improve response time for VALUESIN queries with large ranges by 480% :)...

  def load_chunk(position: Int, dst: ListBuffer[(Long, Double)]) : Int = {

    var read_pos = 0

    // we read the data back in 540 byte blocks (30 samples per block)

    var chunk_size = BLOCK_SIZE * 30

    // we read the data back in 65535 byte blocks (3640 samples per block)

    var chunk_size = BLOCK_SIZE * 3640

    val chunk = ByteBuffer.allocate(chunk_size)

    if (position < chunk_size)

  def execute : String = {

    val metric = MetricFactory.get_metric(key)

    val values = metric.values_in(time1 * 1000, time0 * 1000)

    val resp = new StringBuffer

    // we estimate that every tuple will need around 25 byte in its

    // ascii representation

    val resp = new StringBuffer(values.size * 25)

    if (values.size == 0)

      return "null"

    for (ind <- (0 until values.size)) {

      resp.append(values(ind)._1)

    values.foreach { cur =>

      resp.append(cur._1)

      resp.append(":")

      resp.append(FnordMetric.number_format.format((

        values(ind)._2)))

      resp.append(FnordMetric.number_format.format(cur._2))

      if (ind < values.size - 1)

      // if this is the last item, we write an extra whitespace

      // character as checking for the last item would be expensive

      resp.append(" ")

    }

    // we have to remove the last whitespace as the if statement

    // in the loop would be to expensive

    resp.setLength(resp.length - 1)

    resp.toString

  }