الفهرس | Only 14 pages are availabe for public view |
Abstract Abstract The importance of test automation in the software industry is receiving growing attention in recent years. This attention is due to the significantimprovements made by test automation in the software development life cycle; these improvementsinclude reliability,stability, and cost reductions,especially for projects with a longlife span.Automation has reached all levels of testing, starting from unit tests to Graphical User Interface (GUI) tests. However, a key drawback of test automation is thecontinuous need for maintenance; hence, a higher cost. This maintenance is a crucial factor limiting the adoption of such automation. In addition, the problem becomes aggravated with GUI Tests.This aggravation is mostly due to the failure of the used locators,which are very crucial in all GUI tests.These locators are used to select elements from web pages to perform actions on them. In this thesis, multiple algorithms are proposed to solve the fragility problems of GUI test cases for web applications. The proposed techniques are used to automatically repair the locators used to select elements from web pages. The proposed algorithmsare evaluated using several web applications that differ in the complexity of the page structure, the number of elements per web page, and the number of attributes per element. The final results indicate the effectiveness of our methods by obtaining an average repair percentage of 91.8% in applications with a few elements per page, which is an improvement compared to the highest existing percentage, which is85%. Also, the obtained repair percentage in applications with a high number of elements per page was 79.8%,which is higher than the existing percentage that is reported to be 70%. Key Words: Software Testing, GUI Testing, Test Automation, Test Repair, Genetic Algorithm, Tabu Search. |