Enhanced numeric support.

* Single and double precision floating point numbers are now supported.
* 64 bit integers are implemented.
* An endian keyword has been added to reduce typing and enhance clarity.

Here's an example of the endian keyword in action:

class A < BinData::Struct
. endian :little
.
. uint16 :a
. uint32 :b
. double :c
. uint32be :d
. array :e, :type => :int16
end

is equivalent to:

class A < BinData::Struct
. uint16le :a
. uint32le :b
. double_le :c
. uint32be :d
. array :e, :type => :int16le
end