Um estudo de caracterização de contribuidores e suas contribuições de teste em projetos de código aberto

• Ano de defesa: 2023
• Autor(a) principal: Souza, Hugo Henrique Fumero de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Campo Mourao; Brasil; Programa de Pós-Graduação em Ciência da Computação; UTFPR
• Palavras-chave em Português: Software livre; GitHub (Programa de computador); Software - Testes; Free computer software; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO; Ciência da Computação
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/33615

Even though open-source software projects exhibit some distinct characteristics from industry projects, the dedication of maintainers and contributors to achieving a high level of software quality remains constant. To this end, testing is among the primary practices in these communities. During the development activity, most repositories emphasize the need to submit, along with the developed code, tests that ensure the quality and accuracy of the coded item. We observed that the contribution of tests during development is not as common. In this work, we aim to characterize test contributors in open-source software projects and determine the percentage of pull requests containing bug fixes that also include automated tests. For this purpose, we conducted a quantitative study - characterizing projects, contributors, and their contributions - and a qualitative study with selected contributors identified from the characterization. We found that only 7.87% of bug fix PRs contain some form of test scenario improvement. Furthermore, we identified five test contributor profiles: Contributor in Motion, Scenario Updater, Library Specialist, Comment Analyst and the strategic Evolution Contributor. To reach these conclusions, we developed the PRAnalyzer tool, which allows for the analysis of code in PRs, extracting critical information such as the quantity of asserts, imports, blank lines, and contributed code. Additionally, as a valuable resource for the academic community, we provide a comprehensive dataset containing information on over 400,000 PRs and their characteristics.