to_thread.run_sync() doesn't work with the tornado event loop

Hi there!

jupyter_server uses the anyio v2 run_sync_in_worker_thread function , and I noticed a possible bug after upgrading to anyio v3.

This is the error I see when using to_thread.run_sync():

HTTPServerRequest(protocol='http', host='127.0.0.1:38403', method='GET', uri='/a%40b/api/contents/Directory%20with%20spaces%20in', version='HTTP/1.1', remote_ip='127.0.0.1')
Traceback (most recent call last):
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/anyio/_backends/_asyncio.py", line 697, in run_sync_in_worker_thread
    work_queue = _threadpool_work_queue.get()
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/anyio/lowlevel.py", line 120, in get
    raise LookupError(f'Run variable "{self._name}" has no value and no default set')
LookupError: Run variable "_threadpool_work_queue" has no value and no default set

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/tornado/web.py", line 1704, in _execute
    result = await result
  File "/home/runner/work/jupyter_server/jupyter_server/jupyter_server/services/contents/handlers.py", line 111, in get
    path=path, type=type, format=format, content=content,
  File "/home/runner/work/jupyter_server/jupyter_server/jupyter_server/utils.py", line 176, in ensure_async
    result = await obj
  File "/home/runner/work/jupyter_server/jupyter_server/jupyter_server/services/contents/filemanager.py", line 692, in get
    model = await self._dir_model(path, content=content)
  File "/home/runner/work/jupyter_server/jupyter_server/jupyter_server/services/contents/filemanager.py", line 574, in _dir_model
    dir_contents = await run_sync_in_worker_thread(os.listdir, os_dir)
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/anyio/to_thread.py", line 37, in run_sync_in_worker_thread
    return await run_sync(func, *args, cancellable=cancellable, limiter=limiter)
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/anyio/to_thread.py", line 29, in run_sync
    limiter=limiter)
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/anyio/_backends/_asyncio.py", line 707, in run_sync_in_worker_thread
    find_root_task().add_done_callback(_loop_shutdown_callback)
  File "/opt/hostedtoolcache/Python/3.7.10/x64/lib/python3.7/site-packages/anyio/_backends/_asyncio.py", line 56, in find_root_task
Error:     raise RuntimeError('Cannot find root task for setting cleanup callback')

It looks like the find_root_task function is failing because the callback is a tornado.ioloop and not an uvloop.loop.

Jupyter server is using the tornado event loop. For more context, this is the github issue: https://github.com/jupyter-server/jupyter_server/issues/487

Possible Solution

I was able to fix it by updating the following code, but I’m not sure if this is a proper fix.
src/anyio/_backends/_asyncio.py line 50

if cb is _run_until_complete_cb or cb.__module__ in ('uvloop.loop', 'tornado.ioloop')

[agronholm]

I’ve released v3.0.1 with this fix now.

[agronholm]

I will try to resolve this over the coming weekend.