Implement AStarGrid2D class with jump-point pathfinding

[Chaosus]

After some discussion with Juan I decided to try to implement the class which containing a pathfinding grid. Compared to AStar classes it should be more effecient solution for the many users, since that class would support Jump-point search pathfinding algorithm (see https://en.wikipedia.org/wiki/Jump_point_search).

Compared to AStar(right) the performance of this algorithm should be slighty faster in many situations (Elapsed Time is system ticks before call get_point_path function):

It also provides three different Diagonal modes which are:

Always

Never

At Least One Walkable

Only If No Obstacles

I based my implementation on modified version of https://github.com/juhgiyo/EpPathFinding.cs C# library (MIT licensed) which itself based on https://github.com/qiao/PathFinding.js (also MIT licensed).

Also I've included four heuristics modes which are: Manhattan, Euclidean, Octile and Chebyshev which are originally included in the original JS repository (https://github.com/qiao/PathFinding.js/blob/master/src/core/Heuristic.js). If its not desired I could cut it off and leave only Euclidean (which are closest to the existed AStar heuristic).

If you want to test my PR you can download a demo project with it:
TestAStarGrid.zip (UPDATED: 14/08/2022)

I would leave it as a draft for some time, to be successfully checked by @Calinou, @akien-mga, Juan or other engine members.

[Chaosus]

@reduz Fixed all your remarks. Check again. New demo project version -
TestAStarGrid.zip

[Chaosus]

Note: I've removed the point weight parameter from the field - because I'm not sure how it should interact with the jumps correctly. If this feature would be required by the users - it can be implemented later. As the first iteration I think it's good and Juan think so too.