Chunking data results in different hash to using the data non-chunked

[kcrypt]

Chunking data results in different hash to using the data non-chunked

[mattmook]

I think updating ChunkState.update to the following fixes the problem although I've not tested this in Scala:

  def update(bytes: Array[Byte], from: Int, to: Int): Unit = {
    var i = from

    while (i < to) {
        compressIfRequired()

        val consume = Math.min(BLOCK_LEN - blockLen, to - i)

        System.arraycopy(bytes, i, block, blockLen, consume)
        blockLen += consume
        i += consume
    }
}

I also updated the update(byte) function to use the above so all code paths go through the same thing:

val oneByte: Array[Byte] = new Array[Byte](1)

def update(byte: Byte): Unit = {
    oneByte(0) = byte
    update(oneByte, 0, 1)
  }

[catap]

@mattmook thanks for report and sorry for very long reply. I haven't received any notification about it (!) and have discovered it right now.

[catap]

@mattmook about your code: compressIfRequired() walking over provided array, and System.arraycopy does the same way.

My implementation quite complicated but it allows to never copy a data from a huge block that can be easy compressed from source :)