C代码中的数组索引混淆

Question

I'm porting some C Code to C#. I'm stuck at a piece where I don't quite understand the Author's intention of writing code in unfamiliar fashion.

The Code is:

typedef struct{
  Int32 window[2][8];     
  Int32 windowF[2][8];
  short Index; 
}BLOCK_SWITCHING_CONTROL;

maxWindow = SrchMaxWithIndex( &blockSwitchingControl->window[0][8-1],
                                      &blockSwitchingControl->Index, 8);

*****************************************************************************
*
* function name: SrchMaxWithIndex
* description:  search for the biggest value in an array
* returns:      the max value
*
**********************************************************************************/
static Int32 SrchMaxWithIndex(const Int32 in[], Int16 *index, Int16 n)
{
  Int32 max;
  Int32 i, idx;

  /* Search maximum value in array and return index and value */
  max = 0;                                                       
  idx = 0;                                                       

  for (i = 0; i < n; i++) {

    if (in[i+1]  > max) {
      max = in[i+1];                                             
      idx = i;                                                   
    }
  }
  *index = idx;                                                  

  return(max);
}

As you can see, when SrchMaxWithIndex is being called, not an array but a single Int32 is being passed as its first parameter which is of course wrong. But because I know for sure the C code has nothing wrong with it, I'm convinced that I'm missing something here. What am I missing? What was the Author's intention to pass a single Int32 instead of an array?

So far I've ported the above to C# in the following manner:

static class BLOCK_SWITCHING_CONTROL{
  Int32[][] window = new int[2]{new int[8], new int[8]};     
  Int32[][] windowF = new int[2]{new int[8], new int[8]};
  short Index; 
};


 maxWindow = SrchMaxWithIndex( blockSwitchingControl.window[0]/*[8-1]*/,
                                      out blockSwitchingControl.Index);

*****************************************************************************
*
* function name: SrchMaxWithIndex
* description:  search for the biggest value in an array
* returns:      the max value
*
**********************************************************************************/
static Int32 SrchMaxWithIndex(Int32 _in[], out Int16 index)
{
  Int32 max;
  Int32 i, idx;

  /* Search maximum value in array and return index and value */
  max = 0;                                                       
  idx = 0;                                                       

  for (i = 0; i < _in.Length; i++) {

    if (in[i+1]  > max) {
      max = in[i+1];                                             
      idx = i;                                                   
    }
  }
  index = idx;                                                  

  return(max);
}

But it is just to remove the errors in C#.

Answer 1

The C code is not passing a single integer. It's passing the address of an integer, using the & prefix operator.

There does seem to be some kind of typo though, since the C code references a windowN member in the struct which does not seem to exist.

Assuming it means windowF , this code:

maxWindow = SrchMaxWithIndex(&blockSwitchingControl->windowF[0][8-1],
                                  &blockSwitchingControl->Index, 8);

Tells the called function to treat the given address as an array of 8 integers. I think this will overflow into windowF[1][] . It's very scary code, but if it were as wrong as you state, it wouldn't compile. You can't in general pass an integer to a function expecting a pointer, in C.

Answer 2

The address of an element of an array is being passed (note the & in &blockSwitchingControl->windowN[0][8-1] ).

So in[i+1] will be equivalent to blockSwitchingControl->windowN[0][8] , which is, presumably, a valid item in the array.

Answer 3

OK, first, I have to assume you have a mistake in the code:

blockSwitchingControl->windowN

Does not exist in the BLOCK_SWITCHING_CONTROL structure. So I'll answer this assuming you ment something like windowF

Now to your question, the author did pass an array... sort of. Presumably blockSwitchingControl is an structure of type BLOCK_SWITCHING_CONTROL. What we're doing here:

&blockSwitchingControl->windowF[0][8-1]

is passing the address of windowF's [0][7]'th element. A multi-dimensional array is linear (continuous) memory so Int32 windowF[2][8] is properly sized for 16 Int32s to be stored in a "row" in memory. Something like:

windowF[2][8] => [0][1][2][3][4][5][6][7][8][9][A][B][C][D][E][F][10]

Thus, if I were to pass the address of windowF's [0][7] element, I'm really passing part of the array:

windowF[0][7]    [0][1][2][3][4][5][6][7][8][9][A][B][C][D][E][F][10]  //full array
          -----------------------------^

So now inside the SrchMaxWithIndex() I have _in[] = an array of Int32's which is equivalent to part of windowF's array. So you can see they're passing an "array's worth of values", even if it's not how you'd expect.