Tracking issue for integer types conversion to and from byte arrays

[SimonSapin]

A previous incarnation of this API that went though FCP in #49792, but stabilization was reverted while still in Nightly in favor of this API.

Implemented in #51919, conversions between all integer types and appropriately-sized arrays of u8, using the specified endianness:

impl $Int {
    pub fn to_ne_bytes(self) -> [u8; mem::size_of::<Self>()] {…}
    pub fn to_le_bytes(self) -> [u8; mem::size_of::<Self>()] {…}
    pub fn to_be_bytes(self) -> [u8; mem::size_of::<Self>()] {…}
    pub fn from_ne_bytes(bytes: [u8; mem::size_of::<Self>()]) -> Self {…}
    pub fn from_le_bytes(bytes: [u8; mem::size_of::<Self>()]) -> Self {…}
    pub fn from_be_bytes(bytes: [u8; mem::size_of::<Self>()]) -> Self {…}
}

ne, le, and be mean native-endian, little-endian, and big-endian respectively. Native-endian means the target platform’s endianness. Those conversions are implemented as literally transmute and nothing else (but safe). The other conversions use swap or do not swap the byte order depending on the target’s endianness.

[SimonSapin]

I think this feature is ready for stabilization:

@rfcbot fcp merge

This is very soon after landing in Nightly in its current form, but the same functionality already went through FCP for stabilization in #49792 after a couple months in Nightly. The new API contains methods identical to the previous API (only renamed), plus new methods that implement what was previously documented as a typical usage of the previous API.

So there isn’t really anything new here, I expect this final period to be a formality except for possible last-minute naming bikeshed. (The methods landed with consistent abbreviation, another option that was considered was fully expansion with for example to_native_endian_bytes and from_big_endian_bytes.)

[rfcbot]

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

• @Kimundi
• @SimonSapin
• @alexcrichton
• @dtolnay
• @sfackler
• @withoutboats

Concerns:

• missing_impls resolved by Tracking issue for integer types conversion to and from byte arrays #52963 (comment)

Once a majority of reviewers approve (and none object), 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.

[rfcbot]

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

[Kimundi]

@rfcbot concern missing_impls

I just noticed that we are still missing half the implementation for this.

#51919 did just add the changes for signed integers, but not for unsigned.

Compare:

https://doc.rust-lang.org/nightly/std/primitive.u8.html#method.to_bytes
https://doc.rust-lang.org/nightly/std/primitive.i8.html#method.to_ne_bytes

cc @tbu-