Yoan

No pressure, free time is hard to find, especially during this period. I must confess I haven't taken any time to dig further, also knowing that it turns out not...

CI checks only failed because of `cargo test --features "serde repr_simd"` on Nightly. Sad thing I didn't notice it beforehand, but at least `cargo build` is not broken... I've commited...

By the way, it's always nice hearing about `vek` being used for cool stuff! :+1: I'm also all for more erognomics in the API.

That reminds me of something. Very recently, for a reason I don't remember, I ended up reading the definition of FMA on Wikipedia, and I realized something important : Strictly...

Btw, since the fix for #73 that I've published just now, `vek` just re-export's `num-trait`'s `MulAdd` instead. I think it's for the better... Both `std` and `num-traits` seem to agree...

Yes, I believe `micromath` would need some changes in order to be usable here, or at least, in a way that doesn't require rewriting `vek` not to depend on `num-traits`....

I'm not against the idea, I just wonder if it could cause issues since vectors already implement `AsRef` and `AsMut`. I also find myself using `Option`'s `as_ref` quite often, but...

Hi, well not really, it's only because inverting Mat4 types is all I needed. I'm totally OK with implementing this though, or accepting PRs that add this. In practice I...

The issues I see : - `div_euc` and `mod_euc` in the std are nightly only. Do we use the std with a feature gate, or do we roll our own?...

I wouldn't mind some kind of `EucOps` trait in `vek::ops`; however, this "feels" very much like an `std` feature. `Clamp`, `Lerp` and friends were OK because they're not in `std`,...