From: Junio C Hamano <gitster@pobox.com>
Date: Thu, 28 May 2020 00:43:18 GMT
Subject: Re: jn/experimental-opts-into-proto-v2, was Re: What's cooking in git.git (May 2020, #09; Tue, 26)
Message-ID: <xmqqeer46g9l.fsf@gitster.c.googlers.com>
In-Reply-To: <20200527235808.GD56118@google.com>

Jonathan Nieder <jrnieder@gmail.com> writes:

>> I have been wondering about the same thing, and if it were not using
>> its own custom way to read the configuration, it would have been a
>> non-brainer to merge it down before the release.
>
> Hm, do you have more detail about this?  git_config_get_bool feels
> very standard and non-custom.
>
> Do you mean that you would like it to go in repo-settings.c?

No.  I fully accept your reasoning in the proposed log message why a
handcrafted query to the config system is done in the location the
patch adds a call.  There is nothing wrong with the patch.

But that means that the way it uses the "experimental" variable is
different from the battle-tested way it has been used, which makes
it less than "non-brainer".  It may not be risky after all, but
still.  This late in the cycle, the fewer things we need to worry
about, the better.
>
>>                                                   But this late in
>> the cycle, I'd rather not.
>
> Sure, I can understand.
>
> Would this be something to put on "master" soon after the release?

That was exactly what I had in mind.