Enabling Testing of Android Apps

Abstract

Existing approaches for automated testing of An- droid apps are designed to achieve different goals and exhibit some pros and cons that should be carefully considered by developers and testers. For instance, random testing (RT) provides a high ratio of infeasible inputs or events, and test cases generated with RT and systematic exploration-based testing (SEBT) are not representative of natural (i.e., real) application usage scenarios. In addition, collecting test scripts for automated testing is expensive. We address limitations of existing tools for GUI-based testing of Android apps in a novel hybrid approach called T+. Our approach is based on a novel framework, which is aimed at generating actionable test cases for different testing goals. The framework also enables GUI-based testing without expensive test scripts collection for the stakeholders.