|Topic was automatically imported from the old Question2Answer platform.
I am trying to spawn multiple Area2D nodes with random coordinates while preventing overlapping spawns using a CollisionShape2D. I am spawning the nodes one by one.
I’m using the area_entered signal to check for overlaps; if an overlap is detected, reposition the node (this has the effect that it repositions both the previously spawned node, and the new one - that’s fine). However it seems that after a while the area_entered signal stops firing for some nodes, despite the engine detecting overlapping areas. See here on the bottom-right:
The scene is not frozen by the way, I can spawn more boxes and they start moving around again, i.e. the signal starts firing again.
The code is straightforward:
self.position = Vector2(rand_range(0, get_viewport().size.x),
Any idea why the area_entered signals stops firing despite still detecting an overlapping area?
I forgot to mention that to detect if the engine sees overlaps I’m using
get_overlapping_areas() on each of the spawned boxes in
_process - the result of which is displayed in the label.
dilettante | 2020-04-28 13:30
While this may not be an answer to your question, have you considered trying it a different way? It looks like each box is the same width and height, so you could keep a list of the position of all boxes on screen. Then when you randomly spawn a new one, you can perform a check to see if it overlaps with another box. It’s not the most efficient algorithm, O(n^2) runtime, as you have to check each box against every other box for each one spawned, but if you’re not spawning a large number of boxes and you notice no performance hit, it’s much simpler to do.
Trollend | 2020-04-28 18:08
I actually solved the problem of avoiding overlapping spawns by implementing an overlap check in
_physics_process, but I’d still like understand why the
area_entered signal fails sometimes. Whether it’s a problem with my implementation or a bug in godot, it’d be good to know for future reference.
(btw your idea would work as well, but as you said it’s probably not the most efficient approach)
dilettante | 2020-04-28 20:42