Patrick Lodder 3a700cde38
net: use txrequest for transaction request logic
The major changes are:

* Announcements from outbound (and whitelisted) peers are now
  always preferred over those from inbound peers. This used to be
  the case for the first request (by delaying the first request
  from inbound peers), and a bias after. The 2s delay for requests
  from inbound peers still exists, but after that, if viable
  outbound peers remain for any given transaction, they will
  always be tried first.

* No more hard cap of 100 in flight transactions per peer, as
  there is less need for it (memory usage is linear in the number
  of announcements, but independent from the number in flight,
  and CPU usage isn't affected by it). Furthermore, if only one
  peer announces a transaction, and it has over 100 in flight and
  requestable already, we still want to request it from them. The
  cap is replaced with an additional 2s delay (possibly combined
  with the existing 2s delays for inbound connections).

Backported from: 242d1647
                 173a1d2d
Original Author: Pieter Wuille <pieter@wuille.net>

Conflicts:
  - replaced GenTx with uint256 because no segwit
  - removed additional 2s penalty for non-segwit peers
  - used int64_t instead of std::chrono::microseconds per utiltime
  - implemented TxRequest as g_txrequest instead of as a member of
    PeerManager, which we don't have
  - removed the Dogecoin-specific strict max inflight test
  - make exceptions for whitelisted nodes as there is no fine-
    grained permission system
2024-07-14 21:57:16 -04:00
..
2024-03-12 17:12:49 -04:00
2022-07-16 20:18:50 +02:00
2017-01-14 22:17:23 +00:00
2016-07-23 01:02:53 +09:00
2016-07-23 01:02:53 +09:00
2016-07-23 01:02:53 +09:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2021-08-17 08:51:18 +01:00
2017-01-13 18:55:03 +01:00
2023-08-04 11:24:02 -04:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2016-09-30 02:48:09 +08:00
2022-06-22 17:05:03 +02:00
2024-02-25 13:12:26 -05:00
2024-07-03 11:21:58 -04:00
2017-01-13 19:03:20 +01:00
2021-09-03 10:52:24 +01:00
2024-02-25 13:12:26 -05:00
2022-06-22 22:34:42 -07:00
2022-07-16 20:18:50 +02:00
2017-02-08 10:24:02 -05:00
2024-02-25 13:12:26 -05:00
2024-02-25 13:12:26 -05:00
2022-07-16 20:18:50 +02:00
2024-02-25 13:12:26 -05:00
2024-07-10 22:09:00 -04:00
2017-01-13 19:03:20 +01:00

Regression tests

test_framework/authproxy.py

Taken from the python-bitcoinrpc repository.

test_framework/test_framework.py

Base class for new regression tests.

test_framework/util.py

Generally useful functions.

test_framework/mininode.py

Basic code to support p2p connectivity to a dogecoind.

test_framework/comptool.py

Framework for comparison-tool style, p2p tests.

test_framework/script.py

Utilities for manipulating transaction scripts (originally from python-bitcoinlib)

test_framework/blockstore.py

Implements disk-backed block and tx storage.

test_framework/key.py

Wrapper around OpenSSL EC_Key (originally from python-bitcoinlib)

test_framework/bignum.py

Helpers for script.py

test_framework/blocktools.py

Helper functions for creating blocks and transactions.

P2P test design notes

Mininode

  • mininode.py contains all the definitions for objects that pass over the network (CBlock, CTransaction, etc, along with the network-level wrappers for them, msg_block, msg_tx, etc).

  • P2P tests have two threads. One thread handles all network communication with the dogecoind(s) being tested (using python's asyncore package); the other implements the test logic.

  • NodeConn is the class used to connect to a dogecoind. If you implement a callback class that derives from NodeConnCB and pass that to the NodeConn object, your code will receive the appropriate callbacks when events of interest arrive.

  • You can pass the same handler to multiple NodeConn's if you like, or pass different ones to each -- whatever makes the most sense for your test.

  • Call NetworkThread.start() after all NodeConn objects are created to start the networking thread. (Continue with the test logic in your existing thread.)

  • RPC calls are available in p2p tests.

  • Can be used to write free-form tests, where specific p2p-protocol behavior is tested. Examples: p2p-accept-block.py, maxblocksinflight.py.

Comptool

  • Testing framework for writing tests that compare the block/tx acceptance behavior of a dogecoind against 1 or more other dogecoind instances, or against known outcomes, or both.

  • Set the num_nodes variable (defined in ComparisonTestFramework) to start up 1 or more nodes. If using 1 node, then --testbinary can be used as a command line option to change the dogecoind binary used by the test. If using 2 or more nodes, then --refbinary can be optionally used to change the dogecoind that will be used on nodes 2 and up.

  • Implement a (generator) function called get_tests() which yields TestInstances. Each TestInstance consists of:

    • a list of [object, outcome, hash] entries
      • object is a CBlock, CTransaction, or CBlockHeader. CBlock's and CTransaction's are tested for acceptance. CBlockHeaders can be used so that the test runner can deliver complete headers-chains when requested from the dogecoind, to allow writing tests where blocks can be delivered out of order but still processed by headers-first dogecoind's.
      • outcome is True, False, or None. If True or False, the tip is compared with the expected tip -- either the block passed in, or the hash specified as the optional 3rd entry. If None is specified, then the test will compare all the dogecoind's being tested to see if they all agree on what the best tip is.
      • hash is the block hash of the tip to compare against. Optional to specify; if left out then the hash of the block passed in will be used as the expected tip. This allows for specifying an expected tip while testing the handling of either invalid blocks or blocks delivered out of order, which complete a longer chain.
    • sync_every_block: True/False. If False, then all blocks are inv'ed together, and the test runner waits until the node receives the last one, and tests only the last block for tip acceptance using the outcome and specified tip. If True, then each block is tested in sequence and synced (this is slower when processing many blocks).
    • sync_every_transaction: True/False. Analogous to sync_every_block, except if the outcome on the last tx is "None", then the contents of the entire mempool are compared across all dogecoind connections. If True or False, then only the last tx's acceptance is tested against the given outcome.
  • For examples of tests written in this framework, see invalidblockrequest.py and p2p-fullblocktest.py.