pytest: fix flake in test_onchain_reestablish_reply

We can ask for the state too fast, before WIRE_ERROR is processed:

```
2025-01-24T14:17:57.4799255Z ________________________ test_onchain_reestablish_reply ________________________
2025-01-24T14:17:57.4800383Z [gw2] linux -- Python 3.10.16 /home/runner/.cache/pypoetry/virtualenvs/cln-meta-project-AqJ9wMix-py3.10/bin/python
2025-01-24T14:17:57.4801246Z 
2025-01-24T14:17:57.4802068Z node_factory = <pyln.testing.utils.NodeFactory object at 0x7f66ec144400>
2025-01-24T14:17:57.4802645Z bitcoind = <pyln.testing.utils.BitcoinD object at 0x7f66ec145ab0>
2025-01-24T14:17:57.4803483Z executor = <concurrent.futures.thread.ThreadPoolExecutor object at 0x7f66ec146da0>
2025-01-24T14:17:57.4804029Z 
2025-01-24T14:17:57.4804340Z     def test_onchain_reestablish_reply(node_factory, bitcoind, executor):
2025-01-24T14:17:57.4805013Z         l1, l2, l3 = node_factory.line_graph(3, opts={'may_reconnect': True,
2025-01-24T14:17:57.4805572Z                                                       'dev-no-reconnect': None})
2025-01-24T14:17:57.4805990Z     
2025-01-24T14:17:57.4806262Z         # Make l1 close unilaterally.
2025-01-24T14:17:57.4806697Z         l1.rpc.disconnect(l2.info['id'], force=True)
2025-01-24T14:17:57.4807200Z         l1.rpc.close(l2.info['id'], unilateraltimeout=1)
2025-01-24T14:17:57.4807622Z     
2025-01-24T14:17:57.4808025Z         # l2 doesn't know, reconnection should get REESTABLISH *then* error.
2025-01-24T14:17:57.4808637Z         l2.rpc.connect(l1.info['id'], 'localhost', l1.port)
2025-01-24T14:17:57.4809066Z     
2025-01-24T14:17:57.4809334Z         # We should exchange messages
2025-01-24T14:17:57.4809835Z         l2.daemon.wait_for_logs(["peer_out WIRE_CHANNEL_REESTABLISH",
2025-01-24T14:17:57.4810399Z                                  "peer_in WIRE_CHANNEL_REESTABLISH"])
2025-01-24T14:17:57.4811197Z         # It should be OK
2025-01-24T14:17:57.4811642Z         l2.daemon.wait_for_log("Reconnected, and reestablished.")
2025-01-24T14:17:57.4812116Z     
2025-01-24T14:17:57.4812378Z         # Then we get the error, close.
2025-01-24T14:17:57.4812824Z         l2.daemon.wait_for_log("peer_in WIRE_ERROR")
2025-01-24T14:17:57.4813576Z >       assert only_one(l2.rpc.listpeerchannels(l1.info['id'])['channels'])['state'] == 'AWAITING_UNILATERAL'
2025-01-24T14:17:57.4814446Z E       AssertionError: assert 'CHANNELD_NORMAL' == 'AWAITING_UNILATERAL'
2025-01-24T14:17:57.4815014Z E         - AWAITING_UNILATERAL
2025-01-24T14:17:57.4815362Z E         + CHANNELD_NORMAL
```

```
2025-01-24T14:17:57.5760435Z lightningd-2 2025-01-24T14:04:28.398Z DEBUG   0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-channeld-chan#1: peer_in WIRE_ERROR
2025-01-24T14:17:57.5760674Z lightningd-2 2025-01-24T14:04:28.401Z TRACE   lightningd: Calling rpc_command hook of plugin cln-xpay
2025-01-24T14:17:57.5760935Z lightningd-2 2025-01-24T14:04:28.403Z TRACE   lightningd: Plugin cln-xpay returned from rpc_command hook call
2025-01-24T14:17:57.5761176Z lightningd-2 2025-01-24T14:04:28.408Z TRACE   lightningd: Calling rpc_command hook of plugin cln-xpay
2025-01-24T14:17:57.5761430Z lightningd-2 2025-01-24T14:04:28.414Z TRACE   lightningd: Plugin cln-xpay returned from rpc_command hook call
2025-01-24T14:17:57.5762598Z {'github_repository': 'ElementsProject/lightning', 'github_sha': '390503b31e1949cb84e5a5257bf29a1b80d5cd20', 'github_ref': 'refs/pull/8027/merge', 'github_ref_name': 'HEAD', 'github_run_id': 12950414032, 'github_head_ref': 'guilt/fix-flakes12', 'github_run_number': 12051, 'github_base_ref': 'master', 'github_run_attempt': '1', 'testname': 'test_onchain_reestablish_reply', 'start_time': 1737727408, 'end_time': 1737727468, 'outcome': 'fail'}
2025-01-24T14:17:57.5763430Z lightningd-2 2025-01-24T14:04:28.426Z DEBUG   0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-channeld-chan#1: billboard perm: Received ERROR channel 69d744488de7ec7f4c018da98a8a9d726ee3b47a97809608e4ce6f17747b5f09: Forcibly closed by `close` command timeout
2025-01-24T14:17:57.5764512Z lightningd-2 2025-01-24T14:04:28.433Z UNUSUAL 0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-chan#1: Peer permanent failure in CHANNELD_NORMAL: channeld: received ERROR channel 69d744488de7ec7f4c018da98a8a9d726ee3b47a97809608e4ce6f17747b5f09: Forcibly closed by `close` command timeout (reason=protocol)
2025-01-24T14:17:57.5764997Z lightningd-2 2025-01-24T14:04:28.439Z INFO    0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c03518-chan#1: State changed from CHANNELD_NORMAL to AWAITING_UNILATERAL
```

Signed-off-by: Rusty Russell <[email protected]>

Author: rustyrussell

tests/test_closing.py

@@ -4168,7 +4168,7 @@ def test_onchain_reestablish_reply(node_factory, bitcoind, executor):
 
     # Then we get the error, close.
     l2.daemon.wait_for_log("peer_in WIRE_ERROR")
-    assert only_one(l2.rpc.listpeerchannels(l1.info['id'])['channels'])['state'] == 'AWAITING_UNILATERAL'
+    wait_for(lambda: only_one(l2.rpc.listpeerchannels(l1.info['id'])['channels'])['state'] == 'AWAITING_UNILATERAL')
     # Mine it now so we don't confuse the code below.
     bitcoind.generate_block(1, wait_for_mempool=1)