Understanding Predictive Factors for Merge Conflicts

Understanding Predictive Factors for Merge Conflicts

Abstract

Merge conflicts often occur when developers change the same code artifacts. Such conflicts might be frequent in practice, and resolving them might be costly and is an error-prone activity. To minimize these problems by reducing merge conflicts, it is important to better understand how conflict occurrence is affected by technical and organizational factors. With that aim,we investigate seven factors related to modularity, size, and timing of developers contributions. By analyzing 73504 merge scenarios in GitHub repositories (Ruby and Python MVC projects), we find evidence that the likelihood of merge conflict occurrence significantly increases when contributions to be merged are not modular in the sense that they involve files from the same MVC slice (related model, view, and controller files).We also find bigger contributions involving more developers, commits, and changed files more likely lead to merge conflicts. Regarding the timing factors, we observe contributions developed over longer periods of time more likely lead to conflicts. No evaluated factor shows predictive power concerning both the number of merge conflicts and the number of files with conflicts. Our results could be used to derive recommendations for development teams and merge conflict prediction models. Project management and assistive tools could benefit from these models.

Scripts