impl AsRef<[T; N]> and AsMut<[T; N]> for [T; N]

[GrigorenkoPV]

Fixes #129849.

Apparently .as_ref() is widely used on arrays to mean .as_slice(), so some Edition sorcery similar to #84147 or #124097 is required.

TODO

• Add the impl.
• Wait for rustc_skip_during_method_dispatch: decouple receiver & edition #129871 to land and add #[rustc_skip_during_method_dispatch(..)]
• Implement an edition lint (probably suggesting .as_slice())
• Add some tests similar to Cautiously add IntoIterator for arrays by value #84147.
• Fix any existing problems (do we have any edition 2024 crates internally yet?) and bless the tests.
• The bitter tip of a cucumber: write some docs.

[rust-log-analyzer]

The job x86_64-gnu-tools failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

   Compiling crypto-common v0.1.6
error[E0282]: type annotations needed
   --> /cargo/registry/src/index.crates.io-6f17d22bba15001f/itertools-0.12.1/src/tuple_impl.rs:366:30
    |
336 | / macro_rules! impl_tuple_collect {
337 | |     ($dummy:ident,) => {}; // stop
338 | |     ($dummy:ident, $($Y:ident,)*) => (
339 | |         impl_tuple_collect!($($Y,)*);
    | |         |
    | |         in this macro invocation (#2)
    | |         in this macro invocation (#3)
...   |
...   |
366 | |                     if i < s.len() {
...   |
399 | |     )
400 | | }
    | | -
    | | -
    | | |
    | | in this expansion of `impl_tuple_collect!` (#1)
    | |_in this expansion of `impl_tuple_collect!` (#2)
    |   in this expansion of `impl_tuple_collect!` (#3)
401 |   impl_tuple_collect!(dummy, a, b, c, d, e, f, g, h, i, j, k, l,);

   Compiling num_cpus v1.16.0
error[E0282]: type annotations needed
   --> /cargo/registry/src/index.crates.io-6f17d22bba15001f/itertools-0.12.1/src/tuple_impl.rs:366:30
   --> /cargo/registry/src/index.crates.io-6f17d22bba15001f/itertools-0.12.1/src/tuple_impl.rs:366:30
    |
336 | / macro_rules! impl_tuple_collect {
337 | |     ($dummy:ident,) => {}; // stop
338 | |     ($dummy:ident, $($Y:ident,)*) => (
339 | |         impl_tuple_collect!($($Y,)*);
...   |
...   |
366 | |                     if i < s.len() {
...   |
399 | |     )
400 | | }
    | | -
    | | -
    | | |
    | |_in this expansion of `impl_tuple_collect!` (#1)
    |   in this expansion of `impl_tuple_collect!` (#2)
401 |   impl_tuple_collect!(dummy, a, b, c, d, e, f, g, h, i, j, k, l,);

error[E0282]: type annotations needed
   --> /cargo/registry/src/index.crates.io-6f17d22bba15001f/itertools-0.12.1/src/tuple_impl.rs:366:30
    |
    |
336 | / macro_rules! impl_tuple_collect {
337 | |     ($dummy:ident,) => {}; // stop
338 | |     ($dummy:ident, $($Y:ident,)*) => (
339 | |         impl_tuple_collect!($($Y,)*);
...   |
366 | |                     if i < s.len() {
...   |
399 | |     )
400 | | }
    | |_- in this expansion of `impl_tuple_collect!`
    | |_- in this expansion of `impl_tuple_collect!`
401 |   impl_tuple_collect!(dummy, a, b, c, d, e, f, g, h, i, j, k, l,);

   Compiling block-buffer v0.10.4
   Compiling rustc-rayon-core v0.5.0
   Compiling memmap2 v0.2.3

[compiler-errors]

I am not on T-libs-api, but I don't believe this implementation carries its weight, for the record. Also, it's almost certainly too late to introduce anything new to edition 2024, especially with such weak justification as a single issue report on the tracker.

[GrigorenkoPV]

Also, it's almost certainly too late to introduce anything new to edition 2024

Well, fair enough. I've overestimated this a bit.

I am not on T-libs-api, but I don't believe this implementation carries its weight, for the record.

From what I can tell, there is an intent to eventually add a blanket AsRef<T> for T where T: ?Sized, but it is currently not possible because of overlapping, so in the meantime there are some concrete impls as a replacement. So I'd argue this impl should exist in the meantime.

But anyways, there seems to be no sense in rushing to get this in 2024 edition and there is quite a bit of time before the next (2027?) one, so I guess I will go through the proper ACP process first, even though it will take quite a while.

Thank you for the cool-headed perspective and quick feedback on this! ❤️

[alexkazik]

@GrigorenkoPV Yes, AsRef<[T]> for [T; N] exists, but not AsRef<[T; N]> for [T; N], as well for AsMut, which was the reason I opened the ticket linked it the top (and the name of this issue).