Android App Testing: A Model for Generating Automated Lifecycle Tests

Android is currently the dominating OS in the market. An immense number of Android apps is deployed to the Google Play store every year. Android apps are no longer merely focusing on entertainment or socialization. In fact, the literature shows that apps specializing in critical domains such as health, education and even the military are growing in numbers. This puts more pressure on app developers to produce quality apps. Research shows that current Android app testing approaches rely heavily on manual testing. Research in automatic test generation for Android apps focuses mostly on automated GUI testing, with some approaches introducing model-based testing for test case inputs. However, no studies focus on generating lifecycle tests automatically, especially for testing lifecycle method conformance. In this research, we present a model-based solution approach as a tool to conduct assertion lifecycle tests automatically for Android activity lifecycle callback methods. Our objective is to build a framework to generate such a model. Finally, we evaluated our proposed framework in two ways. First using a group case study. Then we evaluate our work using 10 real world open source Android applications. The results of our evaluation are promising and show that our proposed framework is useful for detecting errors.