Why Would I ever use expect() When Writing Tests With QUnit?

Question

I've recently started using QUnit to unit test my JavaScript and I'm a little confused by a feature in there documentation: expect().

According to the docs, expect() is designed to:

[s]pecify how many assertions are expected to run within a test.

And here's the example they give:

test( "a test", function() {
  expect( 2 );

  function calc( x, operation ) {
    return operation( x );
  }

  var result = calc( 2, function( x ) {
    ok( true, "calc() calls operation function" );
    return x * x;
  });

  equal( result, 4, "2 square equals 4" );
});

The only thing I see here is maintenance nightmare. Every time you add an assertion to a test, you have to update that number or the test will fail. Is there a practical application for this kind of feature?

Answer 1

The only thing I see here is maintenance nightmare... Is there a practical application for this kind of feature?

Well, the way I think expect is meant to be used is with grouped meaningful tasks. It's useful for testing events or callbacks, for example:

test('trigger an event', function() {
  expect(1);

  $('div')
    .on('click', function() { ok(1) });
    .trigger('click');
});

It doesn't become a nightmare if you keep meaningful tasks grouped in small tests, where only 2 or 3 assertions are expected.

Answer 2

It can be used as a safeguard to ensure that you haven't somehow written a test that can't be run. If you get into the habit of writing the expected number of tests, should you ever somehow write a test suite where one test is hidden from QUnit for some reason, QUnit will pick this up before you would.