Wrong name.encodingLength() for "" and "example.com." -> extra, trailing byte

[jablko]

name.encodingLength() returns the wrong length (off by one) for "" (empty string) and "example.com." (absolute, ends in a dot) whereas name.encode() correctly handles both. Is this by design?

name.encodingLength() handles "example.com" (doesn't end in a dot) but not "", and "." (dot) but not "example.com.". name.encode() handles all four cases. The upshot is that answer.encode({ name: "", type: "A", data: "1.2.3.4" }) returns one extra, trailing byte. I've only ever observed it return a zero byte, but I guess it could be anything, because it uses Buffer.allocUnsafe()?

This simple commit just adds "" to the cases that name.encodingLength() supports. I'm happy to fix the "example.com." case in the same commit, if I'm on the right track?

[thenewwazoo]

I think this might be related to the extra trailing byte I'm seeing in mafintosh/multicast-dns#13