Second part of #11518.
Moves the primitive Bitcoin datastructures and message classes into their own module, and tidies up the mininode.py module.
One travis failure in net.py:
02017-11-10 01:10:17.553000 TestFramework.node0 (DEBUG): RPC successfully started
12017-11-10 01:10:17.555000 TestFramework.node1 (DEBUG): RPC successfully started
22017-11-10 01:10:18.043000 TestFramework (ERROR): Assertion failed
3Traceback (most recent call last):
4 File "/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/test/functional/test_framework/test_framework.py", line 120, in main
5 self.run_test()
6 File "/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/test/functional/net.py", line 27, in run_test
7 self._test_getnettotals()
8 File "/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/test/functional/net.py", line 54, in _test_getnettotals
9 assert_equal(before['bytesrecv_per_msg']['pong'] + 32, after['bytesrecv_per_msg']['pong'])
10 File "/home/travis/build/bitcoin/bitcoin/build/bitcoin-x86_64-unknown-linux-gnu/test/functional/test_framework/util.py", line 38, in assert_equal
11 raise AssertionError("not(%s)" % " == ".join(str(arg) for arg in (thing1, thing2) + args))
12AssertionError: not(64 == 32)
132017-11-10 01:10:18.044000 TestFramework (INFO): Stopping nodes
142017-11-10 01:10:18.044000 TestFramework.node0 (DEBUG): Stopping node
mininode.py wildcard imports all names from primitives.py. This is
to avoid having to change all test scripts that import from mininode.py.
Mostly move only. Adds a few extra comments.
I think we need more categorization here; bitcoin P2P message types are clearly a different kind of primitives than consensus data structures
I tried splitting this further into primitives.py (for the primitive consensus data structures) and messages.py (for the p2p message data structures), but the resulting layout didn’t make much sense to me so I reverted it. The main motivation for this PR was to split out the mininode functionality from the underlying data structures. I’ve renamed the new module messages.py as suggested by @ryanofsky.
If you think it’s worth splitting further into consensus primitives and p2p message data structures, I can look at doing that in a future PR. It would end up looking something like https://github.com/jnewbery/bitcoin/tree/pr11648.2.
32+MY_RELAY = 1 # from version 70001 onwards, fRelay should be appended to version messages (BIP37)
33+
34+MAX_INV_SZ = 50000
35+MAX_BLOCK_BASE_SIZE = 1000000
36+
37+COIN = 100000000 # 1 btc in satoshis
It’s used in:
CTxOut.__repr__CTransaction.is_validIt’s arguable where this belongs, the ideal place might be some bitcoin-specific constants header, but for now it makes sense to have it here.
utACK 1135c796a0e1151b03933f2cd36ce95a8e136b7a
Verified that it is indeed move-only and/or fixups due to rename. I am neutral on the overall change. Some prefer a huge file where it is easy to jump back and forth, others prefer multiple files that bundle into smaller sets of related classes and functions.
I think there is no single true way of organizing the files, but if the refactoring helps some greater goal, I am happy with the changes.
52+
53+def hash256(s):
54+ return sha256(sha256(s))
55+
56+def ser_compact_size(l):
57+ r = b""
IMO both ways are acceptable, as long as the commits that move code are separate from those that make changes.
In any case, moving code around doesn’t obligate someone to also fix all the problems it has, in style or other things. Can be done later.