55555db055b59dd529526915dfc59e5a13e43160 doc: Add missing --platform=linux to docker build command (MarcoFalke)
fa0ce4c1486bb441e6e48d0a397334cf36cc8140 ci: Re-enable LINT_CI_SANITY_CHECK_COMMIT_SIG (MarcoFalke)
faa0973de2966a610e47ba4b6d6edf3c5509d52e ci: [refactor] Rename CIRRUS_PR env var to LINT_CI_IS_PR (MarcoFalke)
fa1dacaebe5d326ff8736ab9a4475f8a99ce4bc3 ci: Move lint exec snippet to stand-alone py file (MarcoFalke)
Pull request description:
The sanity check to check the last few merge commit signatures on the main branch was accidentally and silently disabled while moving from the `cirrus-ci.com` platform to the GHA platform.
So fix that by re-enabling it.
Also, contains a few other lint cleanup commits.
ACKs for top commit:
janb84:
re ACK 55555db055b59dd529526915dfc59e5a13e43160
willcl-ark:
ACK 55555db055b59dd529526915dfc59e5a13e43160
Tree-SHA512: e623dc88035ee4d1c6a8efa5fad33c35cface87f54e78c7ebfe5d468d28d8d8097150344d276f90f8ed52a89e61609ce95380476ea0151b50f73ad5919233933
99d012ec80a4415e1a37218fb4933550276b9a0a refactor: return reference instead of pointer (Andrew Toth)
f743e6c5dd386b7535e6c9442923a6ee54341994 refactor: add missing LIFETIMEBOUND annotation for parameter (Andrew Toth)
141117f5e8b41eb27539d217aa4e6c407c067d90 refactor: remove incorrect LIFETIMEBOUND annotations (Andrew Toth)
Pull request description:
The [developer-notes say](https://github.com/bitcoin/bitcoin/blob/master/doc/developer-notes.md#lifetimebound):
> You can use the attribute by adding a `LIFETIMEBOUND`
annotation defined in `src/attributes.h`; please grep the codebase for examples.
While grepping, I found an incorrect usage of the `LIFETIMEBOUND` annotation on `BlockManager::CheckBlockDataAvailability`. This could be misleading about usage for other greppers. As I was looking, I also noticed a missing `LIFETIMEBOUND` on `BlockManager::GetFirstBlock`. While looking more closely at that method, it should return a reference instead of a pointer. The only reason to return a pointer is if it can be null.
ACKs for top commit:
maflcko:
review ACK 99d012ec80a4415e1a37218fb4933550276b9a0a 💧
l0rinc:
ACK 99d012ec80a4415e1a37218fb4933550276b9a0a
stickies-v:
ACK 99d012ec80a4415e1a37218fb4933550276b9a0a
optout21:
ACK 99d012ec80a4415e1a37218fb4933550276b9a0a
vasild:
ACK 99d012ec80a4415e1a37218fb4933550276b9a0a
Tree-SHA512: d6c56ee223d6614d52ee6cf5cd66e787125c98c6ae37705a17e51a6e15774e260ac55b3d60f2fc818132e766ad98dd94232d6c8829119f628498e9d0d2bd977f
552eb90071fd246ba40037f74329403b72453047 doc: CI - Describe qemu-user-static usage (Hodlinator)
2afbbddee5504a77843b7b5963d98d02cde5af4f doc: CI - Clarify how important `env -i` is and why (Hodlinator)
Pull request description:
Should at least partially fix#31199
ACKs for top commit:
maflcko:
lgtm ACK 552eb90071fd246ba40037f74329403b72453047
janb84:
ACK 552eb90071fd246ba40037f74329403b72453047
Tree-SHA512: 45807a61d805646384c8162501f432537b7e655aa01434766ffb90ea47da9532387a76fcccac7fe208ad77f4ea5573f60b9be09e1235b9493eaa8795e1d7fbdd
2578e6fc0f4af35f389cd8ff59825c874e0b72ac test: Fix race condition in IPC interface block propagation test (Fabian Jahr)
Pull request description:
CI failed on this condition here: https://github.com/bitcoin/bitcoin/actions/runs/19395398994/job/55494696022?pr=33878#step:9:3983
The check was added not too long ago in https://github.com/bitcoin/bitcoin/pull/33745 and the fix here switches the check to the node which actually produces the block. There are also some comments added to make the checks easier so understand.
Closes#33884
ACKs for top commit:
Sjors:
re-utACK 2578e6fc0f4af35f389cd8ff59825c874e0b72ac
maflcko:
lgtm ACK 2578e6fc0f4af35f389cd8ff59825c874e0b72ac
Tree-SHA512: bfb7ae44aede50a00d4096e1a9922f9b8df31ce4242e12863e329d0d1e714d8cb46c852f694c32314e4bd26b524535e3a6967b7c57861a9b00cf09831a950b99
fa411f938e4796cf3806f234c9787b7c79492cc2 ci: Consistenly only cache on the default branch (MarcoFalke)
Pull request description:
Fixes https://github.com/bitcoin/bitcoin/issues/33685
The general idea for caches is to only save them on pushes to the default branch, because the cache is limited in size and time that the only benefit of the cache can be to speed up pull requests against the default branch.
Backport pull requests to older branches don't benefit from caches, because usually they will be running into a cache miss anyway. Also, they would cause the cache size to overflow and lead to cache misses down the line.
So fix it by consistently applying cache saves only on the default branch.
For reference, the same is already done for the composite action in this repo: 2444488f6a/.github/actions/save-caches/action.yml (L15)
ACKs for top commit:
hebasto:
ACK fa411f938e4796cf3806f234c9787b7c79492cc2.
willcl-ark:
ACK fa411f938e4796cf3806f234c9787b7c79492cc2
Tree-SHA512: 59d3de4201b596e5f7eb1915c4bd5ded80bcd4df217f8f5d62d92fa8977a77e2c4c8602b17755b33ff0dfa87f2896e7c1c8f4da8e6a77c111f2a42ec9cf78ffd
fae83611b8ef358ea7aca7070fd7e82dc06f9755 ci: [refactor] Use --preset=dev-mode in mac_native task (MarcoFalke)
fadb67b4b4e106cc1078172c5996fd6e8d93b4e2 ci: [refactor] Base nowallet task on --preset=dev-mode (MarcoFalke)
6666980e8653d98ef556f71a3e6907d3deda7147 ci: Enable bitcoin-chainstate and test_bitcoin-qt in win64 task (MarcoFalke)
faff7b231246ddd322211e22f636d08d3a45bd39 ci: Enable experimental kernel stuff in i686 task (MarcoFalke)
fa1632eecf5859af975102bb827a2a6f1dc161b2 ci: Enable experimental kernel stuff in mac-cross tasks (MarcoFalke)
fad10ff7c9235332f0e0496f6ee97960889a0241 ci: Enable experimental kernel stuff in armhf task (MarcoFalke)
fa9d67c13d0dd2641d42308507caedf782422b49 ci: Enable experimental kernel stuff in Alpine task (MarcoFalke)
fab3fb83026ef7770dac45f8a466ba7b19fd682d ci: Enable experimental kernel stuff in s390x task (MarcoFalke)
fa7da8a646ede418b823603ef981e112f9de3c56 ci: Enable experimental kernel stuff in valgrind task (MarcoFalke)
fa9c2973d60bca7ff69ee3b99dbdfe4b5ef32e9d ci: Enable experimental kernel stuff in TSan task (MarcoFalke)
fad30d4395022fef7cc4d09d26209e07b68ce29b ci: Enable experimental kernel stuff in MSan task (MarcoFalke)
Pull request description:
Most of the CI tasks have a long list of stuff that they enable. This makes it hard to see what each CI task is actually running.
Also, most of the CI tasks should probably mimic the `dev-mode` CMake preset and run on as much stuff as possible. Usually, changing the `dev-mode` comes with changing those CI tasks as well in the same commit, which is verbose.
Fix both issues, by basing most CI tasks on the `dev-mode`. In the future, this makes it easier to change the `dev-mode` in a single place. If CI tasks explicitly disable something, it will be listed explicitly in them.
As a side-effect this will enable the kernel stuff for some CI task that did not have it enabled, which seems desirable.
ACKs for top commit:
TheCharlatan:
Nice, ACK fae83611b8ef358ea7aca7070fd7e82dc06f9755
janb84:
ACK fae83611b8ef358ea7aca7070fd7e82dc06f9755
hebasto:
ACK fae83611b8ef358ea7aca7070fd7e82dc06f9755, I have reviewed the code and it looks OK.
Tree-SHA512: 58d9d553437b57362e9ec0766bd202482435f263d3f4c6ee7020c5e1e5ba69f8c064630423424f9d754254a66981e670b964a5aee58ef87f30b7d775642255be
Commit b3bf18f0bac0ffe18206ee20642e11264ba0c99d changed the function
signature from Parse(const std::string& descriptor,...) to
Parse(std::span<const char> descriptor,...).
Calling this new version of Parse with a string literal will trigger
a confusing "Invalid characters in payload" due to the trailing "\0".
Switch to string_view and add a test.
Co-authored-by: MarcoFalke <*~=`'#}+{/-|&$^_@721217.xyz>
chain_start can never be null, so it's better to pass it as a reference
rather than a raw pointer
Also slightly reformat HeaderSyncState constructor to make clang-format
happy
Lastly, remove `const` from `chain_start` declaration in
headers_sync_chainwork_tests, to work aroud a false-positive
dangling-reference warning in gcc 13.0
Co-Authored-By: maflcko <6399679+maflcko@users.noreply.github.com>
fa1bf6818f0910f997e5235a197ff51f2e18780d clang-format: Set InsertNewlineAtEOF: true (MarcoFalke)
Pull request description:
Now that the minimum supported clang version is 17, the `InsertNewlineAtEOF` setting can be set to `true` in the clang-format file. (https://releases.llvm.org/16.0.0/tools/clang/docs/ClangFormatStyleOptions.html#insertnewlineateof)
This is in line with the already existing newline linter. Can be tested via:
```
truncate --size=-1 src/init.cpp
git diff
# Should fail:
cargo run --manifest-path ./test/lint/test_runner/Cargo.toml -- --lint=trailing_newline
# Restore newline:
git diff -U0 | ./contrib/devtools/clang-format-diff.py -p1 -i -v
```
ACKs for top commit:
l0rinc:
ACK fa1bf6818f0910f997e5235a197ff51f2e18780d
achow101:
ACK fa1bf6818f0910f997e5235a197ff51f2e18780d
hodlinator:
ACK fa1bf6818f0910f997e5235a197ff51f2e18780d
stickies-v:
ACK fa1bf6818f0910f997e5235a197ff51f2e18780d
hebasto:
ACK fa1bf6818f0910f997e5235a197ff51f2e18780d.
janb84:
ACK fa1bf6818f0910f997e5235a197ff51f2e18780d
Tree-SHA512: 6443f0f9e2710fb31c85e4ece5deb9fa25bf22ef2d5cadad2d54a239194860c3a785ffa5dd97be942e42139336575425499f085aa2e3f29003e529fdc20fce20
6657bcbdb4d0359c1843ca31fb3670c7c0c260d5 kernel: allow null data_directory (stickies-v)
Pull request description:
An empty path may be represented with a `nullptr`. For example, `std::string_view{}.data()` may return nullptr.
Removes the `BITCOINKERNEL_ARG_NONNULL` attribute for `btck_chainstate_manager_options_create` 's `data_directory` parameter, and instead handles such null arguments in the implementation. [Because an empty path is meaningless](https://github.com/bitcoin/bitcoin/pull/33867#discussion_r2523930442), `btck_chainstate_manager_options_create` now treats both null and empty directories as invalid, tightening the interface.
Also documents how `BITCOINKERNEL_ARG_NONNULL` should be used.
Follow-up to https://github.com/bitcoin/bitcoin/pull/33853#pullrequestreview-3454620265
ACKs for top commit:
stringintech:
ACK 6657bcb
maflcko:
review ACK 6657bcbdb4d0359c1843ca31fb3670c7c0c260d5 🐪
achow101:
ACK 6657bcbdb4d0359c1843ca31fb3670c7c0c260d5
TheCharlatan:
ACK 6657bcbdb4d0359c1843ca31fb3670c7c0c260d5
janb84:
ACK 6657bcbdb4d0359c1843ca31fb3670c7c0c260d5
Tree-SHA512: 11c02b221ff19a5357e94355808e3b503b3a336c16fc5186c9c9137931709e880383ed1f4990fc4cc6b0e23961e2e1e03fc90154a3b546b9490ef66bd63688b7
7f318e1dd0496384e7bc6d8754c5a2a618a14b2a test: Add better coverage for Autofile size() (Fabian Jahr)
b7af960eb82f5e3af530014a4f59c48faa3a4109 refactor: Add AutoFile::size (Fabian Jahr)
ec0f75862e6791eade487f0dc7d4d0a8c91b9ba5 refactor: Modernize logging in util/asmap.cpp (Fabian Jahr)
606a251e0a3146eb2d2beb7cef486b2d46a37b40 tests: add unit test vectors for asmap interpreter (Pieter Wuille)
Pull request description:
This contains some commits from #28792 that can be easily reviewed and merged independently. I hope splitting this change off can make this part move a bit faster and reduce frequency of needed rebases for #28792.
The commits in order:
- Add additional unit test vectors to the asmap interpreter (written by sipa). This helps to ensure that the further refactors in #28792 don't change behavior.
- Modernizes the logging in `util/asmap.cpp`, I added this while touching the rest of the file all over anyway.
- Adds an `AutoFile::size` helper function with some additional test coverage in a separate commit
ACKs for top commit:
maflcko:
review ACK 7f318e1dd0496384e7bc6d8754c5a2a618a14b2a 🏀
hodlinator:
tACK 7f318e1dd0496384e7bc6d8754c5a2a618a14b2a
laanwj:
Code review ACK 7f318e1dd0496384e7bc6d8754c5a2a618a14b2a
Tree-SHA512: 45156b74e4bd9278a7ec24521dfdafe4dab1ba3384243c7d589ef17e16ca374ee2af7178c86b7229e80ca262dbe78c4d456d80a6ee742ec31d2ab5243dac8b57
310e4979b36cbcf1e9e01dd90c14e2e9997343a0 qt: Added test coverage for qt gui#901 console history filter (WakeTrainDev)
Pull request description:
Add test coverage for the QT rpc console updated filtered commands in gui#901
ACKs for top commit:
pablomartin4btc:
ACK 310e4979b36cbcf1e9e01dd90c14e2e9997343a0
hebasto:
ACK 310e4979b36cbcf1e9e01dd90c14e2e9997343a0, tested on Fedora 42 by reverting 4e352efa2ce756c668664486c99d003eef530e0c.
Tree-SHA512: 45bb8583311f145353d8265d28f220d2a318c701346f147979c5d33b27811276d5e18586bf58f35e455701495d2cb87ec54dd78f4ca8631a0c7bd2c1d7fe640c
929f69d0ff29cb803769a423035fdcf675f40b78 qt: Remove HD seed reference from blank wallet tooltip (John Moffett)
Pull request description:
Blank descriptor wallets currently do not have HD seeds and none can be added (or 'set') by the user, so remove the reference in the tooltip.
As I understand it, descriptor wallets don't have a global HD seed and don't store the HD seeds for keys they generate. Currently, no new HD seeds can be added by the user (even for old wallets since `sethdseed` was removed), though it may be possible in the future, eg - https://github.com/bitcoin/bitcoin/pull/33043
ACKs for top commit:
maflcko:
lgtm ACK 929f69d0ff29cb803769a423035fdcf675f40b78
Tree-SHA512: 85e9c8e18a92b3da6fd62b70570cef58e03843633300b65aee5789d38c7bcaa46738970f0aea63f4e9b3e8814abb5bf1e1aa93f568a875ad1e0443d4dafb0aab
28a4fcb03c0fb1cd5112eca1eb36dcb13e0b4ff2 test: check listdescriptors do not return a mix of hardened derivation marker (pythcoiner)
975783cb79e929260873c1055d4b415cd33bb6b9 descriptor: account for all StringType in MiniscriptDescriptor::ToStringHelper() (pythcoiner)
Pull request description:
In `MiniscriptDescriptor::ToStringHelper()` only the `StringType::Private` variant of the `type` argument was handled. This PR implements serializing w/ all variants of `StringType` & add a functional test for the descriptor triggering the related issue.
Closes#31694: previously when calling `listdescriptors` RPC on a wallet containing a taproot descriptor w/ a (miniscript) taptree, origins of internal key & taptree were serialized w/ differents hardened derivation markers:
- origin of the internal key were serialized w/ `StringType::Normalized` type (using `h` as marker)
- origins of taptree keys were serialized w/ `StringType::Private` type (using `'` as marker)
Note: Origins in segwit (`wsh()`) miniscript descriptors were also serialized w/ `StringType::Private` type (`'` marker) and are now serialized w/ `StringType::Normalized` type (`h` marker).
ACKs for top commit:
sipa:
Code review ACK 28a4fcb03c0fb1cd5112eca1eb36dcb13e0b4ff2
achow101:
ACK 28a4fcb03c0fb1cd5112eca1eb36dcb13e0b4ff2
rkrux:
Concept ACK 28a4fcb03c0fb1cd5112eca1eb36dcb13e0b4ff2
Tree-SHA512: 15d14000b5951ca69a64a05b9a0b138c48a07b81eaf2fa86b91ac20cc8735533355a787363c64ba88403dd8a56ef5232cba57d34bea80835a0f40774d62fbc2b
f53dbbc5057b6f676db4be9bc720898149f293fc test: Add functional tests for named argument parsing (zaidmstrr)
694f04e2bd34f994d81e27b68e4d7466a9a319f8 rpc: Handle -named argument parsing where '=' character is used (zaidmstrr)
Pull request description:
Addresses [comment](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2091886628) and [this](https://github.com/bitcoin/bitcoin/pull/31375#discussion_r2092039999).
The [PR #31375](https://github.com/bitcoin/bitcoin/pull/31375) got merged and enables `-named` by default in the `bitcoin rpc` interface; `bitcoin rpc` corresponds to `bitcoin-cli -named` as it's just a wrapper. Now, the problem arises when we try to parse the positional paramater which might contain "=" character. This splits the parameter into two parts first, before the "=" character, which treats this as the parameter name, but the other half is mostly passed as an empty string. Here, the first part of the string is an unknown parameter name; thus, an error is thrown. These types of errors are only applicable to those RPCs which might contain the `=` character as a parameter. Some examples are `finalizepsbt`, `decodepsbt`, `verifymessage` etc.
This is the one example of the error in `finalizepsbt` RPC:
```
./bitcoin-cli -named -regtest finalizepsbt cHNidP8BAJoCAAAAAqvNEjSrzRI0q80SNKvNEjSrzRI0q80SNKvNEjSrzRI0AAAAAAD9////NBLNqzQSzas0Es2rNBLNqzQSzas0Es2rNBLNqzQSzasBAAAAAP3///8CoIYBAAAAAAAWABQVQBGVs/sqFAmC8HZ8O+g1htqivkANAwAAAAAAFgAUir7MzgyzDnRMjdkVa7d+Dwr07jsAAAAAAAAAAAA=
error code: -8
error message:
Unknown named parameter cHNidP8BAJoCAAAAAqvNEjSrzRI0q80SNKvNEjSrzRI0q80SNKvNEjSrzRI0AAAAAAD9////NBLNqzQSzas0Es2rNBLNqzQSzas0Es2rNBLNqzQSzasBAAAAAP3///8CoIYBAAAAAAAWABQVQBGVs/sqFAmC8HZ8O+g1htqivkANAwAAAAAAFgAUir7MzgyzDnRMjdkVa7d+Dwr07jsAAAAAAAAAAAA
```
This PR fixes this by updating the `vRPCConvertParams` table that identifies parameters that need special handling in `-named` parameter mode. The parser now recognises these parameters and handles strings with "=" char correctly, preventing them from being incorrectly split as parameter assignments.
ACKs for top commit:
ryanofsky:
Code review ACK f53dbbc5057b6f676db4be9bc720898149f293fc. Just applied comment & test suggestions since last review
kannapoix:
Code review ACK: f53dbbc5057b6f676db4be9bc720898149f293fc
achow101:
ACK f53dbbc5057b6f676db4be9bc720898149f293fc
Tree-SHA512: 1b517144efeff45a4c4256c27a39ddf187f1d6189d133402a45171678214a10ff2925c31edcfd556d67f85bd26d42f63c528b941b68c9880eab443f2c883e681
With the move from cirrus-ci to GHA, the CIRRUS_REPO_FULL_NAME env var
was always unset, never triggering the sanity check.
Fix this by introducing a new vendor-agnostic env var and setting it
properly.
The CIRRUS_PR env var was cirrus-specific and using a provider-agnostic
name makes more sense.
Also, enable pipefail, while touching this file.
This refactor is needed for the next commit.
Changes AddToMempool() helper to only apply changes if the mempool limits are
respected.
Fix package_rbf fuzz target to handle mempool policy violations
Calculating mempool ancestors for a new transaction should not be done until
after cluster size limits have been enforced, to limit CPU DoS potential.
Achieve this by reworking TRUC and RBF validation logic:
- TRUC policy enforcement is now done using only mempool parents of
new transactions, not all mempool ancestors (note that it's fine to calculate
ancestors of in-mempool transactions, if the number of such calls is
reasonably bounded).
- RBF replacement checks are performed earlier (which allows for checking
cluster size limits earlier, because cluster size checks cannot happen until
after all conflicts are staged for removal).
- Verifying that a new transaction doesn't conflict with an ancestor now
happens later, in AcceptSingleTransaction() rather than in PreChecks(). This
means that the test is not performed at all in AcceptMultipleTransactions(),
but in package acceptance we already disallow RBF in situations where a
package transaction has in-mempool parents.
Also to ensure that all RBF validation logic is applied in both the single
transaction and multiple transaction cases, remove the optimization that skips
the PackageMempoolChecks() in the case of a single transaction being validated
in AcceptMultipleTransactions().
