asyncio: hold our own strong refs for tasks and futures

[SomberNight]

see https://docs.python.org/3.13/library/asyncio-task.html#asyncio.create_task :

Important
Save a reference to the result of this function, to avoid a task
disappearing mid-execution. The event loop only keeps weak references
to tasks. A task that isn’t referenced elsewhere may get garbage
collected at any time, even before it’s done. For reliable
“fire-and-forget” background tasks, gather them in a collection

ref python/cpython#91887

[SomberNight]

for example, this kind of usage is not safe (without this PR):

electrum/electrum/lnpeer.py

Line 3079 in b88d9f9

fut = asyncio.ensure_future(wrapped_callback())

[accumulator]

Interesting read...

Does this make this existing specific strong reference storing code redundant?

electrum/electrum/util.py

Lines 1883 to 1893 in 0b3a283

	fut = asyncio.run_coroutine_threadsafe(callback(*args), loop)
	# keep strong references around to avoid GC issues:
	self._running_cb_futs.add(fut)
	def on_done(fut_: concurrent.futures.Future):
	assert fut_.done()
	self._running_cb_futs.remove(fut_)
	if fut_.cancelled():
	self.logger.debug(f"cb cancelled. {event=}.")
	elif exc := fut_.exception():
	self.logger.error(f"cb errored. {event=}. {exc=}", exc_info=exc)
	fut.add_done_callback(on_done)

[SomberNight]

Does this make this existing specific strong reference storing code redundant?

electrum/electrum/util.py

Lines 1883 to 1893 in 0b3a283

	fut = asyncio.run_coroutine_threadsafe(callback(*args), loop)
	# keep strong references around to avoid GC issues:
	self._running_cb_futs.add(fut)
	def on_done(fut_: concurrent.futures.Future):
	assert fut_.done()
	self._running_cb_futs.remove(fut_)
	if fut_.cancelled():
	self.logger.debug(f"cb cancelled. {event=}.")
	elif exc := fut_.exception():
	self.logger.error(f"cb errored. {event=}. {exc=}", exc_info=exc)
	fut.add_done_callback(on_done)

Well, I don't know.
I had an earlier version of this commit where I undid that, yes.
But ultimately I don't know if run_coroutine_threadsafe is affected by this GC issue at all.
I tried to ask in python/cpython#88831 (comment), but never got a reply.

If it is affected, this current PR might not help it. When writing the unit test, I also tried to check if run_coroutine_threadsafe is going through the task factory and it seems it is not...

[SomberNight]

Anyway, I'll merge this. It cannot hurt.