Now, Require.js is my favorite way to program Javascript. It breaks the code into smaller pieces and makes it easier to manage. The Require.js Optimizer can help us disperse a larger application into multiple smaller applications, connect them through dependencies, and finally merge them during compilation and packaging. These reasons lead us to use require.js.
So, let’s take a look at the awesome features of require.js!
Compatible with CommonJS
AMD (Asynchronous Module Definition Specification) emerged from the CommonJS working group. CommonJS aims to create a Javascript ecosystem. An important part of CommonJS is transport/c, the predecessor of AMD, and require.js is an implementation of this specification.
The syntax difference between the CommonJS module and the AMD module is mainly due to the fact that AMD needs to support the asynchronous features of the browser. The CommonJS module needs to be executed synchronously, for example:
In fact, require.js interprets the module content of the callback function through the function .toString, finds its correct dependencies, and turns it into a normal AMD module. It should be noted that if you write a module in this way, it may be incompatible with other AMD loaders, because it violates the AMD specification, but it can understand the writing of this format very well.
What's happening here is that require.js actually does the function.toString callback function to parse the contents of the module and find the correct dependencies, just like it would if it were a normal AMD module. It's important to note that if you choose to write modules this way, they will most likely be incompatible with other AMD module loaders, as this goes against AMD specifications, but it's good to know that this format exists!
CDN fallback
Another hidden gem of require.js is that it supports falling back to loading the local corresponding library when the CDN is not loaded correctly. We can achieve this through require.config:
No dependencies? Object literal? no problem!
When you write a module without any dependencies and just return an object containing some functional functions, then we can use a simple syntax:
Circular dependencies
In some cases, we may need module moduleA and the functions in moduleA need to depend on some applications. This is circular dependency.
If you need to get the address of the module, you can do this...
BaseUrl
Usually, when performing unit testing, your source code may be placed in a folder similar to src, and at the same time, your tests may be placed in a folder similar to tests. This can be difficult to get the test configuration right.
For example, we have an index.html file in the tests folder and need to load tests/spec/*.js locally. And assuming, all source code is in src/js/*.js, and there is a main.js in that folder.
In index.html, do not set data-main when loading require.js.
Here, you can see main.js being loaded. However, since it does not load data in the main script tag, you must specify a base. When the data is primarily for baseURL it is inferred from the location in the main file. By customizing baseUrl we can easily store test code and application code separately.
JSONP
We can handle JSONP terminal like this:
Under many requests, we need to use non-AMD libraries. For example, Backbone and Underscore are not adapted to AMD specifications. And jQuery actually just defines itself as a global variable named jQuery, so there is nothing to do with jQuery.
Fortunately, we can use shim configuration to solve this problem.