Re: [PATCH 1/2] diff --cached: do not report i-t-a entries as "new"

[Junio C Hamano <gitster@pobox.com>]

Nguyễn Thái Ngọc Duy  <pclouds@gmail.com> writes:

> Paths marked by "git add -N" are simply a reminder to the user that
> these files should be staged. If the user does not stage any of them,
> "git commit" will not record them.
>
> Align the behavior of "diff --cached" and "git commit". The most
> prominent result of this patch is "git status" no longer reports i-t-a
> paths as "Changes to be committed".
>
> Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
> ---
>  diff-lib.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/diff-lib.c b/diff-lib.c
> index a85c497..db0e6f8 100644
> --- a/diff-lib.c
> +++ b/diff-lib.c
> @@ -400,6 +400,8 @@ static void do_oneway_diff(struct unpack_trees_options *o,
>  	 * Something added to the tree?
>  	 */
>  	if (!tree) {
> +		if (idx && (idx->ce_flags & CE_INTENT_TO_ADD))
> +			return;
>  		show_new_file(revs, idx, cached, match_missing);
>  		return;
>  	}

This hunk by itself feels like it is going in the right direction.
The HEAD does not have it, and even though there is idx, it merely
is an i-t-a entry.

Don't you need to special case the call to show_modified() at the
end of this function as well, though?  idx is i-t-a, and tree has a
concrete object.  Should it appear as if the path already exists in
the index, or should we pretend as if idx is not yet there?

For example, after this sequence:

    $ git init
    $ >void
    $ git add void
    $ git commit -m void
    $ git rm --cached void
    $ git add -N void

HEAD has "void" with an empty blob, the index has i-t-a.  I am not
sure if we should say something about path "void" when asked:

    $ git diff --cached

Perhaps something like this to cover that case?

	/*
	 * Something removed from the tree?
	 */
-	if (!idx) {
+	if (!idx || (ix->ce_flags & CE_INTENT_TO_ADD)) {
		diff_index_show_file(revs, "-", tree, tree->sha1, 1, tree->ce_mode, 0);
		return;
	}