Ed Oswald Go

Ed Oswald Go

Software developer, hobbyist photographer, audio, film and tv buff

QUnit Assert Equality with Machine Epsilon

4 minute read

Overview

Unit testing is a great process for producing high-quality software when done correctly. I am a developer that is not used to creating test cases for automated testing or develop software via Test / Behaviour Driven Development. Little by little, I want to be a developer that is very productive even by using TDD / BDD.

I recently tried adding QUnit to my Stock Profit Calculator project in order to explore unit testing in JavaScript. Integration with QUnit is a breeze as the quick start guide is simple and easy to understand.

  • HTML Code (test.html):

    <!DOCTYPE html>
<html>
<head>
  <meta charset="utf-8">
  <title>QUnit Example</title>
  <link rel="stylesheet" href="http://code.jquery.com/qunit/qunit-1.18.0.css">
</head>
<body>
  <div id="qunit"></div>
  <div id="qunit-fixture"></div>
  <!-- CDN -->
  <script src="http://code.jquery.com/jquery-1.11.3.min.js"></script>
  <script src="http://code.jquery.com/qunit/qunit-1.18.0.js"></script>
  <script src="http://cdnjs.cloudflare.com/ajax/libs/mathjs/1.6.0/math.min.js"></script>
  <!-- Libraries to be tested -->
  <script src="../js/broker.js"></script>
  <script src="../js/brokers/broker-col.js"></script>
  <script src="../js/stock.calculator.js"></script>
  <!-- Test Suites -->
  <script src="../js/test/tests.js"></script>
</body>
</html>

  • JavaScript Code (tests.js):

    (function() {
  QUnit.test( "hello test", function( assert ) {
    assert.ok( 1 == "1", "Passed!" );
  });
})();

Problem

As I ran my created test suite, I encountered an error when comparing the equality of two floating point numbers. Based on the example below, 0.3333 is really not equal to 0.33333333333333333333 because of the decimal places. One approach I tried is to round the number to four decimal places in order to be equal to 0.3333. But, this approach will not solve all rounding cases. I checked QUnit’s core API and there’s no method with epsilon support. How did I solve my problem?

QUnit.test( "Test Equality", function( assert ) {
  assert.equal( 1/3, 0.3333);
});

image-center

Solution

My initial plan is to extend QUnit to support an epsilon parameter to some of its methods. But luckily, I encountered a QUnit plugin, named Close, that already addressed the problem stated above.

I used the assert.close(actual, expected, maxDifference, message); function of the plugin wherein the function accepts an additional parameter, maxDifference. The maxDifference parameter is the epsilon that tolerates the difference between the actual and expected output.

QUnit.test( "Test Equality with Machine Epsilon", function( assert ) {
  assert.close( 1/3, 0.3333, 0.00005);
});

image-center

Leave a comment

You may also enjoy