cargo 1.89.0 can't compile the version of metatensor featomic requests by the camakelists.txt

[MilitaoLucas]

Basically the title. I get:

error: implicit autoref creates a reference to the dereference of a raw pointer
   --> src/c_api/blocks.rs:276:17
    |
276 |         *data = (*block).values.raw_copy();
    |                 ^^-----^^^^^^^^
    |                   |
    |                   this raw pointer has type `*mut mts_block_t`
    |
    = note: creating a reference requires the pointer target to be valid and imposes aliasing requirements
    = note: references are created through calls to explicit `Deref(Mut)::deref(_mut)` implementations
    = note: `#[deny(dangerous_implicit_autorefs)]` on by default
help: try using a raw pointer method instead; or if this reference is intentional, make it explicit
    |
276 |         *data = (&(*block)).values.raw_copy();

[FlorianKleemiss]

+1; building featomic fails on Windows using the current rust version. Any reccomendation?

[Luthaf]

This is strange, I recently published updates that should be picked up by cargo. Can you try removing Cargo.lock in the repository and building again? If it fails again, can you share the full compilation output?

---

Alternatively, if you are using rustup, you can also downgrade the default version of cargo/rustc with

rustup toolchain add 1.88
rustup default 1.88

And build again. Once the underlying issue is fully fixed, you can undo the above with

rustup default stable
rustup toolchain remove 1.88

[MilitaoLucas]

CMake uses externalpackege_add for adding metatensor. Isn't that the issue? The latest version is perfectly good and works fine with featomic.

[Luthaf]

Ah, right. Ok, then this should be fixed by #401, and you can downgrade the rust version used to build the code in the mean time!

[MilitaoLucas]

Yes, that seems good. Thank you! I already fixed the actions for it to also build metatensor. So I already have a workaround. I think this is better because it makes so we can test updates downstream and fix such issues faster.

[Luthaf]

This should be fixed by featomic v0.6.2: https://github.com/metatensor/featomic/releases/tag/featomic-v0.6.2, let me know if you have an issue with this release!