A patch for asyncio.tasks.wait_for for Python 3.8-3.11

[kozlovsky]

The patch fixes #7570 by monkey-patching asyncio.tasks.wait_for with a version that did not suppress CancelledError. Also, the patched version adds base_lost_result_handler that writes an error if a result of a completed task is lost due to a race condition. Right now, I don't think we have some lost resources due to the original wait_for bug, but later we can override base_lost_result_handler to properly close such resources if necessary.

[synctext]

Impressive as always! Just a question to understand how 'deep' this is.
How much time would it take to try to test base_lost_result_handler within our application tester infrastructure. For instance, letting it run for a week with a modified Tribler core or IPv8 test community with multiple asyncio.tasks.wait_for stress testers. If its more than a day of work, I guess we should be focusing on fixing the production errors that Tribler clients are sending us.

[kozlovsky]

@synctext I think setting up the stress testing should require a day of work. I'd like first to spend a day fixing the broken metrics collection and dashboard, and then I can do the stress testing.

[synctext]

OK, then it's a lot of your time. Lets try to get a stable release this week this month...