Re: [PATCH] odb: do not use "blank" substitute for NULL

From: Junio C Hamano <gitster@pobox.com>
Date: Dec 19, 2025, 12:25 UTC
Message-ID: <xmqq4ipmwtea.fsf@gitster.g>
In-Reply-To: <4d084712-dc9a-4824-b840-4d78831d9da9@app.fastmail.com>

"Kristoffer Haugsbakk" <kristofferhaugsbakk@fastmail.com> writes:

Show 20 quoted lines

> On Thu, Dec 18, 2025, at 11:59, Carlo Marcelo Arenas Belón wrote:
>> On Thu, Dec 18, 2025 at 09:02:59AM -0800, Kristoffer Haugsbakk wrote:
>>>[snip]
>>>
>>> This project typically does not use that trailer/tag.
>>
>> While factually correct, I think the "why" is more interesting in this case.
>> anf the answer IMHO is: not, because it is not needed.
>>
>> % git describe 385e18810f10
>> v2.52.0-25-g385e18810f
>>
>> shows that this bug is only present after 2.52.0 was released so unless you
>> are using unreleased version of git (ex: some development version, including
>> ones that are based on "next"), there is no need to "backport" this fix, as
>> the next version you will use will include it.
>
> So the Linux Kernel (presumably) uses `Fixes` for backporting and/or
> does *not* use it for commits that fix changes that have not been
> released yet. Got it.

I do not run, and I am not involved in, the Linux Kernel project. I am not sure if "is this fix something backporting folks should care about?" is the criterion they use in their project, but if it is, I think it does make a certain sense.

I have mentioned my displeasure with use of "Fixes" in _this_ project before, but that was primarily based on the fact that you do not really know if a proposed commit really fixes or makes something else worse until your alleged "fix" cooks sufficiently long in the field, and I find it distasteful to make such an unsure thing easier to mechanically process.

Thanks.