Dbwrapper used GetSerializeSize() to compute the size of the buffer
to preallocate. For some cases (specifically: CCoins) this requires
a costly compression call. Avoid this by just using fixed size
preallocations instead.
To get the advantages of faster GetSerializeSize() implementations
back that were removed in "Make GetSerializeSize a wrapper on top of
CSizeComputer", reintroduce them in the few places in the form of a
specialized Serialize() implementation. This actually gets us in a
better state than before, as these even get used when they're invoked
indirectly in the serialization of another object.
The CSerAction's ForRead() method does not depend on any runtime
data, so guarantee that requests to it can be optimized out by
making it constexpr.
Suggested by Cory Fields.
Remove the nType and nVersion as parameters to all serialization methods
and functions. There is only one place where it's read and has an impact
(in CAddress), and even there it does not impact any of the recursively
invoked serializers.
Instead, the few places that need nType or nVersion are changed to read
it directly from the stream object, through GetType() and GetVersion()
methods which are added to all stream classes.
Given that in default GetSerializeSize implementations created by
ADD_SERIALIZE_METHODS we're already using CSizeComputer(), get rid
of the specialized GetSerializeSize methods everywhere, and just use
CSizeComputer. This removes a lot of code which isn't actually used
anywhere.
For CCompactSize and CVarInt this actually removes a more efficient
size computing algorithm, which is brought back in a later commit.
The current getblocktxn implementation drops and ignores requests for old
blocks, which causes occasional sync_block timeouts during the
p2p-compactblocks.py test as reported in
https://github.com/bitcoin/bitcoin/issues/8842.
The p2p-compactblocks.py test setup creates many new blocks in a short
period of time, which can lead to getblocktxn requests for blocks below the
hardcoded depth limit of 10 blocks. This commit changes the getblocktxn
handler not to ignore these requests, so the peer nodes in the test setup
will reliably be able to sync.
The protocol change is documented in BIP-152 update "Allow block responses
to getblocktxn requests" at https://github.com/bitcoin/bips/pull/469.
The protocol change is not expected to affect nodes running outside the test
environment, because there shouldn't normally be lots of new blocks being
rapidly added that need to be synced.
The stream implementations had two cascading layers (the upper one
with operator<< and operator>>, and a lower one with read and write).
The lower layer's functions are never cascaded (nor should they, as
they should only be used from the higher layer), so make them return
void instead.
Increase wallet-dump RPC timeout from 30 seconds to 1 minute. This avoids a
timeout error that seemed to happen regularly (around 50% of builds) on a
particular jenkins server during the first getnewaddress RPC call made by the
test.
The failing stack trace looked like:
Unexpected exception caught during testing: timeout('timed out',)
File ".../bitcoin/qa/rpc-tests/test_framework/test_framework.py", line 146, in main
self.run_test()
File ".../bitcoin/qa/rpc-tests/wallet-dump.py", line 73, in run_test
addr = self.nodes[0].getnewaddress()
File ".../bitcoin/qa/rpc-tests/test_framework/coverage.py", line 49, in __call__
return_val = self.auth_service_proxy_instance.__call__(*args, **kwargs)
File ".../bitcoin/qa/rpc-tests/test_framework/authproxy.py", line 145, in __call__
response = self._request('POST', self.__url.path, postdata.encode('utf-8'))
File ".../bitcoin/qa/rpc-tests/test_framework/authproxy.py", line 121, in _request
return self._get_response()
File ".../bitcoin/qa/rpc-tests/test_framework/authproxy.py", line 160, in _get_response
http_response = self.__conn.getresponse()
File "/usr/lib/python3.4/http/client.py", line 1171, in getresponse
response.begin()
File "/usr/lib/python3.4/http/client.py", line 351, in begin
version, status, reason = self._read_status()
File "/usr/lib/python3.4/http/client.py", line 313, in _read_status
line = str(self.fp.readline(_MAXLINE + 1), "iso-8859-1")
File "/usr/lib/python3.4/socket.py", line 374, in readinto
return self._sock.recv_into(b)
9027680 net: handle version push in InitializeNode (Cory Fields)
7588b85 net: construct CNodeStates in place (Cory Fields)
440f1d3 net: remove now-unused ssSend and Fuzz (Cory Fields)
5c2169c drop the optimistic write counter hack (Cory Fields)
ea33268 net: switch all callers to connman for pushing messages (Cory Fields)
3e32cd0 connman is in charge of pushing messages (Cory Fields)
b98c14c serialization: teach serializers variadics (Cory Fields)
Three categories of modifications:
1)
1 instance of 'The Bitcoin Core developers \n',
1 instance of 'the Bitcoin Core developers\n',
3 instances of 'Bitcoin Core Developers\n', and
12 instances of 'The Bitcoin developers\n'
are made uniform with the 443 instances of 'The Bitcoin Core developers\n'
2)
3 instances of 'BitPay, Inc\.\n' are made uniform with the other 6
instances of 'BitPay Inc\.\n'
3)
4 instances where there was no '(c)' between the 'Copyright' and the year
where it deviates from the style of the local directory.
- `--help`, `--version` etc should exit with `0` i.e. no error ("not enough args" case should still trigger an error)
- error reading config file should exit with `1`
Slightly refactor AppInitRPC/AppInitRawTx to return standard exit codes (EXIT_FAILURE/EXIT_SUCCESS) or CONTINUE_EXECUTION (-1)
The changes here are dense and subtle, but hopefully all is more explicit
than before.
- CConnman is now in charge of sending data rather than the nodes themselves.
This is necessary because many decisions need to be made with all nodes in
mind, and a model that requires the nodes calling up to their manager quickly
turns to spaghetti.
- The per-node-serializer (ssSend) has been replaced with a (quasi-)const
send-version. Since the send version for serialization can only change once
per connection, we now explicitly tag messages with INIT_PROTO_VERSION if
they are sent before the handshake. With this done, there's no need to lock
for access to nSendVersion.
Also, a new stream is used for each message, so there's no need to lock
during the serialization process.
- This takes care of accounting for optimistic sends, so the
nOptimisticBytesWritten hack can be removed.
- -dropmessagestest and -fuzzmessagestest have not been preserved, as I suspect
they haven't been used in years.
