Using Turbofish with the "into" method?

help
1

Thanks for all the help so far!

When calling into one has to define the expected type. I tried to use the Turbofish syntaxt, but Rust did not like it. Did I do it incorrectly (the commented out code) or is it only available in the long form at the end of the example?

And if this is the case, why?

fn main() {
    let a = "Hello World!";
    println!("{a}");

    let b = String::from(a);
    println!("{b}");

    let c: String = a.into();
    println!("{c}");

    // let d = a.into::<String>();
    // method takes 0 generic arguments but 1 generic argument was supplied
    // println!("{d}");

    let e = Into::<String>::into(a);
    println!("{e}");
}
2

The method into is not generic. The trait Into is generic.

It works the same with types.

Here's what it looks like to move the generic to a method.

3 Likes
3

Because you are putting the generic param in the wrong place. It's Into<T>::into, not Into::into<T>.

It's completely logical, too. If it were the way you wanted, that would mean that once a type implements the (single, non-generic, hypothetical) Into trait, it could be turned into any other type whatsoever! That clearly shouldn't even occur to you as something that could be possible.

4

Into could be defined as

trait Into: Sized {
    fn into<T: From<Self>>(self) -> T {
        T::from(self)
    }
}

impl Into for T where T: Sized {}

and indeed you can write that (and call it To::to or whatever) if you want a turbofishable into. Std’s Into doesn’t look like that, at least due to the need to sometimes impl Into manually, without a corresponding From – I think it’s rare these days but used to be more common before the orphan rules were relaxed.

5

The problem with that is exactly the From<Self> bound – while the blanket impl is useful and pervasive, the definition above is not equivalent with the current definition, and is strictly more restrictive.

(Not to mention the conceptual difference – IMO even if the bound makes this possible, the non-generic Into trait would be the wrong way to model the relationship.)

6

There's tap::Conv blanket-implemented for all (Sized) types, by the way, which works exactly like this.

2 Likes
7

This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.