Remove trait LengthAtMost32

[kpp]

This is a continuation of #74026 preserving the original burrbull's commit.

I talked to @burrbull, he suggested me to finish his PR.

[rust-highfive]

r? @dtolnay

(rust_highfive has picked a reviewer for you, use r? to override)

[lcnr]

One thing which concerns me is Default for arrays, which afaik we can't implement using const generics rn
as the current Default impl does not have a T: Default bound for N == 0.

So I fear that we will end up with Default being the only trait which is not implemented for all arrays.

[kpp]

I suppose we can have impl<T, const N: usize> ConstDefault for [T; N] with #49147 and https://github.com/Centril/rfcs/blob/rfc/const-default/text/0000-const-default.md

[withoutboats]

@rfcbot fcp merge

I propose we merge this, making it so that all traits in std implemented for arrays except Default be implemented for every length of array.

• These traits have been implemented on top of const generics for over a year, demonstrating that the implementation is solid and reliable.
• We have universal consensus that we want a form of const generics that could support at least all of these impls. There is no concern that we are committing ourselves to a feature we would not be able to support in the long term.
• Since we know the implementation is solid and that we are going to stabilize this functionality at some point, we should provide this functionality to users sooner, rather than later.

I also think we should have a discussion about an MVP of const generics which is solid enough that we could stabilize it in the near term, but that's a separate discussion.

[rfcbot]

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

• @Amanieu
• @SimonSapin
• @cramertj
• @dtolnay
• @joshtriplett
• @nikomatsakis
• @pnkfelix
• @scottmcm
• @sfackler
• @withoutboats

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.

[ollie27]

The docs will also need to be updated, for example:

rust/src/libstd/primitive_docs.rs

Lines 463 to 483 in 0cd7ff7

	/// Arrays of sizes from 0 to 32 (inclusive) implement the following traits if
	/// the element type allows it:
	///
	/// - [`Debug`][debug]
	/// - [`IntoIterator`][intoiterator] (implemented for `&[T; N]` and `&mut [T; N]`)
	/// - [`PartialEq`][partialeq], [`PartialOrd`][partialord], [`Eq`][eq], [`Ord`][ord]
	/// - [`Hash`][hash]
	/// - [`AsRef`][asref], [`AsMut`][asmut]
	/// - [`Borrow`][borrow], [`BorrowMut`][borrowmut]
	/// - [`Default`][default]
	///
	/// This limitation on the size `N` exists because Rust does not yet support
	/// code that is generic over the size of an array type. `[Foo; 3]` and `[Bar; 3]`
	/// are instances of same generic type `[T; 3]`, but `[Foo; 3]` and `[Foo; 5]` are
	/// entirely different types. As a stopgap, trait implementations are
	/// statically generated up to size 32.
	///
	/// Arrays of *any* size are [`Copy`][copy] if the element type is [`Copy`][copy]
	/// and [`Clone`][clone] if the element type is [`Clone`][clone]. This works
	/// because [`Copy`][copy] and [`Clone`][clone] traits are specially known
	/// to the compiler.

[nikomatsakis]

@rfcbot reviewed

[rfcbot]

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

[bors]

📌 Commit 8fc7d47 has been approved by dtolnay

[bors]

⌛ Testing commit 8fc7d47 with merge 461707c...

[bors]

☀️ Test successful - checks-actions, checks-azure
Approved by: dtolnay
Pushing 461707c to master...

[elichai]

I just want to note that now the "stable since" annotation on this traits is confusing,
I tried using <[u8; 64]>::try_from(slice) on rust 1.36, assuming it would work because it says stable since "1.34", but it's not because removing the "at most 32" didn't change the stable annotation.

[SimonSapin]

There isn’t an easy fix for that, unfortunately. There’s now a single impl for all lengths and it can only have a single "stable since" attribute. Just changing it to 1.47 would make it incorrect for lengths up to 32.

[SimonSapin]

Though this all could be explained in a doc-comment. I’m sure a PR adding that would be welcome.