Choosing float / double packing precision for each field

[mz8i]

Is it at all possible to determine whether a floating-point number should be serialised in packb as a float or a double?
The use_single_float parameter can force a single-precision serialisation for all fields. For bandwidth reasons I would like to use this for all fields in my data structure apart from one, which I would like to keep as a double-precision float. Is this possible with the current API?

[methane]

No. You need to pack those data separately.