Re: [PATCH v3 2/2] reflog: implement subcommand to drop reflogs

From: Christian Couder <christian.couder@gmail.com>
Date: Mar 19, 2025, 08:17 UTC
Message-ID: <CAP8UFD1vTkPH=5b++6+=koozC31=wD1gQszvdFmzb73B9HM4Aw@mail.gmail.com>
In-Reply-To: <xmqqsena2pzh.fsf@gitster.g>

On Tue, Mar 18, 2025 at 6:44 PM Junio C Hamano <gitster@pobox.com> wrote:

Show 16 quoted lines

>
> Christian Couder <christian.couder@gmail.com> writes:
>
> > It looks like 'single_worktree' is only used here. This means that if
> > a user forgets to add --all and only uses --single-worktree, nothing
> > will happen and it seems to me that the command will exit with code 0.
> > Even if `git reflog expire` already works like that, I think this is a
> > bit unfortunate.
> >
> > Otherwise this patch series looks very well done to me.
>
> In the thread Toon too seems to have noticed the same "what if there
> is no --all and --single-worktree is given?" gotcha.  Together with
> the "current would be better name than single", we can consider that
> these funnies are to be "consistent" with the "expire" thing, and I
> am OK to see us move on.

I am OK with moving on too. We can "fix" the behavior and naming later in a dedicated separate patch series.

> An alternative is to "fix" the behaviour
> and naming of the "expire" first, and then use the same improved
> behaviour and naming when adding "drop" as a new thing.

I would be OK with that too.

Thanks.