`Vec`'s const_make_global method erroring when `Vec` is empty

[izagawd]

I tried this code:

#![feature(const_heap)]
fn main() {
    const {
        let mut inner_vec = Vec::<i32>::new();
        let inner_slice = inner_vec.const_make_global();
    }
}

I expected const_make_global() to just return an empty 'static slice

Instead, it gave me this error:

error[E0080]: pointer not dereferenceable: pointer must point to some allocation, but got 0x4[noalloc] which is a dangling pointer (it has no provenance)
   --> src\main.rs:6:27
    |
  6 |         let inner_slice = inner_vec.const_make_global();
    |                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ evaluation of `main::{constant#0}` failed inside this call
    |
note: inside `Vec::<i32>::const_make_global`
   --> C:\Users\theiz\.rustup\toolchains\nightly-x86_64-pc-windows-msvc\lib/rustlib/src/rust\library\alloc\src\vec\mod.rs:897:18
    |
897 |         unsafe { core::intrinsics::const_make_global(self.as_mut_ptr().cast()) };
    |                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the failure occurred here

But this code:

#![feature(const_heap)]
#![feature(const_trait_impl)]
fn main() {
    const {
        let mut inner_vec = Vec::<i32>::new();
        inner_vec.push(4);
        let inner_slice = inner_vec.const_make_global();
    }
}

works just fine

Meta

rustc --version --verbose:

rustc 1.95.0-nightly (859951e3c 2026-02-24)
binary: rustc
commit-hash: 859951e3c7c9d0322c39bad49221937455bdffcd
commit-date: 2026-02-24
host: x86_64-pc-windows-msvc
release: 1.95.0-nightly
LLVM version: 22.1.0

I am assuming it was because the Vec was empty

[ChrisDenton]

This is a very experimental feature so I'm not sure if it's intended for general use so much as for exploration. cc @oli-obk maybe?

[oli-obk]

Ah yea. This function should just silently do nothing if given an integer instead of a pointer. Same thing likely happens if used on a Box<()>

[izagawd]

cc @oli-obk

[izagawd]

cc? @oli-obk

[oli-obk]

Uh... I already responded? Is github having hickups again?

[izagawd]

Uh... I already responded? Is github having hickups again?

My apologies. Those comments were on the mindset that I made a PR rather than a tracking issue