sio.call() almost hangs when calling from a different thread

[andantei]

Hi there, my project is using socketio in asgi mode, basically the server receives a message from the client, executes a blocking code that also needs to talk to the client, and returns the response to the client.

I'm currently running the blocking code in a different thread with a different loop using code like this:

def _run_coroutine_in_thread(corofn, *args):
  loop = asyncio.new_event_loop()
  try:
    coro = corofn(*args)
    asyncio.set_event_loop(loop)
    return loop.run_until_complete(coro)
  finally:
    loop.close()

async def run_plugin_task(sio, ...):
  # ......
  response = await sio.call(...)
  # ......

async def handle_run_plugin(sid, data):
  # ......
  return await asyncio.get_event_loop().run_in_executor(None,  _run_coroutine_in_thread, functools.partial(run_plugin_task, sio= self._client_sio, ......))

sio.on('run-plugin', handle_run_plugin, namespace='/daw')

Here the run_plugin_task is running in a new thread with a new loop, it calls sio.call to get data from the same client that sends message to trigger handle_run_plugin. However, this sio.call is taking a really long time(almost a few minutes) when it should have finished instantly. I'm guessing it might have to do with using the same sio from different threads but not sure what's the real reason or how should I fix it. Anyone can help here?

[miguelgrinberg]

This is an invalid usage. You cannot use a different loop for the call(), the entire Socket.IO collection of coroutines need to run under the same loop.