[mejrs]

Followup to #164329

It's somewhat inconvenient, for 3rd party lint tool authors, to have to enable a feature that's really an implementation detail of the macro.

[rustbot]

r? @tiif

rustbot has assigned @tiif.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 77 candidates
• Random selection from 21 candidates

[tiif]

Thank you for working on this!

View changes since this review

[tiif]

You will need to write a cross-crate test to test the behavior of using #[allow_internal_unstable(..)] across upstream and downstream crates.

In tests/ui/structs/default-field-values/auxiliary/, you can define the upstream crate, for example aux.rs with something like...

#![feature(default_field_values)]
#![feature(allow_internal_unstable)]
#[allow_internal_unstable(default_field_values)]
#[macro_export]
macro_rules! foo {
    () => { fn foo() { struct Foo {x: usize = 0, } Foo {..}; } };
}

In tests/ui/structs/default-field-values/allow_internal_unstable.rs, you can use the macro with...

//@ aux-build:aux.rs
//@check-pass

extern crate aux;
use aux::foo;

foo!();

fn main(){

You will see different behavior in downstream crate when allow_internal_unstable is used and when it is not used.

[mejrs]

Done, but #![feature(default_field_values)] is not required so that shouldn't be necessary.

[tiif]

The idea that I get from this diff is whenever we use #[allow_internal_unstable(default_field_values)]], we always execute the if path below that suggests adding #![feature(default_field_values)]? I feel I am missing something 🤔

[mejrs]

We are allowed to write trailing .. in a struct expression if either the feature has been enabled or if the span allows it. If not, we emit an error.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[mejrs]

@rustbot ready