Coroutine run in different context

[leameow]

While working on tracing, when I stop a worker I get this error during cleanup:

2022-10-26 14:58:43 [error    ] Failed to detach context       [opentelemetry.context]
Traceback (most recent call last):
  File "/Users/israelhalle/Library/Caches/pypoetry/virtualenvs/saturn-engine-WvplgCTT-py3.9/lib/python3.9/site-packages/opentelemetry/trace/__init__.py", line 573, in use_span
    yield span
  File "/Users/israelhalle/Library/Caches/pypoetry/virtualenvs/saturn-engine-WvplgCTT-py3.9/lib/python3.9/site-packages/opentelemetry/sdk/trace/__init__.py", line 1033, in start_as_current_span
    yield span_context
  File "/Users/israelhalle/devel/saturn/src/saturn_engine/worker/services/tracing/tracer.py", line 37, in on_message_executed
    results = yield
GeneratorExit

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/israelhalle/Library/Caches/pypoetry/virtualenvs/saturn-engine-WvplgCTT-py3.9/lib/python3.9/site-packages/opentelemetry/context/__init__.py", line 157, in detach
    _RUNTIME_CONTEXT.detach(token)  # type: ignore
  File "/Users/israelhalle/Library/Caches/pypoetry/virtualenvs/saturn-engine-WvplgCTT-py3.9/lib/python3.9/site-packages/opentelemetry/context/contextvars_context.py", line 50, in detach
    self._current_context.reset(token)  # type: ignore
ValueError: <Token var=<ContextVar name='current_context' default={} at 0x1037375e0> at 0x10371b040> was created in a different Context

It seems like a the coroutine is being switch from contexts / tasks. Perhaps a bug from TaskGroup or the Scheduler? Would be nice to reproduce in a unittest and ensure that Coroutine remains in the same context / tasks during all their lifecycle.

Note that this only happen during cleanup.