Abstract

Software maintenance is the process of modifying an existing system to ensure that it meets current and future requirements. As a result, performing regression testing becomes an essential but time consuming aspect of any maintenance activity. Regression testing is initiated after a programmer has made changes to a program that may have inadvertently introduced errors. It is a quality control approach to ensure that the newly modified code still complies with its specified requirements and that unmodified code has not been affected by the maintenance activity. In the literature various types of test selection techniques have been proposed to reduce the effort associated with re-executing the required test cases. However, the majority of these approach has been focusing only on sequential programs, and provide no or only very limited support for distributed programs or database-driven applications. The thesis presents a lightweight methodology, which applies Formal Concept Analysis to support a regression test selection analysis, in combination with execution trace collection and external data sharing analysis, for distributed Java RMI programs. Two Eclipse plug-ins were developed to automate the regression test selection process and to evaluate our methodology.