Use const generics for array `Default` impl

[varkor]

View all comments

Currently, we generate array impls for every size up to 32 manually using macros, but with const generics at a suitable level of implementation, we can switch to properly parameterising over all lengths.

• Replace most manual impls with const generic impls (PR: Use const generics for array impls [part 1] #62435)
• Replace Default impl with const generic impl (more difficult due to Switch libcore array implementations to const generics. #60466 (comment)).
• Reimplement impls removed due to bloat using const generic impls (see

  rust/src/libcore/array.rs

  Lines 221 to 227 in 7840a0b

  	// NOTE: some less important impls are omitted to reduce code bloat
  	__impl_slice_eq1! { [A; $N], [B; $N] }
  	__impl_slice_eq2! { [A; $N], [B] }
  	__impl_slice_eq2! { [A; $N], &'b [B] }
  	__impl_slice_eq2! { [A; $N], &'b mut [B] }
  	// __impl_slice_eq2! { [A; $N], &'b [B; $N] }
  	// __impl_slice_eq2! { [A; $N], &'b mut [B; $N] }

  and

  rust/src/liballoc/vec.rs

  Lines 2199 to 2204 in bfdfa85

  	// NOTE: some less important impls are omitted to reduce code bloat
  	// FIXME(Centril): Reconsider this?
  	//__impl_slice_eq1! { [const N: usize] Vec<A>, &mut [B; N], [B; N]: LengthAtMost32 }
  	//__impl_slice_eq1! { [const N: usize] Cow<'a, [A]>, [B; N], [B; N]: LengthAtMost32 }
  	//__impl_slice_eq1! { [const N: usize] Cow<'a, [A]>, &[B; N], [B; N]: LengthAtMost32 }
  	//__impl_slice_eq1! { [const N: usize] Cow<'a, [A]>, &mut [B; N], [B; N]: LengthAtMost32 }

  )

[Centril]

As I noted in #60466 (comment) and #60466 (comment) I will not sign off on using const generics in stable Rust until such time that const generics are stable. However, we can provide unstable wrapper types meanwhile.

[petrochenkov]

We can probably both keep the exact observable behavior (impls for 0-32) and reduce metadata bloat by using a trick similar to #60466 (comment) (an empty marker trait implemented using macros + the real impl using const generics and a where clause with the marker trait).

This way const generics would be an implementation detail, so their stabilization wouldn't be required.

[varkor]

The results from the crater run and perf run for using const generics for array impls are in (#60466).

• cargo check performance is increased: Switch libcore array implementations to const generics. #60466 (comment). Otherwise, unchanged.
• No regressions: Switch libcore array implementations to const generics. #60466 (comment). (Actually fixes a couple of crates that were relying on impls for larger arrays 😄)

I think there's good reason to go with @petrochenkov's suggestion (#61415 (comment)) and potentially consider lifting the restriction for array sizes.

[pnkfelix]

does T-compiler actually need to be tagged on this issue? Unless its exposing bugs for const generics themselves, I would think this is solely a T-libs issue, maybe T-lang, but not T-compiler?