Yahoo! Developer Network Home - Help

YUI Library Examples: YUI Test Utility: Advanced Test Options

YUI Test Utility: Advanced Test Options

This example shows how to use advanced test options, which allow you to specify additional information about how a test should be run. Each TestCase can specify up to three different options for tests, including tests that should be ignored, tests that should throw an error, and tests that should fail.

Advanced Test Options

This example begins by creating a namespace and a TestCase object:

This TestCase serves as the basis for this example.

Using _should

Immediately after the name of the TestCase is defined, there is a _should property. This property specifies information about how tests should behave and is defined as an object literal with one or more of the following properties: fail, error, and ignore.Each of these three is also defined as an object literal whose property names map directly to the names of test methods in the TestCase. This example uses all three properties:

This Test specifies one test that should fail, six that should throw an error, and one that should be ignored.

In the fail section, the test method testFail() is specified as one that should fail. By adding the property testFail and settings its value to true, the TestRunner knows that this test is expected to fail. If the test were to be run without failing, it would be considered a failure of the test. This feature is useful when testing YUI Test itself or addon components to YUI Test.

Moving on to the error section, there are six tests specified that should throw an error. There are three different ways to indicate that a test is expected to throw an error. The first is simply to add a property with the same name as the test method and set its value equal to true (similar to specifying tests that should fail). In this example, the testGenericError() method is specified this way. When specified like this, the test passes regardless of the type of error that occurs. This can be dangerous since unexpected errors will also cause the test to pass. To be more specific, set the property value for the test method to an error message string. When a string is used instead of the Boolean true, the test passes only when an error is thrown and that error message matches the string. In this example, testStringError() and testStringError2() expect an error to be thrown with an error message of "I'm a specific error message." If any other error occurs inside of the these methods, the test will fail because the error message doesn't match. The last way to specify an error should occur is to create an actual error object, which is the case with testObjectError(), testObjectError2(), and testObjectError3(). When specified in this way, a test will pass only when an error is thrown whose constructor and error message match that of the error object.

The last section is ignore, which determines tests that should be ignored. In this example, the method testIgnore() is set to be ignored when the TestCase is executed. Test in the ignore section are specified the same way as those in the fail section, by adding the name as a property and setting its value to true.

Creating the test methods

The next part of the example contains the actual test methods. Since each test method is specified as having a certain behavior in _should, they each do something to show their particular functionality.

The first method is testFail(), which does nothing but purposely fail. Since this method is specified as one that should fail, it means that this test will pass:

This method uses Assert.fail() to force the test to fail. This type of method is helpful if you are creating your own type of assert methods that should fail when certain data is passed in.

Next, the test methods that should error are defined. The testGenericError() method is specified as needing to throw an error to pass. In the error section, testGenericError is set to true, meaning that any error causes this method to pass. To illustrate this, the method simply throws an error:

The fact that this method throws an error is enough to cause it to pass (the type of error and error message don't matter). The next two methods, testStringError() and testStringError2() are specified as throwing an error with a specific message ("I'm a specific error message."):

The testStringError() method will pass when executed because the error message matches up exactly with the one specified in the error section. The testStringError2() method, however, will fail because its error message is different from the one specified.

To be more specific, testObjectError(), testObjectError2(), and testObjectError3(), specified an error type (TypeError) and an error messsage ("Number expected."):

Of the these three methods, only testObjectError() will pass because it's the only one that throws a TypeError object with the message, "Number expected." The testObjectError2() method will fail because the type of error being thrown (Error) is different from the expected type (TypeError), as specified in the error section. The last method, testObjectError3(), also fails. Though it throws the right type of error, the error message doesn't match the one that was specified.

The last method in the TestCase is testIgnore(), which is specified to be ignored. To be certain, this method pops up a message:

If this test weren't ignored, then the alert should be displayed. Since it is ignored, though, you will never see the alert. Additionally, there is a special message displayed in the TestLogger when a test is ignored.

Running the tests

With all of the tests defined, the last step is to run them. This initialization is assigned to take place when the document tree has been completely loaded by using the Event utility's onDOMReady() method:

Before running the tests, it's necessary to create a TestLogger object to display the results (otherwise the tests would run but you wouldn't see the results). After that, the TestRunner is loaded with the TestSuite object by calling add() (any number of TestCase and TestSuite objects can be added to a TestRunner, this example only adds one for simplicity). The very last step is to call run(), which begins executing the tests in its queue and displays the results in the TestLogger.

Copyright © 2008 Yahoo! Inc. All rights reserved.

Privacy Policy - Terms of Service - Copyright Policy - Job Openings