Permit uninhabited enums to cast into ints

[Mark-Simulacrum]

This essentially reverts part of #6204; it is unclear why that commit was introduced, and I suspect no one remembers.

The changed code was only called from casting checks and appears to not affect any callers of that code (other than permitting this one case).

Fixes #75647.

[rust-highfive]

r? @nikomatsakis

(rust_highfive has picked a reviewer for you, use r? to override)

[Mark-Simulacrum]

@rust-lang/lang, could someone kick off an fcp merge here?

This permits uninhabited enums to be as cast to integral types. This code is UB to execute due to the presence of the uninhabited value, so we do not need to specify the behavior of such a cast. The primary known use is for macros to not need to special-case this, though I believe we do not currently know of any macros which want this (likely due to no searching having been done).

[RalfJung]

Out of curiosity, what does such a cast compile to?

I'll certainly have to add a Miri test that actually reaches such a cast just to make sure it does not explode in horrible ways.^^

[joshtriplett]

@rfcbot merge

[rfcbot]

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

• @cramertj
• @joshtriplett
• @nikomatsakis
• @pnkfelix
• @scottmcm
• @withoutboats

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!

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

[Mark-Simulacrum]

Pushed up a commit fixing --emit=mir with this PR.

pub enum Void {}
#[no_mangle]
pub fn bar(v: Void) -> usize {
    v as usize
}

MIR:

fn bar(_1: Void) -> usize {
    debug v => _1;                       // in scope 0 at t.rs:3:12: 3:13
    let mut _0: usize;                   // return place in scope 0 at t.rs:3:24: 3:29
    let mut _2: Void;                    // in scope 0 at t.rs:4:5: 4:6

    bb0: {
        StorageLive(_2);                 // scope 0 at t.rs:4:5: 4:6
        _2 = move _1;                    // scope 0 at t.rs:4:5: 4:6
        _0 = move _2 as usize (Misc);    // scope 0 at t.rs:4:5: 4:15
        StorageDead(_2);                 // scope 0 at t.rs:4:14: 4:15
        return;                          // scope 0 at t.rs:5:2: 5:2
    }
}

LLVM IR:

define i64 @bar() unnamed_addr #0 {
start:
  ret i64 undef
}

[RalfJung]

The Miri changes LGTM.
I cannot comment on the other effects is_payloadfree might have.

[nikomatsakis]

@rfcbot fcp 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.

The RFC will be merged soon.

[pickfire]

Should we add a test based on the issue?

[Mark-Simulacrum]

The PR already updates a pre-existing test; did you have something else in mind?

[pickfire]

Test for #75647?

[jplatte]

That's exactly what the uninhabited-enum-cast test is.

[ctrl-z-9000-times]

The primary known use is for macros to not need to special-case this, though I believe we do not currently know of any macros which want this (likely due to no searching having been done).

Can confirm, this is an annoying corner case for my macro which generates custom enums.

[nikomatsakis]

@bors r+

[bors]

📌 Commit 990a395 has been approved by nikomatsakis

[bors]

⌛ Testing commit 990a395 with merge 3f50dea...