search

refactoring-ai / predicting-refactoring-ml

Refactoring recommendation via ML
MIT License
28 stars 8 forks source link

Improving getReverseWalk #167

Closed mauricioaniche closed 4 years ago

mauricioaniche commented 4 years ago

The getReverseWalk for now gets all the commits in the order Git returns. This means whenever there's a branch that was merged, we see commits from both paths.

I'd say that's ok in most cases when it comes to the source code metrics. After all, we'll get the version - 1 and Git will return the right before version of the class, as it correctly knows the parent.

However, for process metrics, this can become tricky. As we will count the process metrics (e.g., authors, commits) from both branches... Or maybe that's ok? Well, maybe that's ok... What do you think?

jan-gerling commented 4 years ago

This issue will be closed for now.

@jan-gerling Future me: For validation sum up the number of duplicate commits appearing in a small number of repositories , in order to check if this issue leads to double counting of commits.