dydycloud / route-separation Goto Github PK

Normal routes in express are in app.js like so app.get('/', routes.index);. The default setup for express routes can be very cumbersome, the larger your app gets the larger your app.js is going to get, making it harder to read and maintain.

Visionmedia's example for route-seporation isn't the greatest (isn't totally clear). Dailyjs has a good article on modules and it gives the example of route-seporation but it isn't very clear.

There are a ton of stack overflow posts on this subject.

• How to include route handlers in multiple files in Express
• Node.js: Configuration and routes in a different file
• Node.js & Express.js: Breaking up the app.js file
• Nodejs/Expressjs app structure

Each directory in this repository is an express app that demonstrates the express routes in a separated way. All the apps were created || tested with express --version 3.0.0rc3. Each app is using a different port number for example, examplel1 uses 3001, example2 uses 3002.

Within apps.js all of the routes are replaced with require('./routes')(app); which fires /routes/index.js by default. From within /routes/index.js we export a module requiring all of the other files in the directory manually like so require('./users')(app); passing the app variable down to each level. Finally each /route file can be wrapped in a module.exports in which we can run app.get, app.post, etc., that's all their is to this method. Attribution ShadowCloud, Alex Young.

This is very similar to example1 the only difference is that /routes/index.js is automatically adding every file in the /routes directory. Attribution Sam Corder, Brad Proctor.

I personally created this method out of desire. The goals that I used as a framework where as follows:

1. Remove all http calls like app.get() from apps.js.

2. Automatically load each of the route files in /routes to the routes object. For example a file like /routes/dog.js and its export of exports.index() would be accessible from routes.dog.index.

3. Include all http calls like app.get() in one file without any logic. Meaning I didn't want things like:

    module.exports = function(app) {	
    	app.get('/', function(req,res) {
    		res.send("/");
    	});
    };
   

or even like

	app.get('/', function(req,res) {
		res.send("/");
	});

all calls should look like

	  app.get('/', routes.main.index);
	  app.get('/mike', routes.main.mike);
	  app.get('/dave', routes.main.dave);
	  app.get('/harry', routes.main.harry);
	  app.get('/users', routes.users.index);

This shows the route definition / or /mike to be separated from the function being executed. I wanted all of the routes to use the route object. I call the file that contains these declarations the /routes/switchboard.js.

The routes/index.js file loads all of the JavaScript files in the /routes directory except index.js and switchboard.js. When everything is loaded with rs.readdir() and validated I require each of the route files into the routes object then load switchboard.js. With this configuration the route files don't have any interaction with the app variable and there are no module.exports in the route level it is all exports. I am new to async and if you see a performance gap I would love to learn what is wrong with this method. Attribution Reggi.

This is a much, much simpler version of the last example. The main difference is that I ditched rs.readdir() automatically generating the routes object which is very clunky and I create it "manually" inside the new /routes/index.js file right above to the app.get() declarations. I figured that you can't really use a new route unless you define it's GET or POST properties manually anyway, so we might as well include the files the the route object manually.