"foo" is &str, so &"foo" should be & &str

[thestinger]

No description provided.

[Kimundi]

Not saying this is a bad idea, but changing it might raise a different consistency problem:
People might now confuse ~"foo" to be ~&'static str - "foo" being sugar for &"foo" is consistent in that regard.

[thestinger]

I don't think it would introduce any confusion. A "foo" literal is already &'static str so there is no change in consistency in regards to ~"foo", just the removal of a useless alternative way of doing the same thing. A no-op & operator on the is definitely confusing... at the moment you need to know that "foo" is a &'static str and also that & is a no-op on it and you're going to need &("foo") to get & &str.

[nikomatsakis]

We made a deliberate decision at some point to make & literals behave inconsistently (e.g., [...] vs &[...] as well). I'm not sure this was a correct decision though.

[thestinger]

I think the distinction for arrays is fine, because both forms are useful. Strings don't really act like arrays either way.

[thestinger]

I'll open a formal RFC about this soon.

[pnkfelix]

(removing I-nominated tag. We do not know what exactly this was nominated for; we believe that the solution to this will fall out of DST work.)

[brson]

This is part of DST: #12938

[ftxqxd]

let x:  &str =  "foo";
let y: &&str = &"foo";

I believe this is done.