arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.09303cs.SE
收藏

可视化GUI测试实践:扩展的工业案例研究

Visual GUI testing in practice: An extended industrial case study

Vahid Garousi, Wasif Afzal, Adem Çağlar, İhsan Berk Işık, Berker Baydan, Seçkin Çaylak, Ahmet Zeki Boyraz, Burak Yolaçan, Kadir Herkiloğlu

摘要:

Visual GUI测试(VGT)被称为最新一代基于GUI的测试。它是一种工具驱动的技术,使用图像识别与被测系统进行交互并断言其行为。受土耳其一家大型软件和系统公司在国防和IT领域提供解决方案的工业需求的推动,最近启动了一个行动研究项目,在该公司的几个团队和项目中实施VGT。目的:为了满足上述需求,我们计划并开展了一项实证调查,目标是使用两种工具(Sikuli和JAutomate)评估VGT。其目的是确定一个合适的方法和工具,为VGT的一个给定的项目(软件产品)在该公司,增加了专门知识,在该公司的测试团队。方法:采用行动研究的案例研究设计,我们调查了VGT在研究组织中的使用。具体而言,使用两个选定的VGT工具,我们进行了定量和定性评价的VGT。结果如下:通过评估以前的工作中提出的挑战,问题和限制(CPL)列表,在我们的实证研究的背景下,我们发现测试工具和SUT相关的CPL与以前的实证研究相当,例如,SUT和测试工具之间的同步并不总是鲁棒的,并且测试工具的图像识别特征存在故障。当评估测试维护活动的类型时,当在SUT的下一个版本上执行自动化测试用例时,对于两个测试工具的情况,我们发现大约一半的测试用例(59.1%和47.8%)在下一个版本中失败。结论:通过我们的研究结果,我们证实了一些以前报道的问题时,进行VGT。此外,我们强调了一些额外的挑战,在测试维护时使用VGT。

英文摘要:

Context: Visual GUI testing (VGT) is referred to as the latest generation GUI-based testing. It is a tool-driven technique, which uses image recognition for interacting with and asserting the behavior of the system under test. Motivated by the industrial need of a large Turkish software and systems company providing solutions in the areas of defense and IT sector, an action-research project was recently initiated to implement VGT in several teams and projects in the company. Objective: To address the above needs, we planned and carried out an empirical investigation with the goal of assessing VGT using two tools (Sikuli and JAutomate). The purpose was to determine a suitable approach and tool for VGT of a given project (software product) in the company, increase the know-how in the company's test teams. Method: Using an action-research case-study design, we investigated the use of VGT in the studied organization. Specifically, using the two selected VGT tools, we conducted a quantitative and a qualitative evaluation of VGT. Results: By assessing the list of Challenges, Problems and Limitations (CPL), proposed in previous work, in the context of our empirical study, we found that test-tool- and SUT-related CPLs were quite comparable to a previous empirical study, e.g., the synchronization between SUT and test tools were not always robust and there were failures in test tools' image recognition features. When assessing the types of test maintenance activities, when executing the automated test cases on next versions of the SUTs, for the case of the two test tools, we found that about half of the test cases (59.1% and 47.8%) failed in the next version. Conclusion: By our results, we confirm some of the previously-reported issues when conducting VGT. Further, we highlight some additional challenges in test maintenance when using VGT.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑