CODEWITHSUNDEEP
rust
matthiascy
matthiascy

Reputation: 85

Method not found using trait

I'm a beginner of Rust. I created an trait named Floating, f32 and f64 implement this trait. A generic struct Vec requiring that T must implement Floating trait. I would like to compute norm of the vector, this won't compile with the error message said that no method named sqrt found for type parameter T in the current scope. Why this is happening and how would it work?

use std::ops::{Mul, Add};

trait Floating: Sized + Copy + Clone + Mul<Output=Self> + Add<Output=Self> {}

impl Floating for f32 {}

impl Floating for f64 {}

struct Vec<T: Floating> {
    x: T,
    y: T,
    z: T,
}

impl<T: Floating> Vec<T> {
    fn norm(&self) -> T {
        (self.x * self.x + self.y * self.y + self.z * self.z).sqrt()
    }
}


fn main() {
    let v: Vec<f32> = Vec {x: 1.0, y: 1.0, z: 1.0};
    println!("norm is {:?}", v.norm());
}

Upvotes: 4

Views: 3025

Answers (1)

vallentin
vallentin

Reputation: 26157

First of all, you probably shouldn't name your type Vec, to avoid conflicts and confusion with the std::vec::Vec container.

The issue is that (your) Vec only knows that T requires Floating, and Floating doesn't have a sqrt method. So you need to define that yourself.

trait Floating: Sized + Copy + Clone + Mul<Output = Self> + Add<Output = Self> {
    fn sqrt(self) -> Self;
}

impl Floating for f32 {
    fn sqrt(self) -> Self {
        f32::sqrt(self)
    }
}

impl Floating for f64 {
    fn sqrt(self) -> Self {
        f64::sqrt(self)
    }
}

This is where the num crate comes in handy, as it defines a Float trait, which includes among others a sqrt() method. Using num you can simplify your example to:

// num = "0.3"
use num::Float;

struct Vec<T: Float> {
    x: T,
    y: T,
    z: T,
}

impl<T: Float> Vec<T> {
    fn norm(&self) -> T {
        (self.x * self.x + self.y * self.y + self.z * self.z).sqrt()
    }
}

fn main() {
    let v: Vec<f32> = Vec {
        x: 1.0,
        y: 1.0,
        z: 1.0,
    };
    println!("norm is {:?}", v.norm());
}

Upvotes: 5

Related Questions