Bas Zalmstra

In rattler we added [some additional fields to the `LockedDependency`s](https://github.com/mamba-org/rattler/pull/221) to be able to completely reconstruct `RepoDataRecord`s from conda lock files. The fields we added are: ```rust /// Experimental: architecture...

To me, it makes sense to have two alternative data structures (`LockedCondaDependency` and `LockedPipDependency`). Conda and Python have relatively different fields and mixing the two seems complicated.

Note that this has recently been implemented in pixi as [`pixi exec`](https://pixi.sh/latest/reference/cli/#exec).

Another option would be to add a `noarch` python package together with a virtual package specifier (e.g. `__linux`). But this is very akward and the `noarch` package channel would grow...

Unfortunately for me this still happens. Even if I added `publish = false` to the crates. See https://github.com/baszalmstra/rattler/actions/runs/8019870583/job/21908517597

Its a file we use in the tests that is pretty big. Therefor it is stored in lfs. I could potentially store it not in lfs but the history would...

I would indeed love some help because Im kinda unsure on where to go from here. Since the problem with `git checkout` only seems to happens in the `checkout_head` function...

Thanks, this issue saved me a lot of trouble.

You could try to using the `LLVM_SYS_140_PREFIX` environment variable.

It looks like the program is still including lld headers that are not for lld 14, but its unclear to me why that would be the case.