Mining Software Repositories for the Characterization of Continuous Integration and Delivery

• Gabriel Augusto Destro Unicamp
• Breno Bernard Nicolau de França Unicamp

Resumo

Continuous Integration (CI) and Delivery (CD) are software development practices increasingly adopted in industry, which aiming at ensuring quality and stability, and making the whole process more efficient and less error-prone. Software engineers that incorporate these practices may have difficulties trying to get an overview of the impact of adopting them and tracking progress. Therefore, performing analysis regarding the level of the application of CI/CD practices in a software repository can help to understand how such practices are adopted, which can result in a better characterization of their benefits and limitations. To support this analysis, we developed Garimpeiro, a web application that supports the characterization of open source repositories hosted on GitHub regarding the level of adoption of CI/CD practices. This paper describes how the tool extracts data from GitHub repositories by showing the execution workflow, high-level architecture, and involved assumptions to define metrics. Finally, this paper also includes running examples of the tool, which demonstrates its capabilities and feasibility.

Palavras-chave: Mining Software Repositories, Continuous Integration, Continuous Delivery