Re: [PATCH v2] blame: allow --contents to work with non-HEAD commit

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

Jacob Keller <jacob.e.keller@intel.com> writes:

> This is because the blame process generates a fake working tree commit
> which always uses the HEAD object as its sole parent.
>
> Enhance fake_working_tree_commit to take the object ID to use for the
> parent instead of always using the HEAD object. Then, always generate a
> fake commit when we have contents provided, even if we have a final
> object. Remove the check to disallow --contents and a final revision.

Around here, probably in between the two paragraphs, it makes sense
to explain why we do this enhancement only for --contents but not
for the case where we take contents from the working tree file (that
would ensure what I wrote will not go waste in my review ;-).

> diff --git a/Documentation/blame-options.txt b/Documentation/blame-options.txt
> index 9a663535f443..6476dd327377 100644
> --- a/Documentation/blame-options.txt
> +++ b/Documentation/blame-options.txt
> @@ -64,11 +64,10 @@ include::line-range-format.txt[]
>  	manual page.
>  
>  --contents <file>::
> +	Pretend the file being annotated has the contents from the named
> +	file instead of using the contents of <rev> or the working tree
> +	copy. You may specify '-' to make the command read from standard
> +	input for the file contents.

Hmph, I can sort of see that "or the working tree copy" refers to
the behaviour when <rev> is not given, but I wonder if it makes it
easier to understand to explicitly say that missing <rev> (or having
no positive end) defaults to "HEAD" when choosing the parent of the
fake commit with the given contents in it.

> @@ -198,10 +198,7 @@ static struct commit *fake_working_tree_commit(struct repository *r,
>  	commit->date = now;
>  	parent_tail = &commit->parents;
>  
> -	if (!resolve_ref_unsafe("HEAD", RESOLVE_REF_READING, &head_oid, NULL))
> -		die("no such ref: HEAD");
> -
> -	parent_tail = append_parent(r, parent_tail, &head_oid);
> +	parent_tail = append_parent(r, parent_tail, oid);

Good.  As fake_working_tree_commit() is no longer about creating a
fake commit based on "HEAD", we do not interpret "HEAD" here, and
instead the caller is now responsible for feeding the fake parent
object name to us.

> @@ -2772,22 +2769,30 @@ void setup_scoreboard(struct blame_scoreboard *sb,
>  		sb->commits.compare = compare_commits_by_reverse_commit_date;
>  	}
>  
> -	if (sb->final && sb->contents_from)
> -		die(_("cannot use --contents with final commit object name"));
> -

OK.

> +	if (sb->contents_from || !sb->final) {
> +		struct object_id head_oid, *parent_oid;
> +
>  		/*
>  		 * "--not A B -- path" without anything positive;
>  		 * do not default to HEAD, but use the working tree
>  		 * or "--contents".
>  		 */

This comment is no longer valid, as this block handles more cases now:

	/*
         * Build a fake commit at the tip of the history, when
         * (1) "git blame ^A ^B -- path", i.e. without any positive
	 *     end of the history range, in which case we build such
	 *     a fake commit on top of the HEAD to blame in-tree
	 *     modifications.
	 * (2) "git blame --contents=file [A] -- path", with or without
	 *     positive end of the history range but with --contents,
	 *     in which case we pretend that there is a fake commit
         *     on top of the positive end (defaults to HEAD) that
         *     has the given contents in the path.
	 */

> +		if (sb->final) {
> +			parent_oid = &sb->final->object.oid;
> +		} else {
> +			if (!resolve_ref_unsafe("HEAD", RESOLVE_REF_READING, &head_oid, NULL))
> +				die("no such ref: HEAD");
> +			parent_oid = &head_oid;
> +		}
> +
>  		setup_work_tree();
>  		sb->final = fake_working_tree_commit(sb->repo,
>  						     &sb->revs->diffopt,
> -						     sb->path, sb->contents_from);
> +						     sb->path, sb->contents_from,
> +						     parent_oid);
>  		add_pending_object(sb->revs, &(sb->final->object), ":");
>  	}

Other than that, looking good.

Thanks.