Re: [PATCH] test-chmtime: use parse_options() to parse command-line flags

From: Shashank S <shashu8660@gmail.com>
Date: Oct 29, 2026, 05:54 UTC
Message-ID: <CANHmUQRO1z1hL-L-To3_5n77P9V=GQAhN5cJ8BdDewP9sNqEYA@mail.gmail.com>
In-Reply-To: <xmqqv76nj7ys.fsf@gitster.g>

On Wed, 28 Oct 2026 at 00:18, Junio C Hamano <gitster@pobox.com> wrote:

Show 16 quoted lines

>
> "Shashank S via GitGitGadget" <gitgitgadget@gmail.com> writes:
>
> >
> >  t/helper/test-chmtime.c | 46 +++++++++++++++++++----------------------
> >  1 file changed, 21 insertions(+), 25 deletions(-)
>
> Forgot to mention one thing.  Do we have any tests to cover the
> behaviour of this test helper?  If not, should we add some (not a
> rhetorical question)?
>
> Also, one thing I missed is that the original ignored --verbose
> when --get was given, since they contradict with each other.
> Shouldn't we make them explicitly incompatible with each other,
> perhaps before where we call timespec_arg()?
>

This is a great catch. While `test-chmtime` is heavily utilized in scripts like `t5304-prune.sh`, there are no tests covering the helper's actual argument parsing.

In my v2, I will add `die_for_incompatible_opt2()` immediately after `parse_options()`, and I will include a second patch in the series that adds a dedicated `t/t00xx-test-chmtime.sh` script to explicitly test the `--get`/`--verbose` incompatibility and the standard usage.

Thanks again for the excellent guidance

Shashank s mogaveera