lnprototest: updating and patch lnprototest

[vincenzopalazzo]

Build on top of #5367

The @niftynei patch finally fixed the lnprototest failure rustyrussell/lnprototest#31 and with the stress test that I added on lnprototest CI that ran lnprototest 20 times, I found another strange failure rustyrussell/lnprototest#56 that I'm no longer able to reproduce, so I added a couple of logging call.

In addition, I enable the CI lnprototest that run the test 2 time, in this way we can try to find the bug before we fall into the lucky case forever. if we fall in the lucky case 2 time? well, we are lucky no?

let's see what the full CI tells us now, and if we finally have the CI clean!

Fixes #5148

[vincenzopalazzo]

@rustyrussell There is some failure related to the gossip test, but this looks like that we can cover in lnprototest? or maybe we did already https://github.com/rustyrussell/lnprototest/blob/master/tests/test_bolt7-01-channel_announcement-success.py

_____________________________ test_node_reannounce _____________________________
[gw3] linux -- Python 3.7.13 /opt/hostedtoolcache/Python/3.7.13/x64/bin/python3
node_factory = <pyln.testing.utils.NodeFactory object at 0x7feabd907c90>
bitcoind = <pyln.testing.utils.BitcoinD object at 0x7feabd3f5c90>
chainparams = {'bip173_prefix': 'bcrt', 'chain_hash': '06226e46111a0b59caaf126043eb5bbf28c34f3a5e332a1fc7b2b73cf188910f', 'elements': False, 'example_addr': 'bcrt1qeyyk6sl5pr49ycpqyckvmttus5ttj25pd0zpvg', ...}
@pytest.mark.developer("Needs fast gossip propagation")
@pytest.mark.openchannel('v1')
@pytest.mark.openchannel('v2')
deftest_node_reannounce(node_factory, bitcoind, chainparams):
"Test that we reannounce a node when parameters change"
        l1, l2 = node_factory.line_graph(2, opts={'may_reconnect': True,
'log-level': 'io'})
        bitcoind.generate_block(5)
        genesis_blockhash = chainparams['chain_hash']
# Wait for node_announcement for l1.
        l2.daemon.wait_for_log(r'\[IN\] 0101.*{}'.format(l1.info['id']))
# Wait for it to process it.
        wait_for(lambda: l2.rpc.listnodes(l1.info['id'])['nodes'] != [])
        wait_for(lambda: 'alias'in only_one(l2.rpc.listnodes(l1.info['id'])['nodes']))
assert only_one(l2.rpc.listnodes(l1.info['id'])['nodes'])['alias'].startswith('JUNIORBEAM')
        lfeatures = expected_node_features()
if l1.config('experimental-dual-fund'):
            lfeatures = expected_node_features(extra=[21, 29])
# Make sure it gets features correct.
assert only_one(l2.rpc.listnodes(l1.info['id'])['nodes'])['features'] == lfeatures
        l1.stop()
        l1.daemon.opts['alias'] = 'SENIORBEAM'
# It won't update within 5 seconds, so sleep.
        time.sleep(5)
        l1.start()
        wait_for(lambda: only_one(l2.rpc.listnodes(l1.info['id'])['nodes'])['alias'] == 'SENIORBEAM')
# Get node_announcements.
        msgs = l1.query_gossip('gossip_timestamp_filter',
                               genesis_blockhash,
'0', '0xFFFFFFFF',
# Filter out gossip_timestamp_filter,
# channel_announcement and channel_updates.
# And pings.
                               filters=['0109', '0107', '0102', '0100', '0012'])
# May send its own announcement *twice*, since it always spams us.
        msgs = list(set(msgs))
assertlen(msgs) == 2
assert (bytes("SENIORBEAM", encoding="utf8").hex() in msgs[0]
orbytes("SENIORBEAM", encoding="utf8").hex() in msgs[1])
# Restart should re-xmit exact same update on reconnect!
        l1.restart()
        msgs2 = l1.query_gossip('gossip_timestamp_filter',
                                genesis_blockhash,
'0', '0xFFFFFFFF',
# Filter out gossip_timestamp_filter,
# channel_announcement and channel_updates.
# And pings.
                                filters=['0109', '0107', '0102', '0100', '0012'])
# May send its own announcement *twice*, since it always spams us.
        msgs2 = list(set(msgs2))
>       assert msgs == msgs2
E       AssertionError: assert ['010152cddd8...000000000000'] == ['010152c2743...000000000000']
E         At index 0 diff: '010152cddd83318c370a74d2a4c411f09574b8a80fa1148267727b2dd9b5dcc4d4096415329507fb40de3cd679306d2720c696cd2b622c54fb756656e5baa8da38590007800000080269a262c1fd73022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59022d2253494c454e544152544953542d643061313932312d6d6f6464656[4000](https://github.com/ElementsProject/lightning/runs/7171067324?check_suite_focus=true#step:5:4001)00000000000' != '010152c2743c00693c0923c0cf6d2036dede2011489b73b8f10e94599987689953061cae9e64086f103b881ee9023e54a0b159f1e114d6b226668f609531bd7cd9110007800000080269a262c1fd7c0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c035180266e453454e494f524245414d000000000000000000000000000000000000000000000000'
E         Full diff:
E           [
E         +  '010152cddd83318c370a74d2a4c411f09574b8a80fa1148267727b2dd9b5dcc4d4096415329507fb40de3cd679306d2720c696cd2b622c54fb756656e5baa8da38590007800000080269a262c1fd73022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59022d2253494c454e544152544953542d643061313932312d6d6f6464656400000000000000',
E            '010152c2743c00693c0923c0cf6d2036dede2011489b73b8f10e94599987689953061cae9e64086f103b881ee9023e54a0b159f1e114d6b226668f609531bd7cd9110007800000080269a262c1fd7c0266e4598d1d3c415f572a8488830b60f7e744ed9235eb0b1ba93283b315c035180266e453454e494f524245414d000000000000000000000000000000000000000000000000',
E         -  '010152cddd83318c370a74d2a4c411f09574b8a80fa1148267727b2dd9b5dcc4d4096415329507fb40de3cd679306d2720c696cd2b622c54fb756656e5baa8da38590007800000080269a262c1fd73022d223620a359a47ff7f7ac447c85c46c923da53389221a0054c11c1e3ca31d59022d2253494c454e544152544953542d643061313932312d6d6f6464656400000000000000',
E           ]
tests/test_gossip.py:1285: AssertionError

I'm pretty sure that if I restart it I will have the test happy, but maybe we can hide this failure

[rustyrussell]

I have a separate fix for this, so restarting...

[niftynei]

@vincenzopalazzo i removed your "run the lnprototests x2" patch in anticipation of getting this merged

[vincenzopalazzo]

ACK 6becf47

CI failure looks like related to ssh failure!

@niftynei thanks to start and finish this