Question

I’m trying to work out a logic for a game (in Flash). In one part, given the following row:

_ _ * _ _ __ *

I need to find if all the spaces between the *‘s are empty (Non-empty values other than * can occur in the row). This operation is done quite frequently.
I was wondering if I could use bit representations of rows to achieve this, instead of looping through and checking the intermediate positions.

For a row _ _ * _ _ _ * represented as xx1xxx1 (x = 1 or 0), I could AND it with 0001110 so that if the answer is 0000000, the intermediate positions are empty.

The question here is, of course, how to find this second map (0001110 above) using bit operations (in Flash AS2)? (Map (1,4) -> 0110, (1,3) -> 0100 etc)
Or is looping through the intermediate positions just the better choice?

Answer 1

You can do this by shifting some bits.

var bits:int = 0x7F; // 7 bits
var mask:int = ((bits >> (bits - end)) << start) & bits;

I’m not sure how you count to the positions, in the example it seems like you did it from the left, starting with a one – I’m going to count from the right, starting with zero, which is more common I guess. So the value of start is the position of the first 1 (from the right, counting up from zero) and end the position of the leftmost 1.

An example of how it works in detail:

// row:          xx1xxx1
// desired mask: 0001110
//               the sequence of ones in the mask starts at 1, ends at 3

bits = 0x7F;        // 1111111
mask = bits >> 4;   // 0000111
mask = mask << 1;   // 0001110

Another way of describing this: bits-(end+1) is the amount of zeroes on the left and start is the amount of zeroes on the right of the sequence of ones.