Using repr C is incompatible with double-precision floats on 32-bit arch. Signed-off-by: Jim Mussared <jim.mussared@gmail.com>