On Thu, Jul 5, 2018 at 8:53 PM Derrick Stolee <stolee@xxxxxxxxx> wrote:
> Create a new multi_pack_index struct for loading multi-pack-indexes into
> memory. Create a test-tool builtin for reading basic information about
> that multi-pack-index to verify the correct data is written.
>
> Signed-off-by: Derrick Stolee <dstolee@xxxxxxxxxxxxx>
> ---
> diff --git a/t/helper/test-read-midx.c b/t/helper/test-read-midx.c
> @@ -0,0 +1,34 @@
> +/*
> + * test-mktemp.c: code to exercise the creation of temporary files
> + */
Meh. Copy/paste botch.
> +static int read_midx_file(const char *object_dir)
> +{
> + struct multi_pack_index *m = load_multi_pack_index(object_dir);
> +
> + if (!m)
> + return 0;
Should this 'return 0' be a die() or BUG() or something?
> + printf("header: %08x %d %d %d\n",
> + m->signature,
> + m->version,
> + m->num_chunks,
> + m->num_packs);
> +
> + printf("object_dir: %s\n", m->object_dir);
> +
> + return 0;
> +}
> diff --git a/t/t5319-multi-pack-index.sh b/t/t5319-multi-pack-index.sh
> @@ -3,9 +3,19 @@
> +midx_read_expect () {
> + cat >expect <<-EOF
I guess you're planning on interpolating some variables here in a
later patch, which is why you used -EOF rather than -\EOF?
> + header: 4d494458 1 0 0
> + object_dir: .
> + EOF
> + test-tool read-midx . >actual &&
> + test_cmp expect actual
> +}
> +
> test_expect_success 'write midx with no packs' '
> git multi-pack-index --object-dir=. write &&
> - test_path_is_file pack/multi-pack-index
> + test_path_is_file pack/multi-pack-index &&
> + midx_read_expect
> '
Kind of a do-nothing change. I wonder if this step would better be
delayed until a later patch. (Not necessarily a big deal.)
