Files
MapLibreParticleLab/android/java/MapboxGLAndroidSDKTestApp
Leith Bade 927b4c6dbe [android] Refactor annotations API
Remove unimplemented properties.
Correct defintions of equals() and hasCode().
Add setOnInfoWindowClickListener and remove old method from Marker.
Refactor showInfoWindow() to remove need for exposing internal method.
Make select/deselectMarker public. Add getSelectedMarker.
Fix bug where you couldn't reselect a closed info window.
Add empty constructor to LatLng and LatLngZoom.
Fixes #2546
Fixes #2631
Fixes #2448
2015-10-22 20:00:43 +11:00
..
2015-10-22 20:00:43 +11:00
2015-09-16 14:03:03 -04:00

Mapbox GL Test App

Testing

Running Espresso tests on a device

This test project comes with all the required Android Testing Support Library dependencies in the Gradle file. Tests are under the app/src/androidTest folder.

Note that before running your tests, you might want to turn off animations on your test device. It's a known issue that leaving system animations turned on in a test device (window animation scale, transition animation scale, animator duration scale) might cause unexpected results, or may lead tests to fail.

To create a new run configuration:

  • Click on Run -> Edit Configurations...
  • Click on the plus sign and then on "Android Tests"
  • Give a name to the configuration, e.g. TestAppTests
  • Choose the MapboxGLAndroidSDKTestApp module
  • Choose android.support.test.runner.AndroidJUnitRunner as the instrumentation runner
  • Click OK to save the new configuration

You can now run this configuration from the main toolbar dropdown menu.

Running Espresso tests on AWS Device Farm

On a terminal, within mapbox-gl-native/android/java, run the tests (cC stands for connectedCheck):

$ ./gradlew cC -p MapboxGLAndroidSDKTestApp

Then:

  • Go to your AWS Console and choose Device Farm.
  • Create a new project, e.g. MapboxGLAndroidSDKTestApp
  • On step 1, upload the APK in mapbox-gl-native/android/java/MapboxGLAndroidSDKTestApp/build/outputs/apk/MapboxGLAndroidSDKTestApp-debug-unaligned.apk
  • On step 2, choose Instrumentation, test filter is com.mapbox.mapboxgl.testapp.MainActivityTest and upload the APK in mapbox-gl-native/android/java/MapboxGLAndroidSDKTestApp/build/outputs/apk/MapboxGLAndroidSDKTestApp-debug-androidTest-unaligned.apk
  • On step 3, choose a device pool. E.g. Top Devices
  • On step 4, customize your device state (if needed)
  • Finally, confirm the configuration and run the tests.

On Step 2, you can also separate by commas different classes: com.mapbox.mapboxgl.testapp.MainActivityTest,com.mapbox.mapboxgl.testapp.MainActivityScreenTest

If you have no tests for your app, or want to test some random user behaviour, you can just choose "Built-in: Fuzz" in step 2.

Running pure JUnit tests

These tests run on a local JVM on your development machine and they are extremely fast to run (as compared to Espresso). These tests are located under src/test/java. To run them you switch to the Unit Tests build variant, then right click the corresponding test class or method and select "Run ...".

You can also have a run configuration:

  • Click on Run -> Edit Configurations...
  • Click on "Junit Tests"
  • Give a name to the configuration, e.g. JUnit tests
  • As "Test Kind", choose "All in directory"
  • As folder, choose the following folder: mapbox-gl-native/android/java/MapboxGLAndroidSDKTestApp/src/test/java
  • Click OK to save the new configuration

You can also run the tests from the command line with:

$ ./gradlew test --continue -p MapboxGLAndroidSDKTestApp

Running the UI/Application Exerciser Monkey

Similar to the "Built-in: Fuzz" test mentioned above, Android provides Monkey, "a program that runs on your emulator or device and generates pseudo-random streams of user events such as clicks, touches, or gestures, as well as a number of system-level events."

To exercise Monkey on the test app, install the package on the device (e.g. via Android Studio) and then:

$ adb shell monkey -p com.mapbox.mapboxgl.testapp -v 500