Skip to content

Conversation

@cruessler
Copy link
Collaborator

This is an early-stage PR to explore how difficult it will be to integrate gix-blame into gitui and how it will compare to the existing implementation, in particular with respect to performance.

A detailed description and measurements will follow.

@extrawurst
Copy link
Collaborator

@cruessler do you think it would be possible to aim to make that the headline feature for 0.29?

@cruessler
Copy link
Collaborator Author

@extrawurst I’m pretty sure that’s a realistic goal!

We would only have to decide how much tolerance we have for gix-blame differing from git blame due to differences in the underlying diff algorithm. (If you haven’t seen yet, there’s this tracking issue for differences in the diff algorithm: GitoxideLabs/gitoxide#2308.)

As far as I can tell, things have improved significantly in that area recently, though. Behind a feature flag, I’ve updated the diff library used by gix-diff, imara-diff, to a new version, 0.2. Using 0.2, gix-blame matches git blame on around 95 % of all lines in the gitoxide repository I tested this with (off the top of my head, I still want to document that somewhere). I’m planning on pushing a commit that uses that feature flag in this PR this weekend.

I hope that makes sense, let me know if you want more context!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Interactive Blame doesn't appear to be as fast as tig or non-interactive git blame

2 participants