Re: [PATCH] maintenance: make unregister idempotent

[Junio C Hamano <gitster@pobox.com>]

"Derrick Stolee via GitGitGadget" <gitgitgadget@gmail.com> writes:

> From: Derrick Stolee <derrickstolee@github.com>
>
> The 'git maintenance unregister' subcommand has a step that removes the
> current repository from the multi-valued maitenance.repo config key.
> This fails if the repository is not listed in that key. This makes
> running 'git maintenance unregister' twice result in a failure in the
> second instance.
>
> Make this task idempotent, since the end result is the same in both
> cases: maintenance will no longer run on this repository.

I am not sure if this is a good idea.  What is the ultimate reason
why we want to allow running it blindly without knowing if it is
necessary?  Is it because there is no easy way to tell if unregister
is needed in the first place?

If this were "unregister X" that takes a parameter that tells what
to unregister, I would be vehemently opposed to the idea, because
I'd expect people make typos and want to be reminded of them when
they happen, but ...

>  	git maintenance unregister &&
>  	git config --global --get-all maintenance.repo >actual &&
> -	test_cmp before actual
> +	test_cmp before actual &&
> +
> +	# Expect unregister to be idempotent.
> +	git maintenance unregister
>  '

... given that the user does not have any control over what to
unregister from the command line (it is decided implicitly by where
the user is), I am halfway OK with it.

A user with two repositories may still be upset after running
"unregister" in the one they did not mean to, and not getting told
about the mistake, though, e.g.

    $ ls *.git
    a.git b.git
    $ cd a.git
    $ git maintenance unregister
    $ cd b.git
    ... this would give you an error message ...
    $ git maintenance unregister
    ... this is silent with the change ...

Surely, the second "cd" should be to ../b.git instead of b.git and
surely the user should have noticed that "cd" gave them an error,
but the unregister that complains would be an extra chance for them
to notice the mistake when they wanted to unregister the maintenance
tasks from all their repositories.

I wonder if something like the "--force" option, i.e.

    $ >cruft
    $ rm curft; echo $?
    rm: cannot remove 'curft': No such file or directory
    1
    $ rm cruft; echo $?
    0
    $ rm cruft; echo $?
    rm: cannot remove 'cruft': No such file or directory
    1
    $ rm -f cruft; echo $?
    0

which gives us both typo detection while supporting blind removal,
is a usable workaround?  I dunno.

But as I said, I am halfway OK with the change.  I just think it is
a bit unfriendly to the users.

Thanks.