Vec2 clamp_length alters vector when length is within specified range

[semihbkgr]

The clamp_length method in the Vec2 incorrectly alters the vector components when the vector's length is within the specified min and max range. This behavior deviates from the expected functionality where the vector should remain unchanged if its length falls within the specified bounds.

let vec = Vec2::new(3., 4.);
let new_vec = vec.clamp_length(-10., 10.);
println!("{}", new_vec);
//the new_vec is [-6, -8]

While it is possible that this behavior is intentional in the current implementation, it seems counterintuitive for the method to alter the vector when its length is already within the specified range. Clarification on whether this is the expected behavior would be appreciated.

[bitshifter]

Thanks for the report, that definitely sounds wrong. I think I can see what the problem is in the implementation, I'll look at fixing that soon.

[semihbkgr]

I've sent a draft PR for the f32 Vec2 method. If it looks good, I can update the other methods afterward.

[bitshifter]

Looking into this some more, I think part of the issue is negative values probably shouldn't be allowed with this function. Basically vector length cannot be less than zero, the clamp function makes sure then length is no less than the given min, but the doesn't really make sense if min (or max for that matter are negative. What should probably happen here is that the docs should say that min and max must be positive and we could add some glam_assets to check the inputs.

[semihbkgr]

I agree. It's much better

[Nul-led]

@bitshifter length shouldn't be zero either, else we'd be dividing by zero which results in NaN values, should prob add check for that in clamp_length & clamp_length_min