Re: should "--recurse-submodule" be "--recurse-submodules"?

From: Stefan Beller <sbeller@google.com>
Date: Feb 14, 2018, 19:12 UTC
Message-ID: <CAGZ79kZ1yYXToC_WvW8e93zvajkw06e8BT_FcJUXWtFBK3Xe4w@mail.gmail.com>
In-Reply-To: <alpine.LFD.2.21.1802131927310.17903@localhost.localdomain>

On Tue, Feb 13, 2018 at 4:30 PM, Robert P. J. Day <rpjday@crashcourse.ca> wrote:

Show 10 quoted lines

>
>   also just noticed the following:
>
> Documentation/RelNotes/2.13.0.txt:489: * A few commands that recently learned the "--recurse-submodule"
> Documentation/RelNotes/2.12.0.txt:226: * "git push --dry-run --recurse-submodule=on-demand" wasn't
> Documentation/RelNotes/2.11.1.txt:27: * "git push --dry-run --recurse-submodule=on-demand" wasn't
> t/t5531-deep-submodule-push.sh:366:             git push --recurse-submodule=check origin master
> t/t5572-pull-submodule.sh:45:test_expect_success 'pull --recurse-submodule setup' '
>
>   should some of those be corrected?

I get the same list via
  git grep -- --recurse-submodule |grep -v -- --recurse-submodules
so there is no missing piece left.

As you can see, the first 3 are Documentation, and the other 2 are tests, the actual code is always --recurse-submodules (with an s!).

Fixing the docs as well as t5572 has the impact of "just fixing typos" (though that is specifically valuable for command line options where exact spelling matters)

The fix in 5531 is an actual bugfix in the test suite, which went unnoticed as the test itself did not fail.

The test did not fail because git autocompleted the command line option silently. But that is not what we want to test there.

Stefan