How is double represented on arduino?

[kosatka]

I wanted to see how arduino represents doubles so I could send them over serial port, so I tried

Serial.begin();
double x = 58.285;
for (int i = 0; i < 8; i++){
   Serial.println((unsigned char)*((char*)&x+i));
}

which prints out

215
35
105
66
0
52
0
91

but no matter how I arrange these bytes, I never get the original number. For getting the original number I used in rust

f64::from_le_bytes([215, 35, 105, 66, 0, 52, 0, 91]);

with some permutations of numbers. For float it works fine without even need to rearrange it, so what's the problem with double?

[wildbill]

Try using sizeof to see how big that double actually is.

[J-M-L]

8 bit or 32 bit Arduino ?

on a 8 bit a double is really just a float (single precision)
on a 32 bit microcontroller then you get 8 bytes and a "real" double

[david_2018]

On most arduino's double is not actually implemented, but will use float instead.

[kosatka]

Thanks a lot, I never considered an option that it is just a float. When I tried that, it worked perfectly.