browserify

require('modules') in the browser

Use a node-style require() to organize your browser code and load modules installed by npm.

browserify will recursively analyze all the require() calls in your app in order to build a bundle you can serve up to the browser in a single <script> tag.

example

Whip up a file, main.js with some require()s in it. You can use relative paths like './foo.js' and '../lib/bar.js' or module paths like 'gamma' that will search node_modules/ using node's module lookup algorithm.

var foo = require('./foo.js');
var bar = require('../lib/bar.js');
var gamma = require('gamma');

var elem = document.getElementById('result');
var x = foo(100) + bar('baz');
elem.textContent = gamma(x);

Export functionality by assigning onto module.exports or exports:

module.exports = function (n) { return n * 111 }

Now just use the browserify command to build a bundle starting at main.js:

$ browserify main.js > bundle.js

All of the modules that main.js needs are included in the bundle.js from a recursive walk of the require() graph using required.

To use this bundle, just toss a <script src="bundle.js"></script> into your html!

install

With npm do:

npm install -g browserify

usage

Usage: browserify [entry files] {OPTIONS}

Standard Options:

    --outfile, -o  Write the browserify bundle to this file.
                   If unspecified, browserify prints to stdout.

    --require, -r  A module name or file to bundle.require()
                   Optionally use a colon separator to set the target.

      --entry, -e  An entry point of your app
  
     --ignore, -i  Replace a file with an empty stub.
 
    --exclude, -u  Omit a file from the output bundle.

   --external, -x  Reference a file from another bundle.
  
  --transform, -t  Use a transform module on top-level files.
 
    --command, -c  Use a transform command on top-level files.
   
  --standalone -s  Generate a UMD bundle for the supplied export name.
                   This bundle works with other module systems and sets the name
                   given as a window global if no module system is found.
  
       --debug -d  Enable source maps that allow you to debug your files
                   separately.

       --help, -h  Show this message

For advanced options, type `browserify --help advanced`.

Specify a parameter.

Advanced Options:

  --insert-globals, --ig, --fast    [default: false]

    Skip detection and always insert definitions for process, global,
    __filename, and __dirname.
                  
    benefit: faster builds
    cost: extra bytes
 
  --insert-global-vars, --igv

    Comma-separated list of global variables to detect and define.
    Default: __filename,__dirname,process,Buffer,global
 
  --detect-globals, --dg            [default: true]

    Detect the presence of process, global, __filename, and __dirname and define
    these values when present.

    benefit: npm modules more likely to work
    cost: slower builds

  --ignore-missing, --im            [default: false]

    Ignore `require()` statements that don't resolve to anything.

  --noparse=FILE

    Don't parse FILE at all. This will make bundling much, much faster for giant
    libs like jquery or threejs.

  --no-builtins

    Turn off builtins. This is handy when you want to run a bundle in node which
    provides the core builtins.

  --no-commondir
 
    Turn off setting a commondir. This is useful if you want to preserve the
    original paths that a bundle was generated with. 
 
  --bare 

    Alias for both --no-builtins, --no-commondir, and sets --insert-global-vars
    to just "__filename,__dirname". This is handy if you want to run bundles in
    node.

  --deps
  
    Instead of standard bundle output, print the dependency array generated by
    module-deps.

  --list
 
    Print each file in the dependency graph. Useful for makefiles.

  --extension=EXTENSION

    Consider files with specified EXTENSION as modules, this option can used
    multiple times.

  --global-transform, --g

    Use a transform module on all files after any ordinary transforms have run.

compatibility

Many npm modules that don't do IO will just work after being browserified. Others take more work.

Many node built-in modules have been wrapped to work in the browser, but only when you explicitly require() or use their functionality.

When you require() any of these modules, you will get a browser-specific shim:

Additionally if you use any of these variables, they will be defined in the bundled output in a browser-appropriate way:

process
Buffer
global - top-level scope object (window)
__filename - file path of the currently executing file
__dirname - directory path of the currently executing file

more examples

external requires

You can just as easily create bundle that will export a require() function so you can require() modules from another script tag. Here we'll create a bundle.js with the through and duplexer modules.

$ browserify -r through -r duplexer -r ./my-file.js:my-module > bundle.js

Then in your page you can do:

<script src="bundle.js"></script>
<script>
  var through = require('through');
  var duplexer = require('duplexer');
  var myModule = require('my-module');
  /* ... */
</script>

external source maps

If you prefer the source maps be saved to a separate .js.map source map file, you may use exorcist in order to archieve that. It's as simple as:

$ browserify main.js --debug | exorcist bundle.js.map > bundle.js

Learn about additional options here.

multiple bundles

If browserify finds a required function already defined in the page scope, it will fall back to that function if it didn't find any matches in its own set of bundled modules.

In this way you can use browserify to split up bundles among multiple pages to get the benefit of caching for shared, infrequently-changing modules, while still being able to use require(). Just use a combination of --external and --require to factor out common dependencies.

For example, if a website with 2 pages, beep.js:

var robot = require('./robot.js');
console.log(robot('beep'));

and boop.js:

var robot = require('./robot.js');
console.log(robot('boop'));

both depend on robot.js:

module.exports = function (s) { return s.toUpperCase() + '!' };

$ browserify -r ./robot > static/common.js
$ browserify -x ./robot.js beep.js > static/beep.js
$ browserify -x ./robot.js boop.js > static/boop.js

Then on the beep page you can have:

<script src="common.js"></script>
<script src="beep.js"></script>

while the boop page can have:

<script src="common.js"></script>
<script src="boop.js"></script>

api example

You can use the API directly too:

var browserify = require('browserify');
var b = browserify();
b.add('./browser/main.js');
b.bundle().pipe(process.stdout);

methods

var browserify = require('browserify')

var b = browserify(files=[] or opts={})

Create a browserify instance b from the entry main files or opts.entries. files can be an array of files or a single file.

For each file in files, if file is a stream, its contents will be used. You should use opts.basedir when using streaming files so that relative requires will know where to resolve from.

You can also specify an opts.noParse array which will skip all require() and global parsing for each file in the array. Use this for giant libs like jquery or threejs that don't have any requires or node-style globals but take forever to parse.

opts.extensions is an array of optional extra extensions for the module lookup machinery to use when the extension has not been specified. By default browserify considers only .js and .json files in such cases.

opts.basedir is the directory that browserify starts bundling from for filenames that start with ..

opts.commondir sets the algorithm used to parse out the common paths. Use false to turn this off, otherwise it uses the commondir module.

opts.builtins sets the list of builtins to use, which by default is set in lib/builtins.js in this distribution.

opts.pack sets the browser-pack implementation to use. The opts.pack() should return a transform stream that accepts objects of the form that module-deps generates. Simplifed, this is roughly:

{"id":"1","source":"console.log('beep boop')","deps":{}}
{"id":"2","source":"require('./boop.js')","deps":{"./boop.js":"1"}}

By default, opts.pack uses browser-pack:

require('browser-pack')({ raw: true, sourceMapPrefix: '//@' });

opts.externalRequireName defaults to 'require' in expose mode but you can use another name.

You can give browserify a custom opts.resolve() function or by default it uses browser-resolve.

Note that if files do not contain javascript source code then you also need to specify a corresponding transform for them.

b.add(file)

Add an entry file from file that will be executed when the bundle loads.

b.require(file[, opts])

Make file available from outside the bundle with require(file).

The file param is anything that can be resolved by require.resolve().

file can also be a stream, but you should also use opts.basedir so that relative requires will be resolvable.

Use the expose property of opts to specify a custom dependency name. require('./vendor/angular/angular.js', {expose: 'angular'}) enables require('angular')

b.bundle(opts, cb)

Bundle the files and their dependencies into a single javascript file.

Return a readable stream with the javascript file contents or optionally specify a cb(err, src) to get the buffered results.

When opts.insertGlobals is true, always insert process, global, __filename, and __dirname without analyzing the AST for faster builds but larger output bundles. Default false.

When opts.detectGlobals is true, scan all files for process, global, __filename, and __dirname, defining as necessary. With this option npm modules are more likely to work but bundling takes longer. Default true.

When opts.debug is true, add a source map inline to the end of the bundle. This makes debugging easier because you can see all the original files if you are in a modern enough browser.

When opts.standalone is a non-empty string, a standalone module is created with that name and a umd wrapper. You can use namespaces in the standalone global export using a . in the string name as a separator. For example: 'A.B.C'

opts.insertGlobalVars will be passed to insert-module-globals as the opts.vars parameter.

b.external(file)

Prevent file from being loaded into the current bundle, instead referencing from another bundle.

b.ignore(file)

Prevent the module name or file at file from showing up in the output bundle.

Instead you will get a file with module.exports = {}.

b.exclude(file)

Prevent the module name or file at file from showing up in the output bundle.

If your code tries to require() that file it will throw unless you've provided another mechanism for loading it.

b.transform(opts={}, tr)

Transform source code before parsing it for require() calls with the transform function or module name tr.

If tr is a function, it will be called with tr(file) and it should return a through-stream that takes the raw file contents and produces the transformed source.

If tr is a string, it should be a module name or file path of a transform module with a signature of:

var through = require('through');
module.exports = function (file) { return through() };

You don't need to necessarily use the through module, this is just a simple example.

Here's how you might compile coffee script on the fly using .transform():

var coffee = require('coffee-script');
var through = require('through');

b.transform(function (file) {
    var data = '';
    return through(write, end);
    
    function write (buf) { data += buf }
    function end () {
        this.queue(coffee.compile(data));
        this.queue(null);
    }
});

Note that on the command-line with the -c flag you can just do:

$ browserify -c 'coffee -sc' main.coffee > bundle.js

Or better still, use the coffeeify module:

$ npm install coffeeify
$ browserify -t coffeeify main.coffee > bundle.js

If opts.global is true, the transform will operate on ALL files, despite whether they exist up a level in a node_modules/ directory. Use global transforms cautiously and sparingly, since most of the time an ordinary transform will suffice. You can also not configure global transforms in a package.json like you can with ordinary transforms.

Global transforms always run after any ordinary transforms have run.

b.deps(opts)

Return a readable stream of the dependency array generated by module-deps.

package.json

browserify uses the package.json in its module resolution algorithm just like node, but there is a special "browser" field you can set to override file resolution for browser-specific versions.

You can specify source transforms in the package.json in the browserify.transform field. There is more information about how source transforms work in package.json on the module-deps readme.

events

b.on('file', function (file, id, parent) {})

When a file is resolved for the bundle, the bundle emits a 'file' event with the full file path, the id string passed to require(), and the parent object used by browser-resolve.

You could use the file event to implement a file watcher to regenerate bundles when files change.

list of source transforms

There is a wiki page that lists the known browserify transforms.

If you write a transform, make sure to add your transform to that wiki page and add a package.json keyword of browserify-transform so that people can browse for all the browserify transforms on npmjs.org.

third-party tools

There is a wiki page that lists the known browserify tools.

If you write a tool, make sure to add it to that wiki page and add a package.json keyword of browserify-tool so that people can browse for all the browserify tools on npmjs.org.

license

MIT

bnkr/node-browserify