Re: [PATCH v2 1/5] submodule: expose the '--for-status' option of summary

From: Christian Couder <christian.couder@gmail.com>
Date: Aug 8, 2020, 20:25 UTC
Message-ID: <CAP8UFD20ORozywSAV+Qayuf_vwve9A21ySAtTZVphwhv5nYWXg@mail.gmail.com>
In-Reply-To: <831df9f2-0663-0dfc-0871-d34864d1ecde@gmail.com>

Le sam. 8 août 2020 à 16:40, Kaartic Sivaraam <kaartic.sivaraam@gmail.com> a écrit :

Show 6 quoted lines

>
> On 06-08-2020 22:10, Shourya Shukla wrote:
> > The 'for-status' option is used to compute the summary of submodule(s)
> > in a superproject by skipping the ignored submdules i.e., those with
> > 'submodule.<name>.ignore' set to 'all' in the '.gitmodules' or
> > '.git/config', with the latter taking precedence over the former.

The above seems to suggest that a name like --skip-ignored could fit, if we wanted to rename --for-status.

Show 6 quoted lines

> > The option was introduced in d0f64dd44d (git-submodule summary:
> > --for-status option, 2008-04-12), refined in 3ba7407b8b (submodule
> > summary: ignore --for-status option, 2013-09-06) and finally perfected
> > in 927b26f87a (submodule: don't print status output with ignore=all,
> > 2013-09-01). But, it was not mentioned in the 'git submodule'
> > Documentation.

After this we would need to tell why it's a good idea to actually document this option (and perhaps rename it if we are going to do that). It could be a good idea, if it could help users to see a summary without the ignored submodules.

So for example a possibly good justification could be that in a repo with many ignored submodules it might be interesting for users to get a summary that contains information only about the non-ignored submodules.

An example output of `git submodule summary` both with and without --for-status (or --skip-ignored) in an interesting case (where there are many ignored submodule) could help convince people that it's a possibly useful option, and that it's worth documenting.

Show 11 quoted lines

> > Expose the '--for-status' option accepted by the command 'git submodule
> > summary'.
> >
>
> I've had one concern about exposing '--for-status'. As of now, the name
> of the option has no relation with the behaviour that we get as a
> consequence. So long, the option has been internal and this wasn't a
> problem. Now that we're considering to expose it in the docs, usage and
> autocomplete, I would say it should be done after renaming it
> appropriately given that it's easy to do now than later. As to name
> suggestions, I really don't have any.

Yeah, I agree that finding a good name and a good use case for the option would surely help.

> Also, as to whether exposing this would be useful at all, I really don't
> know.