So the only way to do as you say would be to used presized arrays, and that would be very very bad for my case. So, any other solutions?

I still could use some help on these...

[This message has been edited by magal (edited December 29, 1999).]

Hmm i belive you can have dynamic multidimension arrays in vc++. I use code like this.

typedef struct {
unsigned int base;
unsigned int fringe;
unsigned int type;
} mapdata;

mapdata **map
I then read in the level/map and get its width and height then call this.

void create_map(int w, int h){
int i;
w = w;
h = h;
if ((map = calloc(w, sizeof(mapdata *))) == NULL){
printf("Out of memory during first dimension calloc()");
exit(-1);}
for(i = 0; i < w; i++){
if ((map = calloc(h, sizeof(mapdata))) == NULL){<BR>printf("Out of memory during second dimension calloc()");<BR>exit(-1);}}}<P>And this gives me a nice dynamic 2 dimensional array for my map. <BR><p>[This message has been edited by Zenroth (edited December 28, 1999).]

These are all good methods and I guess they're not hard to implement, but which one is the most efficient?

Dynamically allocate memory for you map..
ex: MapData = new char[(Map_Width*Map_Height)];

Now, before you check the tile to the left,right,above,or below, first see what the location of the current tile you are on is. If you are at the edge of the map (done by checking the Map_Width & Map_Height variables), then don't attempt to get data from the tile which doesn't exist.

P.S. even though you allocate the MapData as a single dimensional array. You can still easily access the tiles by their location in map coordinates.

ex: MapData[(Y_Loc*Map_Width)+X_Loc]
X_Loc & Y_Loc representing the physical map coordinates.

On the processing for the game, I have to check , for example, if the tile to the right of MyTile has some properties. Thus, I just sum 1 to the subscript of the current tile (MyTile). But if MyTile is in the last column to the right, then the tile I select summing 1 is the leftmost tile on the next line, not what I wanted, of course. This same thing can happen on all borders of the tiled field. How can I solve this?

I have tried to manually check if the tile was in the corners or last columns or rows and handle each case differently (thus there were 8 ifs in the function), but that completely sucked, it was too much code. I guess there must be some other way to solve this.

What I have in mind right now is to create another function that will return a byte containing what are the tile's neighbors, for example, if I call the function with the first tile on the field, it will return me a bitfield (supposedly a class or struct with 8 bitfields of 1 bit each), containing 1s on the members bRight, bBottomRight, bBottom and 0s on the members bBottomLeft, bLeft, bUpLeft, bUp, bUpRight; and then, before I check for the neighboring tiles, I just use a if before, like "if (structTile.bUp)". How does this sound? Or you guys have better ideas to help me out?

This generaly has worked well for me.