Amplificação de testes automatizados para aplicações móveis multiplataformas baseada em test patterns

Abstract

Mobile application developers need to make their products available to numerous devices with distinct features and operating systems (OSs). In this way, multiplatform development is increasingly common, that is, the same application and its functionalities are developed for the different OSs. The search for the quality of mobile applications has grown along with its spread, users require that mobile applications are reliable, robust and efficient; therefore, software developers must adopt quality assurance techniques. Mobile app testing can be challenging considering that crashes may exist due to user interaction such as rotate the device or use various gestures to scroll or magnify the screen while running a mobile application. This allows you to define Test Patterns, strategies used to test events related to the user interface, so that common events in the mobile applications are checked, preventing unexpected failures during the test of a functionality of the mobile application. Objective: This dissertation aims at investigating the amplification of tests in multiplatform mobile applications. Four Test Patterns were applied to check the known characteristics of mobile computing and amplify existing sets of functional tests. Method: To amplify functional test cases in cross-platform mobile applications, we have investigated the use of Test Patterns in mobile applications, raising recurrent failures motivated by events performed on mobile devices. Then, scripts were developed to execute the Test Patterns in an automated way. Finally, an experiment was conducted, and the result was analyzed for the number of failures encountered and the time of execution of the tests. Results: The proposed approach was implemented in a tool able of generating test scripts and was evaluated with nine multiplatform applications, generating automated amplified tests. The data collected in the experiment showed that the amplified set of tests found 23 unique bugs in eight of the nine mobile applications and consumed on average 2.54 times the time spent in a test case without amplification on Android and 1.55 times for iOS.