dash/test/functional/feature_block.py

1333 lines
60 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
# Copyright (c) 2015-2016 The Bitcoin Core developers
# Distributed under the MIT software license, see the accompanying
2015-08-05 23:47:34 +02:00
# file COPYING or http://www.opensource.org/licenses/mit-license.php.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
"""Test block processing."""
import copy
import struct
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
from test_framework.blocktools import create_block, create_coinbase, create_transaction, get_legacy_sigopcount_block
2015-08-05 23:47:34 +02:00
from test_framework.key import CECKey
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
from test_framework.mininode import (
CBlock,
COIN,
COutPoint,
CTransaction,
CTxIn,
CTxOut,
MAX_BLOCK_SIZE,
uint256_from_compact,
uint256_from_str,
)
from test_framework.mininode import P2PDataStore
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
from test_framework.script import (
CScript,
MAX_SCRIPT_ELEMENT_SIZE,
OP_2DUP,
OP_CHECKMULTISIG,
OP_CHECKMULTISIGVERIFY,
OP_CHECKSIG,
OP_CHECKSIGVERIFY,
OP_ELSE,
OP_ENDIF,
OP_EQUAL,
OP_DROP,
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
OP_FALSE,
OP_HASH160,
OP_IF,
OP_INVALIDOPCODE,
OP_RETURN,
OP_TRUE,
SIGHASH_ALL,
SignatureHash,
hash160,
)
from test_framework.test_framework import BitcoinTestFramework
from test_framework.util import assert_equal
MAX_BLOCK_SIGOPS = 20000
2015-08-05 23:47:34 +02:00
class PreviousSpendableOutput():
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
def __init__(self, tx=CTransaction(), n=-1):
2015-08-05 23:47:34 +02:00
self.tx = tx
self.n = n # the output we're spending
# Use this class for tests that require behavior other than normal "mininode" behavior.
# For now, it is used to serialize a bloated varint (b64).
class CBrokenBlock(CBlock):
def initialize(self, base_block):
self.vtx = copy.deepcopy(base_block.vtx)
self.hashMerkleRoot = self.calc_merkle_root()
def serialize(self):
r = b""
r += super(CBlock, self).serialize()
r += struct.pack("<BQ", 255, len(self.vtx))
for tx in self.vtx:
r += tx.serialize()
return r
def normal_serialize(self):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
return super().serialize()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
class FullBlockTest(BitcoinTestFramework):
def set_test_params(self):
2015-08-05 23:47:34 +02:00
self.num_nodes = 1
self.setup_clean_chain = True
# Must set '-dip3params=2000:2000' to create pre-dip3 blocks only
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.extra_args = [['-dip3params=2000:2000']]
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
def setup_nodes(self):
# Very large reorgs cause cs_main to be held for a very long time in ActivateBestChainStep,
# which causes RPC to hang, so we need to increase RPC timeouts
# TODO remove this when bitcoin#13837 gets backported and change it to use self.rpc_timeout
self.add_nodes(self.num_nodes, self.extra_args, timewait=180)
self.start_nodes()
2015-08-05 23:47:34 +02:00
def run_test(self):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
node = self.nodes[0] # convenience reference to the node
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.bootstrap_p2p() # Add one p2p connection to the node
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.block_heights = {}
self.coinbase_key = CECKey()
self.coinbase_key.set_secretbytes(b"horsebattery")
self.coinbase_pubkey = self.coinbase_key.get_pubkey()
self.tip = None
self.blocks = {}
2015-08-05 23:47:34 +02:00
self.genesis_hash = int(self.nodes[0].getbestblockhash(), 16)
self.block_heights[self.genesis_hash] = 0
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.spendable_outputs = []
2015-08-05 23:47:34 +02:00
# Create a new block
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b0 = self.next_block(0)
self.save_spendable_output()
self.send_blocks([b0])
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# Allow the block to mature
blocks = []
for i in range(99):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
blocks.append(self.next_block(5000 + i))
self.save_spendable_output()
self.send_blocks(blocks)
2015-08-05 23:47:34 +02:00
# collect spendable outputs now to avoid cluttering the code later on
out = []
for i in range(33):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
out.append(self.get_spendable_output())
2015-08-05 23:47:34 +02:00
# Start by building a couple of blocks on top (which output is spent is
# in parentheses):
2015-08-05 23:47:34 +02:00
# genesis -> b1 (0) -> b2 (1)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b1 = self.next_block(1, spend=out[0])
self.save_spendable_output()
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b2 = self.next_block(2, spend=out[1])
self.save_spendable_output()
2015-08-05 23:47:34 +02:00
self.send_blocks([b1, b2])
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# Fork like this:
#
2015-08-05 23:47:34 +02:00
# genesis -> b1 (0) -> b2 (1)
# \-> b3 (1)
#
2015-08-05 23:47:34 +02:00
# Nothing should happen at this point. We saw b2 first so it takes priority.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Don't reorg to a chain of the same length")
self.move_tip(1)
b3 = self.next_block(3, spend=out[1])
txout_b3 = PreviousSpendableOutput(b3.vtx[1], 0)
self.send_blocks([b3], False)
2015-08-05 23:47:34 +02:00
# Now we add another block to make the alternative chain longer.
#
2015-08-05 23:47:34 +02:00
# genesis -> b1 (0) -> b2 (1)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reorg to a longer chain")
b4 = self.next_block(4, spend=out[2])
self.send_blocks([b4])
2015-08-05 23:47:34 +02:00
# ... and back to the first chain.
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(2)
b5 = self.next_block(5, spend=out[2])
self.save_spendable_output()
self.send_blocks([b5], False)
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reorg back to the original chain")
b6 = self.next_block(6, spend=out[3])
self.send_blocks([b6], True)
2015-08-05 23:47:34 +02:00
# Try to create a fork that double-spends
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b7 (2) -> b8 (4)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a chain with a double spend, even if it is longer")
self.move_tip(5)
b7 = self.next_block(7, spend=out[2])
self.send_blocks([b7], False)
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b8 = self.next_block(8, spend=out[4])
self.send_blocks([b8], False, reconnect=True)
2015-08-05 23:47:34 +02:00
# Try to create a block that has too much fee
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b9 (4)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block where the miner creates too much coinbase reward")
self.move_tip(6)
b9 = self.next_block(9, spend=out[4], additional_coinbase_value=1)
self.send_blocks([b9], False, 16, b'bad-cb-amount', reconnect=False)
2015-08-05 23:47:34 +02:00
# Create a fork that ends in a block with too much fee (the one that causes the reorg)
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b10 (3) -> b11 (4)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a chain where the miner creates too much coinbase reward, even if the chain is longer")
self.move_tip(5)
b10 = self.next_block(10, spend=out[3])
self.send_blocks([b10], False)
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b11 = self.next_block(11, spend=out[4], additional_coinbase_value=1)
self.send_blocks([b11], False, 16, b'bad-cb-amount', reconnect=False)
2015-08-05 23:47:34 +02:00
# Try again, but with a valid fork first
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b14 (5)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a chain where the miner creates too much coinbase reward, even if the chain is longer (on a forked chain)")
self.move_tip(5)
b12 = self.next_block(12, spend=out[3])
self.save_spendable_output()
b13 = self.next_block(13, spend=out[4])
self.save_spendable_output()
b14 = self.next_block(14, spend=out[5], additional_coinbase_value=1)
self.send_blocks([b12, b13, b14], False, 16, b'bad-cb-amount', reconnect=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# New tip should be b13.
assert_equal(node.getbestblockhash(), b13.hash)
2015-08-05 23:47:34 +02:00
# Add a block with MAX_BLOCK_SIGOPS and one with one more sigop
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5) -> b16 (6)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with lots of checksigs")
lots_of_checksigs = CScript([OP_CHECKSIG] * (MAX_BLOCK_SIGOPS - 1))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(13)
b15 = self.next_block(15, spend=out[5], script=lots_of_checksigs)
self.save_spendable_output()
self.send_blocks([b15], True)
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with too many checksigs")
too_many_checksigs = CScript([OP_CHECKSIG] * (MAX_BLOCK_SIGOPS))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b16 = self.next_block(16, spend=out[6], script=too_many_checksigs)
self.send_blocks([b16], False, 16, b'bad-blk-sigops', reconnect=True)
# Attempt to spend a transaction created on a different fork
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5) -> b17 (b3.vtx[1])
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a spend from a re-org'ed out tx")
self.move_tip(15)
b17 = self.next_block(17, spend=txout_b3)
self.send_blocks([b17], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
# Attempt to spend a transaction created on a different fork (on a fork this time)
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5)
# \-> b18 (b3.vtx[1]) -> b19 (6)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a spend from a re-org'ed out tx (on a forked chain)")
self.move_tip(13)
b18 = self.next_block(18, spend=txout_b3)
self.send_blocks([b18], False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b19 = self.next_block(19, spend=out[6])
self.send_blocks([b19], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
2015-08-05 23:47:34 +02:00
# Attempt to spend a coinbase at depth too low
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5) -> b20 (7)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block spending an immature coinbase.")
self.move_tip(15)
b20 = self.next_block(20, spend=out[7])
self.send_blocks([b20], False, 16, b'bad-txns-premature-spend-of-coinbase')
# Attempt to spend a coinbase at depth too low (on a fork this time)
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5)
# \-> b21 (6) -> b22 (5)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block spending an immature coinbase (on a forked chain)")
self.move_tip(13)
b21 = self.next_block(21, spend=out[6])
self.send_blocks([b21], False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b22 = self.next_block(22, spend=out[5])
self.send_blocks([b22], False, 16, b'bad-txns-premature-spend-of-coinbase')
# Create a block on either side of MAX_BLOCK_SIZE and make sure its accepted/rejected
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5) -> b23 (6)
# \-> b24 (6) -> b25 (7)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block of size MAX_BLOCK_SIZE")
self.move_tip(15)
b23 = self.next_block(23, spend=out[6])
tx = CTransaction()
script_length = MAX_BLOCK_SIZE - len(b23.serialize()) - 69
script_output = CScript([b'\x00' * script_length])
tx.vout.append(CTxOut(0, script_output))
tx.vin.append(CTxIn(COutPoint(b23.vtx[1].sha256, 0)))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b23 = self.update_block(23, [tx])
# Make sure the math above worked out to produce a max-sized block
assert_equal(len(b23.serialize()), MAX_BLOCK_SIZE)
self.send_blocks([b23], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block of size MAX_BLOCK_SIZE + 1")
self.move_tip(15)
b24 = self.next_block(24, spend=out[6])
script_length = MAX_BLOCK_SIZE - len(b24.serialize()) - 69
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
script_output = CScript([b'\x00' * (script_length + 1)])
tx.vout = [CTxOut(0, script_output)]
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b24 = self.update_block(24, [tx])
assert_equal(len(b24.serialize()), MAX_BLOCK_SIZE + 1)
self.send_blocks([b24], False, 16, b'bad-blk-length', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b25 = self.next_block(25, spend=out[7])
self.send_blocks([b25], False, request_block=False, reconnect=True)
# Create blocks with a coinbase input script size out of range
# genesis -> b1 (0) -> b2 (1) -> b5 (2) -> b6 (3)
# \-> b12 (3) -> b13 (4) -> b15 (5) -> b23 (6) -> b30 (7)
# \-> ... (6) -> ... (7)
# \-> b3 (1) -> b4 (2)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with coinbase input script size out of range")
self.move_tip(15)
b26 = self.next_block(26, spend=out[6])
b26.vtx[0].vin[0].scriptSig = b'\x00'
b26.vtx[0].rehash()
# update_block causes the merkle root to get updated, even with no new
# transactions, and updates the required state.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b26 = self.update_block(26, [])
self.send_blocks([b26], False, 16, b'bad-cb-length', reconnect=True)
# Extend the b26 chain to make sure dashd isn't accepting b26
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b27 = self.next_block(27, spend=out[7])
self.send_blocks([b27], False)
# Now try a too-large-coinbase script
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(15)
b28 = self.next_block(28, spend=out[6])
b28.vtx[0].vin[0].scriptSig = b'\x00' * 101
b28.vtx[0].rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b28 = self.update_block(28, [])
self.send_blocks([b28], False, 16, b'bad-cb-length', reconnect=True)
# Extend the b28 chain to make sure dashd isn't accepting b28
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b29 = self.next_block(29, spend=out[7])
self.send_blocks([b29], False)
# b30 has a max-sized coinbase scriptSig.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(23)
b30 = self.next_block(30)
b30.vtx[0].vin[0].scriptSig = b'\x00' * 100
b30.vtx[0].rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b30 = self.update_block(30, [])
self.send_blocks([b30], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# b31 - b35 - check sigops of OP_CHECKMULTISIG / OP_CHECKMULTISIGVERIFY / OP_CHECKSIGVERIFY
#
# genesis -> ... -> b30 (7) -> b31 (8) -> b33 (9) -> b35 (10)
# \-> b36 (11)
# \-> b34 (10)
# \-> b32 (9)
#
# MULTISIG: each op code counts as 20 sigops. To create the edge case, pack another 19 sigops at the end.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with the max number of OP_CHECKMULTISIG sigops")
lots_of_multisigs = CScript([OP_CHECKMULTISIG] * ((MAX_BLOCK_SIGOPS - 1) // 20) + [OP_CHECKSIG] * 19)
b31 = self.next_block(31, spend=out[8], script=lots_of_multisigs)
assert_equal(get_legacy_sigopcount_block(b31), MAX_BLOCK_SIGOPS)
self.send_blocks([b31], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# this goes over the limit because the coinbase has one sigop
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with too many OP_CHECKMULTISIG sigops")
too_many_multisigs = CScript([OP_CHECKMULTISIG] * (MAX_BLOCK_SIGOPS // 20))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b32 = self.next_block(32, spend=out[9], script=too_many_multisigs)
assert_equal(get_legacy_sigopcount_block(b32), MAX_BLOCK_SIGOPS + 1)
self.send_blocks([b32], False, 16, b'bad-blk-sigops', reconnect=True)
# CHECKMULTISIGVERIFY
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with the max number of OP_CHECKMULTISIGVERIFY sigops")
self.move_tip(31)
lots_of_multisigs = CScript([OP_CHECKMULTISIGVERIFY] * ((MAX_BLOCK_SIGOPS - 1) // 20) + [OP_CHECKSIG] * 19)
b33 = self.next_block(33, spend=out[9], script=lots_of_multisigs)
self.send_blocks([b33], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
self.log.info("Reject a block with too many OP_CHECKMULTISIGVERIFY sigops")
too_many_multisigs = CScript([OP_CHECKMULTISIGVERIFY] * (MAX_BLOCK_SIGOPS // 20))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b34 = self.next_block(34, spend=out[10], script=too_many_multisigs)
self.send_blocks([b34], False, 16, b'bad-blk-sigops', reconnect=True)
# CHECKSIGVERIFY
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with the max number of OP_CHECKSIGVERIFY sigops")
self.move_tip(33)
lots_of_checksigs = CScript([OP_CHECKSIGVERIFY] * (MAX_BLOCK_SIGOPS - 1))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b35 = self.next_block(35, spend=out[10], script=lots_of_checksigs)
self.send_blocks([b35], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with too many OP_CHECKSIGVERIFY sigops")
too_many_checksigs = CScript([OP_CHECKSIGVERIFY] * (MAX_BLOCK_SIGOPS))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b36 = self.next_block(36, spend=out[11], script=too_many_checksigs)
self.send_blocks([b36], False, 16, b'bad-blk-sigops', reconnect=True)
# Check spending of a transaction in a block which failed to connect
#
# b6 (3)
# b12 (3) -> b13 (4) -> b15 (5) -> b23 (6) -> b30 (7) -> b31 (8) -> b33 (9) -> b35 (10)
# \-> b37 (11)
# \-> b38 (11/37)
#
# save 37's spendable output, but then double-spend out11 to invalidate the block
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block spending transaction from a block which failed to connect")
self.move_tip(35)
b37 = self.next_block(37, spend=out[11])
txout_b37 = PreviousSpendableOutput(b37.vtx[1], 0)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx = self.create_and_sign_transaction(out[11].tx, out[11].n, 0)
b37 = self.update_block(37, [tx])
self.send_blocks([b37], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
# attempt to spend b37's first non-coinbase tx, at which point b37 was still considered valid
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(35)
b38 = self.next_block(38, spend=txout_b37)
self.send_blocks([b38], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
# Check P2SH SigOp counting
#
#
# 13 (4) -> b15 (5) -> b23 (6) -> b30 (7) -> b31 (8) -> b33 (9) -> b35 (10) -> b39 (11) -> b41 (12)
# \-> b40 (12)
#
# b39 - create some P2SH outputs that will require 6 sigops to spend:
#
# redeem_script = COINBASE_PUBKEY, (OP_2DUP+OP_CHECKSIGVERIFY) * 5, OP_CHECKSIG
# p2sh_script = OP_HASH160, ripemd160(sha256(script)), OP_EQUAL
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Check P2SH SIGOPS are correctly counted")
self.move_tip(35)
b39 = self.next_block(39)
b39_outputs = 0
b39_sigops_per_output = 6
# Build the redeem script, hash it, use hash to create the p2sh script
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
redeem_script = CScript([self.coinbase_pubkey] + [OP_2DUP, OP_CHECKSIGVERIFY] * 5 + [OP_CHECKSIG])
redeem_script_hash = hash160(redeem_script)
p2sh_script = CScript([OP_HASH160, redeem_script_hash, OP_EQUAL])
# Create a transaction that spends one satoshi to the p2sh_script, the rest to OP_TRUE
# This must be signed because it is spending a coinbase
spend = out[11]
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx = self.create_tx(spend.tx, spend.n, 1, p2sh_script)
tx.vout.append(CTxOut(spend.tx.vout[spend.n].nValue - 1, CScript([OP_TRUE])))
self.sign_tx(tx, spend.tx, spend.n)
tx.rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b39 = self.update_block(39, [tx])
b39_outputs += 1
# Until block is full, add tx's with 1 satoshi to p2sh_script, the rest to OP_TRUE
tx_new = None
tx_last = tx
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
total_size = len(b39.serialize())
while(total_size < MAX_BLOCK_SIZE):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx_new = self.create_tx(tx_last, 1, 1, p2sh_script)
tx_new.vout.append(CTxOut(tx_last.vout[1].nValue - 1, CScript([OP_TRUE])))
tx_new.rehash()
total_size += len(tx_new.serialize())
if total_size >= MAX_BLOCK_SIZE:
break
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b39.vtx.append(tx_new) # add tx to block
tx_last = tx_new
b39_outputs += 1
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b39 = self.update_block(39, [])
self.send_blocks([b39], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Test sigops in P2SH redeem scripts
#
# b40 creates 3333 tx's spending the 6-sigop P2SH outputs from b39 for a total of 19998 sigops.
# The first tx has one sigop and then at the end we add 2 more to put us just over the max.
#
# b41 does the same, less one, so it has the maximum sigops permitted.
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with too many P2SH sigops")
self.move_tip(39)
b40 = self.next_block(40, spend=out[12])
sigops = get_legacy_sigopcount_block(b40)
numTxes = (MAX_BLOCK_SIGOPS - sigops) // b39_sigops_per_output
assert_equal(numTxes <= b39_outputs, True)
lastOutpoint = COutPoint(b40.vtx[1].sha256, 0)
new_txs = []
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
for i in range(1, numTxes + 1):
tx = CTransaction()
tx.vout.append(CTxOut(1, CScript([OP_TRUE])))
tx.vin.append(CTxIn(lastOutpoint, b''))
# second input is corresponding P2SH output from b39
tx.vin.append(CTxIn(COutPoint(b39.vtx[i].sha256, 0), b''))
# Note: must pass the redeem_script (not p2sh_script) to the signature hash function
(sighash, err) = SignatureHash(redeem_script, tx, 1, SIGHASH_ALL)
sig = self.coinbase_key.sign(sighash) + bytes(bytearray([SIGHASH_ALL]))
scriptSig = CScript([sig, redeem_script])
tx.vin[1].scriptSig = scriptSig
tx.rehash()
new_txs.append(tx)
lastOutpoint = COutPoint(tx.sha256, 0)
b40_sigops_to_fill = MAX_BLOCK_SIGOPS - (numTxes * b39_sigops_per_output + sigops) + 1
tx = CTransaction()
tx.vin.append(CTxIn(lastOutpoint, b''))
tx.vout.append(CTxOut(1, CScript([OP_CHECKSIG] * b40_sigops_to_fill)))
tx.rehash()
new_txs.append(tx)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.update_block(40, new_txs)
self.send_blocks([b40], False, 16, b'bad-blk-sigops', reconnect=True)
# same as b40, but one less sigop
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with the max number of P2SH sigops")
self.move_tip(39)
b41 = self.next_block(41, spend=None)
self.update_block(41, b40.vtx[1:-1])
b41_sigops_to_fill = b40_sigops_to_fill - 1
tx = CTransaction()
tx.vin.append(CTxIn(lastOutpoint, b''))
tx.vout.append(CTxOut(1, CScript([OP_CHECKSIG] * b41_sigops_to_fill)))
tx.rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.update_block(41, [tx])
self.send_blocks([b41], True)
# Fork off of b39 to create a constant base again
#
# b23 (6) -> b30 (7) -> b31 (8) -> b33 (9) -> b35 (10) -> b39 (11) -> b42 (12) -> b43 (13)
# \-> b41 (12)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(39)
b42 = self.next_block(42, spend=out[12])
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b43 = self.next_block(43, spend=out[13])
self.save_spendable_output()
self.send_blocks([b42, b43], True)
# Test a number of really invalid scenarios
#
# -> b31 (8) -> b33 (9) -> b35 (10) -> b39 (11) -> b42 (12) -> b43 (13) -> b44 (14)
# \-> ??? (15)
# The next few blocks are going to be created "by hand" since they'll do funky things, such as having
# the first transaction be non-coinbase, etc. The purpose of b44 is to make sure this works.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Build block 44 manually")
height = self.block_heights[self.tip.sha256] + 1
coinbase = create_coinbase(height, self.coinbase_pubkey)
b44 = CBlock()
b44.nTime = self.tip.nTime + 1
b44.hashPrevBlock = self.tip.sha256
b44.nBits = 0x207fffff
b44.vtx.append(coinbase)
b44.hashMerkleRoot = b44.calc_merkle_root()
b44.solve()
self.tip = b44
self.block_heights[b44.sha256] = height
self.blocks[44] = b44
self.send_blocks([b44], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a non-coinbase as the first tx")
non_coinbase = self.create_tx(out[15].tx, out[15].n, 1)
b45 = CBlock()
b45.nTime = self.tip.nTime + 1
b45.hashPrevBlock = self.tip.sha256
b45.nBits = 0x207fffff
b45.vtx.append(non_coinbase)
b45.hashMerkleRoot = b45.calc_merkle_root()
b45.calc_sha256()
b45.solve()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.block_heights[b45.sha256] = self.block_heights[self.tip.sha256] + 1
self.tip = b45
self.blocks[45] = b45
self.send_blocks([b45], False, 16, b'bad-cb-missing', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with no transactions")
self.move_tip(44)
b46 = CBlock()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b46.nTime = b44.nTime + 1
b46.hashPrevBlock = b44.sha256
b46.nBits = 0x207fffff
b46.vtx = []
b46.hashMerkleRoot = 0
b46.solve()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.block_heights[b46.sha256] = self.block_heights[b44.sha256] + 1
self.tip = b46
assert 46 not in self.blocks
self.blocks[46] = b46
self.send_blocks([b46], False, 16, b'bad-blk-length', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with invalid work")
self.move_tip(44)
b47 = self.next_block(47, solve=False)
target = uint256_from_compact(b47.nBits)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
while b47.sha256 < target:
b47.nNonce += 1
b47.rehash()
self.send_blocks([b47], False, request_block=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a timestamp >2 hours in the future")
self.move_tip(44)
b48 = self.next_block(48, solve=False)
b48.nTime = int(self.mocktime) + 60 * 60 * 3
b48.solve()
self.send_blocks([b48], False, request_block=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with invalid merkle hash")
self.move_tip(44)
b49 = self.next_block(49)
b49.hashMerkleRoot += 1
b49.solve()
self.send_blocks([b49], False, 16, b'bad-txnmrklroot', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with incorrect POW limit")
self.move_tip(44)
b50 = self.next_block(50)
b50.nBits = b50.nBits - 1
b50.solve()
self.send_blocks([b50], False, request_block=False, reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with two coinbase transactions")
self.move_tip(44)
b51 = self.next_block(51)
cb2 = create_coinbase(51, self.coinbase_pubkey)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b51 = self.update_block(51, [cb2])
self.send_blocks([b51], False, 16, b'bad-cb-multiple', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with duplicate transactions")
# Note: txns have to be in the right position in the merkle tree to trigger this error
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(44)
b52 = self.next_block(52, spend=out[15])
tx = self.create_tx(b52.vtx[1], 0, 1)
b52 = self.update_block(52, [tx, tx])
self.send_blocks([b52], False, 16, b'bad-txns-duplicate', reconnect=True)
# Test block timestamps
# -> b31 (8) -> b33 (9) -> b35 (10) -> b39 (11) -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15)
# \-> b54 (15)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(43)
b53 = self.next_block(53, spend=out[14])
self.send_blocks([b53], False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with timestamp before MedianTimePast")
b54 = self.next_block(54, spend=out[15])
b54.nTime = b35.nTime - 1
b54.solve()
self.send_blocks([b54], False, request_block=False)
# valid timestamp
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(53)
b55 = self.next_block(55, spend=out[15])
b55.nTime = b35.nTime
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.update_block(55, [])
self.send_blocks([b55], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# Test Merkle tree malleability
#
# -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57p2 (16)
# \-> b57 (16)
# \-> b56p2 (16)
# \-> b56 (16)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# Merkle tree malleability (CVE-2012-2459): repeating sequences of transactions in a block without
# affecting the merkle root of a block, while still invalidating it.
# See: src/consensus/merkle.h
#
# b57 has three txns: coinbase, tx, tx1. The merkle root computation will duplicate tx.
# Result: OK
#
# b56 copies b57 but duplicates tx1 and does not recalculate the block hash. So it has a valid merkle
# root but duplicate transactions.
# Result: Fails
#
# b57p2 has six transactions in its merkle tree:
# - coinbase, tx, tx1, tx2, tx3, tx4
# Merkle root calculation will duplicate as necessary.
# Result: OK.
#
# b56p2 copies b57p2 but adds both tx3 and tx4. The purpose of the test is to make sure the code catches
# duplicate txns that are not next to one another with the "bad-txns-duplicate" error (which indicates
# that the error was caught early, avoiding a DOS vulnerability.)
# b57 - a good block with 2 txs, don't submit until end
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(55)
b57 = self.next_block(57)
tx = self.create_and_sign_transaction(out[16].tx, out[16].n, 1)
tx1 = self.create_tx(tx, 0, 1)
b57 = self.update_block(57, [tx, tx1])
# b56 - copy b57, add a duplicate tx
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a duplicate transaction in the Merkle Tree (but with a valid Merkle Root)")
self.move_tip(55)
b56 = copy.deepcopy(b57)
self.blocks[56] = b56
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
assert_equal(len(b56.vtx), 3)
b56 = self.update_block(56, [tx1])
assert_equal(b56.hash, b57.hash)
self.send_blocks([b56], False, 16, b'bad-txns-duplicate', reconnect=True)
# b57p2 - a good block with 6 tx'es, don't submit until end
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(55)
b57p2 = self.next_block("57p2")
tx = self.create_and_sign_transaction(out[16].tx, out[16].n, 1)
tx1 = self.create_tx(tx, 0, 1)
tx2 = self.create_tx(tx1, 0, 1)
tx3 = self.create_tx(tx2, 0, 1)
tx4 = self.create_tx(tx3, 0, 1)
b57p2 = self.update_block("57p2", [tx, tx1, tx2, tx3, tx4])
# b56p2 - copy b57p2, duplicate two non-consecutive tx's
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with two duplicate transactions in the Merkle Tree (but with a valid Merkle Root)")
self.move_tip(55)
b56p2 = copy.deepcopy(b57p2)
self.blocks["b56p2"] = b56p2
assert_equal(b56p2.hash, b57p2.hash)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
assert_equal(len(b56p2.vtx), 6)
b56p2 = self.update_block("b56p2", [tx3, tx4])
self.send_blocks([b56p2], False, 16, b'bad-txns-duplicate', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip("57p2")
self.send_blocks([b57p2], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(57)
self.send_blocks([b57], False) # The tip is not updated because 57p2 seen first
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Test a few invalid tx types
#
# -> b35 (10) -> b39 (11) -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17)
# \-> ??? (17)
#
# tx with prevout.n out of range
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a transaction with prevout.n out of range")
self.move_tip(57)
b58 = self.next_block(58, spend=out[17])
tx = CTransaction()
assert(len(out[17].tx.vout) < 42)
tx.vin.append(CTxIn(COutPoint(out[17].tx.sha256, 42), CScript([OP_TRUE]), 0xffffffff))
tx.vout.append(CTxOut(0, b""))
tx.calc_sha256()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b58 = self.update_block(58, [tx])
self.send_blocks([b58], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# tx with output value > input value
self.log.info("Reject a block with a transaction with outputs > inputs")
self.move_tip(57)
b59 = self.next_block(59)
tx = self.create_and_sign_transaction(out[17].tx, out[17].n, 510 * COIN)
b59 = self.update_block(59, [tx])
self.send_blocks([b59], False, 16, b'bad-txns-in-belowout', reconnect=True)
# reset to good chain
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(57)
b60 = self.next_block(60, spend=out[17])
self.send_blocks([b60], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Test BIP30
#
# -> b39 (11) -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17)
# \-> b61 (18)
#
# Blocks are not allowed to contain a transaction whose id matches that of an earlier,
# not-fully-spent transaction in the same chain. To test, make identical coinbases;
# the second one should be rejected.
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a transaction with a duplicate hash of a previous transaction (BIP30)")
self.move_tip(60)
b61 = self.next_block(61, spend=out[18])
b61.vtx[0].vin[0].scriptSig = b60.vtx[0].vin[0].scriptSig # Equalize the coinbases
b61.vtx[0].rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b61 = self.update_block(61, [])
assert_equal(b60.vtx[0].serialize(), b61.vtx[0].serialize())
self.send_blocks([b61], False, 16, b'bad-txns-BIP30', reconnect=True)
# Test tx.isFinal is properly rejected (not an exhaustive tx.isFinal test, that should be in data-driven transaction tests)
#
# -> b39 (11) -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17)
# \-> b62 (18)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a transaction with a nonfinal locktime")
self.move_tip(60)
b62 = self.next_block(62)
tx = CTransaction()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx.nLockTime = 0xffffffff # this locktime is non-final
assert(out[18].n < len(out[18].tx.vout))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx.vin.append(CTxIn(COutPoint(out[18].tx.sha256, out[18].n))) # don't set nSequence
tx.vout.append(CTxOut(0, CScript([OP_TRUE])))
assert(tx.vin[0].nSequence < 0xffffffff)
tx.calc_sha256()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b62 = self.update_block(62, [tx])
self.send_blocks([b62], False, 16, b'bad-txns-nonfinal')
# Test a non-final coinbase is also rejected
#
# -> b39 (11) -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17)
# \-> b63 (-)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a coinbase transaction with a nonfinal locktime")
self.move_tip(60)
b63 = self.next_block(63)
b63.vtx[0].nLockTime = 0xffffffff
b63.vtx[0].vin[0].nSequence = 0xDEADBEEF
b63.vtx[0].rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b63 = self.update_block(63, [])
self.send_blocks([b63], False, 16, b'bad-txns-nonfinal')
# This checks that a block with a bloated VARINT between the block_header and the array of tx such that
# the block is > MAX_BLOCK_SIZE with the bloated varint, but <= MAX_BLOCK_SIZE without the bloated varint,
# does not cause a subsequent, identical block with canonical encoding to be rejected. The test does not
# care whether the bloated block is accepted or rejected; it only cares that the second block is accepted.
#
# What matters is that the receiving node should not reject the bloated block, and then reject the canonical
# block on the basis that it's the same as an already-rejected block (which would be a consensus failure.)
#
# -> b39 (11) -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18)
# \
# b64a (18)
# b64a is a bloated block (non-canonical varint)
# b64 is a good block (same as b64 but w/ canonical varint)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a valid block even if a bloated version of the block has previously been sent")
self.move_tip(60)
regular_block = self.next_block("64a", spend=out[18])
# make it a "broken_block," with non-canonical serialization
b64a = CBrokenBlock(regular_block)
b64a.initialize(regular_block)
self.blocks["64a"] = b64a
self.tip = b64a
tx = CTransaction()
# use canonical serialization to calculate size
script_length = MAX_BLOCK_SIZE - len(b64a.normal_serialize()) - 69
script_output = CScript([b'\x00' * script_length])
tx.vout.append(CTxOut(0, script_output))
tx.vin.append(CTxIn(COutPoint(b64a.vtx[1].sha256, 0)))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b64a = self.update_block("64a", [tx])
assert_equal(len(b64a.serialize()), MAX_BLOCK_SIZE + 8)
self.send_blocks([b64a], False, 1, b'error parsing message')
# dashd doesn't disconnect us for sending a bloated block, but if we subsequently
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# resend the header message, it won't send us the getdata message again. Just
# disconnect and reconnect and then call sync_blocks.
# TODO: improve this test to be less dependent on P2P DOS behaviour.
node.disconnect_p2ps()
self.reconnect_p2p()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(60)
b64 = CBlock(b64a)
b64.vtx = copy.deepcopy(b64a.vtx)
assert_equal(b64.hash, b64a.hash)
assert_equal(len(b64.serialize()), MAX_BLOCK_SIZE)
self.blocks[64] = b64
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b64 = self.update_block(64, [])
self.send_blocks([b64], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Spend an output created in the block itself
#
# -> b42 (12) -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with a transaction spending an output created in the same block")
self.move_tip(64)
b65 = self.next_block(65)
tx1 = self.create_and_sign_transaction(out[19].tx, out[19].n, out[19].tx.vout[0].nValue)
tx2 = self.create_and_sign_transaction(tx1, 0, 0)
b65 = self.update_block(65, [tx1, tx2])
self.send_blocks([b65], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Attempt to spend an output created later in the same block
#
# -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19)
# \-> b66 (20)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a transaction spending an output created later in the same block")
self.move_tip(65)
b66 = self.next_block(66)
tx1 = self.create_and_sign_transaction(out[20].tx, out[20].n, out[20].tx.vout[0].nValue)
tx2 = self.create_and_sign_transaction(tx1, 0, 1)
b66 = self.update_block(66, [tx2, tx1])
self.send_blocks([b66], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
# Attempt to double-spend a transaction created in a block
#
# -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19)
# \-> b67 (20)
#
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block with a transaction double spending a transaction creted in the same block")
self.move_tip(65)
b67 = self.next_block(67)
tx1 = self.create_and_sign_transaction(out[20].tx, out[20].n, out[20].tx.vout[0].nValue)
tx2 = self.create_and_sign_transaction(tx1, 0, 1)
tx3 = self.create_and_sign_transaction(tx1, 0, 2)
b67 = self.update_block(67, [tx1, tx2, tx3])
self.send_blocks([b67], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
# More tests of block subsidy
#
# -> b43 (13) -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19) -> b69 (20)
# \-> b68 (20)
#
# b68 - coinbase with an extra 10 satoshis,
# creates a tx that has 9 satoshis from out[20] go to fees
# this fails because the coinbase is trying to claim 1 satoshi too much in fees
#
# b69 - coinbase with extra 10 satoshis, and a tx that gives a 10 satoshi fee
# this succeeds
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block trying to claim too much subsidy in the coinbase transaction")
self.move_tip(65)
b68 = self.next_block(68, additional_coinbase_value=10)
tx = self.create_and_sign_transaction(out[20].tx, out[20].n, out[20].tx.vout[0].nValue - 9)
b68 = self.update_block(68, [tx])
self.send_blocks([b68], False, 16, b'bad-cb-amount', reconnect=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block claiming the correct subsidy in the coinbase transaction")
self.move_tip(65)
b69 = self.next_block(69, additional_coinbase_value=10)
tx = self.create_and_sign_transaction(out[20].tx, out[20].n, out[20].tx.vout[0].nValue - 10)
self.update_block(69, [tx])
self.send_blocks([b69], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Test spending the outpoint of a non-existent transaction
#
# -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19) -> b69 (20)
# \-> b70 (21)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block containing a transaction spending from a non-existent input")
self.move_tip(69)
b70 = self.next_block(70, spend=out[21])
bogus_tx = CTransaction()
bogus_tx.sha256 = uint256_from_str(b"23c70ed7c0506e9178fc1a987f40a33946d4ad4c962b5ae3a52546da53af0c5c")
tx = CTransaction()
tx.vin.append(CTxIn(COutPoint(bogus_tx.sha256, 0), b"", 0xffffffff))
tx.vout.append(CTxOut(1, b""))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b70 = self.update_block(70, [tx])
self.send_blocks([b70], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
# Test accepting an invalid block which has the same hash as a valid one (via merkle tree tricks)
#
# -> b53 (14) -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19) -> b69 (20) -> b72 (21)
# \-> b71 (21)
#
# b72 is a good block.
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# b71 is a copy of 72, but re-adds one of its transactions. However, it has the same hash as b72.
self.log.info("Reject a block containing a duplicate transaction but with the same Merkle root (Merkle tree malleability")
self.move_tip(69)
b72 = self.next_block(72)
tx1 = self.create_and_sign_transaction(out[21].tx, out[21].n, 2)
tx2 = self.create_and_sign_transaction(tx1, 0, 1)
b72 = self.update_block(72, [tx1, tx2]) # now tip is 72
b71 = copy.deepcopy(b72)
b71.vtx.append(tx2) # add duplicate tx2
self.block_heights[b71.sha256] = self.block_heights[b69.sha256] + 1 # b71 builds off b69
self.blocks[71] = b71
assert_equal(len(b71.vtx), 4)
assert_equal(len(b72.vtx), 3)
assert_equal(b72.sha256, b71.sha256)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(71)
self.send_blocks([b71], False, 16, b'bad-txns-duplicate', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(72)
self.send_blocks([b72], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Test some invalid scripts and MAX_BLOCK_SIGOPS
#
# -> b55 (15) -> b57 (16) -> b60 (17) -> b64 (18) -> b65 (19) -> b69 (20) -> b72 (21)
# \-> b** (22)
#
# b73 - tx with excessive sigops that are placed after an excessively large script element.
# The purpose of the test is to make sure those sigops are counted.
#
# script is a bytearray of size 20,526
#
# bytearray[0-19,998] : OP_CHECKSIG
# bytearray[19,999] : OP_PUSHDATA4
# bytearray[20,000-20,003]: 521 (max_script_element_size+1, in little-endian format)
# bytearray[20,004-20,525]: unread data (script_element)
# bytearray[20,526] : OP_CHECKSIG (this puts us over the limit)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Reject a block containing too many sigops after a large script element")
self.move_tip(72)
b73 = self.next_block(73)
size = MAX_BLOCK_SIGOPS - 1 + MAX_SCRIPT_ELEMENT_SIZE + 1 + 5 + 1
a = bytearray([OP_CHECKSIG] * size)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
a[MAX_BLOCK_SIGOPS - 1] = int("4e", 16) # OP_PUSHDATA4
element_size = MAX_SCRIPT_ELEMENT_SIZE + 1
a[MAX_BLOCK_SIGOPS] = element_size % 256
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
a[MAX_BLOCK_SIGOPS + 1] = element_size // 256
a[MAX_BLOCK_SIGOPS + 2] = 0
a[MAX_BLOCK_SIGOPS + 3] = 0
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx = self.create_and_sign_transaction(out[22].tx, 0, 1, CScript(a))
b73 = self.update_block(73, [tx])
assert_equal(get_legacy_sigopcount_block(b73), MAX_BLOCK_SIGOPS + 1)
self.send_blocks([b73], False, 16, b'bad-blk-sigops', reconnect=True)
# b74/75 - if we push an invalid script element, all prevous sigops are counted,
# but sigops after the element are not counted.
#
# The invalid script element is that the push_data indicates that
# there will be a large amount of data (0xffffff bytes), but we only
# provide a much smaller number. These bytes are CHECKSIGS so they would
# cause b75 to fail for excessive sigops, if those bytes were counted.
#
# b74 fails because we put MAX_BLOCK_SIGOPS+1 before the element
# b75 succeeds because we put MAX_BLOCK_SIGOPS before the element
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Check sigops are counted correctly after an invalid script element")
self.move_tip(72)
b74 = self.next_block(74)
size = MAX_BLOCK_SIGOPS - 1 + MAX_SCRIPT_ELEMENT_SIZE + 42 # total = 20,561
a = bytearray([OP_CHECKSIG] * size)
a[MAX_BLOCK_SIGOPS] = 0x4e
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
a[MAX_BLOCK_SIGOPS + 1] = 0xfe
a[MAX_BLOCK_SIGOPS + 2] = 0xff
a[MAX_BLOCK_SIGOPS + 3] = 0xff
a[MAX_BLOCK_SIGOPS + 4] = 0xff
tx = self.create_and_sign_transaction(out[22].tx, 0, 1, CScript(a))
b74 = self.update_block(74, [tx])
self.send_blocks([b74], False, 16, b'bad-blk-sigops', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(72)
b75 = self.next_block(75)
size = MAX_BLOCK_SIGOPS - 1 + MAX_SCRIPT_ELEMENT_SIZE + 42
a = bytearray([OP_CHECKSIG] * size)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
a[MAX_BLOCK_SIGOPS - 1] = 0x4e
a[MAX_BLOCK_SIGOPS] = 0xff
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
a[MAX_BLOCK_SIGOPS + 1] = 0xff
a[MAX_BLOCK_SIGOPS + 2] = 0xff
a[MAX_BLOCK_SIGOPS + 3] = 0xff
tx = self.create_and_sign_transaction(out[22].tx, 0, 1, CScript(a))
b75 = self.update_block(75, [tx])
self.send_blocks([b75], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Check that if we push an element filled with CHECKSIGs, they are not counted
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(75)
b76 = self.next_block(76)
size = MAX_BLOCK_SIGOPS - 1 + MAX_SCRIPT_ELEMENT_SIZE + 1 + 5
a = bytearray([OP_CHECKSIG] * size)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
a[MAX_BLOCK_SIGOPS - 1] = 0x4e # PUSHDATA4, but leave the following bytes as just checksigs
tx = self.create_and_sign_transaction(out[23].tx, 0, 1, CScript(a))
b76 = self.update_block(76, [tx])
self.send_blocks([b76], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Test transaction resurrection
#
# -> b77 (24) -> b78 (25) -> b79 (26)
# \-> b80 (25) -> b81 (26) -> b82 (27)
#
# b78 creates a tx, which is spent in b79. After b82, both should be in mempool
#
# The tx'es must be unsigned and pass the node's mempool policy. It is unsigned for the
# rather obscure reason that the Python signature code does not distinguish between
# Low-S and High-S values (whereas the bitcoin code has custom code which does so);
# as a result of which, the odds are 50% that the python code will use the right
# value and the transaction will be accepted into the mempool. Until we modify the
# test framework to support low-S signing, we are out of luck.
#
# To get around this issue, we construct transactions which are not signed and which
# spend to OP_TRUE. If the standard-ness rules change, this test would need to be
# updated. (Perhaps to spend to a P2SH OP_TRUE script)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Test transaction resurrection during a re-org")
self.move_tip(76)
b77 = self.next_block(77)
tx77 = self.create_and_sign_transaction(out[24].tx, out[24].n, 10 * COIN)
b77 = self.update_block(77, [tx77])
self.send_blocks([b77], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
b78 = self.next_block(78)
tx78 = self.create_tx(tx77, 0, 9 * COIN)
b78 = self.update_block(78, [tx78])
self.send_blocks([b78], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b79 = self.next_block(79)
tx79 = self.create_tx(tx78, 0, 8 * COIN)
b79 = self.update_block(79, [tx79])
self.send_blocks([b79], True)
# mempool should be empty
assert_equal(len(self.nodes[0].getrawmempool()), 0)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(77)
b80 = self.next_block(80, spend=out[25])
self.send_blocks([b80], False, request_block=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b81 = self.next_block(81, spend=out[26])
self.send_blocks([b81], False, request_block=False) # other chain is same length
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b82 = self.next_block(82, spend=out[27])
self.send_blocks([b82], True) # now this chain is longer, triggers re-org
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# now check that tx78 and tx79 have been put back into the peer's mempool
mempool = self.nodes[0].getrawmempool()
assert_equal(len(mempool), 2)
assert(tx78.hash in mempool)
assert(tx79.hash in mempool)
# Test invalid opcodes in dead execution paths.
#
# -> b81 (26) -> b82 (27) -> b83 (28)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Accept a block with invalid opcodes in dead execution paths")
b83 = self.next_block(83)
op_codes = [OP_IF, OP_INVALIDOPCODE, OP_ELSE, OP_TRUE, OP_ENDIF]
script = CScript(op_codes)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx1 = self.create_and_sign_transaction(out[28].tx, out[28].n, out[28].tx.vout[0].nValue, script)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx2 = self.create_and_sign_transaction(tx1, 0, 0, CScript([OP_TRUE]))
tx2.vin[0].scriptSig = CScript([OP_FALSE])
tx2.rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b83 = self.update_block(83, [tx1, tx2])
self.send_blocks([b83], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# Reorg on/off blocks that have OP_RETURN in them (and try to spend them)
#
# -> b81 (26) -> b82 (27) -> b83 (28) -> b84 (29) -> b87 (30) -> b88 (31)
# \-> b85 (29) -> b86 (30) \-> b89a (32)
#
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Test re-orging blocks with OP_RETURN in them")
b84 = self.next_block(84)
tx1 = self.create_tx(out[29].tx, out[29].n, 0, CScript([OP_RETURN]))
tx1.vout.append(CTxOut(0, CScript([OP_TRUE])))
tx1.vout.append(CTxOut(0, CScript([OP_TRUE])))
tx1.vout.append(CTxOut(0, CScript([OP_TRUE])))
tx1.vout.append(CTxOut(0, CScript([OP_TRUE])))
tx1.calc_sha256()
self.sign_tx(tx1, out[29].tx, out[29].n)
tx1.rehash()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx2 = self.create_tx(tx1, 1, 0, CScript([OP_RETURN]))
tx2.vout.append(CTxOut(0, CScript([OP_RETURN])))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx3 = self.create_tx(tx1, 2, 0, CScript([OP_RETURN]))
tx3.vout.append(CTxOut(0, CScript([OP_TRUE])))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx4 = self.create_tx(tx1, 3, 0, CScript([OP_TRUE]))
tx4.vout.append(CTxOut(0, CScript([OP_RETURN])))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
tx5 = self.create_tx(tx1, 4, 0, CScript([OP_RETURN]))
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b84 = self.update_block(84, [tx1, tx2, tx3, tx4, tx5])
self.send_blocks([b84], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(83)
b85 = self.next_block(85, spend=out[29])
self.send_blocks([b85], False) # other chain is same length
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b86 = self.next_block(86, spend=out[30])
self.send_blocks([b86], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.move_tip(84)
b87 = self.next_block(87, spend=out[30])
self.send_blocks([b87], False) # other chain is same length
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b88 = self.next_block(88, spend=out[31])
self.send_blocks([b88], True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.save_spendable_output()
# trying to spend the OP_RETURN output is rejected
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
b89a = self.next_block("89a", spend=out[32])
tx = self.create_tx(tx1, 0, 0, CScript([OP_TRUE]))
b89a = self.update_block("89a", [tx])
self.send_blocks([b89a], False, 16, b'bad-txns-inputs-missingorspent', reconnect=True)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
self.log.info("Test a re-org of ~2 days' worth of blocks (1088 blocks)")
self.move_tip(88)
LARGE_REORG_SIZE = 1088
blocks = []
spend = out[32]
for i in range(89, LARGE_REORG_SIZE + 89):
b = self.next_block(i, spend)
tx = CTransaction()
script_length = MAX_BLOCK_SIZE - len(b.serialize()) - 69
script_output = CScript([b'\x00' * script_length])
tx.vout.append(CTxOut(0, script_output))
tx.vin.append(CTxIn(COutPoint(b.vtx[1].sha256, 0)))
b = self.update_block(i, [tx])
assert_equal(len(b.serialize()), MAX_BLOCK_SIZE)
blocks.append(b)
self.save_spendable_output()
spend = self.get_spendable_output()
self.send_blocks(blocks, True, timeout=960)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
chain1_tip = i
# now create alt chain of same length
self.move_tip(88)
blocks2 = []
for i in range(89, LARGE_REORG_SIZE + 89):
blocks2.append(self.next_block("alt" + str(i)))
self.send_blocks(blocks2, False, request_block=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# extend alt chain to trigger re-org
block = self.next_block("alt" + str(chain1_tip + 1))
self.send_blocks([block], True, timeout=960)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# ... and re-org back to the first chain
self.move_tip(chain1_tip)
block = self.next_block(chain1_tip + 1)
self.send_blocks([block], False, request_block=False)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
block = self.next_block(chain1_tip + 2)
self.send_blocks([block], True, timeout=960)
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# Helper methods
################
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
def add_transactions_to_block(self, block, tx_list):
[tx.rehash() for tx in tx_list]
block.vtx.extend(tx_list)
# this is a little handier to use than the version in blocktools.py
def create_tx(self, spend_tx, n, value, script=CScript([OP_TRUE, OP_DROP] * 15 + [OP_TRUE])):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
return create_transaction(spend_tx, n, b"", value, script)
2015-08-05 23:47:34 +02:00
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
# sign a transaction, using the key we know about
# this signs input 0 in tx, which is assumed to be spending output n in spend_tx
def sign_tx(self, tx, spend_tx, n):
scriptPubKey = bytearray(spend_tx.vout[n].scriptPubKey)
if (scriptPubKey[0] == OP_TRUE): # an anyone-can-spend
tx.vin[0].scriptSig = CScript()
return
(sighash, err) = SignatureHash(spend_tx.vout[n].scriptPubKey, tx, 0, SIGHASH_ALL)
tx.vin[0].scriptSig = CScript([self.coinbase_key.sign(sighash) + bytes(bytearray([SIGHASH_ALL]))])
def create_and_sign_transaction(self, spend_tx, n, value, script=CScript([OP_TRUE])):
tx = self.create_tx(spend_tx, n, value, script)
self.sign_tx(tx, spend_tx, n)
tx.rehash()
return tx
def next_block(self, number, spend=None, additional_coinbase_value=0, script=CScript([OP_TRUE]), solve=True):
if self.tip is None:
base_block_hash = self.genesis_hash
block_time = self.mocktime + 1
else:
base_block_hash = self.tip.sha256
block_time = self.tip.nTime + 1
# First create the coinbase
height = self.block_heights[base_block_hash] + 1
coinbase = create_coinbase(height, self.coinbase_pubkey)
coinbase.vout[0].nValue += additional_coinbase_value
coinbase.rehash()
if spend is None:
block = create_block(base_block_hash, coinbase, block_time)
else:
coinbase.vout[0].nValue += spend.tx.vout[spend.n].nValue - 1 # all but one satoshi to fees
coinbase.rehash()
block = create_block(base_block_hash, coinbase, block_time)
tx = create_transaction(spend.tx, spend.n, b"", 1, script) # spend 1 satoshi
self.sign_tx(tx, spend.tx, spend.n)
self.add_transactions_to_block(block, [tx])
block.hashMerkleRoot = block.calc_merkle_root()
if solve:
block.solve()
self.tip = block
self.block_heights[block.sha256] = height
assert number not in self.blocks
self.blocks[number] = block
return block
# save the current tip so it can be spent by a later block
def save_spendable_output(self):
self.log.debug("saving spendable output %s" % self.tip.vtx[0])
self.spendable_outputs.append(self.tip)
# get an output that we previously marked as spendable
def get_spendable_output(self):
self.log.debug("getting spendable output %s" % self.spendable_outputs[0].vtx[0])
return PreviousSpendableOutput(self.spendable_outputs.pop(0).vtx[0], 0)
# move the tip back to a previous block
def move_tip(self, number):
self.tip = self.blocks[number]
# adds transactions to the block and updates state
def update_block(self, block_number, new_transactions):
block = self.blocks[block_number]
self.add_transactions_to_block(block, new_transactions)
old_sha256 = block.sha256
block.hashMerkleRoot = block.calc_merkle_root()
block.solve()
# Update the internal state just like in next_block
self.tip = block
if block.sha256 != old_sha256:
self.block_heights[block.sha256] = self.block_heights[old_sha256]
del self.block_heights[old_sha256]
self.blocks[block_number] = block
return block
def bootstrap_p2p(self):
"""Add a P2P connection to the node.
Helper to connect and wait for version handshake."""
self.nodes[0].add_p2p_connection(P2PDataStore())
# We need to wait for the initial getheaders from the peer before we
# start populating our blockstore. If we don't, then we may run ahead
# to the next subtest before we receive the getheaders. We'd then send
# an INV for the next block and receive two getheaders - one for the
# IBD and one for the INV. We'd respond to both and could get
# unexpectedly disconnected if the DoS score for that error is 50.
self.nodes[0].p2p.wait_for_getheaders(timeout=5)
def reconnect_p2p(self):
"""Tear down and bootstrap the P2P connection to the node.
The node gets disconnected several times in this test. This helper
method reconnects the p2p and restarts the network thread."""
self.nodes[0].disconnect_p2ps()
self.bootstrap_p2p()
def send_blocks(self, blocks, success=True, reject_code=None, reject_reason=None, request_block=True, reconnect=False, timeout=60):
Backport bitcoin#11773: [tests] Change feature_block.py to use BitcoinTestFramework (#3277) * [tests] Change feature_block.py to use BitcoinTestFramework * [tests] Fix flake8 warnings in feature_block.py * [tests] Tidy up feature_block.py - move all helper methods to the end - remove block, create_tx and create_and_sign_tx shortcuts - remove --runbarelyexpensive option, since it defaults to True and it's unlikely that anyone ever runs the test with this option set to false. * [tests] Add logging to feature_block.py * [tests] Improve assert message when wait_until() fails * Merge #13048: [tests] Fix feature_block flakiness c1d742025c [tests] Fix feature_block flakiness (John Newbery) Pull request description: feature_block.py occasionally fails on Travis. I believe this is due to a a race condition when reconnecting to bitcoind after a subtest that expects disconnection. If the test runs ahead and sends the INV for the subsequent test before we've received the initial sync getheaders, then we may end up sending two headers messages - one as a response to the initial sync getheaders and one in response to the INV getheaders. If both of those headers fail validation with a DoS score of 50 or higher, then we'll unexpectedly be disconnected. There is only one validation failure that has a DoS score bewteen 50 and 100, which is high-hash. That's why the test is failing immediately after the "Reject a block with invalid work" subtest. Fix is to wait for the initial getheaders from the peer before we start populating our blockstore. That way we won't have any invalid headers to respond to it with. Tree-SHA512: dc17d795fcfaf0f8c0bf1e9732b5e11fbc8febbfafba4c231b7c13a5404a2c297dcd703a7a75bc7f353c893e12efc87f424f2201abd47ba5268af32d4d2e841f * Temporarely rename MAX_BLOCK_SIZE -> MAX_BLOCK_BASE_SIZE We'll undo this after the next commit. This avoids merge many conflicts and makes reviewing easier. * Rename MAX_BLOCK_BASE_SIZE back to MAX_BLOCK_SIZE * Use DoS score of 100 for bad-blk-sigops This was accidently changed to 10 while backporting bitcoin#7287 and causes test failures in p2p-fullblocktest.py * Use allowOptimisticSend=true when sending reject messages This fixes test failures in p2p-fullblocktest.py which expects reject messages to be sent/received before connections get closed. * Fix p2p-fullblocktest.py - CBlock and friends are still in test_framework.mininode - "-whitelist" causes connections to not be dropped, which in turn causes sync_blocks with reconnect=True to fail - "bad-cb-amount" does not cause a ban in Dash, so reconnect must be False - Dash already bans when a header is received which is a child of an invalid header, causing block requests to never happen * Backport missing changes from bitcoin#13003 bitcoin#13003 was backported out of order which causes missed changes. * Bump p2p-fullblocktest timeouts * Increase RPC timeout in p2p-fullblocktest.py Co-authored-by: John Newbery <jonnynewbs@gmail.com> Co-authored-by: MarcoFalke <falke.marco@gmail.com>
2020-01-11 02:31:25 +01:00
"""Sends blocks to test node. Syncs and verifies that tip has advanced to most recent block.
Call with success = False if the tip shouldn't advance to the most recent block."""
self.nodes[0].p2p.send_blocks_and_test(blocks, self.nodes[0], success=success, reject_code=reject_code, reject_reason=reject_reason, request_block=request_block, timeout=timeout)
if reconnect:
self.reconnect_p2p()
2015-08-05 23:47:34 +02:00
if __name__ == '__main__':
FullBlockTest().main()