On 28/08/2025 23:01, Paul Caprioli wrote:
>
>> For "float" and "double", however, the sizes are almost universally
>> consistent regardless of the "implementation-defined" leeway given in
>> the standards.
>
> Yes, and note that CPython versions 3.11 and later require that the platform C double follows the IEEE 754 binary64 format.
> https://stackoverflow.com/a/70189136
>

That /really/ is not an issue for an 8-bit microcontroller with perhaps
32 KB flash and 8 KB ram :-)