Re: [PATCH] worktree: teach `repair` to fix multi-directional breakage

[Eric Sunshine <sunshine@sunshineco.com>]

On Thu, Dec 17, 2020 at 2:49 PM Junio C Hamano <gitster@pobox.com> wrote:
> Eric Sunshine <sunshine@sunshineco.com> writes:
> > The inference is intentionally simple-minded. There is no path-based
> > inference or other heuristic at play; the entire inference is based
> > upon <id>. The worktree's path is specified as an argument. `git
> > worktree repair` manually reads the ".git" gitfile at that location
> > and, if it is well-formed, extracts the <id>. It then searches for a
> > corresponding <id> in <repo>/worktrees/ and,...
>
> That is exactly the point I got confused.  The worktree's path comes
> as an argument from the user, so we'd trust it.  And it has ".git"
> that is a gitfile that used to point at <repo> but because we are
> trying to deal with a situation where both worktree and repo moved,
> it cannot be used to learn where <repo> is.  Then, even after
> learning what <id> to use, how would we know where to use that <id>
> to find <repo>/worktrees/<id>, when the location of <repo> is unknown?
>
> I think the answer is "where the user starts the 'git worktree'
> command has to be under control of some repository (perhaps found by
> a call to setup_git_directory()), and we'd use that one as <repo>".

Correct. This is why the documentation update:

    If both the main working tree and linked working trees have been
    moved manually, then running `repair` in the main working tree and
    specifying the new `<path>` of each linked working tree will
    reestablish all connections in both directions.

says explicitly that `git worktree repair` must be run in the main
worktree for this particular case. (For a bare repository, the command
would be run in the bare repository instead, but I omitted that bit to
avoid bogging down the documentation, and because the couple preceding
paragraphs already mention the bare repository case, so I figured the
reader would understand.)

I could also have mentioned in the commit message the requirement of
running `git worktree repair` in the main worktree (or bare repo), but
didn't want to repeat what the patch itself was already saying. But I
think I'll update the commit message to mention it when re-rolling
since it's important information for understanding how the repair
works.