It is that you dont have enough bits to fill in the 239 info BYTES of the code that you want to use. So the only way is to fill those places with zeros. When you generate your redudant bytes, all your 1000 bits _are_ encoded, only that 114 zero BYTEs are _also_ encoded.
Of course you need not trasmit those zeros, which means you can send your 125 info bytes, followed by the 16 redundant bytes. At the reciever, you start decoding assuming 114 uncorrupted zero bytes were also recieved in addition. Which is to say, what goes into your decoder is 125 recieved info bytes (possibly corrupted) + 114 uncorrupted zero bytes + 16 recieved redundant bytes (possibly corrupted)
-b