As the number of android applications has increased dramatically, there is a rising concern about their quality and reliability. In particular, the rich gui interactions supported by android should be thoroughly teste...
详细信息
As the number of android applications has increased dramatically, there is a rising concern about their quality and reliability. In particular, the rich gui interactions supported by android should be thoroughly tested in order to ensure if the behavior of an android application is correct. However, manually creating a gui state model can be tedious and error- prone, especially for a nontrivial application. This paper proposes a crawler that can automatically generate the gui state model for an android application. Particularly, a hierarchical state model is employed to represent the intra- and interactivity gui behavior of android applications in order to increase the model readability. Empirical experiments were conducted to evaluate the proposed crawler and the generated model. The results show that the state model generated by the crawler has a promising coverage as compared to the model created manually. The hierarchical state model can greatly improve the model readability to ease the gui behavior analysis and validation for android applications.
model-based gui exploration techniques are widely used to generate test cases for event-driven programs (such as android apps). These techniques traverse the elements of screens during the user interaction and simulta...
详细信息
ISBN:
(纸本)9781450356992
model-based gui exploration techniques are widely used to generate test cases for event-driven programs (such as android apps). These techniques traverse the elements of screens during the user interaction and simultaneously construct the guimodel. Although there are a number of automatic model-based exploration tools, most of them pay more attention to the exploration procedure than the model reusing. This paper presents LAND, an effective and user-friendly test generation tool based on gui exploration of android apps, which constructs an elaborate window transition model "LATTE" that considers more android specific characteristics and provides a customizable test generation interface by reusing the model. Experiments on 20 real-world android apps are conducted to construct their models as well as test cases. The experimental results indicate that LAND can achieve higher code coverage and trigger exceptions in shorter sequence. It is also demonstrated that LATTE can be well reused under different requirements of test suite generation. A demo video of our tool can be found at the website https://***/watch?v=iqtr12eiJ_0.
暂无评论