If python passed None for an optional (i.e. 'null' is
sent), this will lead to the arg being interpreted as not
provided by bitcoind - except for string args, for which the arg is
interpreted as as 'null' string. Bypass this by not sending
named args to bitcoin-cli - so that the default value will
actually be used.
Also drops an unnecessary str() conversion, kwargs keys
are always strings.
Because of the MAX_ARG_STRLEN limit (128kb on most systems)
for args, these would usually fail. As a workaround, use
-stdin for these large calls. Idea by 0xB10C.
Co-authored-by: Ryan Ofsky <ryan@ofsky.org>
f16c8c67bf tests: Expand HTTP coverage to assert libevent behavior (Matthew Zipkin)
Pull request description:
These commits are cherry-picked from #32061 and part of a project to [remove libevent](https://github.com/bitcoin/bitcoin/issues/31194).
This PR only adds functional tests to `interface_http` to cover some HTTP server behaviors we inherit from libevent, in order to maintain those behaviors when we replace libevent with our own HTTP server.
1. Pipelining: The server must respond to requests from a client in the order in which they were received [RFC 7230 6.3.2](https://www.rfc-editor.org/rfc/rfc7230#section-6.3.2)
2. `-rpcservertimeout` config option which sets the amount of time the server will keep an idle client connection alive
3. "Chunked" Transfer-Encoding: Allows a client to send a request in pieces, without the `Content-Length` header [RFC 7230 4.1](https://www.rfc-editor.org/rfc/rfc7230#section-4.1)
ACKs for top commit:
achow101:
ACK f16c8c67bf
vasild:
ACK f16c8c67bf
polespinasa:
ACK f16c8c67bf
fjahr:
utACK f16c8c67bf
Tree-SHA512: 405b59431b4d2bf118fde04b270865dee06ef980ab120d9cc1dce28e5d65dfd880a57055b407009d22f4de614bc3eebdb3e203bcd39e86cb14fbfd62195ed06a
a5ac43d98d doc: Add release notes describing bitcoin wrapper executable (Ryan Ofsky)
258bda80c0 doc: Mention bitcoin wrapper executable in documentation (Ryan Ofsky)
d2739d75c9 build: add bitcoin.exe to windows installer (Sjors Provoost)
ba649c0006 ci: Run multiprocess tests through wrapper executable (Ryan Ofsky)
29bdd743bb test: Support BITCOIN_CMD environment variable (Ryan Ofsky)
9c8c68891b multiprocess: Add bitcoin wrapper executable (Ryan Ofsky)
5076d20fdb util: Add cross-platform ExecVp and GetExePath functions (Ryan Ofsky)
Pull request description:
Intended to make bitcoin command line features more discoverable and allow installing new multiprocess binaries in libexec/ instead of bin/ so they don't cause confusion.
Idea and implementation of this were discussed in https://github.com/bitcoin/bitcoin/issues/30983.
---
Initial implementation of this feature is deliberately minimal so the UX can evolve in response to feedback and there are not too many details to debate and discuss in a single PR. But many improvements are possible or planned:
- Adding manpage and bash completions.
- Showing nicer error messages that detect if an executable isn't installed and suggest how to fix [(comment)](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2073194474)
- Showing wrapper command lines in subcommand in help output [(comment)](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2077800405). This could be done conditionally as suggested in the comment or be unconditional.
- Showing wrapper command lines in subcommand error output. There is a bitcoin-cli error pointed out in [(comment)](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2091152243) that is needlessly confusing.
- Integrating help so `bitcoin help subcommand` invokes `bitcoin subcommand -h`. `bitcoin -h subcommand` should also be supported and be equivalent [(comment)](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2093116725)
- Adding support for `bitcoin-util` subcommands. Ideal interface would probably be more like `bitcoin grind` not `bitcoin util grind` but this has been punted for now. Supporting subcommands directly would require some ArgsManager modifications
- Adding a dedicated python functional test for the wrapper. Right now there is some CI coverage by setting the `BITCOIN_CMD` variable, but this doesn't cover things like the help output and version output, and support for different directory layouts.
- Better `--multiprocess` (`-m`) / `--monolithic` (`-M`) default selection. Right now, default is monolithic but it probably makes sense to chose more intelligently depending on whether -ipc options are enabled and what binaries are available.
- Maybe parsing `bitcoin.conf` and supporting options to control wrapper behavior like custom locations or preferences or aliases.
- Better command command line usability. Allow combining short options like (`-ah`). Allow fuzzy matching of subcommands or suggestions if you misspell. (suggested by stickies in review club)
- Not directly related to this PR but `bitcoin-cli named` implementation used by the wrapper should do a better job disambiguating named arguments from base64 arguments ending in = as pointed out in [(comment)](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2091886628)
---
This PR is part of the [process separation project](https://github.com/bitcoin/bitcoin/issues/28722). A review club meeting for it took place in https://bitcoincore.reviews/31375
ACKs for top commit:
Sjors:
utACK a5ac43d98d
achow101:
ACK a5ac43d98d
vasild:
ACK a5ac43d98d
theStack:
ACK a5ac43d98d
ismaelsadeeq:
fwiw my last review implied an ACK a5ac43d98d
hodlinator:
ACK a5ac43d98d
Tree-SHA512: 570e6a4ff8bd79ef6554da3d01f36c0a7c6d2dd7dace8f8732eca98f4a8bc2284474a9beadeba783114fe2f3dd08b2041b3da7753bae0b7f881ec50668cb821f
Covers:
- http pipelining
- rpcservertimeout
Testing this requires adding an option to TestNode to force
the test framework to establish a new HTTP connection for
every RPC. Otherwise, attempting to reuse a persistent connection
would cause framework RPCs during startup and shutdown to fail.
- "chunked" Transfer-Encoding
b104d44227 test: Remove RPCOverloadWrapper (Ava Chow)
4d32c19516 test: Replace importpubkey (Ava Chow)
fe838dd391 test: Replace usage of addmultisigaddress (Ava Chow)
d314207779 test: Replace usage of importaddress (Ava Chow)
fcc457573f test: Replace importprivkey with wallet_importprivkey (Ava Chow)
94c87bbbd0 test: Remove unnecessary importprivkey from wallet_createwallet (Ava Chow)
Pull request description:
`RPCOverloadWrapper` implemented overloads for legacy wallet only RPCs so that the same function call could be used within tests for both legacy wallets and descriptor wallets. With legacy wallets now removed, there is no need to continue to have these overloads.
For `importaddress`, `addmultisigaddress`, and `importpubkey`, the uses of these are converted to `importdescriptors`.
For `importprivkey`, a new helper function `wallet_imporprivkey` is introduced that does what the overload did. This is mainly to reduce verbosity as `importprivkey` was more widely used throughout the tests.
Some tests that used these RPCs are now also no longer relevant and have been removed.
ACKs for top commit:
Sjors:
ACK b104d44227
pablomartin4btc:
cr ACK b104d44227
rkrux:
ACK b104d44227
w0xlt:
ACK b104d44227
Tree-SHA512: ded2f73829e2ce28466d4a9738eb382783ad990daee5d1859dbc4d354e6f8eec0c483ed5ecb1287fe0dd24ac332065b733a30d71b126b841bd7cd49e9a094b6d
Support new BITCOIN_CMD environment variable in functional test to be able to
test the new bitcoin wrapper executable and run other commands through it
instead of calling them directly.
Co-authored-by: Sjors Provoost <sjors@sprovoost.nl>
importprivkey was a legacy wallet only RPC which had a helper for
descriptor wallets in tests. Add wallet_importprivkey helper and use it
wherever importprivkey is used (other than backward compatibility tests)
10845cd7cc qa: Add feature_framework_startup_failures.py (Hodlinator)
28e282ef9a qa: assert_raises_message() - Stop assuming certain structure for exceptions (Hodlinator)
1f639efca5 qa: Work around Python socket timeout issue (Hodlinator)
9b24a403fa qa: Only allow calling TestNode.stop() after connecting (Hodlinator)
6ad21b4c01 qa: Include ignored errors in RPC connection timeout (Hodlinator)
879243e81f qa refactor: wait_for_rpc_connection - Treat OSErrors the same (Hodlinator)
Pull request description:
Improves handling of startup errors in functional tests and puts tests in place to ensure knock-on errors don't creep in.
- `wait_for_rpc_connection()` now appends specific failures leading up to the `Unable to connect to bitcoind` error to that error message:
`[node 0] Unable to connect to bitcoind after 60s (ignored errors: {'missing_credentials': 1, 'OSError.ECONNREFUSED': 239}, latest error: ConnectionRefusedError(111, 'Connection refused'))`
- Fixes Windows Python issue where `socket.timeout` exceptions end up with unset `errno`-fields.
- Also adds comments, refactors code, improves logging.
The underlying purpose is to ensure developer efficiency in finding root causes of test failures.
Prior iterations of the PR partially focused on fixing the same issue as #31620.
Originally inspired by #30390.
### Testing
Can be tested by reverting either faf2f2c654 or fae3bf6b87 from #31620, or the "qa: Avoid calling stop-RPC if not connected" from this PR, and running *feature_framework_startup_failures.py*.
ACKs for top commit:
l0rinc:
ACK 10845cd7cc
ryanofsky:
Code review ACK 10845cd7cc. Only changes since last review were adding a new commit tweaking assert_raises_message(), extending the new test to have a self-check, and to pass through all options to child tests instead of a hardcoded list of options. I left some cleanup suggestions below but they are not important.
Tree-SHA512: f0235c5cbb6d1bb85d8dc5de492a08a34f6edc83499cbf0a5f9a3824809ff84635888c62c9c01101e3cc9ef9f1cdee2c9ab6537fea6feeb005b29f428caf8b22
Removes all legacy wallet specific functional tests.
Also removes the --descriptor and --legacy-wallet options as these are
no longer necessary with the legacy wallet removed.
Observed on local machine running Windows / Python v3.13.1 when overriding rpc_timeout to small values (5- seconds). Next commit performs such overrides.
(Still tolerate calling it on a no longer (self.)running node, as in a node that has been queried for is_node_stopped() and modified state before returning True).
Tests should not attempt to use the non-functioning RPC interface to call stop() unless wait_for_connections() has succeeded.
No longer log and suppress http.client.CannotSendRequest as a consequence of stop()-RPC, as error conditions causing this knock-on issue are now guarded against before the call.
When an RPC connection attempt with bitcoind times out, include which ignored errors occurred in the exception message.
May provide clues of what has gone wrong.
ConnectionResetError is an OSError as well (ECONNRESET), no reason to have a separate except-block for it.
Also improves comments for other exceptions and make condition above more Pythonic.
In the functional tests there are lots of cases where we assert != which
this new util will replace, we also are adding the imports and the new assertion
fa310cc6f4 test: Fix intermittent issue in p2p_orphan_handling.py (MarcoFalke)
Pull request description:
The test may fail intermittently when the `net` thread is lagging while calling `DeleteNode`. This may result in a split `getdata`, meaning that `peer2.wait_for_parent_requests([int(parent_peekaboo_AB["txid"], 16), int(parent_missing["txid"], 16)])` fails.
Fix it by adding a sync on the `net` thread.
Fixes#31700
ACKs for top commit:
mzumsande:
Code Review ACK fa310cc6f4
Tree-SHA512: e4a58093ab5b9e280c479b845fecb5d228e65519ea3dc2111b393202225fd0feded423e8812452454b6b9348cb37a9c1b01b9d1b1802e9f4aa76b9e56b4b54ef
Add new TestNode.binaries object to manage paths to bitcoin binaries.
Having this object makes it possible for the test framework to exercise the
bitcoin wrapper executable introduced in
https://github.com/bitcoin/bitcoin/pull/31375 and also makes it easier to add
new binaries and options and environment variables controlling how they are
invoked, because logic for invoking them that was previously spread out is now
consolidated in one place.
Co-authored-by: Sjors Provoost <sjors@sprovoost.nl>
Wait until the node's process has fully stopped before starting a new instance.
Since the same code is used in tool_wallet.py, this consolidates the behavior
into a 'kill_process()' function.
faf2f2c654 test: Avoid redundant stop and error spam on shutdown (MarcoFalke)
fae3bf6b87 test: Avoid redundant stop and error spam on startup failure (MarcoFalke)
fa0dc09b90 test: Remove --noshutdown flag (MarcoFalke)
fad441fba0 test: Treat leftover process as error (MarcoFalke)
Pull request description:
The `--noshutdown` flag is brittle, confusing, and redundant:
* Someone wanting to inspect the state after a test failure will likely also want to debug the state on the python side, so the option is redundant with `--pdbonfailure`. If there was a use case to replicate `--pdbonfailure` without starting pdb, a dedicated flag could be added for that use case.
* It is brittle to use the flag for a passing test, because it will disable checks in the test. For example, on shutdown LSan will perform a leak check, and the test framework will check that the node did not crash, and it will check that the node did not print errors to stderr.
Fix all issues by removing it.
Also, tidy up startup error messages to be less confusing as a result.
ACKs for top commit:
hodlinator:
re-ACK faf2f2c654
pablomartin4btc:
re tACK faf2f2c654
Tree-SHA512: 46d7ae59c7be88b93f1f9e0b6be21af0fc101e646512e2c5e725682cb18bfec8aa010e0ebe89ce9ffe239e5caac0da5f81cc97b79e738d26ca5fa31930e8e4e3
Printing to stderr instead of stdout makes the test_runner.py fail on
leftover processes. This is desired and fine, because a leftover process
should only happen on a test failure anyway.
1dd3af8fbc Add release note for #31223 (Martin Zumsande)
997757dd2b test: add functional test for -port behavior (Martin Zumsande)
0e2b12b92a net, init: derive default onion port if a user specified a -port (Martin Zumsande)
Pull request description:
This resolves#31133 (setups with multiple local nodes each using a different `-port` no longer working with v28.0, see the issue description for more details) by deriving the default onion listening port to be the value specified by `-port` incremented by 1 (idea by vasild / laanwj).
Note that with this fix, the chosen `-port` values of two local nodes cannot be adjacent, otherwise there will be port collisions again.
From the discussion in the linked issue, this was the most popular option, followed by doing nothing and telling affected users to change their setups to use `-bind` instead of `-port`. But more opinions are certainly welcome!
I think that if we decide to do something about the problem described in the issue, we should do so soon (in 28.1.), so I opened this PR.
Fixes#31133
ACKs for top commit:
achow101:
ACK 1dd3af8fbc
laanwj:
Tested ACK 1dd3af8fbc
tdb3:
Code review ACK 1dd3af8fbc
Tree-SHA512: 37fda2b23bbedcab5df3a401cf5afce66ae5318fb78f9660f83e3fd075b528e8156d7a0903f9a12ffe97ab5d83860587116b74af28670a1f4c2f0d1be4999f40
111465d72d test: Remove unused attempts parameter from wait_until (Fabian Jahr)
5468a23eb9 test: Add check_interval parameter to wait_until (Fabian Jahr)
16c87d91fd test: Introduce ensure_for helper (Fabian Jahr)
Pull request description:
A repeating pattern in the functional tests is that the test sleeps for a while to ensure that a certain condition is still true after some amount of time has elapsed. Most recently a new case of this was added in #30807. This PR here introduces an `ensure` helper to streamline this functionality.
Some approach considerations:
- It is possible to construct this by reusing `wait_until` and wrapping it in `try` internally. However, the logger output of the failing wait would still be printed which seems irritating. So I opted for simplified but similar internals to `wait_until`.
- This implementation starts for a failure in the condition right away which has the nice side-effect that it might give feedback on a failure earlier than is currently the case. However, in some cases, it may be expected that the condition may still be false at the beginning and then turns true until time has run out, something that would work when the test sleeps without checking in a loop. I decided against this design (and even against adding it as an option) because such a test design seems like it would be racy either way.
- I have also been going back and forth on naming. To me `ensure` works well but I am also not a native speaker, happy consider a different name if others don't think it's clear enough.
ACKs for top commit:
maflcko:
re-ACK 111465d72d🍋
achow101:
ACK 111465d72d
tdb3:
code review re ACK 111465d72d
furszy:
utACK 111465d72d
Tree-SHA512: ce01a4f3531995375a6fbf01b27d51daa9d4c3d7cd10381be6e86ec5925d2965861000f7cb4796b8d40aabe3b64c4c27e2811270e4e3c9916689575b8ba4a2aa
After port collisions are no longer tolerated but lead to
a startup failure in v28.0, local setups of multiple nodes,
each with a different -port value would not be possible anymore
due to collision of the onion default port - even if the nodes
were using tor or not interested in receiving onion inbound connections.
Fix this by deriving the onion listening port to be -port + 1.
(idea by vasild / laanwj)
Co-authored-by: Vasil Dimov <vd@FreeBSD.org>
57529ac4db test: set P2PConnection.p2p_connected_to_node in peer_connect_helper() (Vasil Dimov)
22cd0e888c test: support WTX INVs from P2PDataStore and fix a comment (Vasil Dimov)
ebe42c00aa test: extend the SOCKS5 Python proxy to actually connect to a destination (Vasil Dimov)
ba621ffb9c test: improve debug log message from P2PConnection::connection_made() (Vasil Dimov)
Pull request description:
If requested, make the SOCKS5 Python proxy redirect connections to a set of given destinations. Actually act as a real proxy, connecting the client to a destination, except that the destination is not what the client asked for.
This would enable us to "connect" to Tor addresses from the functional tests.
Plus a few other minor improvements in the test framework as individual commits.
---
These changes are part of https://github.com/bitcoin/bitcoin/pull/29415 but they make sense on their own and would be good to have them, regardless of the fate of #29415. Also, if this is merged, that would reduce the size of #29415, thus the current standalone PR.
ACKs for top commit:
jonatack:
Approach ACK 57529ac4db
achow101:
ACK 57529ac4db
tdb3:
CR and test ACK 57529ac4db
mzumsande:
Code review / tested ACK 57529ac4db
Tree-SHA512: a2892c97bff2d337b37455c409c6136cb62423ce6cc32b197b36f220c1eec9ca046b599135b9a2603c0eb6c1ac4d9795e73831ef0f04378aeea8b245ea733399
Set `P2PConnection.p2p_connected_to_node` in
`P2PConnection.peer_connect_helper()` instead of
`TestNode.add_p2p_connection()` and
`TestNode.add_outbound_p2p_connection()`.
This way tests can create an instance of `P2PConnection` and use
`P2PConnection.peer_connect_helper()` directly.
bca346a970 net: require P2P binds to succeed (Vasil Dimov)
af552534ab net: report an error if unable to bind on the Tor listening addr:port (Vasil Dimov)
9a7e5f4d68 net: don't extra bind for Tor if binds are restricted (Vasil Dimov)
Pull request description:
Make it possible to disable the Tor binding on `127.0.0.1:8334` and stop startup if any P2P bind fails instead of "if all P2P binds fail".
Fixes https://github.com/bitcoin/bitcoin/issues/22726
Fixes https://github.com/bitcoin/bitcoin/issues/22727
ACKs for top commit:
achow101:
ACK bca346a970
cbergqvist:
ACK bca346a970
pinheadmz:
ACK bca346a970
Tree-SHA512: fabef89a957191eea4f3e3b6109d2b8389f27ecc74440a920b0c10f31fff00a85bcfd1eb3c91826c7169c618f4de8a8d0a260e2caf40fd854f07ea9a980d8603
c9dacd958d test: Check that non empty version packet is ignored and no disconnection happens (stratospher)
997cc00b95 test: Check that disconnection happens when AAD isn't filled (stratospher)
b5e6238fdb test: Check that disconnection happens when garbage sent/received are different (stratospher)
ad1482d5a2 test: Check that disconnection happens when wrong garbage terminator is sent (stratospher)
e351576862 test: Check that disconnection happens when >4095 garbage bytes is sent (stratospher)
e075fd131d test: Introduce test types and modify v2 handshake function accordingly (stratospher)
7d07daa623 log: Add V2 handshake timeout (stratospher)
d4a1da8543 test: Make global TRANSPORT_VERSION variable an instance variable (stratospher)
c642b08c4e test: Log when the garbage is actually sent to transport layer (stratospher)
86cca2cba2 test: Support disconnect waiting for add_p2p_connection (stratospher)
bf9669af9c test: Rename early key response test and move random_bitflip to util (stratospher)
Pull request description:
Add tests for the following v2 handshake scenarios:
1. Disconnection happens when > `MAX_GARBAGE_LEN` bytes garbage is sent
2. Disconnection happens when incorrect garbage terminator is sent
3. Disconnection happens when garbage bytes are tampered with
4. Disconnection happens when AAD of first encrypted packet after the garbage terminator is not filled
5. bitcoind ignores non-empty version packet and no disconnection happens
All these tests require a modified v2 P2P class (different from `EncryptedP2PState` used in `v2_p2p.py`) to implement our custom handshake behaviour based on different scenarios and have been kept in a single test file (`test/functional/p2p_v2_misbehaving.py`). Shifted the test in `test/functional/p2p_v2_earlykeyresponse.py` which is of the same pattern to this file too.
ACKs for top commit:
achow101:
ACK c9dacd958d
mzumsande:
ACK c9dacd958d
theStack:
Code-review ACK c9dacd958d
Tree-SHA512: 90df81f0c7f4ecf0a47762d290a618ded92cde9f83d3ef3cc70e1b005ecb16125ec39a9d80ce95f99e695d29abd63443240cb5490aa57c5bc8fa2e52149a0672
In the Tor case, this prevents us from telling the Tor daemon to send
our incoming connections from the Tor network to an address where we
do not listen (we tried to listen but failed probably because another
application is already listening).
In the other cases (IPv4/IPv6 binds) this also prevents unpleasant
surprises caused by continuing operations even on bind failure. For
example, another application may be listening on portX, bitcoind tries
to bind on portX and portY, only succeeds with portY and continues
operation leaving the user thinking that his bitcoind is listening on
portX whereas another application is listening (the error message in
the log could easily be missed).
Avoid having the functional testing framework start multiple `bitcoind`s
that try to listen on the same `127.0.0.1:18445` (Tor listen for
regtest) if `bind_to_localhost_only` is set to `False`.
Also fix a typo in `test-shell.md` related to `bind_to_localhost_only`.
Fixes https://github.com/bitcoin/bitcoin/issues/22727
