Integration of an Automatic Fault Localization Tool in an IDE and its Evaluation

• Autor(a) principal: Leão, João Manuel Madureira de
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/20822
• topic: Fault Localization; Automated Debugging; Debugging; Fault Localization Tools; Fault Localization Techniques

Debugging is one of the most demanding and error-prone tasks in software development. Trying to address bugs has become overall more expensive as the software complexity and size have increased. As a result, several researchers attempted to improve the developers’ debugging experience and efficiency by automating as much of the process as possible. Existing auto-finding tools will assist developers in automatically detecting bugs, however, they are not yet widely available to software engineers. Making such tools available to developers can save debugging time and increase the productivity. Subsequently, the main goal of this dissertation is to incorporate an automatic fault localization tool into an Integrated Development Environment (IDE). The selected IDE was Visual Studio Code, a source-code editor developed by Microsoft for Windows, Linux, and macOS. Visual Studio Code is one of the most used IDEs and is known for its flexible API, which allows nearly every aspect of it to be customized. Furthermore, the chosen automatic fault localization tool was FLACOCO, a recent fault localization tool for Java that supports up to the most recent versions. Nonetheless, this document contains a full overview of several fault localization methodologies and tools, as well as an explanation of the complete planning and development process of the produced Visual Studio Code extension. After the development and deployment were completed, an evaluation was carried out. The extension was evaluated through a user study in which thirty Java professionals took part. The test had two parts: the first involved users using the extension to complete two debugging tasks in previously unknown projects, and the second had them filling out a satisfaction questionnaire for further analysis. Finally, the results show that the extension was a success, with the system being rated positively in all areas. However, it may be revised in light of the questionnaire responses, with the suggestions received being considered for future work.