Comment when there are too many Similar Commit Messages

[julesdepoi]

Very similar Commit Messages in a MR may not always be ideal.

Note: We should consider if commit message analysis belongs in an MR specific bot? There does not seems to be another more appropriate event.

Research

Good use of similar commit messages

• update: add unit tests to feature_branch_1
• update: add unit tests to feature_branch_2
• chore: remove commented-out code from dev
• chore: rebase local branch from dev

Bad use of similar commit messages

• remove files
• remove tests
• working code?
• working test?

Notice: One is indicative of a sense of order and parallelism. The other shows a lazy programmer using the same type of commit messages for a variety of loosely related scenarios`.

A few approaches that might work

1. Check whether any commit message contains any other commit message
2. Calculate the string distance between the two closes strings
3. Calculate the average pairwise string distance between every string.