Fixing UB in fastpow

I perceive that your intent is to ensure that the memory moves are correctly sized, but why not make the simpler fix of including stdint.h and changing the integer to int32_t? The float type is already standardized as 32-bit by IEEE.

Thoughts?

Good point, that would definitely be a cleaner and more readable fix. Using int32_t makes the intent explicit and avoids relying on assumptions about int size. I can update the patch to include <stdint.h> and switch to int32_t instead. Thanks for pointing that out.

Top notch fix, It must be approved!