Why does ref-to-ptr cast need a dereference?

help
1

For instance,

pub fn g(r: &'static u32) -> *const u32 {
    let p = r as *const _;
    p
}

is converted to the following in MIR

fn g(_1: &u32) -> *const u32 {
    debug r => _1;
    let mut _0: *const u32;
    scope 1 {
        debug p => _0;
    }

    bb0: {
        _0 = &raw const (*_1);
        return;
    }
}

Why is the reference _1 reborrowed but directly assigned to a pointer type? I understand implicit reborrowing in the context of mutable references but I am not aware of this happening with shared references.

2

A &raw to the place *_1 is created to avoid the need to create a reference in the general case.

Here's the RFC.

The stable/public facing side of avoiding creating reference (which have validity requirements) are macros like addr_of. Or just the methods on *const _ and *mut _.

I doubt there are many if any stability guarantees phrased in terms of MIR, by the way.

3

This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.