Pleines, Julian (2023) Analysis of Git Repositories for Software Visualization. (Bachelor thesis), Kiel University, Kiel, 71 pp.

Preview

Text bachelor-thesis-julian-pleines.pdf - Published Version Download (4MB) | Preview

Abstract

Software monitoring and analysis solutions are increasingly important throughout the development process. Continuously verifying software designs and reviewing in-production code is part of keeping a software’s code base maintainable. To detect potential weaknesses in the software design or overly complex source code, static program analysis is often used to detect these obstacles and maintain serviceable code. Complex source code makes it particularly difficult for the developer to clarify the facts of the program. Thus, an overview of the software in terms of complexity and interrelationships is always beneficial for program comprehension.
Since source code is often managed in Git repositories, enhancing statical analysis data with Git history metrics aids in the detection of design flaws by revealing frequently accessed and modified files.
This thesis shows the implementation of the first stage to provide ExplorViz with robust, extendable, and easy-to-set-up statical analysis capabilities, which are able to calculate and factor in Git metrics. The analysis service is built to be easily integrable into continuous integration pipelines to provide up-to-date project overviews triggered by Git actions. We will show how such a static analysis is designed and implemented, how Git repositories are handled, and how to design the service with extendability in mind. Additionally, we demonstrate how this service can analyze arbitrary Java projects in a locally running Docker container or integrated into continuous integration based on GitLab. As an accompanying visualization for the generated data is missing at the time of this thesis, the usefulness of the generated data will be examined in upcoming works.

Actions (login required)

View Item