Merge

  •  Feature Engineering for Classification-Based Merge Conflict Resolution |  Current Topics

    Context

    Merge conflict resolution remains a significant challenge in Git-based software development, as manual conflict resolutions slow down collaboration and reduce developer productivity. However, empirical research results suggest that a vast majority of chunk resolutions found in practice can be derived from a fixed set of conflict resolution patterns, combining the ours, theirs, and base parts of a conflicting chunk in a pre-defined way. These findings form the foundation for phrasing merge conflict resolution as a classification problem, and thus using traditional machine learning for predicting conflict resolutions.

     Show more...

  •  Predicting Merge Conflict Resolutions: WSRC vs. Random Forest |  Current Topics

    Context

    Merge conflict resolution remains a significant challenge in Git-based software development, as manual conflict resolutions slow down collaboration and reduce developer productivity. However, empirical research results suggest that a vast majority of chunk resolutions found in practice can be derived from a fixed set of conflict resolution patterns, combining the ours, theirs, and base parts of a conflicting chunk in a pre-defined way. These findings form the foundation for phrasing merge conflict resolution as a classification problem, and thus using traditional machine learning for predicting the correct resolution.

     Show more...