Re: [PATCH 1/2] t7501: Add tests for various index usages, -i and -o, of commit command.
- From
Junio C Hamano <gitster@pobox.com>
- Date
- Jan 9, 2024, 17:35 UTC
- Message-ID
- <xmqqplyaif4l.fsf@gitster.g>
- In-Reply-To
- <20240109060417.1144647-3-shyamthakkar001@gmail.com>
Ghanshyam Thakkar <shyamthakkar001@gmail.com> writes:
> Subject: Re: [PATCH 1/2] t7501: Add tests for various index usages, -i and -o, of commit command.
Overly long subject that has an unusual capitalization after "t7501:" (see "git log --no-merges --format=%s -20 v2.43.0" for example and try to write something that blends better).
Show 6 quoted lines
> +test_expect_success 'commit with -i fails with untracked files' ' > + test_when_finished "rm -rf testdir" && > + git init testdir && > + echo content >testdir/file.txt && > + test_must_fail git -C testdir commit -i file.txt -m initial > +'
In addition to "why a new repository???" comment raised already, I do not want to see the last command spelled like so. Always write dashed options (and their parameters) before non-option arguments, i.e.
git commit -i -m initial file.txt git -C testdir commit -i -m initial file.txt test_must_fail git -C testdir commit -i -m initial file.txt
The command line parser does rotate the unrecognized arguments to the end and keeps looking for recognisable option (possibly followed by its parameter), but that is purely to help lazy writers (i.e., interactive command users). When writers know "-i" does not take any parameter, it may be convenient if the writer who forgot to say "-m" can just append "-m initial" to what has already be written.
When writing source (be it the production code or test), however, we write for readers. What you wrote at a first glance, especially given that "-i" (or "-o" for that matter) is a relatively less commonly used option, would confuse less experienced readers by making them wonder what "-i file.txt" means (e.g., "is that taking input from the contents of file.txt?").