Supertest .expect(200) vs. res.status.should.equal(200);

Question

Do both of these serve the same purpose? Why are they both used in, for example, this tutorial https://codeforgeek.com/2015/07/unit-testing-nodejs-application-using-mocha/ ?

Edit, Looking at the following code:

var supertest = require("supertest");
var should = require("should");

// This agent refers to PORT where program is runninng.

var server = supertest.agent("http://localhost:3000");

// UNIT test begin

describe("SAMPLE unit test",function(){

  // #1 should return home page

  it("should return home page",function(done){

    // calling home page api
    server
    .get("/")
    .expect("Content-type",/json/)
    .expect(200) // THis is HTTP response
    .end(function(err,res){
      // HTTP status should be 200
      res.status.should.equal(200);
      // Error key should be false.
      res.body.error.should.equal(false);
      done();
    });
  });

});

Is it necessary to have

.expect(200)

and

res.status.should.equal(200);

? What is the difference?

Answer 1

The .expect(200) part is using the supertest facility for verifying data. the object.should.equal(value) part is using shouldJS for verification.

I prefer utilizing shouldJS in the .end() because it allows me to do a bit of data manipulation, testing, logging, etc, as needed.

Do note the following from: https://www.npmjs.com/package/supertest

If you are using the .end() method .expect() assertions that fail will not throw - they will return the assertion as an error to the .end() callback.

So, in the example code you show above, if .expect("Content-type",/json/) or .expect(200) fails, there is nothing in the .end() to catch it. A better example would be:

var supertest = require("supertest");
var should = require("should");

// This agent refers to PORT where program is runninng.

var server = supertest.agent("http://localhost:3000");

// UNIT test begin

describe("SAMPLE unit test",function(){

  // #1 should return home page

  it("should return home page",function(done){

    // calling home page api
    server
      .get("/")
      .expect("Content-type",/json/)
      .expect(200) // THis is HTTP response
      .end(function(err,res){
        // NOTE: The .expect() failures are handled by err and is  
        //       properly passed to done. You may also add logging
        //       or other functionality here, as needed.
        if (err) {
            done(err);
        }

        // Error key should be false.
        res.body.error.should.equal(false);
        done();
      });
  });

});

Update to answer the question in the comment and provide a prettier response here:

Question: Would doing something like .expect(200, done) catch the error then?

Answer: The short answer is, "Yes". On the same page I quoted above, it has the following:

Here's an example with mocha, note how you can pass done straight to any of the .expect() calls:

describe('GET /user', function() { 
  it('respond with json', function(done) { 
    request(app) 
      .get('/user') 
      .set('Accept', 'application/json') 
      .expect('Content-Type', /json/) 
      .expect(200, done); 
  }); 
});