ipnet

Merge @imp's Default implementation. See #18. Ipv4Net and Ipv6Net now default to 0.0.0.0/0 and ::/0 respectively. IpNet defaults to the 0/0 Ipv4Net.

Add #[allow(arithmetic_overflow)] for Ipv4AddrRange::count() and Ipv6AddrRange::count(). Since 1.43.0-nightly it gives a build error but this panic behavior is desired. In future it may be replaced with explicit use of panic!. See #21.

Implement FusedIterator for IpAddrRange, Ipv4AddrRange, Ipv6AddrRange, IpSubnets, Ipv4Subnets, and Ipv6Subnets.

Implement DoubleEndedIterator for IpAddrRange, Ipv4AddrRange, Ipv6AddrRange.

Implement custom count(), last(), max(), min(), nth(), and size_hint() for IpAddrRange, Ipv4AddrRange, Ipv6AddrRange.

Fix bug where IpAddrRange never ends when start and end are both 0 #11

Fix warning about missing 'dyn'

The Emu128 module has been removed. This provided an emulated 128-bit integer for supporting IPv6 addresses. As of Rust 1.26 the built-in 128-bit integers have been marked stable and this library has been updated to use these instead of Emu128.

The with-serde feature name shim has been removed. The serde feature should now be used using the bare serde feature name per the Rust API Guidelines.

The Deref on Ipv4Net and Ipv6Net has been removed. This dereferenced to the Ipv4Addr and Ipv6Addr contained in the type. To use these methods call them directly on the contained IP address of interest, which may be accessed using the addr() or network() methods.

In prior versions it was necessary to use the Contains trait to access the contains() methods. These are now inherited in public methods on the IpNet, Ipv4Net, and Ipv6Net types so are always available.

The implementations of IpAdd<u32> and IpSub<u32> for IpAddr have been removed.

The implementations of IpAdd<u32> and IpSub<u32> for Ipv6Addr have been removed.