Decode decimal-encoded numbers

[andrewthad]

Because of how GHC works, bytesmith always boxes the result of parse. This is normally fine. The result of a parse is typically some kind of boxed data that can live happily on the heap. However, in several different projects, I've ended up needed to define

decodeWord64 :: Bytes -> Maybe Word64
decodeWord64 = Parser.parseBytesMaybe (Latin.decWord64 ())

It's sad to have to allocate for the W# data constructor. I've been avoiding adding Char8/Ascii/Latin module to this library, but I think that Data.Bytes.Latin would be a good place for these:

decWord64 :: Bytes -> Maybe Word64
splitDecWord64 :: Bytes -> Maybe (Word64, Bytes)

The second variant allows extra bytes to be leftover. The first does not.