Pathfinding bug?

I think that pathfinding is broken.

Please check this video from my prototype https://drive.google.com/file/d/1fitbWivSSj-H_2pCqnkQ5EE33KeYRQbx/view

I could replicate the same problem with Pathfinding basic example https://drive.google.com/file/d/1EgCajrGTWpfB48W1SxII5rz0Wc6vWddu/view

Is this a bug?

It’s because the grid on the pathfinding behavior is not enough small for move the player at the location.

i was going to mention the same thing about the virtual cell grid… it says this in the wiki
virtual cell height and width: our path is generated using a virtual grid. We can change the size of the cells here, the smaller the cell size, the smoother the movement is going to be. A smaller cell size implies more calculations so try to still keep the size as large as possible.

http://wiki.compilgames.net/doku.php/gdevelop5/behaviors/pathfinding

ok, but if you see in first video, the sprite can go from point A to B, but not from B to A.

If it is a problem with the grid, should not be consistence? I can go from A to B, shouldn’t not be able to go from B to A?

And… I can’t go from B to A, but if I make stops in the path, I definitly can go from B to A. If there is a problem with grid, shouldn’t not be able to get from point B to A in any way?

I see what you mean but it’s maybe worth trying different grid size to see if it changes things?

Also I wonder if the cost or impassable factors are coming into play because from the point of your first click, it has to get through 3 different barriers of obstacles. If it’s calculating the “accumulated cost” of all 3 barriers perhaps it’s giving the “i can’t get there” response for that reason? Maybe it’s worth trying to adjust those figures too to see if it rectifies?

I guess you could test the cost if you remove one box from the first row that you are nearest to, and see if the problem persists, or if the character is able to walk through the new opening that you created?

It looks like a bug indeed. Can you share the modified example project? It will make our job easier. :slight_smile:

Thanks @Gruk

Here you have a new simple video with the problem https://drive.google.com/file/d/1k9mjomFgMabygsq46ubDbqG_83wuYG19/view and the test project https://drive.google.com/file/d/1YcsoRjLACoB6ezy-m2AuKky9Gym-PBQW/view?usp=sharing

1 Like

Have the same issues with pathfinding so hope will be fixed very soon :frowning:

I know that lower virtual grid values aren’t recomended for the calculations time but just for test have tried to put their value to 0 and the result is that the sprite sometimes is stuck and don’t reach the target position. I don’t know if it depends from the bug or not.

Working in 5.0.120, due to the changes made by Davy on the precision.

2 Likes

This topic was automatically closed after 8 days. New replies are no longer allowed.