Re: [PATCH v2 00/10] Change midx.c and midx-write.c to not use global variables

From: karthik nayak <karthik.188@gmail.com>
Date: Nov 22, 2024, 10:25 UTC
Message-ID: <CAOLa=ZTMk-xDF1o3jfowLeghxa3uAPdZo2_Fmz2chz9PxAOUaw@mail.gmail.com>
In-Reply-To: <xmqqh681fhve.fsf@gitster.g>

Junio C Hamano <gitster@pobox.com> writes:

Show 12 quoted lines

> Karthik Nayak <karthik.188@gmail.com> writes:
>
>> Since the `packfile.c` cleanup is still in flight, this series is based
>> on top of master: b31fb630c0 (Merge https://github.com/j6t/git-gui,
>> 2024-11-11) with those patches merged in. This applies cleanly on top
>> of next, but conflicts with `bf/set-head-symref` in seen, the conflict
>> is mostly straight forward. I'll merge the topic in if it is merged into
>> next soon.
>
> I think set_head() part is fairly trivial.  But there seems to be a
> much bigger conflicts with the incremental midx topic.
>

I see that you've merged the topic to seen and the merge resolution looks ok to me, I'll also merge in `tb/incremental-midx-part-2` for my next version.

Show 8 quoted lines

> I actually do not understand offhand why doing anything to the midx
> layer needs to change the calling convention of set_head() or many
> other things that are unrelated to what midx layer does, and that do
> *not* use the new parameter *anyway*.  Shouldn't the "subsubcommand
> can inherit the repository from the builtin subcommand
> implementation" be split out as a separate topic, which the midx
> thing later builds on?
>

That was just the progression of how I worked on these patches. My goal was to remove global variables, the subsubcommand can inherit just was a dependency which I bundled along, since it was a single patch.

There is no reason it cannot be split, Let me go ahead and do that.