ipaddress 0.5.1

Addition/subtraction of large integers is now much faster (#64)
ip_to_bytes() now returns a list of raw vectors instead of a blob object (#65)
- Result can still be converted using blob::as_blob()
- Removes excessive blob package dependency
Refactored C++ code to reduce dependence on Asio library (#64)
- Package developers can now use LinkingTo: ipaddress without LinkingTo: AsioHeaders

ipaddress 0.5.0

This release achieves feature parity with the Python ipaddress module (#46).

Additional checks for reserved addresses/networks: is_private(), is_reserved(), is_site_local() (#49)
New is_global() checks if addresses or networks are not reserved (#49)
New collapse_networks() collapses contiguous and overlapping networks (#54)
New exclude_networks() removes networks from others (#59)
New reverse_pointer() returns the record used by reverse DNS (#57)
New datasets: iana_ipv4 and iana_ipv6 contain registries of allocated blocks (#55)
New bitwise shift operators: %<<% and %>>% (#52)
Reorganized C++ code to expose IpAddressVector and IpNetworkVector classes to other packages (use LinkingTo: AsioHeaders, ipaddress, Rcpp). This does not affect the public R interface. (#51)

format.ip_address(), format.ip_network() and format.ip_interface() gain an exploded parameter to display leading zeros for IPv6 addresses (#56)
ip_to_integer() gains a base parameter to select between decimal, hexadecimal and binary outputs (#47)
Integer-like doubles are now accepted everywhere in place of an integer
Update vignette("ipaddress-examples") to use the fuzzyjoin package (#61)
Update vignette("ipaddress-examples") with how to randomly generate public IP addresses (#62)
Fixed comparison of ip_address() vectors to account for machine endianness (#53)
Comparison of ip_network() and ip_interface() vectors is now consistent with the Python ipaddress module
- ip_network(): network address compared before prefix length
- ip_interface(): network compared before host address

Add support for constructing ip_network() vectors without using CIDR notation
- New common_network() function finds the smallest network containing two addresses (#39)
- New summarize_address_range() function lists the constituent networks of an address range (#41)
- prefix_length() can now infer the prefix length from an ip_address() vector of netmasks and/or hostmasks. This makes it possible to construct an ip_network() vector like so: (#36)
  - ip_network(ip_address("192.0.2.0"), prefix_length(ip_address("255.255.255.0")))
  - ip_network(ip_address("192.0.2.0"), prefix_length(ip_address("0.0.0.255")))
New supernet() and subnets() functions for traversing the network hierarchy (#42)
Added vignette("ipaddress-examples") to describe some typical usage patterns (#43)

Check for user interrupts when processing large data sets
integer_to_ip() now accepts integerish doubles
Addition/subtraction operators now accept integerish doubles
Addition/subtraction operators now catch missing values in 2nd argument
netmask() and hostmask() now raise an error if the prefix_length and is_ipv6 arguments are not both specified (#38)
Fix is_within_any() to correctly catch when IPv6 addresses are in zero networks
Rename vignette("ipaddress") as vignette("ipaddress-classes")

Renamed functions for converting to and from alternative representations of IP addresses: (#30)
- as_packed() and from_packed() become ip_to_bytes() and bytes_to_ip()
- as_binary() and from_binary() become ip_to_binary() and binary_to_ip()
- as_hostname() and from_hostname() become ip_to_hostname() and hostname_to_ip()

New ip_to_integer() and integer_to_ip() functions to encode and decode addresses as integers. Note that the integers are contained within a character vector. Please see the function documentation for an explanation. (#30)

Support pretty printing in tibbles (#27)
- CIDRs aligned on forward slash separator
- Prefix length displayed in green to increase prominence
- Dots and colons displayed in gray to reduce prominence
- IPv6 addresses can be automatically truncated, to fit more columns into the console width
Added hex logo
Compatible with vctrs 0.3.0 (#31, #34)

Added vignette("ipaddress") to introduce the data classes provided by ipaddress (#24)
New ip_interface() class to simultaneously store the address and the network it is on (#15)
ip_address() vectors now support bitwise operations (!, &, | and ^)
ip_address() vectors now support addition and subtraction of integers (#14)
Added an ip_network() constructor that accepts ip_address and prefix length vectors (#9)

New as_packed() and from_packed() functions to encode and decode addresses as raw bytes (#13)
New as_binary() and from_binary() functions to encode and decode addresses as binary strings (#18)
New as_hostname() and from_hostname() functions translate addresses to and from hostnames (#22)

New max_prefix_length() function to get the size of the address space (32-bit for IPv4 and 128-bit for IPv6)
New functions to check if addresses or networks are reserved for special use: is_multicast(), is_unspecified(), is_loopback(), is_link_local()
New functions for IPv6 transition mechanisms: (#16)
- is_ipv4_mapped(), is_6to4(), is_teredo()
- extract_ipv4_mapped(), extract_6to4(), extract_teredo_server(), extract_teredo_client()

New functions about the size of a network: num_addresses(), network_address() and broadcast_address()
New overlaps() function to check for any overlap between networks
netmask() and hostmask() are now generics, so they can now accept:
- an ip_network() vector (existing behavior)
- an ip_interface() vector
- a vector of prefix lengths

New seq.ip_network() and hosts() functions to list all addresses within a network
New functions to randomly sample addresses: sample_ipv4(), sample_ipv6() and sample_network() (#11, #19)

Improved input validation warnings. They now report the row number involved and no longer mention internal function names.
Many functions now support tidy vector recyling
Performance improvements

Fixed decoding bug which meant that some IP addresses were interpreted as NA
- In IPv4, this affected 0.0.0.128
- In IPv6, this affected any address in the 0:80::/32 network
is_subnet() and is_supernet() now check the prefix length, to ensure they are a true subnet or supernet.

Hotfix for CRAN errors on Solaris.

First CRAN release

Added ip_address() and ip_network() classes
- Full support for both IPv4 and IPv6 address spaces
- Efficiently store data in their native representations (i.e. a sequence of bits)
- Classes based on vctrs package for full tidyverse compatibility
Added functions to determine network membership
- is_within() and is_within_any() check if addresses are within networks
- is_subnet() and is_supernet() check if networks are within other networks
Added convenience functions to extract IP metadata
- is_ipv4() and is_ipv6() classify the address space
- prefix_length(), netmask() and hostmask() yield different representations of the network mask
Calculations performed in C++ (using Rcpp and AsioHeaders packages) for improved performance