Re: [PATCH v1 0/7] diff: fix -s and --no-patch

From: Sergey Organov <sorganov@gmail.com>
Date: May 12, 2023, 09:32 UTC
Message-ID: <87r0rlj3od.fsf@osv.gnss.ru>
In-Reply-To: <645df6e614f00_215cec29462@chronos.notmuch>

Felipe Contreras <felipe.contreras@gmail.com> writes:

Show 13 quoted lines

> Felipe Contreras wrote:
>> This fixes an issue Sergey Organov reported.
>
> Sergey, as you can see this series fixes the issue you reported.
>
> First, I think these should remain working the same, simply for convenience:
>
>  * git diff         # default output
>  * git diff --patch # patch output
>  * git diff --raw   # raw output
>  * git diff --stat  # stat output
>
> I don't think there's a way I can be convinced otherwise.

Fine with me.

Show 7 quoted lines

>
> But there's many changes:
>
>  1. git diff -s --raw                 # before: nil, after: raw
>  2. git diff --no-patch --raw         # before: nil, after: raw
>  3. git diff --patch --no-patch --raw # before: nil, after: raw
>  4. git diff --raw --patch --no-patch # before: nil, after: raw

Fine as well.

>
> I don't think there's any way you can say my 174 changes make the code work
> "exactly the same".

I said that in the context where we discussed entirely separate issue "handling of defaults by Git commands". Irrelevant to these series as they don't touch this aspect as visible from outside, even though you do change the implementation for better.

>
> And this is better than Junio's solution, because #4 outputs a raw format,
> while in Junio's solution it doesn't output anything.

Yes, and that's where I agreed from the very beginning.

> Even if you don't agree with everything, this solution is better than the
> status quo, and it's better than Junio's solution as it fixes --no-patch
> immediately.

Yep, it fixes "--no-patch" semantics indeed, and as I already said, I do vote in favor of this change, for what it's worth.

Thanks, -- Sergey Organov