a2a35b58cb doc: fix LIBRARY_PATH comment (fanquake)
Pull request description:
Now that we build capnp, qt isn't the only native package.
ACKs for top commit:
l0rinc:
code review ACK a2a35b58cb
willcl-ark:
ACK a2a35b58cb
Tree-SHA512: 7b4a54b53b4c673e31dea011e569c6ef9e7238a56a7f09c7f5a7bc521fb0ffdda3794f2b65764908005333a30f7c8c7b4d6b45d985bc92a70dc2daf8c3d1ff3d
The std::move in coinstatsindex was not necessary since it was passed as a const reference argument.
The other change in the utxo supply fuzz test changes a line that seems to have triggered a false alarm.
The coinstatsindex currently looks for block data at a hash key if the prev block in CustomAppend is different than expected. This is not needed since base index should always prevent us ending up in this scenario since it should rewind the index before calling CustomAppend in this case. But even if we run into this and our belt-and-suspenders code is getting hit, the index could not recover properly from the hash key index data so it can be removed without any real impact.
This is practically irrelevant due to the unlikeliness of a re-org
reaching so deep that it would drop the BIP30 blocks from the chain
(91842 and 91880). However this serves as documentation and ensures that
the functions RevertBlock and CustomAppend are consistent.
The index originally stored cumulative values in a CAmount type but this allowed for
potential overflow issues which were observed on Signet. Fix this by
storing the values that are in danger of overflowing in a arith_uint256.
Also turns an unnecessary copy into a reference in RevertBlock and
CustomAppend and gets
rid of the explicit total unspendable tracking which can be calculated
by adding the four categories of unspendables together.
a341e11ac9 ci: test IPC on additional hosts (Sjors Provoost)
6aee573bfc ci: enable IPC tests in CI (Pieter Wuille)
8d2ee88fa2 tests: add functional tests for IPC interface (Pieter Wuille)
3cc9a06c8d test: Add TestNode ipcbind option (Ryan Ofsky)
3cceb60a71 test: Provide path to `bitcoin` binary (Ryan Ofsky)
8c7f005629 test: add is_ipc_compiled() and skip_if_no_ipc() functions (Ryan Ofsky)
Pull request description:
This adds support to the functional test framework to run the multiprocess `bitcoin-node` binary, and then tests it in a new `interface_ipc.py` functional test through the `pycapnp` module.
ACKs for top commit:
Sjors:
ACK a341e11ac9
ryanofsky:
Code review ACK a341e11ac9. Changes since last review: rebasing, switching to miniwallet and expanding wallet test, improving pycapnp install steps in instructions and CI.
TheCharlatan:
ACK a341e11ac9
Tree-SHA512: 98330283cf0d66d5537eec1219345b8aec5740dbc2171bd8b70680d7a282e0962fcdf0588a75518110761e8bc95af57d6097c93937cac710805b0df10837093c
Install pycapnp on all (active) CI hosts which have IPC enabled and
run the functional tests.
Except for previous_releases, which uses an older version of pip
that doesn't support --break-system-packages.
With this change, tests can specify `self.extra_init = [{ipcbind: True}]` to
start a node listening on an IPC socket, instead of needing to choose which
node binary to invoke and what `self.extra_args=[["-ipcbind=..."]]` value to
pass to it.
The eliminates boilerplate code #30437 (interface_ipc_mining.py), #32297
(interface_ipc_cli.py), and #33201 (interface_ipc.py) previously needed in
their test setup.
Set new `BitcoinTestFramework.binary_paths.bitcoin_bin` property with path to
the `bitcoin` wrapper binary. This allows new tests for `bitcoin-mine` in
#30437 and `bitcoin-cli` in #32297 to find the `bitcoin` binary and call
`bitcoin -m` to start nodes with IPC support. This way the new tests can run
whenever the ENABLE_IPC build option is enabled, instead of only running when
the `BITCOIN_CMD` environment variable is set to `bitcoin -m`
61ec8866c6 [doc] archive v29.1 release notes (glozow)
Pull request description:
Copied from https://github.com/bitcoin/bitcoin/blob/v29.1/doc/release-notes.md
This is needed for announcement links and so we can see historical release notes in master.
ACKs for top commit:
l0rinc:
review ACK 61ec8866c6
Tree-SHA512: da9692c8cd8de54e848caab19da41975e8e75049b4fd3e1c6475ee86bf9947132597ceb4bf2e217710a73178b54c05b8f27668c67da202ba5fb1799b582fb15d
c9d5f211c1 depends: strip when installing qt (fanquake)
Pull request description:
Otherwise we end up with ~1.5GB of binaries (Linux) when `DEBUG=1`. This isn't great generally, but is worse in the CI, where disk may be limited (#33293).
ACKs for top commit:
TheCharlatan:
ACK c9d5f211c1
hebasto:
ACK c9d5f211c1.
Tree-SHA512: bf83e0d8c41c64aaa6d841e24c4f25bbe33034ae54a32f34ca14aca59eaa1a004809d48acf171414ed43b99f7d3d1f4b973aee0b272475bd7cc2ca708718b8da
Currently, if the user inadvertently starts the node with duplicate bind options,
such as `-bind=0.0.0.0 -bind=0.0.0.0`, it will cause a fatal error with the
misleading message "Bitcoin Core is probably already running".
This commit adds early validation to detect duplicate bindings across all binding
configurations (-bind, -whitebind, and onion bindings) before attempting to bind.
When duplicates are detected, the node terminates with a clear, specific error
message: "Duplicate binding configuration for address <addr>. Please check your
-bind, -bind=...=onion and -whitebind settings."
The validation catches duplicates both within the same option type (e.g.,
`-bind=X -bind=X`) and across different types (e.g., `-bind=X -whitebind=Y@X`),
helping users identify and fix configuration mistakes.
4f1a4cbccd net: Quiet down logging when router doesn't support natpmp/pcp (laanwj)
Pull request description:
When the router doesn't support natpmp and PCP, one'd normally expect the UDP packet to be ignored, and hit a time out. This logs a message that is already in the debug category. However, there's also the case in which sending an UDP packet causes a ICMP response (type 3, code 3 "port unreachable"). This is returned to user space as "connection refused" (despite UDP having no concept of connections).
Move the warnings from `Send` and `Recv` to debug level too, to reduce log spam in that case.
Closes#33301.
ACKs for top commit:
willcl-ark:
utACK 4f1a4cbccd
sipa:
utACK 4f1a4cbccd
davidgumberg:
Tested ACK 4f1a4cbccd
achow101:
ACK 4f1a4cbccd
darosior:
utACK 4f1a4cbccd
mzumsande:
utACK 4f1a4cbccd
Tree-SHA512: 2c99a5679720482ece47af33616b6b207509fb58ba1962a1c2d30f8d0e68554f8f5ef25224313d93f4c5a1cc702183fcf8e6119abc411209c9884119ef680aad
The warnings are false positive and have been fixed upstream.
See: https://github.com/capnproto/capnproto/pull/2334.
This change disables the `UndefinedBinaryOperatorResult` clang-tidy
check for source files generated by the `mpgen` tool.
When the router doesn't support natpmp and PCP, one'd normally expect
the UDP packet to be ignored, and hit a time out. This logs a warning
that is already in the debug category. However, there's also the case in
which sending an UDP packet causes a ICMP response. This is returned to
user space as "connection refused" (despite UDP having no concept of
connections).
Move the warnings from `Send` and `Recv` to debug level too, to reduce
log spam in that case.
Closes#33301.
fae610d858 ci: Remove redundant RUN_UNIT_TESTS_SEQUENTIAL (MarcoFalke)
Pull request description:
`RUN_UNIT_TESTS_SEQUENTIAL` is useful to detect cases where global state is left dirty in the test process and leads to subsequent unit test cases failing. However, one CI task is sufficient to catch this.
As there already is one, add docs there and remove this env var (and extra logic).
ACKs for top commit:
fanquake:
ACK fae610d858
Tree-SHA512: b7ace1257d039f144cb0acb08d5d19d641028464517e6a2468e248ed79b2511512dc904867dacd66157b7483ec8041c95cce00f8ce3c89f3a2c3bb47939d7ff9
88db09bafe net: handle multi-part netlink responses (willcl-ark)
42e99ad773 net: skip non-route netlink responses (willcl-ark)
57ce645f05 net: filter for default routes in netlink responses (willcl-ark)
Pull request description:
...for default route in pcp pinholing.
Currently we only make a single recv call, which trucates results from large routing tables, or in the case the kernel may split the message into multiple responses (which may happen with `NLM_F_DUMP`).
We also do not filter on the default route. For IPv6, this led to selecting the first route with an `RTA_GATEWAY` attribute, often a non-default route instead of the actual default. This caused PCP port mapping failures because the wrong gateway was used.
Fix both issues by adding multi-part handling of responses and filter for the default route.
Limit responses to ~ 1MB to prevent any router-based DoS.
ACKs for top commit:
achow101:
ACK 88db09bafe
davidgumberg:
Code Review re-ACK 88db09b
Sjors:
re-utACK 88db09bafe
Tree-SHA512: ea5948edebfad5896a487a61737aa5af99f529fad3cf3da68dced456266948238a7143383847e79a7bb90134e023eb173c25116d8eb80ff57fa4c4a0377ca1ed
af4156ab75 build: set ENABLE_IPC to OFF when fuzzing (fanquake)
Pull request description:
A `BUILD_FOR_FUZZING` build will currently failure to configure, with missing `capnp`.
ACKs for top commit:
Crypt-iQ:
tACK af4156ab75
marcofleon:
ACK af4156ab75
dergoegge:
utACK af4156ab75
janb84:
ACK af4156ab75
Tree-SHA512: e3c5238cb5823116a958502eab84ee72a94cac0853fc3908ef97b6b6dc037db27806be0726f321d70ab706c37924dec526b46a3a46ea3f3f3684ce48da46a803
Handle multi-part netlink responses to prevent truncated results from
large routing tables.
Previously, we only made a single recv call, which led to incomplete
results when the kernel split the message into multiple responses (which
happens frequently with NLM_F_DUMP).
Also guard against a potential hanging issue where the code would
indefinitely wait for NLMSG_DONE for non-multi-part responses by
detecting the NLM_F_MULTI flag and only continue waiting when necessary.
m_cached_from_me is used to track whether a transaction is "from me", i.e. has
any inputs which belong to the wallet. This is held in memory only in
the same way that a transaction's balances are.
Instead of checking whether the total amount of inputs known by the
wallet is greater than 0, we should be checking for whether the input is
known by the wallet. This enables us to determine whether a transaction
spends an of output with an amount of 0, which is necessary for marking
0-value dust outputs as spent.
