impl From<T> for [T; 1]

[jyn514]

I would have found this useful for writing functions like this: https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=b9c1e08152d1cd3709edf65587d87fc7

[rust-highfive]

r? @joshtriplett

(rust_highfive has picked a reviewer for you, use r? to override)

[sfackler]

How would this implementation allow your example to compile?

[jyn514]

@sfackler looking back at it, it probably wouldn't have. I think this feature stands on its own, though.

impl AsRef<[T]> for T is closer to what I want, but I get an error I don't really understand when I try to implement it:

error[E0119]: conflicting implementations of trait `convert::AsRef<[&_]>` for type `&_`:
    --> src/libcore/slice/mod.rs:5532:1
     |
5532 |   impl<T: ?Sized> AsRef<[T]> for T {
     |   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ conflicting implementation for `&_`
     | 
    ::: src/libcore/convert/mod.rs:504:1
     |
504  | / impl<T: ?Sized, U: ?Sized> AsRef<U> for &T
505  | | where
506  | |     T: AsRef<U>,
507  | | {
...    |
510  | |     }
511  | | }
     | |_- first implementation here

[tesuji]

What about accepting impl Iterator<usize> and using slice::from_ref on callsite?

[jyn514]

@lzutao the idea was to avoid requiring additional effort from the caller. In any case I solved it a different way that still only took one argument.

[clarfonthey]

I feel like an example like this would require From<T> for (T,) too. And honestly, I feel like adding implementations of things for singleton tuples and arrays can start to get very excessive as time goes on, but I'm not personally against it if people have use cases.

[Dylan-DPC-zz]

r? @Amanieu

[Amanieu]

@jyn514 You said you've solved your issue. Does that mean that you don't need this feature any more?

[jyn514]

I guess not. I'm ok with closing this if you think it's not worth the hassle. #68692 would still be super useful though.