Part 4 - asyncio: removes all deprecated patterns and adopts modern asyncio idioms.

[ping-ke]

Summary

Python 3.10 deprecated the loop= parameter in most asyncio APIs, and Python 3.12 removed it entirely. This removes all deprecated patterns across the codebase and adopts modern asyncio idioms.

Key changes:

• Remove loop= parameter from asyncio.Event, asyncio.wait, asyncio.ensure_future, asyncio.start_server, open_connection, etc.

• Replace asyncio.ensure_future() with asyncio.create_task()

• Replace get_event_loop().run_until_complete() with asyncio.run()

• Replace asyncio.get_event_loop() with asyncio.get_running_loop() where inside a running event loop

• Add _get_or_create_event_loop() helper in utils.py for call sites that run before the loop starts (SlaveServer, SlaveConnectionManager)

• Replace asyncio.Future used as a one-shot signal with asyncio.Event (active_future → active_event, ping_received_future → ping_received_event)

• Make start()/shutdown() methods async where they previously called loop.run_until_complete() internally

• Update eth_hash import: BasePreImage → PreImageAPI (eth-hash >=0.5 API)

• Update cryptography API: EllipticCurvePublicNumbers.from_encoded_point → EllipticCurvePublicKey.from_encoded_point

• Add conftest.py to cancel pending asyncio tasks between tests

• Replace assertRaisesRegexp (removed in Py 3.12) with assertRaisesRegex

  Test plan

  • pytest quarkchain/cluster/tests/ passes
  • pytest quarkchain/p2p/ passes
  • No DeprecationWarning about event loop parameters

[qzhodl]

Code review

Found 2 issues:

1. self.upnp_service.discover() does not exist on UPnPService — will raise AttributeError at runtime. The correct method is add_nat_portmap(), which was the original call. UPnPService in quarkchain/p2p/nat.py defines add_nat_portmap() (line 81) and _discover_upnp_devices() (private, line 153), but has no public discover() method.

pyquarkchain/quarkchain/p2p/p2p_server.py

Lines 104 to 106 in e720240

	if self.upnp_service:
	mapped_external_ip = await self.upnp_service.discover()
	external_ip = mapped_external_ip or "0.0.0.0"

2. asyncio.create_task() in Logger.send_log_to_kafka() requires a running event loop, but this is a sync @classmethod called from sync methods like Logger.error() and Logger.error_exception(). If logging is triggered outside an async context (e.g., during startup or shutdown), this will raise RuntimeError: no running event loop. The previous asyncio.ensure_future() had the same limitation in Python 3.10+, but this is worth noting as a potential runtime crash when Kafka logging is configured.

pyquarkchain/quarkchain/utils.py

Lines 383 to 388 in e720240

	}
	asyncio.create_task(
	cls._kafka_logger.log_kafka_sample_async(
	cls._kafka_logger.cluster_config.MONITORING.ERRORS, sample
	)
	)

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}

[ping-ke]

Code review

Found 2 issues:

1. self.upnp_service.discover() does not exist on UPnPService — will raise AttributeError at runtime. The correct method is add_nat_portmap(), which was the original call. UPnPService in quarkchain/p2p/nat.py defines add_nat_portmap() (line 81) and _discover_upnp_devices() (private, line 153), but has no public discover() method.

pyquarkchain/quarkchain/p2p/p2p_server.py

Lines 104 to 106 in e720240

	if self.upnp_service:
	mapped_external_ip = await self.upnp_service.discover()
	external_ip = mapped_external_ip or "0.0.0.0"

2. asyncio.create_task() in Logger.send_log_to_kafka() requires a running event loop, but this is a sync @classmethod called from sync methods like Logger.error() and Logger.error_exception(). If logging is triggered outside an async context (e.g., during startup or shutdown), this will raise RuntimeError: no running event loop. The previous asyncio.ensure_future() had the same limitation in Python 3.10+, but this is worth noting as a potential runtime crash when Kafka logging is configured.

pyquarkchain/quarkchain/utils.py

Lines 383 to 388 in e720240

	}
	asyncio.create_task(
	cls._kafka_logger.log_kafka_sample_async(
	cls._kafka_logger.cluster_config.MONITORING.ERRORS, sample
	)
	)

🤖 Generated with Claude Code

• If this code review was useful, please react with 👍. Otherwise, react with 👎.

For issue 1:
discover() has been defined in Part 6, move this change to Part 6 now.
For issue 2:
Fix it as following

      384 -            asyncio.create_task(
      384 +            try:
      385 +                loop = asyncio.get_running_loop()
      386 +            except RuntimeError:
      387 +                # No running event loop (e.g., during startup/shutdown).
      388 +                # Silently skip Kafka logging to avoid crashing the caller.
      389 +                return
      390 +            loop.create_task(
      391                  cls._kafka_logger.log_kafka_sample_async(
      392                      cls._kafka_logger.cluster_config.MONITORING.ERRORS, sample
      393                  )