Kovacsics Robert

For b), that's why I created https://github.com/KoviRobi/nixpkgs/commit/308f84ff16cb962de509dc9f2333cc7ddc79aeac but it's too ugly to merge back I think. Seems like https://github.com/SuperSandro2000/nixpkgs/commit/449114c6240520433a650079c0b5440d9ecf6156 solves the problem by moving `"${flakeFlags[@]}"` after the sub-command to `nix`/`nom`....

As an added bonus, I've also played around with phantom types to add context to the error, see the branch [error-context-with-phantom-types](https://github.com/KoviRobi/rustler_raise_error/blob/error-context-with-phantom-types/native/nif/src/lib.rs) or the [difference between that branch and main](https://github.com/KoviRobi/rustler_raise_error/pull/1/files). I...

Initially I was trying to get rid of the two wrappers, but I have had the realisation that if we have the pattern ```rust fn _getenv(key: String) -> Result {...

True, it could do that, and it's easy in Elixir to just do a compile-time `for` I guess. I agree about string errors not being great, but I just want...

Or if we could have an `encode` method on a custom, crate-local error type that could return `RaiseTerm` or `RaiseAtom` not just `Term` or `Atom`

Oh hang on, I realised timers work differently than I thought (manual millisecond timer), I'll fix that soon. (The timers are being incremented N times, where N is the number...

Since I think it's in a good state now, I'll pop the first couple `cargo fmt` commits (this comment is partly for me to know when I did that if...

I actually needed the two cores because I am planning on using one of the cores to debug the other one! And still need USB for the other core's stuff...

Of course, feel free to split and modify – my git fu is also reasonably advanced so happy to split things up too, but currently just busy with other stuff...

Also, this doesn't yet include support for mutexes being multicore (I have a draft branch somewhere), I'm just using the multicore FIFO. And as I was thinking about this code...