2dede9f675 Adjust RPCTypeCheckObj error string (Leonardo Araujo)
Pull request description:
Unifies the JSON type error strings as mentioned in #26214. Also refer to #25737.
ACKs for top commit:
furszy:
ACK 2dede9f6
Tree-SHA512: c918889e347ba32cb6d0e33c0de5956c2077dd40c996151e16741b0c4983ff098c60258206ded76ad7bbec4876c780c6abb494a97e4f1e05717d28a59b9167a6
fa09525751 univalue: string_view test (MacroFake)
1111c7e3f1 univalue: Avoid std::string copies (MacroFake)
Pull request description:
This shouldn't matter too much, unless a really large string is pushed into a json struct, but I think it also clarifies the code.
ACKs for top commit:
martinus:
Code review ACK fa09525751
aureleoules:
reACK fa09525751
ryanofsky:
Code review ACK fa09525751
Tree-SHA512: 74c441912bd0b00cdb9ea7890121f71ae5d62a7594e7d29aa402c9e3f033710c5d3afb27a37c552e6513804b249aa37e375ce013a3db853a25d1fd7b6e6cd3a8
737c285f69 test: Don't pass add_to_wallet option to walletcreatefundedpsbt (Ryan Ofsky)
Pull request description:
It's not a documented option. Noticed while working on #19762
ACKs for top commit:
achow101:
ACK 737c285f69
Tree-SHA512: 1bf4186fae4390233b2f23389eb6c515c7f0209f12553592df5166e75c452ccd1fb125d9246047c08cff0b869fdda7793812d15da01441e2c4777514446f3ed6
This change allows to simplify CI tests, and makes it easier to
integrate the `bench_bitcoin` binary into CMake custom targets or
commands, as `COMMAND` does not support output redirection
Review note: The changes are complete, because self.options.descriptors
is set to None in parse_args (test_framework.py).
A value of None implies -disablewallet, see the previous commit.
So if a call to add_wallet_options is missing, it will lead to a test
failure when the wallet is compiled in.
The bool is only used to call a public helper, which some tests already
do. So use the public helper in all tests consistently and make the
confusingly named bool private.
887d85e43d test: add missing bech32m / BIP86 test-cases to wallet_descriptor.py (Sebastian Falbesoner)
Pull request description:
This small PR adds missing "bech32m" address type / BIP86 checks w.r.t. to the `getnewaddress`/`getrawchangeaddress` RPC and descriptor export functionality to the functional test `wallet_descriptor.py`.
ACKs for top commit:
shaavan:
ACK 887d85e43d
kristapsk:
ACK 887d85e43d
Tree-SHA512: 05b443ae14138769dc3c87a0178f21db2698fa5bcbeaa953c50ed0c9cf5dcd1effcf4afd09551ca9f4ce73898a7882caaf4c57078767beb6a6a65eb3a662726d
c8f91478c1 test: Avoid collision with valid path names in `getarg_tests/logargs` (Hennadii Stepanov)
Pull request description:
This PR prevents test failure when "private" is a part of a valid path.
For example, `/private/var` is a valid path on macOS for temporary files, which in turn causes test failure on CI for tests managed by the [CTest](https://github.com/bitcoin/bitcoin/pull/25797) framework.
ACKs for top commit:
MarcoFalke:
ACK c8f91478c1
Tree-SHA512: 09d257f8fa6be903ec8092b2ae92887a4bec2d05085c76c110637657f4a4bfe2714bf87e2e4727719b3624c8fa4c835ce2ca259c2c9c93033837f997b2057e4f
In the wallet key-value-loading routine, most legacy type entries
require a LegacyScriptPubKeyMan instance after successful
deserialization. On a descriptor wallet, creating that (via method
`GetOrCreateLegacyScriptPubKeyMan`) fails and then leads to a
null-pointer dereference crash. Fix this by throwing an error if
if the wallet flags indicate that we have a descriptor wallet and there
is a legacy entry found.
It's nice to be able to use named options and positional arguments together.
Most shell tools accept both, and python functions combine options and
arguments allowing them to be passed with even more flexibility. This change
adds support for python's approach so as a motivating example:
bitcoin-cli -named createwallet wallet_name=mywallet load_on_startup=1
Can be shortened to:
bitcoin-cli -named createwallet mywallet load_on_startup=1
JSON-RPC standard doesn't have a convention for passing named and positional
parameters together, so this implementation makes one up and interprets any
unused "args" named parameter as a positional parameter array.
25ef049d60 log: mempool: log removal reason in validation interface (James O'Beirne)
Pull request description:
Currently the exact reason a transaction is removed from the mempool isn't logged. It is sometimes detectable from context, but adding the `reason` to the validation interface logs (where it is already passed) seems like an easy way to disambiguate.
For example in the case of mempool expiry, the logs look like this:
```
[validationinterface.cpp:220] [TransactionRemovedFromMempool] [validation] Enqueuing TransactionRemovedFromMempool: txid=<txid> wtxid=<wtxid>
[txmempool.cpp:1050] [RemoveUnbroadcastTx] [mempool] Removed <txid> from set of unbroadcast txns before confirmation that txn was sent out
[validationinterface.cpp:220] [operator()] [validation] TransactionRemovedFromMempool: txid=<txid> wtxid=<wtxid>
[validation.cpp:267] [LimitMempoolSize] [mempool] Expired 1 transactions from the memory pool
```
There is no context-free way to know $txid was evicted on the basis of expiry. This change will make that case (and probably others) clear.
ACKs for top commit:
0xB10C:
ACK 25ef049d60
Tree-SHA512: 9890f9fa16f66c8a9296798d8c28993e1b81da17cf592946f2abc22041f0b30b0911ab86a0c48d4aa46b9a8b3f7f5de67778649ac48c97740b0a09aa6816e0af
c3b1fe59db rpc: doc: add missing option "bech32m" for `change_type` parameters (Sebastian Falbesoner)
Pull request description:
Affects the help of the `fundrawtransaction`, `send` and `walletcreatefundedpsbt` RPCs.
This was found by manually inspecting the results of `$ git grep p2sh-segwit.*bech32`.
ACKs for top commit:
achow101:
ACK c3b1fe59db
Tree-SHA512: a3f1f8fde5905c80e1b95bd042ca0bc73d08c1c0e79c52ab0d6d12d7afdd4aa288afb41e12279fcea328a396f3d0a5564018170c0a11c5aa26dc6d44d2a62b1c
0de30ed509 tests: Test Taproot PSBT signing with keys in other descriptor (Andrew Chow)
6efcdf6b7f tests: Use new wallets for each test in wallet_taproot.py (Andrew Chow)
8781a1b6bb psbt: Include output pubkey in additional pubkeys to sign (Andrew Chow)
323890d0d7 sign: Fill in taproot pubkey info for all script path sigs (Andrew Chow)
Pull request description:
A user reported on [stackexchange](https://bitcoin.stackexchange.com/q/115742/48884) that they were unable to sign for a `multi_a` script using a wallet that only had the corresponding keys (i.e. it did not have the `multi_a()` descriptor). This PR fixes this issue.
Additionally, `wallet_taproot.py` is modified to test for this scenario by having another wallet in `do_test_psbt` which contains descriptors that only have the keys involved in the descriptor being tested. `wallet_taproot.py` was also modified to create new wallets for each test case rather than sharing wallets throughout as the sharing could result in the signing wallet having the keys in a different descriptor and accidentally result in failing to detect a test failure.
The changes to the test also revealed a similar issue with `rawtr()` descriptors, which has also been fixed by checking if a descriptor can produce a `SigningProvider` for the Taproot output pubkey.
ACKs for top commit:
instagibbs:
crACK 0de30ed509
darosior:
ACK 0de30ed509
Tree-SHA512: 12e131dd8afd93da7b1288c9054de2415a228d4477b97102da3ee4e82ce9de20b186260c3085a4b7b067bd8b74400751dcadf153f113db83abc59e7466e69f14
74d975318a test: fix intermittent failure in p2p_sendtxrcncl.py (Martin Zumsande)
Pull request description:
`p2p_sendtxrcncl.py` currently fails intermittently in the CI, see e.g. https://cirrus-ci.com/task/5511952184115200?logs=ci#L4024
I believe that this is related to the reuse of the parameter `p2p_idx=2` of `add_outbound_p2p_connection` in this test: When we call `peer_disconnect`, we don't wait until the node has completed the disconnection. So there is a race between setting up the next connection (next `addconnection` RPC), and if the old one hasn't been removed and has an identical port like the new one (because we didn't increment `p2p_idx`), `CConnman::OpenNetworkConnection` just [returns](5274f32437/src/net.cpp (L1976)) without establishing a connection, and the test fails.
Fix this by using distinct `disconnect_p2ps` instead of `peer_disconnect`, which waits for the disconnect to complete. We can then use the same value for `p2p_idx` everywhere.
ACKs for top commit:
MarcoFalke:
review ACK 74d975318a
Tree-SHA512: f99f2550b6b320c0a2416a475c1cf189c009fce3a5abf1d4462486e1bfe309e2c3fd4228a4009b0ca38cb77465ce85e3d22298719eb07302fa0a72fbab0e0668
Currently the exact reason a transaction is removed from the mempool isn't
logged. It is sometimes detectable from context, but adding the `reason` to
the validation interface logs (where it is already passed) seems like an easy
way to disambiguate.
For example, in the case of mempool expiry, the logs look like this:
```
[validationinterface.cpp:220] [TransactionRemovedFromMempool] [validation] Enqueuing TransactionRemovedFromMempool: txid=<txid> wtxid=<wtxid>
[txmempool.cpp:1050] [RemoveUnbroadcastTx] [mempool] Removed <txid> from set of unbroadcast txns before confirmation that txn was sent out
[validationinterface.cpp:220] [operator()] [validation] TransactionRemovedFromMempool: txid=<txid> wtxid=<wtxid>
[validation.cpp:267] [LimitMempoolSize] [mempool] Expired 1 transactions from the memory pool
```
There is no context-free way to know $txid was evicted on the basis of expiry.
This change will make that case (and probably others) clear.
Using disconnect_p2ps instead of peer_disconnect makes
the node wait for the disconnect to complete. As a result,
we can reuse p2p_idx=0 in the add_outbound_p2p_connection calls.
29fa38a41a .python-version: bump patch version to 3.6.15 (Sjors Provoost)
Pull request description:
I'm unable to build Python 3.6.12 using PyEnv on macOS 13.0 as well as Ubuntu 22.10. Bumping the patch version fixes that issue on both systems.
A workaround is to add `.python-version` to your local git excludes and then do `pyenv local 3.6.15`, but this won't persist when you switch branches. Another workaround is to disable `pyenv`, but then you're potentially not running the test suite against the oldest supported Python version.
ACKs for top commit:
MarcoFalke:
concept ACK 29fa38a41a
Tree-SHA512: 60ecf76c957456fe4ece9e92c30866b27aba32d5bdb1696495281c2f7ef7a1fc30226e8992d55bf6f45e430b7588e05fbe49de7138aefab5597dd52bd4f021c0
This change allows to use the `test-{security,symbol}-check.py` scripts
when building out of source tree with no need to link scripts into the
build directory.
fa3ea81c3e refactor: Add LIFETIMEBOUND / -Wdangling-gsl to Assert() (MacroFake)
Pull request description:
Currently compiles clean, but I think it may still be useful.
Can be tested by adding an `&`:
```diff
diff --git a/src/test/util_tests.cpp b/src/test/util_tests.cpp
index 5766fff92d..300c1ec60f 100644
--- a/src/test/util_tests.cpp
+++ b/src/test/util_tests.cpp
@@ -125,7 +125,7 @@ BOOST_AUTO_TEST_CASE(util_check)
// Check -Wdangling-gsl does not trigger when copying the int. (It would
// trigger on "const int&")
- const int nine{*Assert(std::optional<int>{9})};
+ const int& nine{*Assert(std::optional<int>{9})};
BOOST_CHECK_EQUAL(9, nine);
}
```
Output:
```
test/util_tests.cpp:128:29: warning: object backing the pointer will be destroyed at the end of the full-expression [-Wdangling-gsl]
const int& nine{*Assert(std::optional<int>{9})};
^~~~~~~~~~~~~~~~~~~~~
./util/check.h:75:50: note: expanded from macro 'Assert'
#define Assert(val) inline_assertion_check<true>(val, __FILE__, __LINE__, __func__, #val)
^~~
1 warning generated.
ACKs for top commit:
jonatack:
ACK fa3ea81c3e
theuni:
ACK fa3ea81c3e
Tree-SHA512: 17dea4d75f2ee2bf6e1b6a6f6d8f439711c777df0390574e8d8edb6ac9ee807a135341e4439050bd6a15ecc4097a1ba9a7ab15d27541ebf70a4e081fa6871877
201b9a02fd test: fix intermittent failure in feature_index_prune.py (Martin Zumsande)
Pull request description:
I can't reproduce the error from #26630 locally, but from analying the logs I think the problem is the following:
After calling `sync_blocks`, we didn't check that the indexes have caught up to the tip before performing the manual pruning. This could possibly lead to prune blockers with a lower height than the expected 2489, which do appear in the logs of the failed CI runs, e.g.
- `2022-10-27T21:14:17.703920Z [C:\Users\ContainerAdministrator\AppData\Local\Temp\cirrus-ci-build\src\validation.cpp:2395] [FlushStateToDisk] [prune] coinstatsindex limited pruning to height 2488` ([Cirrus](https://cirrus-ci.com/task/5443742333665280?logs=functional_tests#L2506))
So, this should be fixed by a call to `sync_index`.
Fixes#26330
ACKs for top commit:
brunoerg:
crACK 201b9a02fd
Tree-SHA512: fb7023c9eb2ba6d0e69e059a401453cbdf63abc6804543dffcf36ba9f93c9cd13209e57aa5536d94b2e420c9d4cd0b1a7eff1adadd19aa7b3c33f592502e1bc0
fa24239a1c net: Avoid SetTxRelay for feeler connections (MacroFake)
Pull request description:
Seems odd to reserve memory for the struct (the heaviest member being `m_tx_inventory_known_filter`) when it is never used.
This also avoids sending out `msg_sendtxrcncl` before disconnecting. This shouldn't matter, as other messages, such as `msg_wtxidrelay`, `msg_sendaddrv2`, `msg_verack` or `msg_getaddr` are still sent. Though, it allows to test the changes here as a side-effect.
ACKs for top commit:
naumenkogs:
ACK fa24239a1c
vasild:
ACK fa24239a1c
jonatack:
ACK fa24239a1c
mzumsande:
ACK fa24239a1c
Tree-SHA512: d7604c7eb4df8f2de811e600bdd312440ee03e508d3a0f09ae79f7f2d3eeec663bfd47a2d079fa50b756d61e35dfa998de068a7b9afaf35378fa0e62a538263d
f1ee974e8e test: remove unused `CHANGE_{XPRV,XPUB}` constants (Sebastian Falbesoner)
Pull request description:
These constants exist since the introduction of the functional test wallet_taproot.py (2667366aaa), but they have never been used.
ACKs for top commit:
instagibbs:
ACK f1ee974e8e
brunoerg:
ACK f1ee974e8e
Tree-SHA512: 2c6f8cd537c35af59c4cbe6abb07386b7fb9e4662c7bed27ad8935e004f9bf452705fb1caef4f5ea1f4ad509745afba966005f95a18cf2b59e00b7739a5df7f8
