alameda

An AMD loader, like requirejs, but with the following implementation changes:

Assumes Promises are available in the JS environment.
Targets "modern" web browsers that implement standardized script.onload behavior: execute load listener right after script execution, something IE9 and below did not do.
Assumes browser support for Array.isArray, array extras, ES5 features.
Does not support a couple of less-used APIs (see tests section below).

These changes means alameda is around 35% smaller than requirejs, 4.1 KB vs 6.4 KB, minified+gzipped sizes.

Browser support: browsers that natively provide Promises. If you need to support IE 10 and 11, the alameda-prim project includes a private promise shim.

You can continue to use requirejs and the r.js optimizer for other scenarios. The r.js optimizer works well with alameda-based projects.

Install

Latest release information

If using a package manager:

npm install alameda

# or

[npm | bower | volo] install requirejs/alameda

API

alameda supports the requirejs API. It even declares requirejs, to make passing the requirejs tests easier. alameda also has a good chance of becoming requirejs in a far-future requirejs version.

There are some differences with requirejs though:

require promise

require([]) will return a promise. The success callback passed to require([]) should return a value if you want a value to be passed to the next .then() in the promise chain.

require(['a', 'b'], function(a, b) {
  // succes callback. Return a value for the next part in the promise chain.
  return [a, b];
}).then(function(mods) {
  //mods[0] is the 'a' module, mods[1] is the 'b' module in this case.
});

config.defaultErrback

In requirejs and alameda, with this sort of call, the errback will be called if there is an error in either loading ['a', 'b'] or if the success callback throws an error.

require(['a', 'b'], function(a, b) {
  // success callback
}, function(err) {
  // errback, called if 'a', 'b' do not load, or
  // if the success callback is called.
});

So, the errback operates like:

require([], function() {}).catch(function(err) {})`;

If you do not pass an errback into the require() call, and instead use a .then() or .catch() to deal with the error, you still may see the error surface outside. This is done because browsers do not all show unhandled errors in a promise chain, and the require() call itself does not know if an error handler was chained on the end, so it generates an error to make debugging and development easier.

However, if you are properly chaining error handlers but do not pass an errback as the third arg to the require([]) call, then you can turn off this extra error surfacing by doing:

requirejs.config({
  defaultErrback: null
});

If you pass a function for the defaultErrback value, then that will be used instead of the default "delayedError" handler used by alameda to surface the error.

onError is context-specific

When using contexts, in requirejs, all top level errors would bubble up to requirejs.onError, but in alameda, the context's onError is called instead. Example:

var fooReq = requirejs.config({ context: 'foo' });
requirejs.onError = function () { console.log('requirejs.onError'); };
fooReq.onError = function () { console.log('fooReq.onError'); };

fooReq(['nonexistent']);

// In alameda, fooReq.onError() is called, in requirejs, requirejs.onError is called.

onResourceLoad

requirejs supports a hook into its internals, onResourceLoad. alameda supports an onResourceLoad function too, but the arguments passed to the function are objects that have different property names than the ones in requirejs.

This is the general signature, which is the same between alameda and requirejs:

alameda.onResourceLoad = function (context, map, depArray) {};

The differences between property names in the different argument objects is described below. See the onResourceLoad page for the description of the arguments.

context

alameda	requirejs
`id`	`contextName`

map

alameda	requirejs
`pr`	`prefix`
`n`	`name`
n/a	`parentMap`
`url`	`url`
n/a	`originalName`
`id`	`fullName`

depArray

An array of map objects with the same properties as the map listing above.

License

MIT

Code of Conduct

jQuery Foundation Code of Conduct.

Running tests

The tests are pulled from almond and requirejs. All tests should be served through a local web server, as the text loader plugin is used for some tests, and some browsers restrict local XHR usage when the files are served from a file:// URL.

Bundled tests

To run the tests that are just part of this repo, open tests/index.html in a web browser.

requirejs tests

To run the requirejs tests, first make sure the following projects have been cloned and are siblings to the the alameda repo:

Then do the following:

symlink alameda.js to require.js
./copyrequirejstests.sh

requirejs tests that do not pass

require.undef()-related tests.
onResourceLoadNestedRequire: depends on implementing requirejs.onResourceLoad hook used for builds/some third party tools. This API is not required for normal module loading.

How to get help

Open issues in the issue tracker.
Contact the requirejs list.

ejdaly/alameda