AETing : an automated exploratory testing strategy based on code evolutioncoverage

• Autor(a) principal: OLIVEIRA, Marlom Jobsom Dias de
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Institucional da UFPE
• dARK ID: ark:/64986/001300000hgn5
• Texto Completo: https://repositorio.ufpe.br/handle/123456789/40070
• topic: Engenharia de Software e Linguagens de Programação; Teste de interface gráfica; Teste de Android; Geração de casos de testes automáticos

Testers face challenges to have their test suites up-to-date with respect to the applica tion source code evolution to be tested. These challenges are greater in a global distributed software development context. The growing daily testing demand also makes it difficult to maintain such suites. These test suites are also continuously automated to save the time of human testers, but they require maintenance as well. Exploratory testing comes as a trade-off between test case maintenance, human expertise, and flexibility. Unfortunately, it is a man ual task in general. In this work, we specifically designed a strategy called AETing. AETing receives Test Scenarios generated by ArcWizard, which are the results of code static analysis that produces screen navigations suggestion that covers code change, and AETing maps such static screen navigations suggestion into concrete code implementation based on Page Objects maintained by Motorola in its testing framework called Page Browser. AETing combines these two resources to generate automated test cases that perform screen navigations and Monkey testing aiming to maximize coverage of code changes evolution between two versions of a given Android application. We developed and evaluated our approach in a real testing opera tion environment related to Motorola Mobility, through a partnership between CIn-UFPE and this company supported by the Informatics Law. The evaluation consisted of testing four dif ferent Motorola Android applications. Through the evaluation, we obtained promising results concerning the comparison between AETing and expert exploratory testers’ code coverage. We discuss in detail how AETing works and the results achieved.