Test with context android

Android unit tests explained

Feb 24, 2016 · 3 min read

Have you ever tried unit testing in Android? Haven’t you ended up so frustrated that you thought it wasn’t worth it? In this article we’re going to go through 3 simple examples that might help you get started and discover some neat tricks.

In Android there are two types of unit tests:

• Instrumented: they are slow, only run on a real device or an emulator and live in the app/src/androidTest folder by default
• Local or non-instrumented: they are very fast, run on your computer’s JVM and live in the app/src/test folder by default

Non-instrumented tests are always preferred, but they are not always possible: you can’t unit test code in your JVM that uses classes provided by the Android framework.

This means that you c a nnot use classes from the the android.* or org.json.* packages in your unit tests if you want to run them on the JVM. You can still unit test these classes on your JVM with Roboelectric (which is a great tool), but I didn’t want to use it here for the sake of simplicity.

Only if you are careful enough while programming you will be able to separate your code from the framework’s code. Read about clean code and hide/wrap implementation details. And please, pull your business logic out of activities/fragments to make it more testable!

In this article, we are going to see how to test these 3 examples:

This project is in Github. In most of these tests we will use a very simple Item object that is basically:

1. JsonParserTest

JsonParser just serializes an Item object to a String and then deserializes it back to an Item object. JsonParser uses classes from the org.json.* package that are provided by the Android framework, so this is an instrumented unit test and needs to run on a device or emulator.

There is a way of testing the JsonParser object on the JVM: adding the json classes to your gradle dependencies (only for testing purposes), so these classes will be used instead of the Android framework ones.

Now, if you move your JsonParserTest class from ‘androidTest’ to the ‘test’ folder, it will work. You just made it much faster! You can see it working here.

2. FileSerializerTest

File operations in Android usually need a context so that the path is relative to your app. But in this case we are going to use a trick to make it run on the JVM! (the java.io.* package comes from the JDK, not the Android framework). The code for the FileSerializer is here.

As you can see, we are using the JUnit rule TemporaryFolder that gets the Android’s context.getCacheDir() for us. This folder will be implicitly created and destroyed for every test (in the @Before and @After methods).

If you needed another relative folder you might need a context, and the test will become instrumented. Next example (shared preferences) uses a context.

3. SharedPreferencesTest

This is undoubtedly an instrumented test, since it needs Android’s classes. So I’m going to show you how to get hold of the Android context for instrumented unit tests.

The the important line is:

It will retrieve your application’s context. There is another method in the InstrumentationRegistry class called getContext(). Be careful and don’t use this one since it will retrieve the test application’s context.

Useful official documentation:

Part 2 is about reading assets. This project is in Github. There are instructions in the README file. Happy unit testing!

Источник

Developing Android unit and instrumentation tests — Tutorial

This tutorial describes how to write unit and instrumentation tests for your Android application. It describes how to execute these tests via Android studio and Gradle. This tutorial assumes that you are familiar with Android programming in general.

1. Introduction into Android testing

1.1. Testing Android applications

Android applications run on a variety of devices. Also the Android framework and the surrounding open source frameworks evolve at a high speed. To ensure that you application works well, it is import to write software tests. This helps you to enhance and maintain the Android application.

Unit testing for Android can be classified into:

Local unit tests — tests which can run on the JVM.

Instrumented unit tests — tests which require the Android system.

Local unit tests run much faster compared to the time required to deploy and run tests on an Android device. Prefer writing local unit tests and only run tests on Android, if you require a real Android system.

If you write local unit test and have dependencies to Android API, you need to replace them, e.g., via a mocking framework like Mockito.

1.2. What to test on Android applications

Your test should focus on the business logic of your application. A good rule of thumb is to have the following distribution of tests:

70-80 % unit tests to ensure stability of your code basis

20-30 % functional tests to ensure that the application really works

some cross functional tests if your application integrates intensively with other Application components

You should test your application at least on one device with the lowest possible configuration. In addition you should test on one device with the highest available configuration, e.g., pixel density, screen resolution to ensure that it works fine on these devices.

1.3. Tooling support for Android testing

The Android Testing Support library (ATSL) project from Google provides tooling for Android testing. For example, it ( AndroidJUnitRunner ).

If you run local unit tests, a special version of the android.jar (also known as the Android mockable jar ) is created by the tooling. This modified JAR file is provided to the unit test so that all fields, methods and classes are available. Any call to the Android mockable JAR results, by default, in an exception, but you can configure Android to return default values. See Activating default return values for mocked methods in android.jar for details.

The library provides a JUnit 4-compatible test runner ( AndroidJUnitRunner ), the Espresso test framework and the UI Automator test framework. Espresso test framework can be used to test the User Interface of your application. UI Automator allows to write cross application functional tests.

AndroidJunitRunner provides access to the instrumentation API, via the InstrumentationRegistery .

InstrumentationRegistry.getInstrumentation() , returns the Instrumentation currently running.

InstrumentationRegistry.getContext() , returns the Context of this Instrumentation’s package.

InstrumentationRegistry.getTargetContext() , returns the application Context of the target application.

InstrumentationRegistry.getArguments() , returns a copy of arguments Bundle that was passed to this Instrumentation. This is useful when you want to access the command line arguments passed to the instrumentation for your test.

It also gives access to the life cycle via the ActivityLifecycleMonitorRegistry .

1.4. Android project organization for tests

The following is the default directory structure for your application and test code:

app/src/main/java — for your source code of your main application build

app/src/test/java — for any unit test which can run on the JVM

app/src/androidTest/java — for any test which should run on an Android device

If you follow this conversion, the Android build system runs your tests on the correct target (JVM or Android device).

If you receive the following error message: «error duplicate files in path. Path in archive: LICENSE.txt» you can add the following to your app/gradle.build file.

2. Android unit testing

2.1. Unit testing in Android

A unit test verifies in isolation the functionality of a certain component. For example, assume a button in an Android activity is used to start another activity. A unit test would determine, if the corresponding intent was issued, not if the second activity was started

A unit tests for an Android application can be:

local unit test — which runs on the JVM

Android unit test — which runs on the Android runtime

If they run on the JVM, they are executed against a modified version of the android.jar Android library. In this version all final modifiers have been stripped off. This makes it easier to use mocking libraries, like Mockito.

The local unit tests of an Android project should be located in the app/src/test folder.

2.2. Required dependencies in the Gradle build file

To use JUnit tests for your Android application, you need to add it as dependency to your Gradle build file.

2.3. Running the unit tests

2.3.1. Using Gradle

Run your unit tests with the gradlew test command.

2.3.2. Using Android Studio

To run a unit test, right-click on your test class in the Project window and select Run.

2.4. Location of test reports

The Test reports are created in the app/build/reports/tests/debug/ directory. The index.html gives an overview and links to the individual test pages.

2.5. Activating default return values for mocked methods in android.jar

By default, all calls to methods in the modified android.jar file throw exceptions. This default should ensure that your unit tests only test your code and do not depend on any particular behavior of the Android platform. If you want to configure a certain behavior, you can use a mocking framework to replace these call.

You can also instruct the Gradle build system to return default values for method calls in the `android.jar`with the following configuration in your Gradle build file.

3. Exercise: Create unit test

In this exercise you learn how to create a simple JUnit4 test for an Android project.

3.1. Preparation: Create Android project

Create the Android project as described in Android temperature converter.

3.2. Add JUnit dependency

Ensure you have the dependency to Junit in your app/build.gradle file.

3.3. Create test

In your app/src/test directory create the following two test methods for the ConverterUtil class.

3.4. Run unit tests

Ensure your unit tests are correctly implemented by running test tests. They should run successfully.

4. Writing tests to run on the Android device

4.1. Instrumentation tests

The Android testing API provides hooks into the Android component and application life cycle. These hooks are called the instrumentation API and allow your tests to control the life cycle and user interaction events.

Under normal circumstances your application cannot control the life cycle events and the user drives the application. For example, if Android creates your activity the onCreate() method is called. Or if the user presses a button your corresponding code is called. Via instrumentation you can control these events via your test code. For example, your instrumentation test can start the activity. Afterwards, it can call the finish() and restart the activity to test if the instance state of the activity is correctly restored.

Instrumented unit tests are unit tests that run on Android devices and emulators instead of running on the Java virtual machine. These tests have access to the real device and its resources and are useful to unit test functionality which cannot be easily mocked by mocking frameworks. An example is a test which validates a Parcelable implementation.

An instrumentation-based test class allows you to send key events (or touch events) to the application under test.

With user interface testing framework like Espresso, the developer rarely has to use the instrumentation API directly.

4.2. How the Android system executes tests

The InstrumentationTestRunner is the base test runner for Android tests. This test runner starts and loads the test methods. Via the instrumentation API it communicates with the Android system. If you start a test for an Android application, the Android system kills any process of the application under test and then loads a new instance. It does not start the application, this is the responsibility of the test methods. The test method controls the life cycle of the components of the application.

The test runner also calls the onCreate() method of the application and activity under test during its initialization.

4.3. Mocking objects in Android

The mocking framework Mockito can also be used for instrumentation tests. This allows you to replace parts of the Android system which are not interesting for the test. The Android framework provided in the past specialized mocking classes but these are not necessary anymore.

4.4. Location of instrumentation tests

As described in Android project organization for tests the unit tests of an Android project should be located in the app/src/androidTest/java folder.

4.5. Define dependencies and testInstrumentationRunner in the Gradle build file

To use JUnit tests for your Android application you need to add the dependency to JUnit to your Gradle build file. You also need to specify android.support.test.runner.AndroidJUnitRunner as testInstrumentationRunner in the build file.

4.6. Using the @RunWith(AndroidJUnit4.class)

It is also recommended annotating the test with the @RunWith(AndroidJUnit4.class) annotation. AndroidJUnit4 extends JUnit4, so if you use pure Junit4 syntax and ActivityTestRule it is not required. But you need it, if you want to run, i.e., Espresso tests with ActivityTestRule and JUnit4.

4.7. Run the tests on Android

Run your local unit tests via Gradle with the gradlew connectedCheck command.

To run your local unit tests via Android Studio, right-click on your test class in the Project window and select Run.

4.8. Location of test reports

The test reports are created in the app/build/reports/androidTests/connected/ directory. The index.html gives an overview and links to the individual test pages.

4.9. How to replace the application for instrumentation tests

You can replace the application class for the instrumentation tests by overriding the AndroidJUnitRunner and its newApplication method.

Источник