Stabilize extern_crate_self

[mehcode]

Fixes #56409

[bors]

☔ The latest upstream changes (presumably #57332) made this pull request unmergeable. Please resolve the merge conflicts.

[Centril]

Looks good; the version needs to be adjusted tho.

[Centril]

Stabilization report

r? @Centril

I propose that we stabilize #[feature(extern_crate_self)] with 1.34 as the earliest stable version.

@rfcbot merge

History

• Proposed in: experiment: Support aliasing local crate root in extern prelude #55275 and accepted by fcp-merge experiment: Support aliasing local crate root in extern prelude #55275 (comment).
• Implementation PR: experiment: Support aliasing local crate root in extern prelude #55275 (reviewed by: @eddyb)
• Days since in nightly: 39 (~5.57 weeks)
• Tracking issue: Tracking issue for feature extern_crate_self #56409
• Feature gate test being removed
• Documentation issue in reference: Document #56409 (extern crate self as ALIAS) reference#501

Motivation

To resolve #54647 (and so that proc macro crates can migrate to 2018). Also see #56409 (comment).

As noted in #55275 (comment), this solution doesn't introduce any new syntax and has minimal maintenance cost.

What is stabilized

Suppose that we have the following item in the current crate:

fn the_answer() -> usize { 42 }

It now becomes semantically possible (test) (it already was syntactically possible) to write extern crate self as ALIAS; where ALIAS is an identifier. The semantics of extern crate self as ALIAS; is ALIAS is added to the extern prelude and will refer to the current crate. Since ALIAS is now in the extern prelude you may now write ::ALIAS::the_answer and ALIAS::the_answer under the rules due to #54658.

Since ALIAS is any identifier, you may now for example write: extern crate self as serde; and then refer to the_answer by serde::the_answer.

You may also write:

• macro_rules! alias_self {
      ($alias:ident) => { extern crate self as $alias; }
  }
  
  alias_self!(the_alias);
  
  fn main() {
      assert_eq!(the_alias::the_answer(), 42);
  }

• macro_rules! extern_something {
      ($alias:ident) => { extern crate $alias as the_alias; }
  }
  
  extern_something!(self);
  
  fn main() {
      assert_eq!(the_alias::the_answer(), 42);
  }

  (@mehcode: please extend the set of tests to include both as // run-pass tests..) Done.

What is not stabilized:

• extern crate self; is semantically rejected (test) since it would add self to the extern prelude thus coming into conflict with self::.

  However, it is already syntactically valid and will remain so.
  For example:

  macro_rules! accept_item { ($x:item) => {} }
  
  accept_item! {
      extern crate self;
  }

  (@mehcode: please extend the set of tests to include this as a // compile-pass test..) Done.

[rfcbot]

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

• @Centril
• @aturon
• @cramertj
• @eddyb
• @joshtriplett
• @nikomatsakis
• @nrc
• @pnkfelix
• @scottmcm
• @withoutboats

No concerns currently listed.

Once a majority of reviewers approve (and none object), 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!

See this document for info about what commands tagged team members can give me.

[mehcode]

Thanks for walking me through this @Centril.

[Centril]

Some minor style nits; it looks great otherwise.

[rfcbot]

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

[mehcode]

@Centril What's the next step here?

[Centril]

@mehcode Waiting ~two days; after that I'll r+ :)

[mehcode]

Thanks for the clarification. Read somewhere FCP was 7 days so I was confused.

[Centril]

Hehe ;) It's 10 days. The bot will add a comment when FCP is done.

[rfcbot]

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

[Centril]

@bors r+

[bors]

📌 Commit 09d073a has been approved by Centril

[scottmcm]

@mehcode you're not the only one: rust-lang/rfcs#1981