We check that the given microsecond fraction is in the valid range [0, 1000000[, and otherwise mark it as invalid. We also drop the useless do loop; a plain block is sufficient here.
