[Bug Report] Waterworld random freezes

[jjshoots]

Describe the bug
Not quite sure what it is yet, but when waterworld is continuously stepped and reset, sometimes it just freezes.

Code example

from pettingzoo.sisl import waterworld_v4

env = waterworld_v4.parallel_env(
    render_mode="rgb_array",
    n_obestacles=2,
    obstacle_coord=None,
)

iters = 0
while True:
    env.reset()
    terms = [False]
    truncs = [False]
    print(iters)
    while not (any(terms) or any(truncs)):
        iters += 1

        # sample_actions
        action_dict = {
            a: env.action_space(a).sample() for a in env.possible_agents
        }

        # step env
        next_obs, rew, terms, truncs, _ = env.step(action_dict)

        terms = terms.values()
        truncs = truncs.values()

Managed to replicate it on 3 different machines, my guess is that there's a rogue while loop somewhere according to the diagnosis below.

The logs of the run are here

Notice by the 2 hour mark everything grinds to a halt. It's not a memory leak AFAICT.

What's weird is that when running the above script, when it freezes and you ctrl-c, it continues running after that. Here's the stack trace:

0
500
1000
1500
2000
2500
3000
3500
4000
^CException ignored from cffi callback <function CollisionHandler._set_separate.<local
s>.cf at 0x7f060d4d52d0>:
Traceback (most recent call last):
  File "/home/jet/Sandboxes/waterworld_sweep/venv/lib/python3.10/site-packages/pymunk/
collision_handler.py", line 199, in cf
    func(Arbiter(_arb, self._space), self._space, self._data)
  File "/home/jet/Sandboxes/waterworld_sweep/venv/lib/python3.10/site-packages/petting
zoo/sisl/waterworld/waterworld_base.py", line 697, in pursuer_evader_separate_callback
    x, y = self._generate_coord(evader_shape.radius)
  File "/home/jet/Sandboxes/waterworld_sweep/venv/lib/python3.10/site-packages/petting
zoo/sisl/waterworld/waterworld_base.py", line 250, in _generate_coord
    while (
KeyboardInterrupt:
4500
5000
5500
6000
6500
7000
7500