Unit testing allows you to check the quality of your code after you've written it, but you can also use unit testing to improve your development process as you go along. Instead of writing tests after you finish developing your application, consider writing the tests as you go. This helps you design small, maintainable, reusable units of code. It also makes it easier for you to test your code thoroughly and quickly.
When you do local unit testing, you run tests that stay inside your own development environment without involving remote components. App Engine provides testing utilities that use local implementations of datastore and other App Engine services . This means you can exercise your code's use of these services locally, without deploying your code to App Engine, by using service stubs.
A service stub is a method that simulates the behavior of the service. For example, the datastore service stub shown in Writing Datastore and Memcache Tests allows you to test your datastore code without making any requests to the real datastore. Any entity stored during a datastore unit test is held in memory, not in the datastore, and is deleted after the test run. You can run small, fast tests without any dependency on datastore itself.
This document gives some information about setting up a testing framework, then describes how to write unit tests against several local App Engine services.
- Setting up a testing framework
- Introducing the Java testing utilities
- Writing Datastore and memcache tests
- Writing High Replication Datastore tests
- Writing task queue tests
- Writing deferred task tests
- Writing local service capabilities tests
- Writing tests for other services
- Writing tests with authentication expectations
Setting up a testing framework
Even though the SDK's testing utilities are not tied to any specific framework, we'll use JUnit for our examples so you have something concrete and complete to work from. Before you begin writing tests you'll need to add the appropriate JUnit 4 JAR to your testing classpath. Once that's done you're ready to write a very simple JUnit test.
import static org.junit.Assert.*;
public class MyFirstTest {
@Test
public void testAddition() {
assertEquals(4, 2 + 2);
}
}
If you're running Eclipse 3.5, select the source file of the test to run. Select the Run menu > Run As > JUnit Test . The results of the test appear in the Console window.
For information on running your tests with Apache Ant, please read the Ant JUnit Task documentation.
Introducing the Java testing utilities
MyFirstTest
demonstrates the simplest possible test setup, and for tests that have no dependency on App Engine APIs or local service implementations, you may not need anything more. However, if your tests or code under test have these dependencies, add the following JAR files to your testing classpath:
-
${SDK_ROOT}/lib/impl/appengine-api.jar
-
${SDK_ROOT}/lib/impl/appengine-api-labs.jar
-
${SDK_ROOT}/lib/impl/appengine-api-stubs.jar
These JARs make the runtime APIs and the local implementations of those APIs available to your tests.
App Engine services expect a number of things from their execution environment, and setting these things up involves a fair amount of boilerplate code. Rather than set it up yourself, you can use the utilities in the
com.google.appengine.tools.development.testing
package. To use this package, add the following JAR file to your testing classpath:
-
${SDK_ROOT}/lib/testing/appengine-testing.jar
Take a minute to browse the
javadoc
for the
com.google.appengine.tools.development.testing
package. The most important class in this package is
LocalServiceTestHelper
, which handles all of the necessary environment setup and gives you a top-level point of configuration for all the local services you might want to access in your tests.
To write a test that accesses a specific local service:
-
Create an instance of
LocalServiceTestHelper
with aLocalServiceTestConfig
implementation for that specific local service. -
Call
setUp()
on yourLocalServiceTestHelper
instance before each test andtearDown()
after each test.
Writing Datastore and memcache tests
The following example tests the use of the datastore service.
import com.google.appengine.api.datastore.DatastoreService;
import com.google.appengine.api.datastore.DatastoreServiceFactory;
import com.google.appengine.api.datastore.Entity;
import static com.google.appengine.api.datastore.FetchOptions.Builder.withLimit;
import com.google.appengine.api.datastore.Query;
import com.google.appengine.tools.development.testing.LocalDatastoreServiceTestConfig;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import static org.junit.Assert.*;
public class LocalDatastoreTest {
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalDatastoreServiceTestConfig());
@Before
public void setUp() {
helper.setUp();
}
@After
public void tearDown() {
helper.tearDown();
}
// run this test twice to prove we're not leaking any state across tests
private void doTest() {
DatastoreService ds = DatastoreServiceFactory.getDatastoreService();
assertEquals(0, ds.prepare(new Query("yam")).countEntities(withLimit(10)));
ds.put(new Entity("yam"));
ds.put(new Entity("yam"));
assertEquals(2, ds.prepare(new Query("yam")).countEntities(withLimit(10)));
}
@Test
public void testInsert1() {
doTest();
}
@Test
public void testInsert2() {
doTest();
}
}
In this example,
LocalServiceTestHelper
sets up and tears down the parts of the execution environment that are common to all local services, and
LocalDatastoreServiceTestConfig
sets up and tears down the parts of the execution environment that are specific to the local datastore service. If you read the
javadoc
you'll learn that this involves configuring the local datastore service to keep all data in memory (as opposed to flushing to disk at regular intervals) and wiping out all in-memory data at the end of every test. This is just the default behavior for a datastore test, and if this behavior isn't what you want you can change it.
Changing the example to access memcache instead of datastore
To create a test that accesses the local memcache service, you can use the code that is shown above, with a few small changes.
Instead of importing classes related to datastore, import those related to memcache. You still need to import
LocalServiceTestHelper
.
import com.google.appengine.api.memcache.MemcacheService;
import com.google.appengine.api.memcache.MemcacheServiceFactory;
import com.google.appengine.tools.development.testing.LocalMemcacheServiceTestConfig;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
Change the name of the class that you are creating, and change the instance of
LocalServiceTestHelper
, so that they are specific to memcache.
public class LocalMemcacheTest {
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalMemcacheServiceTestConfig());
And finally, change the way you actually run the test so that it is relevant to memcache.
private void doTest() {
MemcacheService ms = MemcacheServiceFactory.getMemcacheService();
assertFalse(ms.contains("yar"));
ms.put("yar", "foo");
assertTrue(ms.contains("yar"));
}
As in the datastore example, the
LocalServiceTestHelper
and the service-specific
LocalServiceTestConfig
(in this case
LocalMemcacheServiceTestConfig
) manage the execution environment.
Writing High Replication Datastore tests
If your app uses the High Replication Datastore (HRD), you may want to write tests that verify your application's behavior in the face of eventual consistency.
LocalDatastoreServiceTestConfig
exposes options that make this easy:
import com.google.appengine.api.datastore.DatastoreService;
import com.google.appengine.api.datastore.DatastoreServiceFactory;
import com.google.appengine.api.datastore.Entity;
import static com.google.appengine.api.datastore.FetchOptions.Builder.withLimit;
import com.google.appengine.api.datastore.Key;
import com.google.appengine.api.datastore.KeyFactory;
import com.google.appengine.api.datastore.Query;
import com.google.appengine.tools.development.testing.LocalDatastoreServiceTestConfig;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import static org.junit.Assert.*;
public class LocalHighRepDatastoreTest {
// maximum eventual consistency
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalDatastoreServiceTestConfig()
.setDefaultHighRepJobPolicyUnappliedJobPercentage(100));
@Before
public void setUp() {
helper.setUp();
}
@After
public void tearDown() {
helper.tearDown();
}
@Test
public void testEventuallyConsistentGlobalQueryResult() {
DatastoreService ds = DatastoreServiceFactory.getDatastoreService();
Key ancestor = KeyFactory.createKey("foo", 3);
ds.put(new Entity("yam", ancestor));
ds.put(new Entity("yam", ancestor));
// global query doesn't see the data
assertEquals(0, ds.prepare(new Query("yam")).countEntities(withLimit(10)));
// ancestor query does see the data
assertEquals(2, ds.prepare(new Query("yam", ancestor)).countEntities(withLimit(10)));
}
}
By setting the unapplied job percentage to 100, we are instructing the local datastore to operate with the maximum amount of eventual consistency. Maximum eventual consistency means writes will commit but always fail to apply, so global (non-ancestor) queries will consistently fail to see changes. This is of course not representative of the amount of eventual consistency your application will see when running in production, but for testing purposes, it's very useful to be able to configure the local datastore to behave this way every time. If you want more fine-grained control over which transactions fail to apply, you can register your own
HighRepJobPolicy
:
// ...
import com.google.appengine.api.datastore.dev.HighRepJobPolicy;
public class LocalCustomPolicyHighRepDatastoreTest {
private static final class CustomHighRepJobPolicy implements HighRepJobPolicy {
static int count = 0;
@Override
public boolean shouldApplyNewJob(Key entityGroup) {
// every other new job fails to apply
return count++ % 2 == 0;
}
@Override
public boolean shouldRollForwardExistingJob(Key entityGroup) {
// every other existing job fails to apply
return count++ % 2 == 0;
}
}
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalDatastoreServiceTestConfig()
.setAlternateHighRepJobPolicyClass(CustomHighRepJobPolicy.class));
@Before
public void setUp() {
helper.setUp();
}
@After
public void tearDown() {
helper.tearDown();
}
@Test
public void testEventuallyConsistentGlobalQueryResult() {
DatastoreService ds = DatastoreServiceFactory.getDatastoreService();
ds.put(new Entity("yam")); // applies
ds.put(new Entity("yam")); // does not apply
// first global query only sees the first Entity
assertEquals(1, ds.prepare(new Query("yam")).countEntities(withLimit(10)));
// second global query sees both Entities because we "groom" (attempt to
// apply unapplied jobs) after every query
assertEquals(2, ds.prepare(new Query("yam")).countEntities(withLimit(10)));
}
}
The testing APIs are useful for verifying that your application behaves properly in the face of eventual consistency, but please keep in mind that the local High Replication read consistency model is an approximation of the production High Replication read consistency model, not an exact replica. In the local environment, performing a
get()
of an
Entity
that belongs to an entity group with an unapplied write will always make the results of the unapplied write visible to subsequent global queries. In production this is not the case.
Writing task queue tests
Tests that use the local task queue are a bit more involved because, unlike datastore and memcache, the task queue API does not expose a facility for examining the state of the service. We need to access the local task queue itself to verify that a task has been scheduled with the expected parameters. To do this, we need
com.google.appengine.api.taskqueue.dev.LocalTaskQueue
.
import com.google.appengine.api.taskqueue.dev.LocalTaskQueue;
import com.google.appengine.api.taskqueue.dev.QueueStateInfo;
import com.google.appengine.api.taskqueue.QueueFactory;
import com.google.appengine.api.taskqueue.TaskOptions;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
import com.google.appengine.tools.development.testing.LocalTaskQueueTestConfig;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import static org.junit.Assert.*;
public class TaskQueueTest {
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalTaskQueueTestConfig());
@Before
public void setUp() {
helper.setUp();
}
@After
public void tearDown() {
helper.tearDown();
}
// Run this test twice to demonstrate we're not leaking state across tests.
// If we _are_ leaking state across tests we'll get an exception on the
// second test because there will already be a task with the given name.
private void doTest() throws InterruptedException {
QueueFactory.getDefaultQueue().add(TaskOptions.Builder.withTaskName("task29"));
// give the task time to execute if tasks are actually enabled (which they
// aren't, but that's part of the test)
Thread.sleep(1000);
LocalTaskQueue ltq = LocalTaskQueueTestConfig.getLocalTaskQueue();
QueueStateInfo qsi = ltq.getQueueStateInfo().get(QueueFactory.getDefaultQueue().getQueueName());
assertEquals(1, qsi.getTaskInfo().size());
assertEquals("task29", qsi.getTaskInfo().get(0).getTaskName());
}
@Test
public void testTaskGetsScheduled1() throws InterruptedException {
doTest();
}
@Test
public void testTaskGetsScheduled2() throws InterruptedException {
doTest();
}
}
Notice how we ask the
LocalTaskqueueTestConfig
for a handle to the local service instance, and then we investigate the local service itself to make sure the task was scheduled as expected. All
LocalServiceTestConfig
implementations expose a similar method. You may not always need it, but sooner or later you'll be glad it's there.
Writing deferred task tests
If your application code uses Deferred Tasks , the Java Testing Utilities make it easy to write an integration test that verifies the results of these tasks.
import static org.junit.Assert.*;
import com.google.appengine.api.taskqueue.DeferredTask;
import com.google.appengine.api.taskqueue.QueueFactory;
import com.google.appengine.api.taskqueue.TaskOptions;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
import com.google.appengine.tools.development.testing.LocalTaskQueueTestConfig;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import java.util.concurrent.TimeUnit;
public class DeferredTaskTest {
// Unlike CountDownLatch, TaskCountDownlatch lets us reset.
private final LocalTaskQueueTestConfig.TaskCountDownLatch latch =
new LocalTaskQueueTestConfig.TaskCountDownLatch(1);
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalTaskQueueTestConfig()
.setDisableAutoTaskExecution(false)
.setCallbackClass(LocalTaskQueueTestConfig.DeferredTaskCallback.class)
.setTaskExecutionLatch(latch));
private static class MyTask implements DeferredTask {
private static boolean taskRan = false;
@Override
public void run() {
taskRan = true;
}
}
@Before
public void setUp() {
helper.setUp();
}
@After
public void tearDown() {
MyTask.taskRan = false;
latch.reset();
helper.tearDown();
}
@Test
public void testTaskGetsRun() throws InterruptedException {
QueueFactory.getDefaultQueue().add(
TaskOptions.Builder.withPayload(new MyTask()));
assertTrue(latch.await(5, TimeUnit.SECONDS));
assertTrue(MyTask.taskRan);
}
}
As with our first Local Task Queue example, we are using a
LocalTaskqueueTestConfig
, but this time we are initializing it with some additional arguments that give us an easy way to verify not just that the task was scheduled but that the task was executed: We call
setDisableAutoTaskExecution(false)
to tell the Local Task Queue to automatically execute tasks. We call
setCallbackClass(LocalTaskQueueTestConfig.DeferredTaskCallback.class)
to tell the Local Task Queue to use a callback that understands how to execute Deferred tasks. And finally we call
setTaskExecutionLatch(latch)
to tell the Local Task Queue to decrement the latch after each task execution. This configuration allows us to write a test in which we enqueue a Deferred task, wait until that task runs, and then verify that the task behaved as expected when it ran.
Writing local service capabilities tests
Capabilities testing involves changing the status of some service, such as datastore, blobstore, memcache, and so forth, and running your application against that service to determine whether your application is responding as expected under different conditions. The capability status can be changed using the LocalCapabilitiesServiceTestConfig class.
The following code snippet changes the capability status of the datastore service to disabled and then runs a test on the datastore service. (You can substitute other services for datastore as needed.)
import com.google.appengine.api.capabilities.Capability;
import com.google.appengine.api.capabilities.CapabilityState;
import com.google.appengine.api.capabilities.CapabilityStatus;
import com.google.appengine.api.datastore.DatastoreService;
import com.google.appengine.api.datastore.DatastoreServiceFactory;
import com.google.appengine.api.datastore.FetchOptions;
import com.google.appengine.api.datastore.Query;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
import com.google.appengine.tools.development.testing.LocalCapabilitiesServiceTestConfig;
import com.google.apphosting.api.ApiProxy;
import org.junit.After;
import org.junit.Test;
import static org.junit.Assert.*;
public class ShortTest {
private LocalServiceTestHelper helper;
@After
public void tearDown() {
helper.tearDown();
}
@Test(expected = ApiProxy.CapabilityDisabledException.class)
public void testDisabledDatastore() {
Capability testOne = new Capability("datastore_v3");
CapabilityStatus testStatus = CapabilityStatus.DISABLED;
//Initialize
LocalCapabilitiesServiceTestConfig config =
new LocalCapabilitiesServiceTestConfig().setCapabilityStatus(testOne, testStatus);
helper = new LocalServiceTestHelper(config);
helper.setUp();
FetchOptions fo = FetchOptions.Builder.withLimit(10);
DatastoreService ds = DatastoreServiceFactory.getDatastoreService();
assertEquals(0, ds.prepare(new Query("yam")).countEntities(fo));
}
}
The sample test first creates a
Capability
object initialized to datastore, then creates a
CapabilityStatus
object set to DISABLED. The
LocalCapabilitiesServiceTestConfig
is created with the capability and status set using the
Capability
and
CapabilityStatus
objects just created.
The
LocalServiceHelper
is then created using the
LocalCapabilitiesServiceTestConfig
object. Now that the test has been set up, the
DatastoreService
is created and a Query is sent to it to determine whether the test generates the expected results, in this case, a
CapabilityDisabledException
.
Writing tests for other services
Testing utilities are available for blobstore and other App Engine services. For a list of all the services that have local implementations for testing, see the
LocalServiceTestConfig
documentation.
Writing tests with authentication expectations
If your application behaves differently when a user is logged in, or if your application has admin-only functionality, you can write tests to verify that this logic is behaving as expected. Here's how to do it:
import com.google.appengine.api.users.UserService;
import com.google.appengine.api.users.UserServiceFactory;
import com.google.appengine.tools.development.testing.LocalServiceTestHelper;
import com.google.appengine.tools.development.testing.LocalUserServiceTestConfig;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import static org.junit.Assert.*;
public class AuthenticationTest {
private final LocalServiceTestHelper helper =
new LocalServiceTestHelper(new LocalUserServiceTestConfig())
.setEnvIsAdmin(true).setEnvIsLoggedIn(true);
@Before
public void setUp() {
helper.setUp();
}
@After
public void tearDown() {
helper.tearDown();
}
@Test
public void testIsAdmin() {
UserService userService = UserServiceFactory.getUserService();
assertTrue(userService.isUserAdmin());
}
}
In this example, we're configuring the
LocalServiceTestHelper
with the
LocalUserServiceTestConfig
so we can use the
UserService
in our test, but we're also configuring some authentication-related environment data on the
LocalServiceTestHelper
itself.