core: Make Debug impl of raw pointers print metadata if present

[Enselic]

Make Rust pointers appear less magic by including metadata information in their Debug output.

This does not break Rust stability guarantees because Debug impl are explicitly exempted from stability:
https://doc.rust-lang.org/std/fmt/trait.Debug.html#stability

Stability

Derived Debug formats are not stable, and so may change with future Rust versions. Additionally, Debug implementations of types provided by the standard library (std, core, alloc, etc.) are not stable, and may also change with future Rust versions.

Note that a regression test is added as a separate commit to make it clear what impact the last commit has on the output.

Closes #128684 because the output of that code now becomes:

thread 'main' panicked at src/main.rs:5:5:
assertion `left == right` failed
  left: Pointer { addr: 0x7ffd45c6fc6b, metadata: 5 }
 right: Pointer { addr: 0x7ffd45c6fc6b, metadata: 3 }
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[rustbot]

r? @BoxyUwU

rustbot has assigned @BoxyUwU.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[BoxyUwU]

r? libs
lol

[bors]

☔ The latest upstream changes (presumably #135937) made this pull request unmergeable. Please resolve the merge conflicts.

[Enselic]

Amanieu seems busy. Re-rolling reviewer.

r? libs

[thomcc]

This needs an FCP.

@rustbot fcp merge

[thomcc]

Oh right, wrong bot.

@rfcbot fcp merge

[rfcbot]

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

• @Amanieu
• @cuviper
• @joshtriplett
• @m-ou-se
• @the8472
• @thomcc

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.

[Amanieu]

It seems a bit strange to check the size of the metadata. Could we instead explicitly check that the metadata type is ()?

[Enselic]

I tried doing that at first but couldn't figure out how. I tried a bit more now but without success. I can't find any other place in the compiler/libraries that does that either.

Maybe you had some particular technique in mind?

Personally I get peace of mind thinking of this logic as "if the metadata has no size there is nothing of interest to show".

[cuviper]

If we added a 'static requirement to Pointee::Metadata, then you could compare TypeId. I'm not sure if there's any good reason it doesn't have that already.

[Enselic]

Thanks, that seems to work. I think I can push a commit for that to see if CI passes without disrupting the FCP?

Edit: I'm pretty sure, so I'll do it shortly.

[lukas-code]

Pointee::Metadata can't be 'static, because the metadata of dyn Trait + 'a is DynMetadata<dyn Trait + 'a> which is not 'static.

To me it seems safer to drop the 'static bound from the type_id intrinsic rather than potentially causing a trait system unsoundness by adding a 'static bound that isn't actually satisfied by the compiler-builtin impls.

[Enselic]

That makes sense. Out of interest I'll let CI continue to see if it catches it , and then go back to the code I had when the FCP started.

Trying to drop the 'static bound from the type_id intrinsic sounds like out of scope for this PR.

[cuviper]

You could use a local specialization:

trait IsUnit {
    fn is_unit() -> bool;
}

impl<T: ?Sized> IsUnit for T {
    #[inline(always)]
    default fn is_unit() -> bool { false }
}

impl IsUnit for () {
    #[inline(always)]
    fn is_unit() -> bool { true }
}

(It would be better as an associated const, but I don't think min_specialization allows that.)

[Enselic]

Thanks. I now

• check for () with IsUnit instead of size_of() == 0
• have added a comment for future adventurers that explains why Pointee::Metadata can't be 'static
• have rebased on master to resolve the latest conflict

[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.

This will be merged soon.