Re: [proposal][RFC] Improve the new git repo command
- From
Lucas Seiki Oshiro <lucasseikioshiro@gmail.com>
- Date
- Feb 22, 2026, 21:14 UTC
- Message-ID
- <3C0852FD-59FE-496D-9521-E123181901B3@gmail.com>
- In-Reply-To
- <CA+rGoLdSR=NPoD7XEbYPoRTt0VS5M0QhzHcy-OmyuZMMVN-H5w@mail.gmail.com>
> A list of my past activities in Git:
Looking the Git history (`git log --author='K Jayatheerth'`), there are many meaningful patches that you didn't listed here.
Show 10 quoted lines
> 3. Project Proposal > > 3.1 Why "Improve the new git repo command"? > This project is particularly compelling because I have closely > followed its development since its inception. Consistently reading > the weekly updates (https://lucasoshiro.github.io/gsoc-en/) and > following the mailing list patches for git repo info has deepened > my ongoing interest in this specific initiative since GSoC 2025. > This continuous engagement has provided a strong understanding of > why the command exists and exactly what needs to be done.
Thanks for your interest in my work :-).
Show 34 quoted lines
> 3.3 Proposed Solution and Objectives > The main objective of this project is to implement the changes and > additions discussed in the introduction to make git repo a complete, > modern replacement for parts of rev-parse. My proposed solutions are: > > - Removing the global state: The builtin/repo.c file currently opts > into using global state by declaring > #define USE_THE_REPOSITORY_VARIABLE at the top of the file. My > goal is to remove this macro entirely to align with Git's > libification efforts. To achieve this, I will refactor functions > that implicitly rely on this global state instead of the passed > repository context. For example, in get_layout_bare(), the repo > argument is currently marked as UNUSED because the function calls > is_bare_repository() (which checks global state). I will update > this and similar functions to evaluate the explicit repo struct > instead. > > - Implementing category keys: I will add a way to map specific > categories to a group of values. For example, if a user types > git repo info paths, the internal logic will look up the paths > category and return git-dir, common-dir, and other related > values all at once instead of requiring manual queries for each. > > - Adding path values: I will integrate the missing path values > currently obtained through git rev-parse (like toplevel and > superproject-working-tree) and --git-path (like index and hooks). > Since initial work on this has already started, my goal is to > take over the effort, lead the necessary design choices on the > mailing list, and complete the implementation. > > - Enhancing repo structure: I will study the external git-sizer > tool to figure out which of its repository analysis and > statistics features can be natively implemented into the git > repo structure sub-command.
It looks to me that you're proposing too much here. I mean, I agree with everything that you proposed here, but maybe you won't have enough time to do that given the pace of the reviewing process. For example, my first GSoC patch series took 11 versions until it was accepted.