#![allow(unexpected_cfgs)] does not work for modules about rust

[lukas-code]

I have never been able to find where that cfg might be set by rustfmt so I (wrongly) assumed it was anymore, looks this I was wrong.

It looks like rustfmt will just always treat cfg_attr(anything, rustfmt::skip) as rustfmt::skip, regardless of the actual cfg condition, so this will also work: (playgound)

mod m {
    #![cfg_attr(any(), rustfmt::skip)]
    
    fn dont_format_me ()  {   }
}

[Urgau]

Another thing that isn't right here is that the rustfmt cfg is not a well known cfg.

I have never been able to find where that cfg might be set by rustfmt so I (wrongly) assumed it was anymore, looks this I was wrong.

@rustbot label -needs-triage +F-check-cfg
@rustbot claim

[taiki-e]

they can (stably) be applied directly to the mod, like this:

#[rustfmt::skip]
mod a;

Unfortunately, that is not an option available to me due to a rust-analyzer bug: rust-lang/rust-analyzer#10826 (comment)

[Urgau]

I would also expect the #[allow] to work here.

However, I don't think you need the cfg_attr anymore, it can be removed safely.

- #![cfg_attr(rustfmt, rustfmt::skip)]
+ #![rustfmt::skip]

[taiki-e]

it can be removed safely.

No. Custom inner attributes are unstable.

error[E0658]: custom inner attributes are unstable
 --> src/lib.rs:1:4
  |
1 | #![rustfmt::skip]
  |    ^^^^^^^^^^^^^
  |
  = note: see issue #54726 <https://github.com/rust-lang/rust/issues/54726> for more information
  = help: add `#![feature(custom_inner_attributes)]` to the crate attributes to enable
  = note: this compiler was built on 2024-05-04; consider upgrading it if it is out of date

playground

[Urgau]

Oh, good point, but they can (stably) be applied directly to the mod, like this:

#[rustfmt::skip]
mod a;

[fmease]

It looks like rustfmt will just always treat cfg_attr(anything, rustfmt::skip) as rustfmt::skip, regardless of the actual cfg condition

Oh, no! >.<

[Urgau]

The rustfmt cfg is now (as of nighlty-2024-05-07) recognized has a well known cfg.
Fixing one of the issue reported by this issue.

[pnkfelix]

cc #82450