I don't get the shortest path

[gyril]

Here is my initial grid, startpoint, endpoint and astar path:

However, if I add an obstacle (4 blocks of the graph are turned to 0s), this is the path I get:

It is obviously a sub-optimal path (it shouldn't go all the way down before the obstacle), however A* is supposed to return the optimal path and I don't get how/why it fails there.

Tell me if you need more data.

[gyril]

Hm after toying with it, the number of nodes is constant, so it's most likely because of the weighting of the diagonal moves (making this issue a duplicate of #10)

[bgrins]

Are you using the diagonal heuristic? I think this might actually be due to missing documentation / not assuming diagonal heuristic with a diagonal grid. Can you try adding { heuristic: astar.heuristics.diagonal } as a final parameter to the search function?

[gyril]

I've edited your library now so I think I broke things, but yeah, from what I understand, applying the diagonal heuristic would have worked; however, it is not documented and I expected it to be applied automatically when I chose a diagonal pathing.

[bgrins]

however, it is not documented and I expected it to be applied automatically when I chose a diagonal pathing.

Sorry about that, I've added a note to the README about the heuristic.