Use a default lifetime of 'static in associated consts

[tmandry]

Expand the interpretation of rust-lang/rfcs#1623 "static_lifetime_in_statics" to include associated consts in traits and trait impls when there are no other lifetimes in scope. This means the following code will compile:

struct MyType;

impl MyType {
    // Warns today (elided_lifetimes_in_associated_constant) but won't after this FCP
    const FOO: &i32 = &4;
}

trait Trait {
    // Errors today (missing lifetime specifier) but won't after this FCP
    const BAR: &i32;
    // You will also be able to provide a default:
    // const BAR: &i32 = &1;
}

impl Trait for MyType {
    // Warns today (elided_lifetimes_in_associated_constant) but won't after this FCP
    const BAR: &i32 = Self::FOO;
}

However, if there are any lifetime parameters on the item declaration or const, we will continue to require an explicit choice of lifetime. We may revisit this decision in the future. Examples are listed in the appendix below.

Rationale: There is a high level of churn suggested by the backreferences to #115010. The primary intention of this FCP is to reduce the churn associated with this change. The lang team discussed #124211 and thought it would be good to decide on what we want in the most obvious (and common) cases, which should also reduce the level of churn significantly.

Appendix: Examples with lifetimes

struct MyType<'a>(PhantomData<&'a ()>);

impl<'a> MyType<'a> {
    // Continues to warn (elided_lifetimes_in_associated_constant), may be a hard error later
    const FOO: &i32 = &4;
}

trait Trait<'a> {
    // Continues to error:
    //const BAR: &i32;
    // (but this is fine)
    const BAR: &'a i32;
}

impl Trait<'static> for i32 {
    // Okay, because there are no lifetime parameters on this item.
    // If the trait lifetime were not 'static, you would still get a
    // "const not compatible with trait" error.
    const BAR: &i32 = &42;
}

impl<'a> Trait<'static> for MyType<'a> {
    // Continues to warn (elided_lifetimes_in_associated_constant), may be a hard error later
    const BAR: &i32 = Self::FOO;
}

struct Thing;
impl Thing {
    // Nightly-only feature, but would also error in this case because of
    // the lifetime parameter on the const impl item:
    const FOO<'a>: &i32 = &4;
}

fn foo<'a>(_: &'a ()) {
    impl Thing {
        // Okay; the lifetime of the function cannot be named here,
        // so it does not affect whether there is a default.
        const BAZ: &i32 = &5;
    }
}

[tmandry]

@rfcbot fcp merge

[rfcbot]

Team member @tmandry has proposed to merge this. The next step is review by the rest of the tagged team members:

• @joshtriplett
• @nikomatsakis
• @pnkfelix
• @scottmcm
• @tmandry

No concerns currently listed.

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

cc @rust-lang/lang-advisors: FCP proposed for lang, please feel free to register concerns.
See this document for info about what commands tagged team members can give me.

[compiler-errors]

I can put up a PR to implement this adjustment, pending FCP approval.

[nikomatsakis]

@rfcbot reviewed

[rfcbot]

🔔 This is now entering its final comment period, as per the review above. 🔔

[rfcbot]

The final comment period, with a disposition to merge, as per the review above, is now complete.

As the automated representative of the governance process, I would like to thank the author for their work and everyone else who contributed.

This will be merged soon.

[compiler-errors]

This is done