Re: [PATCH v2 09/17] Do support Windows again after requiring Rust

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 17, 2025, 15:57 UTC
Message-ID: <xmqqjz32kkjx.fsf@gitster.g>
In-Reply-To: <DB9P250MB0692F39EA259A4B31845C6B2A537A@DB9P250MB0692.EURP250.PROD.OUTLOOK.COM>

Matthias Aßhauer <mha1993@live.de> writes:

Show 6 quoted lines

>> It seems that Dscho did not respond on the list to your initial
>> objection in the discussion you cited.
>
> He didn't, but from various interactions surrounding Git for Windows,
> I do think he's currently in favour of keeping Windows 8.1 supported
> in Git for Windows.

OK, as long as folks with stakes in Git for Windows are in agreement, I have no problem (except that in principle we should avoid doing disservice to the end user population by doing things that encourage their prolonged use of out-of-security-support platforms).

> We don't do that in git.git, no. Git for Windows very explicitly
> spells out which versions of Windows are supported (though usually we
> just mention the Desktop versions and imply the corresponding Windows
> Server versions).

Yup, thanks for clarifying it to me. Could you do the same for future readers of the updated version of the commit 09/17 by telling the author about that in your review comment, so that the log message can talk about the reasons why a specific CARGO_BUILD_TARGET was chosen (e.g. "as described in Git for Windows documentation at $URL, we support Windows versions X or newer, so we use this cargo build target to ensure we still work with that version").

> I'm not completely opposed, but I do think it should be a concious
> decision and not an unintended side effect of some change that our CI
> didn't catch.

Oh, absolutely.

Thanks for clarification.