Re: [PATCH] diff-files: treat "i-t-a" files as "not-in-index"

From: Srinidhi Kaushik <shrinidhi.kaushik@gmail.com>
Date: Jun 11, 2020, 23:28 UTC
Message-ID: <20200611232851.GA55419@death-star>
In-Reply-To: <xmqq4krhjql1.fsf@gitster.c.googlers.com>

Thanks for replying!

On Thu, Jun 11, 2020 at 01:27:22PM -0700, Junio C Hamano wrote:

Show 20 quoted lines

> Srinidhi Kaushik <shrinidhi.kaushik@gmail.com> writes:
> 
> > The `diff-files' command and related commands which call `cmd_diff_files()',
> > consider the "intent-to-add" files as a part of the index when comparing the
> > work-tree against it. This was previously addressed in [1] and [2] by turning
> > the option `--ita-invisible-in-index' (introduced in [3]) on by default.
> >
> > For `diff-files' (and `add -p' as a consequence) to show the i-t-a files as
> > as new, `ita_invisible_in_index' will be enabled by default here as well.
> >
> > [1] 0231ae71d3 (diff: turn --ita-invisible-in-index on by default, 2018-05-26)
> > [2] 425a28e0a4 (diff-lib: allow ita entries treated as "not yet exist in
> >                 index", 2016-10-24)
> > [3] b42b451919 (diff: add --ita-[in]visible-in-index, 2016-10-24)
> 
> Is there any place where we still run the internal diff machinery to
> compare the index and the working tree without setting the
> ita_invisible_in_index bit on with this patch applied, and if so,
> why?  Does the justification why that other place needs to leave
> the bit off apply to this codepath as well?

Yes, I believe that there exist some use cases for `ita_invisible_in_index' to be unset. For instance, `index_differs_from' which is used in a quite a few places -- like "commit", "revert", and "rebase" -- which require a "no change" to be returned.

This commit: [1] addressed the issue where the cache-tree was producing the same tree as the current commit when it involved "intent-to-add" entries, instead of aborting.

[1] 018ec3c820 (commit: fix empty commit creation when there's no changes
                but ita entries, 2016-10-24)

> What I am trying to get at is if this is helping only one usecase
> for "diff-files" while breaking other usecases.

Currently, `run_add_p' (for "add"; which this patch addresses
the fix), and `push_to_deploy' (in "receive-pack"; where this
is the intended behavior), call "diff-files" as a subprocess,
in which case the `ita_invisible_in_index' bit is explicitly
set. For all other cases, calls are made directly
to `run_diff_files' and will be unaffected by this change.

> On the other hand, if there is no longer anybody who wants
> ita_invisible_in_index off, perhaps we can get rid of the bit and
> lose many conditionals.