Resetting random numbers on the same variable...

Alright, so im storing a random number to a variable using the GetRand function from the tutorial. What im wanting to do is get a different random number assigned to that variable..

Code:

#include <cstdio>
#include <cstdlib>
#include <iostream>
#include <ctime>

int GetRand(int min, int max)
{
  static int Init = 0;
  int rc;
  
  if (Init == 0)
  {
    /*
     *  As Init is static, it will remember it's value between
     *  function calls.  We only want srand() run once, so this
     *  is a simple way to ensure that happens.
     */
    srand(time(NULL));
    Init = 1;
  }

  /*
   * Formula:  
   *    rand() &#37; N   <- To get a number between 0 - N-1
   *    Then add the result to min, giving you 
   *    a random number between min - max.
   */  
  rc = (rand() % (max - min + 1) + min);
  
  return (rc);
}

int newnumber;
int number = getrand(1,12345);

int main()
{
  cout<<"Your current random number is: "<<number<<endl;
  cout<<"Enter '1' for a new number: ";
  cin>>newnumber;
  if(newnumber==1)
  { 
     cout<<"Your new random number is: "<<number<<endl;
  }
  system("PAUSE");
  return 0;
}

Please forgive the crude example program, im sure im missing a header or something. Id really appreciate any help that you all are willing to offer.

To get a new random number you are going to hve to call your function again. EG:

Code:

  if(newnumber==1)
  {
     number = getrand(1, 12345); 
     cout<<"Your new random number is: "<<number<<endl;
  }

Also, you might want to move your variable declarations into main; its good practice to avoid using globals as much as possible. And, I would seed the randomisation at the start of main as it would save having all that extra code in your function to check if it has been done or not, plus it would be better if functions you add later use randomisation as well. In effect this would leave your rand function something like:

Code:

int GetRand(int min, int max)
{
    return (rand() % (max - min + 1) + min);
}

Wow. Thats just ridiculously simple. Thanks alot for the help and tips, Mike.

Originally Posted by mike_g

And, I would seed the randomisation at the start of main as it would save having all that extra code in your function to check if it has been done or not, plus it would be better if functions you add later use randomisation as well.

It is certainly debatable, but there is something to be said for interfaces which automatically initialize themselves. Having main() be responsible for calling srand() increasing coupling between the main program and the RNG -- why should main() care about initializing it?

And pushing all these kinds of initializations into a catch-all "init" function isn't necessarily an improvement either. Especially in large programs, it can be a pain to add initialization calls for each new library layer. If a library requires you to call an XXX_init() function before using it, that opens the possibility that someone will forget to initialize it. And so you wind up writing code which looks a lot like this anyway -- except at a higher layer, where it is more distracting.

Being that this is the C++ forum, I suppose I should suggest encapsulating all this into a RNG class, the constructor of which calls srand().

But I'm going off on a tangent again.