Grant: Discovery Grants Program - Individual

Software systems are evolving rapidly to cope with ever-changing requirements and business goals. Hence, having shorter release cycles helps improve user experience by providing the latest features or bug fixes, and improves developer productivity by reducing integration overhead. To achieve fast releases while maintaining high software quality, developers adopt the continuous integration (CI) practice. In CI, every few code changes (e.g., a few commits) trigger an automated process of integrating the code with the central code repository, building the system, testing, and deploying to production. In particular, automated test execution is the cornerstone in CI that ensures the central repository remains stable and up-to-date, and helps developers detect faults before releasing the code changes to production. CI has changed how software practitioners develop, test, and release their software. However, decades of research in software quality assurance focuses mostly on traditional development settings and has not yet been exercised, evaluated, and renovated in CI settings. The goal of this research program is to improve software quality assurance practices in the era of CI. We plan to address the challenges in CI and propose techniques to utilize fine-grained code evolution information. In particular, we plan to work on three research objectives. RO1: To address the challenges associated with frequent test execution, we plan to propose automated techniques to enhance test cases. We will propose techniques to automatically generate test code to cover the newly added and changed code. We will also conduct studies and propose techniques to detect performance issues and improve test execution time. RO2: To address the challenges of lacking fault localization research in CI settings, we will develop benchmarks for fault localization in CI and propose tailored fault localization techniques that utilize fine-grained code evolution information. RO3: To further utilize fine-grained evolution information, we plan to assist developers in diagnosing production issues by combining production data (e.g., logs) and rich CI data for automated diagnosis of production failure. The outcome of this research will significantly advance the research community by enabling researchers to adopt current and future techniques for CI settings. The proposed techniques will significantly improve software quality and reduce software development costs among not only Canadian companies but also companies around the world that follow CI practices. The proposed research will also train several highly qualified personnel (HQP) and allow them to contribute to state-of-the-art Software Engineering research and practice.