The PINF JavaScript Loader
is part of the PINF Platform for JavaScript
and is a tiny optimized (intended for production use) CommonJS package mappings
based JavaScript module loader for the browser.
- Mailing list: groups.google.com/group/pinf
Online Demo: pinf.js.org/pinf-loader-js/ (Code Coverage)
Namespace isolation is essential for modular development when integrating arbitrary JavaScript libraries.
To achieve namespace isolation you need JavaScript libraries written in conventions that:
- do not pollute the global namespace and
- expose the library's API consistently
There are two evolving standards that specify such conventions:
When coding using these standards you need to keep in mind the two primary environments that the application will run in:
-
Development - Needs a loader that will, on demand, locate in the source tree, assemble and transport module source files to the browser for rapid development.
-
Production - Needs a build step that collects modules from the source tree and generates static optimized bundles that will be fetched by a loader optimized for production runtime performance.
The PINF JavaScript Loader
is intended to run your application in development and production and may be fed
by various Bundlers
(see above).
The PINF JavaScript Loader
provides a minimal CommonJS environment that requests optimized static JavaScript code files
called Bundles from a server using SCRIPT injection (GET requests) and boots these into sandboxes in the browser identified by the requested URL.
- Browser:
- Firefox 3.6+
- Google Chrome
- Internet Explorer 6+
- Safari 5+
- Opera
- iPad
- iPhone
- Android
- Web Worker
- Load bundled JavaScript programs from static URLs
- Asynchronously load more program code bundles as needed
- Load bundles cross-domain
- Isolated module scopes
- Isolated package namespaces
- Isolated sandbox namespaces
- Nested and circular dependency trees
- Consistent mapping of static application resource URLs to loader namespaces
- CommonJS/Modues/1.1
function(require, exports, module) {}
var ModuleAPI = require("./Module")
- CommonJS/Packages/Mappings/C (proposal)
package.json ~ {mappings:{"PackageAlias": "PackageIdentifier"}}
var ModuleAPI = require("PackageAlias/Module")
- CommonJS/Modues/2.0draft8 (draft)
global.require.memoize("PackageIdentifier/ModuleIdentifier", ModuleInitializer)
(no dependency argument)require.id(ModuleIdentifierString)
(returns PackageIdentifier/ModuleIdentifier)
- (Un)CommonJS(kriskowal)/Modules
require.async(ModuleIdentifierString, function loaded(ModuleAPI) {}, function error(e) {})
- Proposed:
[global.]require.sandbox(SandboxURI, SandboxOptions, function loaded(sandbox) {}, function error(e) {})
[global.]require.sandbox.id
to hold SandboxURIsandbox.main()
require.bundle("BundleIdentifier", function ConsistentModuleSet(require) {})
<script src="dist/loader-core.js"></script>
<script>
PINF.sandbox("app.js", function (sandbox) {
sandbox.main();
});
</script>
PINF.bundle("", function (require) {
require.memoize("/main.js", function (require, exports, module) {
exports.main = function () {
console.log("HelloWorld!");
}
});
});
There are various feature examples that double as tests in ./features
.
For an online demo of these loader features see pinf.js.org/pinf-loader-js/.
There are adapters for streamlined integration with other tools:
- RequireJS plugin: github.com/pinf/pinf-for-requirejs (status: DEV)
- NodeJS package: github.com/pinf/pinf-for-nodejs (status: DEV)
- jQuery plugin: github.com/pinf/pinf-for-jquery (status: DEV)
Applications may be coded directly in the bundle format. Alternatively the bundle format may be treated as a compile target (typical).
The following tools can generate PINF JavaScript Loader
compatible bundles:
-
PINF.it Bundler (status: DEV)
Supports:
- NodeJS Modules
- CommonJS Modules (CJS)
- Asynchronous Module Definition (AMD)
- Plain old JavaScript
- Various other formats
- When testing an application use the
./dist/loader-full.browser.js
file to get all error messages. - When deploying a standalone loader us the
./loader-core.browser.min.gz
or./loader-core.browser.min.br
files for optimum performance. - When using a different loader during development make sure only supported API features of this loader are used. Load extra features along with your application by augmenting a sandbox.
- When writing or generating bundles make sure one consistent set of statically linked modules
is contained in each bundle file. Dynamic links to other modules or bundles must be made using
require.async()
orrequire.sandbox()
respectively. The hierarchy of how your application nests these dynamic links will determine which modules must be included in subsequently loaded bundles to avoid sending the same modules twice. - A module can only be memoized once for each Canonical Identifier (comprising of SandboxIdentifier/PackageIdentifier/ModuleIdentifier). When placing modules into bundles make sure bundle filenames do not overlap with module filenames (and the reverse) as these have the potential to conflict (modules and bundles share the same logical file hierarchy). The idea is that a set of statically linked modules can always be combined into one file which is placed into the file that first requires the dependencies and represents the entry point into the bundle.
This loader is pretty much complete in terms of what needs to be implemented at the core loader level. Convenience features can be loaded along with the application by augmenting a sandbox.
Why does the loader not support AMD-style Loader Plugins?
Because loader plugins that are invoked by modifying the string literal passed to require()
are not necessary
and combine two concepts that should really be separate and implemented differently.
The AMD-style Loader Plugins can be replaced by:
- Augmenting a sandbox
- Loading helper modules within the application.
- Using a loader that can run package-declared plugins.
- Using a server helper to run plugins as modules are requested.
NOTE: Modules using some of the RequireJS loader plugins can be automatically converted to run on this loader using github.com/pinf-it/pinf-it-bundler.
How does the loader compare to almond?
While the RequireJS + almond combination focuses on loading of optimized AMD formatted modules this loader focuses on loading of optimized CJS formatted modules.
The AMD Specification is a small subset combining several CommonJS Concepts in a different form.
CommonJS represents a more pure and modular approach to devising arbitrary JavaScript application architectures by carefully layering a few core concepts into a framework that provides one small existential foundation for all other concepts. It allows for isolated namespaces, nested package dependency structures and runtime sandboxes as well as automatic conversion from source trees to optimized bundles. This loader is one existential foundation implementation and fully compatible with the CommonJS Concepts.
In contrast RequireJS + almond focuses on optimally loading (primarily into the browser) a list of packages containing JavaScript modules and resource files into a single namespace. In optimized form (for almond), several key RequireJS features are not supported.
Influential Specifications:
- CommonJS/Modues/1.1 (approved)
- (Un)CommonJS(kriskowal)/Modules
- CommonJS/Modues/2.0draft8 (draft)
- CommonJS/Packages/1.1 (draft)
- CommonJS/Packages/Mappings/C (proposal)
- Asynchronous Module Definition (AMD)
Prior Art:
- https://github.com/unscriptable/curl
- https://github.com/jrburke/almond
- https://github.com/jrburke/requirejs
- http://code.google.com/p/bravojs/
- https://github.com/NobleJS/Noble-Modules
- https://github.com/pinf/loader-js
- https://github.com/kriszyp/nodules
Requirements:
Install:
nvm use
npm install
Run tests:
# Headless Chrome
npm test
# Firefox, Chrome, Safari
npm run test-all
Launch development workspace (uses puppeteer):
npm run dev
Then make changes to source code and browser page will reload.
To make a release use:
npm run release
This will first build the dist/
files (via npm run build
) and then increment, tag, push and publish.
Original source logic under Zero-Clause BSD by Christoph Dorn since 2011.