Re: [PATCH v3] git-send-email: Respect core.hooksPath setting

From: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date: Mar 24, 2021, 00:44 UTC
Message-ID: <87zgytpehg.fsf@evledraar.gmail.com>
In-Reply-To: <xmqqeeg51ojs.fsf@gitster.g>

On Tue, Mar 23 2021, Junio C Hamano wrote:

Show 27 quoted lines

> Ævar Arnfjörð Bjarmason  <avarab@gmail.com> writes:
>
>> From: Robert Foss <robert.foss@linaro.org>
>>
>> get-send-email currently makes the assumption that the
>> 'sendemail-validate' hook exists inside of the repository.
>>
>> Since the introduction of 'core.hooksPath' configuration option in
>> 867ad08a261 (hooks: allow customizing where the hook directory is,
>> 2016-05-04), this is no longer true.
>>
>> Instead of assuming a hardcoded repo relative path, query
>> git for the actual path of the hooks directory.
>>
>> Signed-off-by: Robert Foss <robert.foss@linaro.org>
>> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
>> ---
>>
>> Here's a v3 that fixes various issues with Robert's v2. Range-diff &
>> updated patch below.
>>
>> The advice I had in the v1 feedback about GetHooksPath was bad, just
>> having it be a new accessor is better. It's not like anyone is calling
>> this in a loop.
>
> How urgent is this "fix".  I am wondering if Emily's "git hook"
> automatically fix this for us when it comes.

We've had iterations of that topic for almost a year now (since 2019 counting RFC discussions).

While I'd like to see it land I'm skeptical of parts of that approach[1] and expect we'll have more re-rolls of it, and in any case the conflict in send-email[2] will be trivial to resolve. So I think it makes sense to queue up this narrow fix and not have this wait on the larger topic.

1. https://lore.kernel.org/git/87mtv8fww3.fsf@evledraar.gmail.com/
2. https://lore.kernel.org/git/20210311021037.3001235-36-emilyshaffer@google.com/