def test_alice_can_grant_even_when_the_first_nodes_she_tries_are_down(federated_alice, federated_bob, federated_ursulas):
m, n = 2, 3
policy_end_datetime = maya.now() + datetime.timedelta(days=5)
label = b"this_is_the_path_to_which_access_is_being_granted"
federated_alice.known_nodes._nodes = {}
federated_alice.network_middleware = NodeIsDownMiddleware()
# OK, her first and only node is down.
down_node = list(federated_ursulas)[0]
federated_alice.remember_node(down_node)
federated_alice.network_middleware.node_is_down(down_node)
# Here's the command we want to run.
alice_grant_action = partial(federated_alice.grant, federated_bob, label, m=m, n=n, expiration=policy_end_datetime, timeout=.1)
# Try a first time, failing because no known nodes are up for Alice to even try to learn from.
with pytest.raises(down_node.NotEnoughNodes):
alice_grant_action()
# Now she learn about one node that *is* up...
reliable_node = list(federated_ursulas)[1]
federated_alice.remember_node(reliable_node)
# ...amidst a few others that are down.
more_nodes = list(federated_ursulas)[2:10]
# Alice still only knows aboot two nodes (the one that is down and the new one).
assert len(federated_alice.known_nodes) == 2
for node in more_nodes:
federated_alice.network_middleware.node_is_down(node)
# Alice can't verify enough nodes to complete the Policy.
with pytest.raises(Learner.NotEnoughNodes):
alice_grant_action()
# Now we'll have a situation where Alice knows about all 10,
# though only one is up.
# She'll try to learn about more, but there aren't any.
# Because she has successfully completed learning, but the nodes about which she learned are down,
# she'll get a different error.
for node in more_nodes:
federated_alice.remember_node(node)
with pytest.raises(Policy.MoreKFragsThanArrangements):
alice_grant_action()
# Now let's let a few of them come up.
for node in more_nodes[0:4]:
federated_alice.network_middleware.node_is_up(node)
# Now the same exact action works.
policy = alice_grant_action() # I'm a little surprised this doesn't fail once in a while, if more then 2 of the nodes selected are among those still down.
# The number of accepted arrangements at least the number of Ursulas we're using (n)
assert len(policy._accepted_arrangements) >= n
# The number of actually enacted arrangements is exactly equal to n.
assert len(policy._enacted_arrangements) == n
def test_node_has_changed_cert(federated_alice, federated_ursulas):
federated_alice.known_nodes._nodes = {}
federated_alice.network_middleware = NodeIsDownMiddleware()
federated_alice.network_middleware.client.certs_are_broken = True
firstula = list(federated_ursulas)[0]
federated_alice.remember_node(firstula)
federated_alice.start_learning_loop(now=True)
federated_alice.learn_from_teacher_node()
def test_bob_does_not_let_a_connection_error_stop_him(enacted_federated_policy, federated_ursulas, federated_bob,
federated_alice):
assert len(federated_bob.known_nodes) == 0
ursula1 = list(federated_ursulas)[0]
ursula2 = list(federated_ursulas)[1]
federated_bob.remember_node(ursula1)
federated_bob.network_middleware = NodeIsDownMiddleware()
federated_bob.network_middleware.node_is_down(ursula1)
with pytest.raises(TreasureMap.NowhereToBeFound):
federated_bob.get_treasure_map(federated_alice.stamp, enacted_federated_policy.label)
federated_bob.remember_node(ursula2)
map = federated_bob.get_treasure_map(federated_alice.stamp, enacted_federated_policy.label)
assert sorted(list(map.destinations.keys())) == sorted(
list(u.checksum_address for u in list(federated_ursulas)))
def test_blockchain_ursulas_reencrypt(blockchain_ursulas, blockchain_alice,
blockchain_bob, policy_value):
label = b'bbo'
# TODO: Make sample selection buffer configurable - #1061
m = n = 10
expiration = maya.now() + datetime.timedelta(days=5)
_policy = blockchain_alice.grant(bob=blockchain_bob,
label=label,
m=m,
n=n,
expiration=expiration,
value=policy_value)
enrico = Enrico.from_alice(blockchain_alice, label)
message = b"Oh, this isn't even BO. This is beyond BO. It's BBO."
message_kit, signature = enrico.encrypt_message(message)
blockchain_bob.join_policy(label, bytes(blockchain_alice.stamp))
plaintext = blockchain_bob.retrieve(
message_kit,
alice_verifying_key=blockchain_alice.stamp,
label=label,
enrico=enrico)
assert plaintext[0] == message
# Let's consider also that a node may be down when granting
blockchain_alice.network_middleware = NodeIsDownMiddleware()
blockchain_alice.network_middleware.node_is_down(blockchain_ursulas[0])
with pytest.raises(Policy.Rejected):
_policy = blockchain_alice.grant(bob=blockchain_bob,
label=b'another-label',
m=m,
n=n,
expiration=expiration,
value=policy_value)
def test_alice_can_grant_even_when_the_first_nodes_she_tries_are_down(
federated_alice, federated_bob, federated_ursulas):
m, n = 2, 3
policy_end_datetime = maya.now() + datetime.timedelta(days=5)
label = b"this_is_the_path_to_which_access_is_being_granted"
federated_alice.known_nodes._nodes = {}
federated_alice.network_middleware = NodeIsDownMiddleware()
# OK, her first and only node is down.
down_node = list(federated_ursulas)[0]
federated_alice.remember_node(down_node)
federated_alice.network_middleware.node_is_down(down_node)
# Here's the command we want to run.
alice_grant_action = partial(federated_alice.grant,
federated_bob,
label,
m=m,
n=n,
expiration=policy_end_datetime,
timeout=.1)
# Try a first time, failing because no known nodes are up for Alice to even try to learn from.
with pytest.raises(down_node.NotEnoughNodes):
alice_grant_action()
# Now she learn about one node that *is* up...
reliable_node = list(federated_ursulas)[1]
federated_alice.remember_node(reliable_node)
# ...amidst a few others that are down.
more_nodes = list(federated_ursulas)[2:10]
for node in more_nodes:
federated_alice.network_middleware.node_is_down(node)
# Alice still only knows about two nodes (the one that is down and the new one).
assert len(federated_alice.known_nodes) == 2
# Now we'll have a situation where Alice knows about all 10,
# though only one is up.
# She'll try to learn about more, but there aren't any.
# Because she has successfully completed learning, but the nodes about which she learned are down,
# she'll get a different error.
for node in more_nodes:
federated_alice.remember_node(node)
with pytest.raises(Policy.Rejected):
alice_grant_action()
# Now let's let a few of them come up.
for node in more_nodes[0:4]:
federated_alice.network_middleware.node_is_up(node)
# Now the same exact action works.
# TODO: This action only succeeds here because we are forcing
# grant to accept the ursulas that just came back online (handpicked_ursulas).
# Since there are now enough Ursulas online, this action *can* succeed without forcing sample.
policy = alice_grant_action(handpicked_ursulas=more_nodes[:3])
# TODO: This is how it's actually done. How can we measure such random results?
# The below line will fail with ? probability, if more then 2 of the nodes selected
# are among those still down.
# policy = alice_grant_action()
# The number of accepted arrangements at least the number of Ursulas we're using (n)
assert len(policy._accepted_arrangements) >= n
# The number of actually enacted arrangements is exactly equal to n.
assert len(policy._enacted_arrangements) == n