RFC 1498: ipv6addr-octets

libs (net)

Summary

Add constructor and conversion functions for std::net::Ipv6Addr and std::net::Ipv4Addr that are oriented around arrays of octets.

Motivation

Currently, the interface for std::net::Ipv6Addr is oriented around 16-bit "segments". The constructor takes eight 16-bit integers as arguments, and the sole getter function, segments, returns an array of eight 16-bit integers. This interface is unnatural when doing low-level network programming, where IPv6 addresses are treated as a sequence of 16 octets. For example, building and parsing IPv6 packets requires doing bitwise arithmetic with careful attention to byte order in order to convert between the on-wire format of 16 octets and the eight segments format used by std::net::Ipv6Addr.

Detailed design

The following method would be added to impl std::net::Ipv6Addr:

pub fn octets(&self) -> [u8; 16] {
	self.inner.s6_addr
}

The following From trait would be implemented:

impl From<[u8; 16]> for Ipv6Addr {
	fn from(octets: [u8; 16]) -> Ipv6Addr {
		let mut addr: c::in6_addr = unsafe { std::mem::zeroed() };
		addr.s6_addr = octets;
		Ipv6Addr { inner: addr }
	}
}

For consistency, the following From trait would be implemented for Ipv4Addr:

impl From<[u8; 4]> for Ipv4Addr {
	fn from(octets: [u8; 4]) -> Ipv4Addr {
		Ipv4Addr::new(octets[0], octets[1], octets[2], octets[3])
	}
}

Note: Ipv4Addr already has an octets method that returns a [u8; 4].

Drawbacks

It adds additional functions to the API, which increases cognitive load and maintenance burden. That said, the functions are conceptually very simple and their implementations short.

Alternatives

Do nothing. The downside is that developers will need to resort to bitwise arithmetic, which is awkward and error-prone (particularly with respect to byte ordering) to convert between Ipv6Addr and the on-wire representation of IPv6 addresses. Or they will use their alternative implementations of Ipv6Addr, fragmenting the ecosystem.

Unresolved questions