Integration of Fault Localization into your GitHub Repository

• Autor(a) principal: Hugo Filipe Ribeiro Paiva de Almeida
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: https://hdl.handle.net/10216/153042
• topic: Outras ciências da engenharia e tecnologias; Other engineering and technologies

With the increased complexity and scale of software, there is a strong demand for techniques to guide software engineers to locate faults with less human intervention as possible. The purpose of this dissertation is to look into the usage of Spectrum-based Fault Localization approaches to help discover faults in Java programs, as well as the use of bots in the software development lifecycle. Spectrum-based Fault Localization techniques were found to be chosen in the research area of software fault localization due to their low execution costs and popularity. Three tools (GZoltar, FLACOCO, and Jaguar) stood out as the top choices for spectrum-based fault localization in Java according to the research, and even though all produced comparable outcomes, GZoltar was preferred. A GitHub Action was created that, when integrated with GZoltar, allows analysis of Spectrum-based Fault Localization reports in any Java repository on GitHub. The outcome of it is a detailed report of potentially faulty lines of code, customizable by the user. This action is tested in both a sample repository and several open-source projects. While successful integration is achieved with the sample repository, limitations of GZoltar hinder its integration with most open-source projects, highlighting the need for updates and further compatibility testing.