Union precision values

Question

Why union float values are not precise?
Is it just a bug here?
https://code.sololearn.com/c87OG0c7PQyG/?ref=app

Answer 1

You can't store all possible numbers in a float, because you would need infinite storage.

So there is a gap between a float and the next bigger float. If your calculations end up somewhere in between, the result will be rounded. Floating-point maths is never exact!

Answer 2

That's true, however the rounding happens straight away. When you write down a number literal, you can make that any precision you want. C will happily take

float x = 0.120917581976109238109121;

But that is way more precision than what fits into a single-precision float. The number will have to be truncated so it fits inside a float.

All that happens before the printing, by the way :)

Answer 3

But I was not calculating in that float yet. It was just saved into memory and then called out to print...