Stabilize some MaybeUninit behavior as const

[jhpratt]

This stabilizes the MaybeUninit::as_ptr, MaybeUninit::assume_init, and MaybeUninit::assume_init_ref as const fn. MaybeUninit::as_mut_ptr has been moved to a new flag: const_maybe_uninit_as_mut_ptr, which is blocked on #57349. MaybeUninit::slice_assume_init_ref can be const fn when the method is stabilized in general.

The relevant intrinsic has been stabilized as const as well, though this isn't user-visible. Due to the seemingly unrelated feature name I performed rg const_assert_type and found no other instances of it being used.

r? @oli-obk

@rustbot label: +A-const-fn +S-waiting-on-review +T-libs-api

[oli-obk]

Seems like we never added a test for these. Please make sure they are tested in a const context with some reasonablish use case (can we perform examples from the docs in const contexts now?)

[jhpratt]

My personal use case for field-by-field initialization of a struct. This would have to be done in a single method for a test, as mut refs aren't yet const stable.

[oli-obk]

My personal use case for field-by-field initialization of a struct. This would have to be done in a single method for a test, as mut refs aren't yet const stable.

tests can use unstable features, it seems ok to me to use const mut refs for such a test

[jhpratt]

I knew that; I wasn't sure because you alluded to having a doctest. I'll add a regular UI test momentarily.

[oli-obk]

I knew that; I wasn't sure because you alluded to having a doctest. I'll add a regular UI test momentarily.

Ah sorry. I meant whether the doc examples contain something that we could copy into a test and make const. But your example is a great test, too, so let's roll with that

[oli-obk]

r? rust-lang/libs-api

cc @rust-lang/wg-const-eval

[jhpratt]

Ah gotcha. Either way works 🤷🏽

[RalfJung]

Can this be removed now?

[jhpratt]

No, as the method takes a mutable parameter. const_mut_refs hasn't yet been stabilized.

[RalfJung]

This is for assume_init, which is being stabilized... is it not? I am confused.

[jhpratt]

Whoops, when I responded I thought misread where the code was. I'll check the necessity of this.

[jhpratt]

Removing this results in the following errors:

error: `maybe_uninit::MaybeUninit::<T>::as_mut_ptr` is not yet stable as a const fn
   --> library/core/src/mem/maybe_uninit.rs:915:19
    |
915 |             &mut *self.as_mut_ptr()
    |                   ^^^^^^^^^^^^^^^^^
    |
    = help: add `#![feature(const_maybe_uninit_as_mut_ptr)]` to the crate attributes to enable

error: `maybe_uninit::MaybeUninit::<T>::as_mut_ptr` is not yet stable as a const fn
   --> library/core/src/ptr/mod.rs:374:32
    |
374 |         copy_nonoverlapping(x, tmp.as_mut_ptr(), 1);
    |                                ^^^^^^^^^^^^^^^^
    |
    = help: add `#![feature(const_maybe_uninit_as_mut_ptr)]` to the crate attributes to enable

error: `maybe_uninit::MaybeUninit::<T>::as_mut_ptr` is not yet stable as a const fn
   --> library/core/src/ptr/mod.rs:485:17
    |
485 |         let t = t.as_mut_ptr() as *mut u8;
    |                 ^^^^^^^^^^^^^^
    |
    = help: add `#![feature(const_maybe_uninit_as_mut_ptr)]` to the crate attributes to enable

error: `maybe_uninit::MaybeUninit::<T>::as_mut_ptr` is not yet stable as a const fn
   --> library/core/src/ptr/mod.rs:511:17
    |
511 |         let t = t.as_mut_ptr() as *mut u8;
    |                 ^^^^^^^^^^^^^^
    |
    = help: add `#![feature(const_maybe_uninit_as_mut_ptr)]` to the crate attributes to enable

error: `maybe_uninit::MaybeUninit::<T>::as_mut_ptr` is not yet stable as a const fn
   --> library/core/src/ptr/mod.rs:701:34
    |
701 |         copy_nonoverlapping(src, tmp.as_mut_ptr(), 1);
    |                                  ^^^^^^^^^^^^^^^^
    |
    = help: add `#![feature(const_maybe_uninit_as_mut_ptr)]` to the crate attributes to enable

error: `maybe_uninit::MaybeUninit::<T>::as_mut_ptr` is not yet stable as a const fn
   --> library/core/src/ptr/mod.rs:793:47
    |
793 |         copy_nonoverlapping(src as *const u8, tmp.as_mut_ptr() as *mut u8, mem::size_of::<T>());
    |                                               ^^^^^^^^^^^^^^^^
    |
    = help: add `#![feature(const_maybe_uninit_as_mut_ptr)]` to the crate attributes to enable

error: could not compile `core` due to 6 previous errors

As such I'll be leaving the attribute in.

[RalfJung]

I was asking about the const_maybe_uninit_assume_init, not the const_maybe_uninit_as_mut_ptr. (Note the line this comment is attached to.)

[jhpratt]

Yeah, that probably should have been explicitly stated 🙂 I'll check when I'm on my laptop and send a follow-up if necessary.

[RalfJung]

Thanks. :)

[jhpratt]

Removing that line results in an error on line 496 of the same file (inside MaybeUninit::write). That's for the assume_init_mut method.

So...I think this means that const stability isn't currently subject to the same checks, in that a feature can be both simultaneously stable and unstable? If that's the case we should fix that.

[RalfJung]

Ah I see, makes sense. I opened an issue for this: #91357

[m-ou-se]

@rfcbot merge

[dtolnay]

@rust-lang/libs-api
@rfcbot fcp merge

This stabilizes const for the following already stable signatures:

impl<T> MaybeUninit<T> {
    pub const fn as_ptr(&self) -> *const T;
    pub const unsafe fn assume_init(self) -> T;
    pub const unsafe fn assume_init_ref(&self) -> &T;
}

[Amanieu]

What about assume_init_mut? Is there any reason not to const-stabilize it as well?

[dtolnay]

You can't have &mut in a const fn signature on stable.

error[E0658]: mutable references are not allowed in constant functions
 --> src/lib.rs:1:12
  |
1 | const fn f(_: &mut u8) {}
  |            ^
  |
  = note: see issue #57349 <https://github.com/rust-lang/rust/issues/57349> for more information

[rfcbot]

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

• @Amanieu
• @BurntSushi
• @dtolnay
• @joshtriplett
• @m-ou-se
• @yaahc

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.

[rfcbot]

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

[jhpratt]

Bumped stabilization to 1.59 and moved to a unit test. The other review comment I marked resolved as the feature gate is necessary (see my response for the error).

@rustbot label: -S-waiting-on-author +S-waiting-on-review

[oli-obk]

@bors r+ rollup

[bors]

📌 Commit 44b5b83 has been approved by oli-obk