Today I was playing around with Backbone and Javascript, and came upon an interesting problem. Consider the following:
var App = {};
App.model = Backbone.Model.Extend({/* Omitted for brevity */});
App.view = Backbone.View.Extend({/* Omitted for brevity */});
App.obj = new App.view();
At this point I wanted to refactor for readability and mantainability's sake:
var App = {
model: Backbone.Model.Extend({}),
view: Backbone.View.Extend({}),
obj: new view() // Here comes trouble
}
The snippet above is what I wanted to obtain, but obviously the initialization doesn't work:
view is not in scope
App is not yet initialized, so App.view isn't usable.
this refers to window, so this.view isn't usable.
At this point, with the help of this answer I concocted a solution:
var App = {
model: Backbone.Model.Extend({}),
view: Backbone.View.Extend({}),
get obj() { delete this.obj; this.obj = new this.view(); return this.view; }
}
Using the getter I'm able to delay the creation of the object instance until used (thus after completing App initialization), and then I replace the getter from within with the object instance.
Everything works as expected, but since I'm not exactly fluent with modern JS I was wondering if there was a different or proper way to achieve what I wanted. I was also wondering if this sort of implementation could cause drawbacks or unexpected problems I'm not considering.
First and foremost I want to thank everyone for the insights. For the sake of completion, I wanted to post what seems to be the best and less obnoxious way to obtain what I wanted in the first place. While my hack works, it should never be used, and converting the literal into a constructor gives the same result with no hacks:
function App() {
this.model: Backbone.Model.Extend({});
this.view: Backbone.View.Extend({});
this.obj = new this.view();
}
var app = new App();
This version keeps itself dry, and has two added benefits: instantiation, and "private" members. It could easily be rewritten as such:
function App() {
var model: Backbone.Model.Extend({});
var view: Backbone.View.Extend({});
this.obj = new view();
}
var app = new App();
This would keep both model and view out of reach, while having easily accessible app.obj. Of course, if instantiation is an unwanted effect, nothing beats var App = {}.
Also to be noted that since my example was featuring Backbone, the constructor way is probably suggested when extensibility is welcome, and prototypes for the class can be added via _.extend(). This is the the way Backbone does it:
function App() {
// […]
}
_.extend(App.prototype, {
method1: function() {}
method2: function() {}
});
References: _.extend(), object constructors, literals vs. constructors.
Related
I have a JavaScript MVC design, implemented using prototypes, where different items may be displayed in different ways as decided by the controller. For example an 'Event' item may be displayed with the views ViewTabs or ViewSimple. Class hierarchy:
ViewBase
- ViewTabs
-- EventViewTabs, which implements EventViewTabs.Validate
- ViewSimple
-- EventViewSimple, which implements EventViewSimple.Validate
Deciding whether to use EventViewTabs or EventViewSimple is done by a EventController. My problem is: I have a Validate method for checking inputs from the Event views, but this method is identical for the EventViewTabs and the EventViewSimple views. Where should I put Validate in order to avoid duplication? I cannot put it in ViewBase, as other items (e.g. User) also inherit from this class.
Seems I need multiple inheritance for this, but is there a smarter way to do it? I have a feeling I'm overlooking something obvious.
You're missing composition. Inheritance isn't the answer to all issues about code reuse to avoid copy-paste programming.
Let's say you've a View base prototype:
function View() { }
If you want this view to support validation, you can inject the validation dependency in the constructor function:
function View(validator) {
this.validator = validator;
}
View.prototype = {}; // A lot of functions here
That is, now any view which inherits View's prototype will have an associated validator. In other words: you don't need to derive two prototypes in your concrete view (you don't need and you can't do it anyway).
In the other hand, in terms of object-oriented programming, it wouldn't make sense to derive from Validator to create a View.
When you say a view has a validator, since you're using has as verb, you're talking about an association (a form of composition). Alternatively, when you say my main screen is as view, we're talking about an inheritance, because a specific view must be also a view, so it needs base view's members to act like a view.
Basically your validator could be tailor-made with the type it has to work with. In UML, it's called composition. I figure out your code as follows:
function Validator {}
Validator.prototype.validate = function(arg) {
//arg is no longer inputs
return true|false; //the ultimate output along with additional information;
}
function EventViewTabsValidator() {}
EventViewTabsValidator.prototype = Object.extend(Validator.prototype); //inheritance
EventViewTabsValidator.prototype.constructor = EventViewTabsValidator; //enforce the constructor to point to your derived type
EventViewTabsValidator.prototype.validate = function() {
var inputs = $('inputs');
var param = 'do some stuff specific to EventViewTabsValidator based on the inputs';
return Validator.prototype.validate.call(this, param); //pass param, not inputs
}
function EventViewSimpleValidator() {}
EventViewSimpleValidator.prototype = Object.extend(Validator.prototype); //inheritance
EventViewSimpleValidator.prototype.constructor = EventViewSimpleValdiator; //enforce the constructor to point to your derived type
EventViewSimpleValidator.prototype.validate = function() {
var inputs = $('inputs');
var param = 'do some stuff specific to EventViewSimpleValidator based on the inputs';
return Validator.prototype.validate.call(this, param); //pass param, not inputs
}
function EventViewTabs() {
this.validator = null; //see init
}
EventViewTabs.prototype.init = function() {
this.validator = new EventViewTabsValidator();
}
function EventViewSimple() {
this.validator = null; //see init
}
EventViewSimple = function() {
this.validator = new EventViewSimpleValidator();
}
Your could abstract up both types to a base EventView, which could expose this.validator.
Your instance of EventController will call:
var simple = new EventViewSimple();
simple.validator.validate();
var tabs = new EventViewTabs();
tabs.validator.validate();
Whatever the EventView instance, they implement their own specific validator that can be called in a generic way.
One approach is to use mixins to add the other behavior (this is the ruby approach, and is also used by react.js and react.rb) You can google for javascript+mixins and find some excellent tutorials like this one: http://raganwald.com/2014/04/10/mixins-forwarding-delegation.html
For your specific case validate (or perhaps validator) would be the mixin.
Why not to do something like this:
ViewBase
ViewBase
-EventValidator, which implements Validate
--ViewTabs
---EventViewTabs
--ViewSimple
---EventViewSimple.
Also consider to use composition over inheritance see this video
Will keep it short: I have the need to build different components for my app. What I'm calling a "component" here is a collection of methods and constructors that repeats itself in many places of my app but is not necessarily exactly the same.
For the sake of an example, consider a pagination component. It is made of a few methods, a view and a template:
var Pagination = function() {
this.View = Backbone.View.extend({});
this.method = function() {};
this.method2 = function() {};
};
// Create instance
var myComponent = new Pagination();
Heavily abstracted code, but gives the idea. Now, I've created a "Component" constructor to aid me in my task. It is fairly simple:
Component = function() {
this.initialize.apply(this, arguments);
};
_.extend(Component.prototype, {
initialize: function() {}
});
Component.extend = Backbone.Model.extend;
This allows me to build multiple instances of the same component with slightly different methods or contents, without having to rebuild the same thing over and over again:
var Pagination = Component.extend({
View: Backbone.View.extend({}),
method: function() {},
method2: function() {}
}};
// Create instance
var myComponent = new Pagination();
So far so good. myComponent has the list of methods as well as the constructor for the View. Here is where the problem lies: for some of my Components's instances I need to extend not only the component itself, but the Backbone constructors inside it as well. Say, for example, that I want to extend (not replace) the View inside my pagination component. That's were everything falls down.
This is the last approach I've had, but the basic idea is: I have components that repeat themselves frequently in my app but with slight differences between them. These components may include methods, backbone constructors and primitive values inside of them. I need to be able to extend these but ALSO be able to extend it's individual parts (Backbone constructors) if the need arises.Anyone has an idea on how to accomplish this?
I am building an application using Durandal and I have the need to share some functionality across view models.
I have 5 screens to build and they are all virtually the same screen except that in the activate function they will call to a different api end points but otherwise the view and view models will be identical.
Is there a pattern that I should be following to structure this correctly to promote code reuse?
If the views and the view models are identical except for calling different api actions, what about just taking in a parameter as part of the route? Then in the activate function, you can switch on the parameter. The route values can be designated so that your url is relevant, like [http://site/page/subtype], where subtype is the parameter (instead of using numeric values)
Regarding inheritance, depending on the features you need, there's so many ways to do JavaScript inheritance it can be a little confusing. There are some full-featured inheritance models provided by libraries such as base2 and Prototype. John Resig also has an inheritance model that I've used successfully.
In general, I prefer to stick to simpler solutions when it comes to JS inheritance. If you need a pretty much the full set of inheritance features, those libraries are good to consider. If you only really care about accessing a set of properties and functions from a base class, you might be able to get by with just defining the view model as a function, and replacing the function's prototype with the desired base class. Refer to Mozilla's Developer Docs for good info on inheritance.
Here's a sample:
//viewModelBase
define(function (require) {
"use strict";
function _ctor() {
var baseProperty = "Hello from base";
function baseFunction() {
console.log("Hello from base function");
}
//exports
this.baseProperty = baseProperty;
this.baseFunction = baseFunction;
};
//return an instance of the view model (singleton)
return new _ctor();
});
//view model that inherits from viewModelBase
define(function (require) {
"use strict";
function _ctor() {
var property1 = "my property value";
function activate() {
//add start up logic here, and return true, false, or a promise()
return true;
}
//exports
this.activate = activate;
this.property1 = property1;
};
//set the "base"
var _base = require("viewModelBase");
_ctor.prototype = _base;
_ctor.prototype.constructor = _ctor;
//return an instance of the view model (singleton)
return new _ctor();
});
Keep in mind this example all results in what effectively is a singleton (i.e. you'll only get the same instance, no matter how many times you require() it)
If you want a transient (non-singleton) just return _ctor. Then you'll need to instantiate a new instance after you require() it.
One more note, in general, functions should be defined on the prototype, not within the constructor function itself. See this link for more information on why. Because this example results in only a single instance, it's a moot point, so the functions are inside the constructor for improved readability and also the ability to access the private vars and functions.
Could someone explain the fundamental difference between:
define(['backbone'], function(Backbone) {
MyModel = Backbone.Model.extend({
});
});
define(['backbone', 'models/mymodel'], function(Backbone){
var app = Backbone.View.extend({
initialize: function() {
var model = new MyModel();
}
});
});
and:
define(['backbone'], function(Backbone) {
var MyModel = Backbone.Model.extend({
});
return MyModel;
});
define(['backbone', 'models/mymodel'], function(Backbone, MyModel){
var app = Backbone.View.extend({
initialize: function() {
var model = new MyModel();
}
});
});
In the former, the first module simply defines MyModel. In the latter, it's created as a variable and returned, and the second module needs to have it put in the parameters when imported.
RequireJS examples I see around seem to vary between the two, but I don't really understand the difference - does one return an instance and the other a constructor?
In my application I didn't even notice that I was actually using both ways in different places, and I think it was causing problems. I was using a lot of
self = this
self.model.doSomething
inside my views and models, and as my app got bigger, I started getting errors because there were conflicts with definitions of self.
Short Version: 1st version == wrong.
Medium Version: The first one bypasses Require entirely by using global variables, while the second one actually uses Require.
Long version:
The way Backbone modules work is that you run "define", pass it a function (and usually an array of dependencies also), and whatever gets returned from that function is defined as that module. So if I do:
// Inside foo.js
define([], function() {
return 1;
});
I've defined the "foo" module to be 1, so if elsewhere I do:
define(['foo'], function(foo) {
alert(foo); // alerts 1
});
Your first version doesn't return anything, so it's not actually creating a Require module at all.
How does it work then? Well, in that version you do:
MyModel = Backbone.Model.extend({
NOT:
var MyModel = Backbone.Model.extend({
So that's really the same as doing:
window.MyModel = Backbone.Model.extend({
Then when the second part of the code runs, it access window.MyModel, and works ... but it's completely bypassing Require.js in the process.
I think the most important thing to takeaway is: ALWAYS DECLARE (ie. var) YOUR JAVASCRIPT VARIABLES. I don't agree with everything Crockford says, but he's dead right on this one. You will get lots of bugs (with Require and without) if you don't make this a habit.
Beyond that, the next most important thing is probably: ALWAYS RETURN SOMETHING FROM THE FUNCTION YOU PASS TO define. There are certain special cases where you don't want to return anything, but unless you are deliberately trying to solve one of those cases you should always return something to define the module.
Finally, if you're using Require, every variable in your code should either:
Come from the define function (ie. it should be an argument variable from the function that you pass to define), or
It should be declared (ie. var-ed ) inside that file
If you use JSLint or 'use strict'; (as Valentin Nemcev suggested), or if you use an editor like Eclipse, your tools can help you ensure this (and in fact make it easy to ensure).
MyModel = Backbone.Model.extend({});
Here you are not returning a constructor, you are defining a global variable and accessing it later in different module.
Actually it is wrong, it works by accident. You should return your modules from define and access them via parameters in other modules.
Like this:
return Backbone.Model.extend({});
You should use strict mode to avoid problems with global variables in JS.
Also, constructor in JS is just a function that is meant to be run with new. Backbone extend always returns a constructor function, and you create a model instance by calling the constructor with new, like you are doing in both examples.
How much can I stretch RequireJS to provide dependency injection for my app? As an example, let's say I have a model that I want to be a singleton. Not a singleton in a self-enforcing getInstance()-type singleton, but a context-enforced singleton (one instance per "context"). I'd like to do something like...
require(['mymodel'], function(mymodel) {
...
}
And have mymodel be an instance of the MyModel class. If I were to do this in multiple modules, I would want mymodel to be the same, shared instance.
I have successfully made this work by making the mymodel module like this:
define(function() {
var MyModel = function() {
this.value = 10;
}
return new MyModel();
});
Is this type of usage expected and common or am I abusing RequireJS? Is there a more appropriate way I can perform dependency injection with RequireJS? Thanks for your help. Still trying to grasp this.
This is not actually dependency injection, but instead service location: your other modules request a "class" by a string "key," and get back an instance of it that the "service locator" (in this case RequireJS) has been wired to provide for them.
Dependency injection would involve returning the MyModel constructor, i.e. return MyModel, then in a central composition root injecting an instance of MyModel into other instances. I've put together a sample of how this works here: https://gist.github.com/1274607 (also quoted below)
This way the composition root determines whether to hand out a single instance of MyModel (i.e. make it singleton scoped) or new ones for each class that requires it (instance scoped), or something in between. That logic belongs neither in the definition of MyModel, nor in the classes that ask for an instance of it.
(Side note: although I haven't used it, wire.js is a full-fledged dependency injection container for JavaScript that looks pretty cool.)
You are not necessarily abusing RequireJS by using it as you do, although what you are doing seems a bit roundabout, i.e. declaring a class than returning a new instance of it. Why not just do the following?
define(function () {
var value = 10;
return {
doStuff: function () {
alert(value);
}
};
});
The analogy you might be missing is that modules are equivalent to "namespaces" in most other languages, albeit namespaces you can attach functions and values to. (So more like Python than Java or C#.) They are not equivalent to classes, although as you have shown you can make a module's exports equal to those of a given class instance.
So you can create singletons by attaching functions and values directly to the module, but this is kind of like creating a singleton by using a static class: it is highly inflexible and generally not best practice. However, most people do treat their modules as "static classes," because properly architecting a system for dependency injection requires a lot of thought from the outset that is not really the norm in JavaScript.
Here's https://gist.github.com/1274607 inline:
// EntryPoint.js
define(function () {
return function EntryPoint(model1, model2) {
// stuff
};
});
// Model1.js
define(function () {
return function Model1() {
// stuff
};
});
// Model2.js
define(function () {
return function Model2(helper) {
// stuff
};
});
// Helper.js
define(function () {
return function Helper() {
// stuff
};
});
// composition root, probably your main module
define(function (require) {
var EntryPoint = require("./EntryPoint");
var Model1 = require("./Model1");
var Model2 = require("./Model2");
var Helper = require("./Helper");
var entryPoint = new EntryPoint(new Model1(), new Model2(new Helper()));
entryPoint.start();
});
If you're serious about DI / IOC, you might be interested in wire.js: https://github.com/cujojs/wire
We use a combination of service relocation (like Domenic describes, but using curl.js instead of RequireJS) and DI (using wire.js). Service relocation comes in very handy when using mock objects in test harnesses. DI seems the best choice for most other use cases.
Not a singleton in a self-enforcing getInstance()-type singleton, but
a context-enforced singleton (one instance per "context").
I would recommend it only for static objects. It's perfectly fine to have a static object as a module that you load using in the require/define blocks. You then create a class with only static properties and functions. You then have the equivalent of the Math Object that has constants like PI, E, SQRT and functions like round(), random(), max(), min(). Great for creating Utility classes that can be injected at any time.
Instead of this:
define(function() {
var MyModel = function() {
this.value = 10;
}
return new MyModel();
});
Which creates an instance, use the pattern for a static object (one where values are always the same as the Object never gets to be instantiated):
define(function() {
return {
value: 10
};
});
or
define(function() {
var CONSTANT = 10;
return {
value: CONSTANT
};
});
If you want to pass an instance (the result of using a Module that have return new MyModel();), then, within an initialize function, pass a variable that capture the current state / context or pass on the Object that contains information on state / context that your modules needs to know about.