From: Eric Sunshine <ericsunshine@gmail.com>
Date: Thu, 29 Jan 2026 22:51:40 GMT
Subject: Re: What's cooking in git.git (Jan 2026, #09)
Message-ID: <CAPig+cStjN15WEKSN3b5VGLK-+N5pzj-dUDQ2rm15ne_9Q07FQ@mail.gmail.com>
In-Reply-To: <aXvYnCsXPV0yefyD@fruit.crustytoothpaste.net>

On Thu, Jan 29, 2026 at 5:01 PM brian m. carlson
<sandals@crustytoothpaste.net> wrote:
> On 2026-01-28 at 22:50:50, brian m. carlson wrote:
> > On 2026-01-27 at 17:42:00, Junio C Hamano wrote:
> > > * bc/sha1-256-interop-02 (2025-11-17) 15 commits
> > >
> > >  The code to maintain mapping between object names in multiple hash
> > >  functions is being added, written in Rust.
> > >
> > >  Any progress on CI breakages???
> >
> > Not yet, but I hope to spend some time on that this week or early next.
> > I now have some more time to work on this at work and I can try to poke
> > around and figure out what might be wrong, or at least try to beg a
> > colleague to try things on their Windows VM.
>
> I did not get a Windows VM working but I did discover the problem and
> I'm working on a patch and will send it out once I have CI green.
>
> The problem is roughly that both the MSVC linker and MSYS linker are
> both called link.exe.  When compiling with Rust against MSVC, the MSVC
> linker must be used, but the MSYS linker is first in PATH (because I
> believe installing the Git for Windows SDK causes that to be the case),
> causing build.rs to fail to compile.

Didn't Ezekiel already tackle[*] this problem or am I misunderstanding?

[*]: https://lore.kernel.org/git/CAPig+cSBEX5QGnzpBnVs_hKM2iUqcmA4-DzKDgkwpG9ZzWZ__w@mail.gmail.com/