Operator overloads for Graphene

[jgcodes2020]

Currently includes:

• vector addition, subtraction, scaling, component-wise multiply/divide
• matrix scaling, matrix multiplication, matrix-vector multiplication
• quaternion addition, multiplication

Semantics follow GLSL: vector * vector is component-wise multiplication, scalar * vector is scaling.

Only thing I haven't implemented is dividing a vector by a scalar, since Graphene for whatever reason doesn't have that as a function.

[sdroege]

Only thing I haven't implemented is dividing a vector by a scalar, since Graphene for whatever reason doesn't have that as a function.

That's just 1.0/scalar * vector or not?

[jgcodes2020]

Only thing I haven't implemented is dividing a vector by a scalar, since Graphene for whatever reason doesn't have that as a function.

That's just 1.0/scalar * vector or not?

Sure that works, but it can lose accuracy because of the limited precision of floating-point numbers.

[sdroege]

Looks good to me in general. @bilelmoussaoui ?

[sdroege]

Sure that works, but it can lose accuracy because of the limited precision of floating-point numbers.

Indeed. Maybe better to keep that explicit then that there's no direct division and leave the choice to the caller

[sdroege]

The clippy warnings need some fixing

[jgcodes2020]

The clippy warnings need some fixing

Do I just explicitly specify the method names (i.e. Self::add(&self, &rhs)?) If I remove the borrow, Rust selects add from std::ops::Add instead of add from graphene::Vec2, since it prefers to take by value.

[sdroege]

Do I just explicitly specify the method names (i.e. Self::add(&self, &rhs)?

I think that would make it most clear