An address in IPv4 space uses 32-bits. It is usually represented
as 4 groups of 8 bits, each shown as decimal digits (e.g. 192.168.0.1).
This is known as dot-decimal notation.
An address in IPv6 space uses 128-bits. It is usually represented
as 8 groups of 16 bits, each shown as hexadecimal digits
(e.g. 2001:0db8:85a3:0000:0000:8a2e:0370:7334). This representation can
also be compressed by removing leading zeros and replacing consecutive
groups of zeros with double-colon (e.g. 2001:db8:85a3::8a2e:370:7334).
Finally, there is also the dual representation. This expresses the final
two groups as an IPv4 address (e.g. 2001:db8:85a3::8a2e:3.112.115.52).
The ip_address() constructor accepts a character vector of IP addresses
in these two formats. It checks whether each string is a valid IPv4 or IPv6
address, and converts it to an ip_address object. If the input is invalid,
a warning is emitted and NA is stored instead.
When casting an ip_address object back to a character vector using
as.character(), IPv6 addresses are reduced to their compressed representation.
A special case is IPv4-mapped IPv6 addresses (see is_ipv4_mapped()), which
are returned in the dual representation (e.g. ::ffff:192.168.0.1).
Integers can be added to or subtracted from ip_address vectors.
This class also supports bitwise operations: ! (NOT), & (AND),
| (OR) and ^ (XOR).