> If we use "istate->repo", we don't even need to change this function. Oh you're absolutely right about that. I actually just got to learn more about `index_state` from another thread where Junio brought it up. But now with his suggestion that attributes may not belong in the repository struct at all, I'm a bit unsure how to move the patch forward. One thing I did take away from this is that we shouldn't be cramming environment variables into the repository struct. This discussion has definitely helped me think more clearly about the design, and I think it'll guide me take better decisions going forward. I’d really appreciate your thoughts on how you think we should approach this from here. Thanks, Ayush:)
