From: Taylor Blau <me@ttaylorr.com>
Date: Thu, 03 Aug 2023 15:24:33 GMT
Subject: Re: [QUESTION] how to diff one blob with nothing
Message-ID: <ZMvGsYSystLu6oBY@nand.local>
In-Reply-To: <CAOLTT8S=_1Vd2Y4rBqq03JsJ1megRBcx9v-HYAWXe88jDek98Q@mail.gmail.com>

On Thu, Aug 03, 2023 at 01:16:02PM +0800, ZheNing Hu wrote:
> > Here, it is necessary to compare multiple versions of blobs while also
> > considering the situations of creation and deletion.
> >
> > Well, what I need is the "diff" content, with lines in the diff indicating
> > '+' or '-' signs. This can be achieved by manually adding them, but it is
> > not very compatible with the original logic.
>
> The native diff command itself supports comparison with an empty file.
>
> #diff -u  /dev/null a
> --- /dev/null 2023-07-25 16:47:50.270094301 +0800
> +++ a 2023-08-03 13:14:16.980262362 +0800
> @@ -0,0 +1 @@
> +a
>
> So I believe this feature would also be useful in git.

Sure, you can easily diff any file against any other, including if
either one or both are empty. I think the main difference here is that
/dev/null exists on your system without additional configuration and
the empty blob does not exist in a Git repository without additional
configuration (in this case, `git hash-object -w -t blob --stdin
</dev/null`).

TBH, I don't know if /dev/null existing by default is necessarily a
solid argument in favor of having Git repositories come initialized with
the empty blob by default.

(To be clear, when I say "initialized", I mean that a Git repository
would recognize the empty blob object's hash for any value of
`the_hash_algo`, not that every repository would be prepared with a
loose object by default.)

Thanks,
Taylor