Let me start with the code:

    #include <iostream>  unsigned reverse0(unsigned);  union FourBytes {     unsigned value;     unsigned b1 : 8;     unsigned b2 : 8;     unsigned b3 : 8;     unsigned b4 : 8; }; unsigned reverse1(unsigned);  int main(void) {     std::cout << reverse0(2) << std::endl         << reverse1(2) << std::endl;     return 0; }  unsigned reverse0(unsigned p) {     unsigned res(0);     for (int i(0); i < sizeof(p)*8; ++i)     {         unsigned ithBit( (p & (1<<i)) >> i); //get ith bit          int byteNumber(sizeof(p) - (i / 8) - 1); // get the current byte we are working with         res |= ithBit << byteNumber*8+i; // set corresponding bit in res     }     return res; }  unsigned reverse1(unsigned p) {     FourBytes fb;     fb.value = p;      FourBytes res;     res.b4 = fb.b1;     res.b3 = fb.b2;     res.b2 = fb.b3;     res.b1 = fb.b4;      return res.value; } 

My intention is to figure out how bit manipulations in C++ work. I am trying to implement two methods of reversing the machine word (as far as I get it – the machine word is the same as any data type which consists of 4 bytes). But so far it seems like no one of my methods work. I am very confused. Help me, please.

First and foremost I would like to get comments which would explain what I am doing wrong and how I can fix it. And then I would like to get comments on how I can utilize the best coding practices here and how I can improve performance (if possible). Will be grateful for any other kinds of useful comments.