Junio C Hamano <gitster@xxxxxxxxx> writes: > Nguyễn Thái Ngọc Duy <pclouds@xxxxxxxxx> writes: > >> v4 adds a new patch, 2/5, which makes the hashmap related functions in >> refs.c generic, so I could add a new map for worktree ref stores and >> not abuse submodule hashmap. >> >> I mentioned about moving these hashmap back to submodule.c and >> worktree.c where it can map more than just ref stores (in worktree >> case, we can cache 'struct worktree', for example). But I'm not doing >> it now to keep things small. >> >> The commit message in 1/5 is rephrased a bit, hopefully clearer. > > Michael, does this look good to replace what has been queued? > > I did not spot any discussion on this topic after v4 when I skimmed > the backlog messages, so I am hoping that I can pick this up without > having to worry about seeing another reroll immediately after that > (in which case I'd have been better off if I tended other topics in > flight in the meantime). > > Thanks, both. Oops, I shouldn't have done that. When applied on top of the files-backend thing (or have you updated that one and is my tree lacking it???), this breaks quite a few tests. t0001#41 dumps core from "git worktree add ../linked-worktree" for example.