Chapter 4. Fixtures

One of the most time-consuming parts of writing tests is writing the code to set the world up in a known state and then return it to its original state when the test is complete. This known state is called the fixture of the test.

In Example 2.1, the fixture was simply the array that is stored in the $stack variable. Most of the time, though, the fixture will be more complex than a simple array, and the amount of code needed to set it up will grow accordingly. The actual content of the test gets lost in the noise of setting up the fixture. This problem gets even worse when you write several tests with similar fixtures. Without some help from the testing framework, we would have to duplicate the code that sets up the fixture for each test we write.

PHPUnit supports sharing the setup code. Before a test method is run, a template method called setUp() is invoked. setUp() is where you create the objects against which you will test. Once the test method has finished running, whether it succeeded or failed, another template method called tearDown() is invoked. tearDown() is where you clean up the objects against which you tested.

In Example 2.2 we used the producer-consumer relationship between tests to share fixture. This is not always desired or even possible. Example 4.1 shows how we can write the tests of the StackTest in such a way that not the fixture itself is reused but the code that creates it. First we declare the instance variable, $stack, that we are going to use instead of a method-local variable. Then we put the creation of the array fixture into the setUp() method. Finally, we remove the redundant code from the test methods and use the newly introduced instance variable, $this->stack, instead of the method-local variable $stack with the assertEquals() assertion method.

Example 4.1: Using setUp() to create the stack fixture

<?php
class StackTest extends PHPUnit_Framework_TestCase
{
    protected $stack;

    protected function setUp()
    {
        $this->stack = array();
    }

    public function testEmpty()
    {
        $this->assertTrue(empty($this->stack));
    }

    public function testPush()
    {
        array_push($this->stack, 'foo');
        $this->assertEquals('foo', $this->stack[count($this->stack)-1]);
        $this->assertFalse(empty($this->stack));
    }

    public function testPop()
    {
        array_push($this->stack, 'foo');
        $this->assertEquals('foo', array_pop($this->stack));
        $this->assertTrue(empty($this->stack));
    }
}
?>


The setUp() and tearDown() template methods are run once for each test method (and on fresh instances) of the test case class.

In addition, the setUpBeforeClass() and tearDownAfterClass() template methods are called before the first test of the test case class is run and after the last test of the test case class is run, respectively.

The example below shows all template methods that are available in a test case class.

Example 4.2: Example showing all template methods available

<?php
class TemplateMethodsTest extends PHPUnit_Framework_TestCase
{
    public static function setUpBeforeClass()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
    }

    protected function setUp()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
    }

    protected function assertPreConditions()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
    }

    public function testOne()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
        $this->assertTrue(TRUE);
    }

    public function testTwo()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
        $this->assertTrue(FALSE);
    }

    protected function assertPostConditions()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
    }

    protected function tearDown()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
    }

    public static function tearDownAfterClass()
    {
        fwrite(STDOUT, __METHOD__ . "\n");
    }

    protected function onNotSuccessfulTest(Exception $e)
    {
        fwrite(STDOUT, __METHOD__ . "\n");
        throw $e;
    }
}
?>
phpunit TemplateMethodsTest
PHPUnit 3.8.0 by Sebastian Bergmann.

TemplateMethodsTest::setUpBeforeClass
TemplateMethodsTest::setUp
TemplateMethodsTest::assertPreConditions
TemplateMethodsTest::testOne
TemplateMethodsTest::assertPostConditions
TemplateMethodsTest::tearDown
.TemplateMethodsTest::setUp
TemplateMethodsTest::assertPreConditions
TemplateMethodsTest::testTwo
TemplateMethodsTest::tearDown
TemplateMethodsTest::onNotSuccessfulTest
FTemplateMethodsTest::tearDownAfterClass


Time: 0 seconds, Memory: 5.25Mb

There was 1 failure:

1) TemplateMethodsTest::testTwo
Failed asserting that <boolean:false> is true.
/home/sb/TemplateMethodsTest.php:30

FAILURES!
Tests: 2, Assertions: 2, Failures: 1.


More setUp() than tearDown()

setUp() and tearDown() are nicely symmetrical in theory but not in practice. In practice, you only need to implement tearDown() if you have allocated external resources like files or sockets in setUp(). If your setUp() just creates plain PHP objects, you can generally ignore tearDown(). However, if you create many objects in your setUp(), you might want to unset() the variables pointing to those objects in your tearDown() so they can be garbage collected. The garbage collection of test case objects is not predictable.

Variations

What happens when you have two tests with slightly different setups? There are two possibilities:

  • If the setUp() code differs only slightly, move the code that differs from the setUp() code to the test method.

  • If you really have a different setUp(), you need a different test case class. Name the class after the difference in the setup.

Sharing Fixture

There are few good reasons to share fixtures between tests, but in most cases the need to share a fixture between tests stems from an unresolved design problem.

A good example of a fixture that makes sense to share across several tests is a database connection: you log into the database once and reuse the database connection instead of creating a new connection for each test. This makes your tests run faster.

Example 4.3 uses the setUpBeforeClass() and tearDownAfterClass() template methods to connect to the database before the test case class' first test and to disconnect from the database after the last test of the test case, respectively.

Example 4.3: Sharing fixture between the tests of a test suite

<?php
class DatabaseTest extends PHPUnit_Framework_TestCase
{
    protected static $dbh;

    public static function setUpBeforeClass()
    {
        self::$dbh = new PDO('sqlite::memory:');
    }

    public static function tearDownAfterClass()
    {
        self::$dbh = NULL;
    }
}
?>


It cannot be emphasized enough that sharing fixtures between tests reduces the value of the tests. The underlying design problem is that objects are not loosely coupled. You will achieve better results solving the underlying design problem and then writing tests using stubs (see Chapter 8), than by creating dependencies between tests at runtime and ignoring the opportunity to improve your design.

Global State

It is hard to test code that uses singletons. The same is true for code that uses global variables. Typically, the code you want to test is coupled strongly with a global variable and you cannot control its creation. An additional problem is the fact that one test's change to a global variable might break another test.

In PHP, global variables work like this:

  • A global variable $foo = 'bar'; is stored as $GLOBALS['foo'] = 'bar';.

  • The $GLOBALS variable is a so-called super-global variable.

  • Super-global variables are built-in variables that are always available in all scopes.

  • In the scope of a function or method, you may access the global variable $foo by either directly accessing $GLOBALS['foo'] or by using global $foo; to create a local variable with a reference to the global variable.

Besides global variables, static attributes of classes are also part of the global state.

By default, PHPUnit runs your tests in a way where changes to global and super-global variables ($GLOBALS, $_ENV, $_POST, $_GET, $_COOKIE, $_SERVER, $_FILES, $_REQUEST) do not affect other tests. Optionally, this isolation can be extended to static attributes of classes.

Note

The implementation of the backup and restore operations for static attributes of classes requires PHP 5.3 (or greater).

The implementation of the backup and restore operations for global variables and static attributes of classes uses serialize() and unserialize().

Objects of some classes that are provided by PHP itself, such as PDO for example, cannot be serialized and the backup operation will break when such an object is stored in the $GLOBALS array, for instance.

The @backupGlobals annotation that is discussed in the section called “@backupGlobals” can be used to control the backup and restore operations for global variables. Alternatively, you can provide a blacklist of global variables that are to be excluded from the backup and restore operations like this

class MyTest extends PHPUnit_Framework_TestCase
{
    protected $backupGlobalsBlacklist = array('globalVariable');

    // ...
}

Note

Please note that setting the $backupGlobalsBlacklist attribute inside the setUp() method, for instance, has no effect.

The @backupStaticAttributes annotation that is discussed in the section called “@backupStaticAttributes” can be used to control the backup and restore operations for static attributes. Alternatively, you can provide a blacklist of static attributes that are to be excluded from the backup and restore operations like this

class MyTest extends PHPUnit_Framework_TestCase
{
    protected $backupStaticAttributesBlacklist = array(
      'className' => array('attributeName')
    );

    // ...
}

Note

Please note that setting the $backupStaticAttributesBlacklist attribute inside the setUp() method, for instance, has no effect.

Please open a ticket on GitHub to suggest improvements to this page. Thanks!