C++ program halts after dynamic memory allocation

Question

I'm having problem with a copying method in a simple C++ program. Everytime I call copy:

Sudoku::SudokuNode** Sudoku::copy(SudokuNode** sudokuBoard)
{
  SudokuNode** tempSudokuBoard = new SudokuNode*[9];
  for(int i = 0; i<9; i++)
  {
   tempSudokuBoard[i] = new SudokuNode[9];
   for(int j = 0; j<9; j++)
   {
    tempSudokuBoard[i][j].currentInteger = sudokuBoard[i][j].currentInteger;
    for(vector<int>::iterator iter = sudokuBoard[i][j].possibleIntegers.begin(); iter!= sudokuBoard[i][j].possibleIntegers.end();)
    {
     tempSudokuBoard[i][j].possibleIntegers.push_back(*iter);
    }
   }
  }
  return tempSudokuBoard;
}

The program seems to completely halt, not returning a a visible error.

If I try to debug the program, the debugger works fine until I arrive at the copy method. Then the debugger displays a dialog box saying:

There is no source code available for the current location.

Any idea what is wrong?

you should format your question to make the code more readable — chollida
– chollida, Commented Oct 5, 2009 at 19:28
Anytime I see a T**, I stop thinking of the code as "simple C++". — MSalters
– MSalters, Commented Oct 7, 2009 at 14:01

Adam Wright · Accepted Answer · 2009-10-05 19:29:06Z

6

    for(vector<int>::iterator iter = sudokuBoard[i][j].possibleIntegers.begin(); iter!= sudokuBoard[i][j].possibleIntegers.end();)

You don't seem to be advancing the iterator in that loop, so it will never end. Add ++iter to the counting expression (after the last ; in the for loop).

As to why your debugger can't find source for that location, that's platform dependent. What debugger are you using?

answered Oct 5, 2009 at 19:29

Adam Wright

49.5k12 gold badges134 silver badges153 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

LiraNuna · Accepted Answer · 2009-10-05 19:37:58Z

2

You do not increase the iterator on the inside loop:

   for(vector<int>::iterator iter = sudokuBoard[i][j].possibleIntegers.begin(); iter!= sudokuBoard[i][j].possibleIntegers.end(); ++iter)

Resulting an infinate for loop (Compiler knows this and "optimized" it for an infinite loop, which is why there is no code available).

edited Oct 5, 2009 at 19:37

answered Oct 5, 2009 at 19:30

LiraNuna

67.7k15 gold badges121 silver badges141 bronze badges

7 Comments

Adam Wright Over a year ago

The answer is mostly correct, but there's no "optimisation" for the infinite loop. It still has meaning, so can't be removed.

LiraNuna Over a year ago

Yes there is, the compiler most likely converted it to vector<int>::iterator iter = sudokuBoard[i][j].possibleIntegers.begin(); while(1) { ... }. It's called branch prediction, and GCC (at least) does it well.

Chris K Over a year ago

The compiler does NOT know this. The semantic value of .end() can change at any time in the loop body to make it equal to iter. (Elements added or removed, etc).

LiraNuna Over a year ago

There is nothing in the loop body that modifies .end(), GCC (at least) uses SSA trees to determine that.

Adam Wright Over a year ago

Well, let's assume this is the case, and the code emitted is equivalent to what you claim. It would still have to be either emitted in debug builds, or debug information to compensate. Otherwise, any compiler optimisations at all would render code undebuggable.

|

ChrisW · Accepted Answer · 2009-10-05 19:30:36Z

0

I suggest you recode it to use ...

class SudokuBoard
{
  SudokuNode nodes[9];
};

answered Oct 5, 2009 at 19:30

ChrisW

56.4k14 gold badges125 silver badges236 bronze badges

Collectives™ on Stack Overflow

C++ program halts after dynamic memory allocation

3 Answers 3

Comments

7 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

7 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related