumption

Bower also makes available a source mapping – this can be used by build tools to easily consume Bower components.

If you pass the option --map to bower's list command, it will generate a JSON with dependency objects. Alternatively, you can pass the --paths flag to the list command to get a simple path to name mapping:

{
  "backbone": "components/backbone/index.js",
  "jquery": "components/jquery/index.js",
  "underscore": "components/underscore/index.js"
}

Authoring packages

To register a new package, it's as simple as specifying a component.json, pushing the package to a Git endpoint, say GitHub, and running:

bower register myawesomepackagename git://github.com/maccmans/face

There's no authentication or user management. It's on a first come, first served basis. Think of it like a URL shortener. Now anyone can run bower install myawesomepackagename, and get your library installed.

There is no direct way to unregister a package yet. Meanwhile you can request it here.

Philosophy

Currently, people are managing dependencies, such as JavaScript libraries, manually. This sucks, and we want to change it.

In a nutshell, Bower is a generic tool which will resolve dependencies and lock packages down to a version. It runs over Git, and is package-agnostic. A package may contain JavaScript, CSS, images, etc., and doesn't rely on any particular transport (AMD, CommonJS, etc.).

Bower then makes available a simple programmatic API which exposes the package dependency model, so that existing build tools (like Sprockets, LoadBuilder, curls.js, Ender, etc.) can consume it and build files accordingly.

Programmatic API

Bower provides a pretty powerful programmatic api. All commands can be accessed through the bower.commands object.

var bower = require('bower');

bower.commands
  .install(paths, options)
  .on('end', function (data) {
    data && console.log(data);
  });

All commands emit three types of events: data, end, and error.

For a better of idea how this works, you may want to check out our bin file.

For the install command, there is an additional package event that is emitted for each installed/uninstalled package.

Completion

experimental

Based on the completion feature and fantastic work done in npm, Bower now has an experimental completion command that works similarly.

This command will output a Bash / ZSH script to put into your ~/.bashrc, ~/.bash_profile or ~/.zshrc file.

bower completion >> ~/.bash_profile

This doesn't work for Windows user, even with Cygwin.

Windows users

A lot of people are experiencing problems using bower on windows because msysgit must be installed correctly. Be sure to check the option shown above, otherwise it will simply not work:

msysgit

FAQ

What distinguishes Bower from Jam, Volo or Ender? What does it do better?

Bower is a lower level component than Jam, Volo, or Ender. These managers could consume Bower as a dependency.

Bower's aim is simply to install Git paths, resolve dependencies from a component.json, check versions, and then provide an API which reports on these things. Nothing more. This is a major diversion from past attempts at browser package management.

Bower is working under the assumption that there is a single, common problem in frontend application development: dependency resolution. Past attempts (Jam, Volo, Ender) try to tackle this problem in such a way that they actually end up alienating and further segregating the JavaScript community around transports (Sprockets, CommonJS, RequireJS, regular script tags).

Bower offers a generic, unopinionated solution to the problem of package management, while exposing an API that can be consumed by a more opinionated build stack.

Volo is an arguably more established project and works with the GitHub search API. Will it take long for Bower to contain a decent number of packages?

Bower (being a Git powered package manager) should, in theory, be capable of consuming every package that Volo does, with the additional benefit of supporting internal networks and other Git repositories not hosted on GitHub.

We recently saw what happened when the main NPM registry went down. Is a single point of failure possible for Bower and if so, do you have redundancy planned?

There's no redundancy planned at the moment, as Bower just installs Git URLs. It's up to the URL provider to establish redundancy.

Isn't having a package.json file going to conflict with my npm's package.json? Will this be a problem?

Don't use a package.json – use a component.json.

Bower is an open-source Twitter project. How well can we expect it to be maintained in the future?

Twitter is in the process of migrating its frontend architecture onto Bower, so it's fairly safe to say it will be maintained and invested in going forward.

Contact

Have a question? Ask on our mailing list!

twitter-bower@googlegroups.com

http://groups.google.com/group/twitter-bower

Authors

Thanks for assistance and contributions:

License

Copyright 2012 Twitter, Inc.

Licensed under the MIT License