All development for curl.js and cram.js has stopped. For the foreseeable future, we will continue to respond to issues on github, as well as in the #cujojs room on freenode. However, keep in mind that there will be no further development, so some issues might not be resolved fully.
Also, if you are interested in becoming the lead maintainer of curl.js and/or cram.js, please let us know on #cujojs!
curl.js is a small and very fast AMD-compliant asynchronous loader. Size: ~4KB (gzipped) using Google's Closure Compiler.
If you'd like to use curl.js for non-AMD modules (ordinary javascript files), you'll want to use the legacy loader.
curl.js, like all async loaders, cannot circumvent browsers' security
restrictions when using the file: protocol. Therefore, you must use
curl from a page served from a web server (i.e. using http: or https:).
Trying to run curl.js from a page loaded from your local file system
will not work correctly in all browsers.
What the heck is "cujoJS"? cujoJS is the JavaScript Architectural Toolkit. More info: cujojs.com
What is "cram"? cram (cujoJS resource assembler) is the build tool companion to curl.js. You use cram to compile all of your modules into a small number of javascript files which are loaded much faster into the browsers.
define())More detailed information below and on the wiki.
It's that easy.
Got more in-depth questions? Browse the cujoJS discussion group or come chat with us on freenode @ #cujojs.
See the wiki for information about using curl.js with jQuery, dojo, or underscore.
define is undefined when wrapping cjs/node modules.bin/make-all.sh.define from cram parser.At it's core, curl.js provides an AMD environment:
define(['dep1', 'dep2', 'dep3' /* etc */], factory);
define(['dep1', 'dep2', 'dep3' /* etc */], module);
define(module);
define(name, ['dep1', 'dep2', 'dep3' /* etc */], factory);
define(name, ['dep1', 'dep2', 'dep3' /* etc */], module);
define(name, module);These all define a module per the AMD specification.
define(function (require, exports, module) {
var dep1 = require('app/foo');
exports.foo2 = function () { return foo() + 2; };
});Defines a module using the AMD-wrapped-CommonJS format. If a factory function
has parameters, but the dependency list is missing, this format is assumed.
The exports and module parameters are optional, but when specified, must
be in this exact order.
define(function (require) {
var dep1 = require('app/foo');
return function () { return foo() + 2; };
});Another variation on AMD-wrapped-CommonJS that uses require() in the
tidy CommonJS manner, but returns the exports as per typical AMD.
define(['require', 'exports', 'module'], function (require, exports, module) {
var dep1 = require('app/foo');
exports.foo2 = function () { return foo() + 2; };
});Another way to gain access to the CommonJS-like variables, require,
exports, module. When specified in the dependency list, these
"pseudo-modules" are provided as arguments to the factory function.
var dep1 = require('app/foo');
exports.foo2 = function () { return foo() + 2; };curl.js also supports unwrapped CommonJS modules (and node.js modules) via the cjsm11 module loader. To use this module loader for a package, say Backbone, you would provide it to a package config, like this:
curl.config({
packages: [
{
name: 'backbone',
location: 'bower_components/backbone',
main: 'backbone.min.js',
config: { moduleLoader: 'curl/loader/cjsm11' } /* <-- hey! */
}
]
});Read the notes in the src/curl/loader folder and the cjsm11.js file for more information about loading CommonJS and node.js modules.
curl.js's global API is for bootstrapping an app. You would typically only use this API once in your application to load the main module of your application.
For a complete description, check out the wiki.
curl(['main', 'other', 'another' /* etc */], callback, errorback);Loads a module named "main" along with two other modules and then executes callback, handing it the exported values of the modules as parameters.
curl(['main', 'other', 'another' /* etc */])
.then(callback, errorback);Promises-based API for executing callbacks.
when(curl(['dep1'])).then(callback); or Promise.resolve(curl(['dep1'])).then(callback);curl(config, ['main' /* etc */], callback, errorback);Specify configuration options, load dependencies, and execute callback.
curl(['main', 'domReady!' /* etc */]).then(
callback,
errorback
);curl(['main', 'domReady!' /* etc */], function (main) {
// do some bootstrapping here
});Executes the callback when the dom is ready for manipulation AND all dependencies have loaded.
curl(['domReady!', 'js!nonAMD.js!order', 'js!another.js!order']), function () {
/* do something cool here */
});Executes the function when the non-AMD javascript files are loaded and the dom is ready. The another.js file will wait for the nonAMD.js file before executing.
Note: Please use curl.js's new legacy loader for much more flexible and sensible loading of non-AMD scripts. Please read the docs in the src/curl/loader folder for more information.
Note: if a file supports AMD or CommonJS module formats, you can not use the js! plugin on that file.
curl(['js!nonAMD.js'])
.next(['dep1', 'dep2', 'dep3'], function (dep1, dep2, dep3) {
// do something before the dom is ready
})
.next(['domReady!'])
.then(
function () {
// do something after the dom is ready
},
function (ex) {
// show an error to the user
}
);Executes callbacks in stages using .next(deps, callback).
Note: .next() does not load resources in parallel. Therefore, it is a
last resort when other options do not satisfy your use case. You should
use the preloads config option and/or the legacy loader
whenever possible.
curl = {
baseUrl: '/path/to/my/js',
pluginPath: 'for/some/reason/plugins/r/here',
paths: {
curl: 'curl/src/curl',
cssx: 'cssx/src/cssx',
my: '../../my-lib/'
},
apiName: 'someOtherName'
};If called before the <script> that loads curl.js, a global curl var will
configure curl.js when it loads. All of the configuration parameters are
optional. curl.js tries to do something sensible in their absence. :)
Some common configuration options:
curl and require for curl.js's global
variablewindow, to place curl on when using
apiNameA more complete list can be found on the wiki.
<script>
// configure curl
curl = {
paths: {
cssx: 'cssx/src/cssx/',
stuff: 'my/stuff/'
}
};
</script>
<script src="https://github.com/cujojs/curl/raw/master/./js/curl.js" type="text/javascript"></script>
<script type="text/javascript">
curl(
// fetch all of these resources ("dependencies")
[
'stuff/three', // an AMD module
'cssx/css!stuff/base', // a css file
'i18n!stuff/nls/strings', // a translation file
'text!stuff/template.html', // an html template
'domReady!'
]
)
// when they are loaded
.then(
// execute this callback, passing all dependencies as params
function (three, link, strings, template) {
var body = document.body;
if (body) {
body.appendChild(document.createTextNode('three == ' + three.toString() + ' '));
body.appendChild(document.createElement('br'));
body.appendChild(document.createTextNode(strings.hello));
body.appendChild(document.createElement('div')).innerHTML = template;
}
},
// execute this callback if there was a problem
function (ex) {
var msg = 'OH SNAP: ' + ex.message;
alert(msg);
}
);
</script>The file structure for this example would look as follows:
js/
curl/
plugin/
i18n.js
text.js
domReady.js
cssx/
src/
cssx/
css.js
my/
stuff/
nls/
strings.js
base.css
template.html
three.js
curl.jsWeb apps, especially large ones, require many modules and resources. Most of these modules and resources need to be loaded at page load, but some may be loaded later, either in the background or "just in time". They also need to be loaded as quickly as possible.
The traditional way to load javascript modules is via a <SCRIPT> element in
an HTML page. Similarly, CSS files are loaded via a <LINK> element, and
text resources are either loaded in the page or via XHR calls.
The problem with <SCRIPT> and <LINK> elements is that a browser must execute
them sequentially since it has no idea if one may depend on another. It just
assumes the developer has placed them in the correct order and that there are
dependencies. (The term "synchronous loading" is used to describe this process
since the elements are executed in a single timeline.)
If there are no dependencies between two files, loading them sequentially is a waste of time. These files could be loaded and executed in parallel (i.e at the same time).
An asynchronous loader does just that: it loads javascript files (and other types of files) in parallel whenever possible.
curl.js has lots of company. Other async loaders include LABjs, Steal.js, yepnope.js, $script.js, the Backdraft loader (bdLoad), and RequireJS.
Asynchronous Module Definition is a de facto standard for javascript modules that can be loaded by asynchronous loaders. It defines a simple API that developers can use to write their javascript modules so that they may be loaded by any AMD-compliant loader.
The AMD proposal follows some parts of the CommonJS Modules proposal. Because of the way browsers load and evaluate scripts, AMD can't follow it completely without causing significant processing overhead. Instead, AMD allows us to place a lightweight wrapper around javascript modules to help work around the shortcomings.
Ultimately, both proposals (AMD and Modules 1.1) are in preparation for an official javascript modules specification and eventual implementation in browsers.
If you don't want to wait for official javascript modules, then don't. The future is now. AMD works now -- and it's awesome.
AMD's API focuses on one globally-available function: define() and some
CommonJS-inspired variables, require(), exports, and module.
require() specifies a list of dependent modules or resources that must be
loaded before running a set of code. This code resides in a callback function
that is executed asynchronously, i.e. it runs later, not in the current
"thread". Specifically, it executes when all of the dependencies are loaded
and ready.
The proposal does not mandate that require() be specified globally. In fact,
at a global level, the concerns of a loader are about application bootstrapping
and not about finding dependencies. To keep the confusion about these two
concepts to a minimum, curl.js uses curl() for the public API. You may rename
this API back to require() by supplying the apiName config param
(apiName: "require"), but this is not recommended.
It's more important that the define() method be consistent. This is the
method that tells the loader what modules have been loaded by a script.
define() also specifies a list of dependencies and a callback function that
defines and/or creates the resource when the dependencies are ready.
Optionally, define() also takes a name parameter, but this is mainly for build
tools and optimizers.
Inside the define(), the require() method acts like other AMD loaders.
AMD's API also helps code reuse by providing compatibility with CommonJS
(server) modules. AMD-compliant loaders support the same require() syntax and
argument signatures as server-side javascript (ssjs) modules.
The beauty of AMD loaders is their ability to remove the drudgery of manually managing dependencies. Since all dependencies are listed within the modules, the loader will ensure that everything is loaded into the browser -- and in the right order.
Yes, but why would you? Once you start using AMD, you'll never go back! :)
You may use non-AMD javascript files by using the legacy loader like this:
curl.config({
paths: {
plainOldJsFile1: {
location: 'js/plainOldJsFile1.js',
config: { loader: 'curl/loader/legacy', exports: 'aGlobal' }
},
anotherPlainOldJsFile: {
location: 'js/anotherPlainOldJsFile.js',
config: {
loader: 'curl/loader/legacy',
exports: 'anotherGlobal',
requires: [ 'plainOldJsFile1' ]
}
}
}
});
curl(['anotherPlainOldJsFile']).then(
function (anotherGlobal) {
/* do something with your plain, boring javascript files */
},
function () {
/* do something if any fail to load */
}
);Please read the docs in the src/curl/loader folder for more information.
Yes, curl.js follows the CommonJS Loader Plugin specification, so you can use any compatible plugin. The following plugins are included:
js! -- loads non-AMD javascript files. more info on the wiki
text! -- loads text files
link! -- loads css files via a link element (simple, fast)
css! -- loads css files (lots of options)
domReady! -- resolves when the dom is ready for manipulation
async! -- resolves when a module signals it's ready
i18n! -- loads text strings and other locale-specific constants
Some plugin docs are on the wiki.
More documentation is available inside the source of these plugins!
curl.js uses <script> element injection rather than XHR/eval for AMD modules.
This allows curl.js to load cross-domain scripts as well as local scripts.
CommonJS modules use XHR/eval so must be wrapped for "transport" if not
on the same domain. Typically, you will bundle your modules before using them
in production, anyways. Most bundlers will wrap CommonJS modules in AMD.
To find scripts and other resources, curl.js uses module names. A module name looks just like a file path, but typically without the file extension. If a module requires a plugin in order to load correctly, it will have a prefix delimited by a "!" and will also often have a file extension when a plugin may load different types of files.
Some examples of module names:
By default, curl.js will look in the same folder as the current document's
location. For instance, if your web page is located at
http://my-domain/apps/myApp.html, curl.js will look for the JsonRest module
at http://my-domain/apps/dojo/store/JsonRest.js.
You can tell curl.js to find modules in other locations by specifying a baseUrl
or individual packages for each of your libraries. For example, if you specify
a baseUrl of /resources/ and the following paths:
packages: [
{ name: "dojo", location: "third-party/dojo" },
{ name: "css", location: "third-party/cssmojo/css" },
{ name: "my", location: "my-cool-app-v1.3" },
{ name: "my/lib/js", location: "old-js-libs" }
]Then the modules listed above will be sought in the following locations:
Note: you will need to create a path to curl.js's plugins and other modules if the curl folder isn't directly under the same folder as your web page. curl.js uses the same mechanism to find its own modules.
AMD supports the notion of plugins. Plugins are AMD modules that can be used to load javascript modules -- or other types of resources. curl comes with several plugins already, including a text plugin (for templates or other text resources), two different css plugins, a dom-ready plugin, and several others.
Plugins are designated by a prefix on the name of the module or resource to be loaded. They are delineated by a ! symbol. The following example shows the use of some plugins:
define(
[
'text!myTemplate.html',
'css!myCssFile'
],
function (templateString, cssLinkNode) {
// do something with the template and css here
}
);Since plugins are just AMD modules, they would typically be referenced using their fully-pathed names. curl provides a pluginPath configuration option that allows you to specify the folder where [most of] your plugins reside so you don't have to specify their full paths. This also helps with compatibility with other AMD loaders.
If one or more of your plugins does not reside in the folder specified by the
pluginPath config option, you can use its full path or you can specify a path
for it in curl's paths config object.
// example of a fully-pathed plugin under the lib/cssx folder
define(['lib/cssx!myCssFile'], function (cssxDef) {
// do some awesome css stuff here
});Plugins can also have configuration options. Global options can be specified on curl's configuration object. Options can also be supplied to plugins via suffixes. Suffixes are also delineated by the ! symbol. Here's an example of a plugin using options:
// don't try to repair IE6-8 opacity issues in my css file
define(['css!myCssFile!ignore:opacity'], function (cssxDef) {
// do some awesome css stuff here
});I have dozens (or hundreds) of modules. Even with parallel loading, the performance sucks! What can I do about that?
True! No parallel loader can lessen the latency required to create an HTTP connection. If you have dozens or hundreds of files to download, it's going to take time to initiate each of the connections.
However, there are tools to that are designed to fix this problem! There are builders and compilers. dojo users are probably already familiar with dojo's build tool and optimizer. RequireJS comes with a build tool and Google's Closure compiler.
The build tool is used to concatenate several modules (and/or resources) into just a few files. It does this by following the dependency chain specified in the define() and require() calls. You can specify which top-level modules or resources are in each file and the build tool finds the rest.
After the build tool creates the concatenated files, the files can be passed into a compiler (also called a shrinker or compressor).
curl.js is compatible with RequireJS's build tool, r.js, but there's also cram. Cram is the cujoJS resource assembler.
cujoJS supports packages. Packages are defined by the packages
configuration parameter:
curl = {
baseUrl: 'path/to/js',
packages: {
'my-package': {
location: 'path/to/my-package',
main: 'main/main-module-file',
config: { /* package-specific configuration options */ }
}
}
};curl = {
baseUrl: 'path/to/js',
packages: [
{
name: 'my-package',
location: 'path/to/my-package',
main: 'main/main-module-file',
config: { /* package-specific configuration options */ }
}
]
};The path property describes where to find the package in relation to the baseUrl parameter. The main and lib properties describe where to find modules inside the package. The main property gives the relative path to the package's main module.
In the example above, the main module of the package can be obtained as follows:
curl(['my-package'], callback);and will be fetched from the following path:
path/to/js/path/to/my-package/main/main-module-file.jsSome other file in the package would be obtained as follows:
curl(['my-package/other-module'], callback);and will be fetched from the following path:
path/to/js/path/to/my-package/other-module.jsNote that a package may also contain it's own embedded set of config options. Most, but not all, config options may be specified here. Specifically, you cannot specify any options that change the path or id of modules, such as paths, packages, plugins, or 'pluginPrefix.
cujoJS is the JavaScript Architectural Toolkit. It employs MVC, IOC, AMD and lots of other TLAs. :) Our goal is to provide architectural tools and guidance. Find out more at cujoJS.com.
Many thanks to Bryan Forbes (@bryanforbes) for helping to clean up my code and for making curl's domReady much more robust and help with dojo compatibility. More about Bryan: http://www.reigndropsfall.net/
Kudos also to James Burke (@jrburke) who instigated the AMD standard and paved the way to create AMD-style loaders. More about James: http://jrburke.com/about/
Shout out to Kris Zyp (@kriszyp) for excellent ideas and feedback and to Kyle Simpson (@getify) who is inarguably the godfather of javascript loading.