Endianness now used in packing bytes

[patrickfuller]

I've been chasing down a weird bug with some Modbus TCP devices. The holding registers read:

40001: 0000, 40002: 3fc0, 40003: 0000, 40004: 4040

or [0, 16320, 0, 16448]. This should decode to two 32-bit floats: 1.5 and 3.0. This code:

from pymodbus.payload import BinaryPayloadDecoder
decoder = BinaryPayloadDecoder.fromRegisters([0, 16320, 0, 16448])
print(decoder.decode_32bit_float())
print(decoder.decode_32bit_float())

prints -2.984375 and 3.0. I found that this was due to a hardcoded bigendian in BinaryPayloadDecoder.fromRegisters. With that replaced with the endian variable, everything decodes properly.