Android Uygulamaları için Kötü Huylu Test Yaratımı

Graphical User Interface (GUI) testing of mobile applications has been an important topic in the literature. Although there are many studies that focus on automated test generation for Android GUI applications, all these studies aim to explore depths of the Application Under Test (AUT) by generating meaningful and well-behaving tests. In this study we propose a fully automated tool, Bad Behaving Test Generator (BBTG), that builds on top of standard test generation tools. BBTG generates bad-behaving test cases from the well-behaving test cases by modifying existing events and adding new events. In this study, we develop BBTG on top of the state-of-the-art test generation approach, QBE (QLearning-Based Exploration). We perform experiments on 100 AUTs randomly selected from the commonly known F-Droid applications. We show that BBTG is effective at detecting novel crashes in a given time budget.