uctypes.FLOAT32 has a special value representation and uctypes_struct_scalar_size() should be used instead of GET_SCALAR_SIZE(). Signed-off-by: Damien George <damien@micropython.org>