A very lightweight and dependency-free Flux implementation
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.md

Build Status npm version npm downloads

nanoflux

PROJECT SITE

nanoflux is a very lightweight (about 3.5 KiB minified, and 1.25 KiB gzipped) dependency-free Flux implementation.

The idea of this implementation is to support a very small, but full Flux implementation (separated Action, Dispatcher, and Store), and also a "fluxy" version, with Action and Dispatcher merged in one unit.

Furthermore, nanoflux uses a pure functional approach as a performant solution.

Features

  • Extremely tiny implementation
  • No dependencies at all
  • Support for full Flux using full stack of ActionProvider/Creator, Dispatcher, and Stores
  • Support for a simplified 'fluxy' concept, where Dispatcher is also ActionProvider
  • Interoperable/Chainable Stores
  • Multiple Dispatchers
  • Built in ActionCreator
  • Quite fast
  • CommonJS, RequireJS ready
  • Middleware support (new, since 1.1.0)

Hint: Have a look at nanoflux-fusion, which is an extension for nanoflux providing a even more comfortable API inspired by redux

Comparison to Facebook's Implementation

From an architectural point of view, the main difference is that Facebook's Flux implementation provides one central dispatcher, while nanoflux supports also multiple dispatchers (if needed). Given that flexibility, it is possible to link multiple stores and multiple dispatchers, but IMHO this would only be a preferable scenario for really large applications. Additionally, it is also possible (as a built in feature) to link stores easily, so they can notify each other on changes (chaining).

For more comfort, nanoflux supports a 'fluxy' way, which means, that a dispatcher provides actions directly without the need of a dedicated ActionProvider. This can be quite handy in less complex applications and reduces much of boilerplate code. Of course, nanoflux supports the original concept with separated ActionProvider.

The verbosity may be one of the 'weakest' aspects of Facebook's Flux: this is due to the fact, that Facebook provides the Dispatcher only. A Store and/or an ActionProvider is not part of their library, and therefore Facebook's Flux implementation is very lightweight, too. And even a bit smaller than nanoflux. The developer gains more liberty on implementation decisions, but for the costs of more work. For example, it is left to the developer how stores and actions may interoperate, p.e. common approaches base on event emitters.

In this point nanoflux offers slightly less flexibility with its convention based the dispatcher-store-binding, but is more comfortable.

Since version 1.1.0 nanoflux provides a middleware, which allows something like a payload data transformation pipeline, or dispatch inspection for logging, or debugging tools.

Size

nanoflux is a really tiny implementation, although it offers much more comfort than the reference implementation from Facebook.

  1. fb.flux.min.js ca. 2 KiB
  2. nanoflux.min.js ca. 3.5 KiB
  3. reflux.min.js ca. 18 KiB
  4. delorean.min.js ca 20 KiB
  5. alt.min.js ca 23 KiB

Performance

nanoflux use synchronous function calls, that makes nanoflux quite fast. Synchronous cycles guarantee consistent dispatch cycles.

Here are some results of benchmarks for entire action-dispatch-notify-cycles:

  1. fbflux-perf: 163983.67 op/s (0.00 op/s) - 100.00%
  2. nanoflux-fluxy-perf: 157380.00 op/s (-6603.67 op/s) - 95.97%
  3. nanoflux-fullflux-perf: 151334.33 op/s (-12649.34 op/s) - 92.29%
  4. reflux-perf: 61861.33 op/s (-102122.34 op/s) - 37.72%
  5. alt-perf: 27704.33 op/s (-136279.34 op/s) - 16.89%
  6. delorean-perf: 9350.33 op/s (-154633.34 op/s) - 5.70%

The benchmark code is available under ./perf.

Currently, all measuring is done server side using nodejs (listed results run on Dell XPS15 i7). I think it is slightly slower than Facebooks implementation, as nanoflux uses a comfortable auto-binding, without verbose switch-case-statements like the Facebook version. Nevertheless, it should be fast enough :)

Example

The following example demonstrates the 'full' Flux approach, using ActionProvider, Dispatcher, and Store

	var NanoFlux = require('nanoflux'); // UMD with browserify!

    var setup = function() {
    
        // Creating a store 'myStore' with functions triggered by dispatched actions
        // The convention for action handlers name is: on<ActionName>
        NanoFlux.createStore('myStore', {
    
            // the handlers signature bases on the users convention
            onAction1: function (test) {
                console.log("Store.onAction1: " + test);
                // this will call the subscribed callbacks
                this.notify({data: test});
            },
    
            onAction2: function (test) {
                console.log("Store.onAction2: " + test);
                this.notify({data: test});
            },
    
            onAction3: function (test) {
                console.log("Store.onAction3: " + test);
                this.notify({data: test});
            }
        });

		// Creating the Dispatcher
		// You may also use the implicit default dispatcher: 
		// var dispatcher = NanoFlux.getDispatcher();
        var dispatcher = NanoFlux.createDispatcher('myDispatcher');
        
        // The full flux concept foresees a separation of actions and dispatcher
        // Here we create actions using the built in action creator
        NanoFlux.createActions('myActions', dispatcher, {
            action1 : function(data){
                console.log("Action 1");
                // this way, the dispatcher establishes dynamically the action binding, calling stores onAction1().
                this.dispatch('action1', data);
            },
    
            action2 : function(data){
                console.log("Action 2");
                this.dispatch('action2', data);
            }
        });    
    };
    
    setup();
    function Component(){
    
        // callback called by Store.notify
        this.onNotify = function(data){
            console.log("Component notified: " + JSON.stringify(data));
        };
    
        this.exec = function(){
    
                
            var dispatcher = NanoFlux.getDispatcher('myDispatcher');
            var store = NanoFlux.getStore('myStore');
            var actions = NanoFlux.getActions('myActions'); 
            
            // Now, connecting Store and Dispatcher
            dispatcher.connectTo(store);
            
            // establishes the link between store's notification mechanism and this component.
            // use the returned object to unsubscribe, if needed!
            var subscription = store.subscribe(this, this.onNotify);
    
			// executing the actions    
            actions.action1("test 1");
            actions.action2("test 2");
        };
    }   

Middleware Example

Applying middleware is as simple as licking ice cream on the beach:

function Logger(){
    var log = [];

    return function(handlerName, args){
        log.push({
            handler: handlerName,
            payload : args
            }
        )
    }
}

// somewhere in your app --- using the fluxy approach for sake of simplicity
// ...
var dispatcher = NanoFlux.createDispatcher(null, ["action1", "action2"]);
NanoFlux.use(new Logger(), dispatcher);

dispatcher.action1({foo:"fromAction1"});
/* Log is:  [{handler: "onAction1", payload: [{foo:"fromAction1"}]}] */

dispatcher.action2({foo:"fromAction2"});
/* Final Log is:
    [
        {handler: "onAction1", payload: [{foo:"fromAction1"}]}
        {handler: "onAction2", payload: [{foo:"fromAction2"}]}
    ]
*/

Getting nanoflux

You may pick the library directly from ./dist, use npm install nanoflux, or use bower install nanoflux

Build your own

  1. Get sources: npm install nanoflux (or fork/clone this repo)
  2. Install dependencies: npm install
  3. Build:
  • gulp to build the minified and non-minified bundle in .\dist

Pronto!

Automated Testing

The gulp build chain runs tests only for the browserified nanoflux module.

All tests can be run using jasmine-node or npm run test.

Available Node Tasks

Use npm run <task> to execute additional task. Available tasks are:

  • test : Runs all tests
  • benchmark : Runs a performance benchmark for different Flux Implementations.