I have 2 2D arrays that represent开发者_如何学JAVA a maze

const char maze1[10][11] and const char maze2[20][21]

I'm trying to create 1 function to handle both mazes like so:

void solveMaze(maze[][])
{
}

and just pass the maze like solveMaze(maze1);

However, I have to supply a size for the array, which is different depending on which maze is being passed in. Without overloading the function or using function templates, how can I have 1 function to handle both arrays?

C++ answer

Use std::vector:

// Initialize the vector with 11 rows of 10 characters
std::vector<std::vector<char> > maze(11, std::vector<char>(10));

void solveMaze(const std::vector<std::vector<char> > &maze) {
    // note that you can access an element as maze[x][y]
}

The boost::multi_array is slightly more efficient (if you're allowed to use boost). I think it goes something like this:

boost::multi_array<char, 2> maze(boost::extents[10][11]);

void solveMaze(const boost::multi_array<char, 2> &maze) {
    // note that you can access an element as maze[x][y]
}

C answer

Use pointers:

const char maze1[10][11];

void solveMaze(char *maze, size_t x_length, size_t y_length) {
    // note that you can access an element as maze[x + (x_length * y)]
}

Std c++ doesn't allow variably sized arrays. Gnu extensions allow this.

given a gnu compiler, you can

 void solvemaze(int w, int h, const char maze[h][w])
 {    //solve it...
 }

otherwise,

 void solvemaze(int w, int h, const char *maze)
 {    //solve it, bearing in mind:
      //maze[y][x] = maze[(w*y)+x];
 }

Actually it can be solved without vector:

template<size_t N, size_t M>
void foo(char (&maze)[N][M])
{
    // do your stuff here
}

On the other hand, I would also prefer to use vectors: it just feels safer.