Age | Commit message (Collapse) | Author |
|
The goal of this improvement is to reduce
TopUp calls which can lead to timeouts.
|
|
|
|
Our `QT_STATICPLUGIN` macro is effectively equivalent to the Qt's
`QT_STATIC` macro. No need to handle both of them.
|
|
favour of transaction_identifier::FromHex()
f553e6d86fe114e90585ea6d4b8e345a209cae5d refactor: remove TxidFromString (stickies-v)
285ab50ace4c04209f331ccaf121152b977cc490 test: replace WtxidFromString with Wtxid::FromHex (stickies-v)
9a0b2a69c4d836d7382f6fe7703de40288fc7421 fuzz: increase FromHex() coverage (stickies-v)
526a87ba6b4f20592ad3c090b82e83ecff2107fc test: add uint256::FromHex unittest coverage (stickies-v)
Pull request description:
Since https://github.com/bitcoin/bitcoin/pull/30482/commits/fab6ddbee64e50d5e2f499aebca35b5911896ec4, `TxidFromString()` has been deprecated because it is less robust than the `transaction_identifier::FromHex()` introduced in [the same PR](https://github.com/bitcoin/bitcoin/pull/30482). Specifically, it tries to recover from length-mismatches, recover from untrimmed whitespace, 0x-prefix and garbage at the end, instead of simply requiring exactly 64 hex-only characters.
In this PR, `TxidFromString` is removed completely to clean up the code and prevent further unsafe usage. Unit and fuzz test coverage on `uint256::FromHex()` and functions that wrap it is increased.
Note: `TxidFromSring` allowed users to prefix strings with "0x", this is no longer allowed for `transaction_identifier::FromHex()`, so a helper function for input validation may prove helpful in the future _(this overlaps with the `uint256::uint256S()` vs `uint256::FromHex()` future cleanup)_. It is not relevant to this PR, though, besides the fact that this unused (except for in tests) functionality is removed.
The only users of `TxidFromString` are:
- `test`, where it is straightforward to drop in the new `FromHex()` methods without much further concern
- `qt` coincontrol. There is no need for input validation here, but txids are not guaranteed to be 64 characters. This is already handled by the existing code, so again, using `FromHex()` here seems quite straightforward.
Addresses @maflcko's suggestion: https://github.com/bitcoin/bitcoin/pull/30482#discussion_r1691826934
Also removes `WtxidFromString()`, which is a test-only helper function.
### Testing GUI changes
To test the GUI coincontrol affected lines, `regtest` is probably the easiest way to quickly get some test coins, you can use e.g.
```
alias cli="./src/bitcoin-cli -regtest"
cli createwallet "coincontrol"
# generate 10 spendable outputs on 1 address
cli generatetoaddress 10 $(cli -rpcwallet=coincontrol getnewaddress)
# generate 10 spendable outputs on another address
cli generatetoaddress 10 $(cli -rpcwallet=coincontrol getnewaddress)
# make previous outputs spendable
cli generatetoaddress 100 $(cli -rpcwallet=coincontrol getnewaddress)
```
ACKs for top commit:
maflcko:
re-ACK f553e6d86fe114e90585ea6d4b8e345a209cae5d 🔻
hodlinator:
ACK f553e6d86fe114e90585ea6d4b8e345a209cae5d
paplorinc:
ACK f553e6d86fe114e90585ea6d4b8e345a209cae5d
TheCharlatan:
Nice, ACK f553e6d86fe114e90585ea6d4b8e345a209cae5d
Tree-SHA512: c1c7e6ea4cbf05cf660ba178ffc4f35f0328f7aa6ad81872e2462fb91a6a22e4681ff64b3d0202a5a9abcb650c939561585cd309164a69ab6081c0765ee271ef
|
|
89b1d5c818a33440c0c497aa78721fb9b0af35a5 depends: Fix `zeromq` build on OpenBSD (Hennadii Stepanov)
Pull request description:
On the master branch @ 66e82dc90c598c9c42ff980693ef5367a845e1d0, the `zeromq` package fails to build on OpenBSD 7.5:
```
[ 19%] Building CXX object CMakeFiles/objects.dir/src/io_thread.cpp.o
/home/hebasto/bitcoin/depends/work/build/amd64-unknown-openbsd7.5/zeromq/4.3.5-df5b1b9f936/src/io_thread.cpp:14:22: error: static_cast from 'std::nullptr_t' to 'poller_t::handle_t' (aka 'int') is not allowed
_mailbox_handle (static_cast<poller_t::handle_t> (NULL))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.
```
This [regression](https://github.com/bitcoin/bitcoin/pull/29723#issuecomment-2261513105) was overlooked by me in https://github.com/bitcoin/bitcoin/pull/29723.
This PR fixes the issue by backporting an upstream commit from https://github.com/zeromq/libzmq/pull/4659.
ACKs for top commit:
theStack:
tACK 89b1d5c818a33440c0c497aa78721fb9b0af35a5
Tree-SHA512: 48d22ea99dfd44c5adf858c74e64082390da27b8ccad8c0d5a91d4dabfa3d12267cef98e4bb8c088e4cd0ec477c242cb1d47aace5c88cd86f796715bba957ed8
|
|
handshake
afd237bb5d85923273a69f7b45dc6aae6aa1680e [fuzz] Harness for version handshake (dergoegge)
a90ab4aec9cccd44d867842d6a1161d5758fcb2c scripted-diff: Rename lazily initialized bloom filters (dergoegge)
82de1bc478d54bea6125b459febfb2fb23929c61 [net processing] Lazily initialize m_recent_confirmed_transactions (dergoegge)
fa0c87f19c1eca47ee7052f3c988ff7273801ff3 [net processing] Lazily initialize m_recent_rejects_reconsiderable (dergoegge)
662e8db2d3efc651951315b295952a2eebb822cd [net processing] Lazily initialize m_recent_rejects (dergoegge)
Pull request description:
This adds a fuzzing harness dedicated to the version handshake. To avoid determinism issues, the harness creates necessary components each iteration (addrman, peerman, etc). A harness like this would have easily caught https://bitcoincore.org/en/2024/07/03/disclose-timestamp-overflow/.
As a performance optimization, this PR includes a change to `PeerManager` to lazily initialize various filters (to avoid large unnecessary memory allocations each iteration).
ACKs for top commit:
brunoerg:
ACK afd237bb5d85923273a69f7b45dc6aae6aa1680e
marcofleon:
Tested ACK afd237bb5d85923273a69f7b45dc6aae6aa1680e. I compared the coverage of `net_processing` from this harness to the `process_message` and `process_messages` harnesses to see the differences. This target hits more specific parts of the version handshake. The stability looks good as well, at about 94%.
glozow:
utACK afd237bb5d85923273a69f7b45dc6aae6aa1680e lazy blooms look ok
mzumsande:
Code Review ACK afd237bb5d85923273a69f7b45dc6aae6aa1680e
Tree-SHA512: 62bba20aec0cd220e62368354891f9790b81ad75e8adf7b22a76a6d4663bd26aedc4cae8083658a75ea9043d60aad3f0e58ad36bd7bbbf93ff1d16e317bf15cc
|
|
|
|
and link to closed PR
3cd24aadb25c1fbdca725c906e0fca79d090ebdd doc: remove obsolete mention and link to closed PR (Marnix)
Pull request description:
Remove the mention and link as the PR (https://github.com/bitcoin/bitcoin/pull/22341) is closed and the description is wrong/outdated anyway.
ACKs for top commit:
BrandonOdiwuor:
ACK 3cd24aadb25c1fbdca725c906e0fca79d090ebdd
tdb3:
ACK 3cd24aadb25c1fbdca725c906e0fca79d090ebdd
Tree-SHA512: 5cd97029337f0cdfe81b6be9401adc4fe51ae2868f8fcadcb03828531a38380a587c32840850a924b6428f62df7d20a1e16ef7414d4078e7bb2c4e359b1fae40
|
|
TxidFromString was deprecated due to missing 64-character length-check
and hex-check, replace it with the more robust Txid::FromHex.
|
|
The newly introduced Wtxid::FromHex is more robust and removes
the need for a WtxidFromString helper function
|
|
|
|
Simultaneously cover transaction_identifier::FromHex()
|
|
test again
fa2269dd65bac09f22c7cee8a2dfc46694e9a986 test: Try previously intermittently broken USDT test again (MarcoFalke)
Pull request description:
Seems fine to try it again, given that the infra changed in the meantime.
Should be trivial to disable again, on the first failure.
Ref: https://github.com/bitcoin/bitcoin/issues/27380#issuecomment-1637971779
ACKs for top commit:
fanquake:
ACK fa2269dd65bac09f22c7cee8a2dfc46694e9a986 - seems fine to re-try this now.
Tree-SHA512: e203625b49019517def8e92cc465dd947d5aa85a080d370aabce7a45442efc3900b5c3783f6cc15720f81e5aaa3ff193a26be3d33048f20a73cd7b1fce320118
|
|
This does not cause any issues, because CScript in the tests are const.
However, this change allows to enable the
"function-call-in-default-argument (B008)" lint rule.
|
|
`doc/design/libraries.md`
f70eb0eeefcd69aa233af4259f42b85763b9ff9b doc: Remove reference to resolved issue (Daniela Brozzoni)
b27ef8ec7f9a06318c7287e8cf245827a0731cc2 doc: Update issue reference for libbitcoinkernel (Daniela Brozzoni)
Pull request description:
- The discussion of libbitcoinkernel has moved from 24303 to 27587
- Issue 15732 has been resolved, removing it from the document
ACKs for top commit:
maflcko:
ACK f70eb0eeefcd69aa233af4259f42b85763b9ff9b
Tree-SHA512: 11b597d9710504010945aae66f7e488403895aa8e1e091f3a8f6737dc128a4fde185daff8d4709cbbb69f454d3a649c4217e82a6bfc8ee2b25c8a1c047b57f1b
|
|
|
|
-BEGIN VERIFY SCRIPT-
sed -i 's/m_recent_confirmed_transactions/m_lazy_recent_confirmed_transactions/g' $(git grep -l 'm_recent_confirmed_transactions')
sed -i 's/m_recent_rejects_reconsiderable/m_lazy_recent_rejects_reconsiderable/g' $(git grep -l 'm_recent_rejects_reconsiderable')
sed -i 's/m_recent_rejects/m_lazy_recent_rejects/g' $(git grep -l 'm_recent_rejects')
-END VERIFY SCRIPT-
|
|
Issue #15732 has been resolved, this commit removes it from
the WIP section of the libraries document.
|
|
|
|
|
|
|
|
This reverts commit faf8be7c32be00f660eba90d3f07313fb25d5d1c.
|
|
93fb0e7897000072ea790a91816aea876718ad27 kernel: Only setup kernel context globals once (TheCharlatan)
Pull request description:
The globals setup by the function calls when creating a new kernel context only need to be setup once. Calling them multiple times may be wasteful and has no apparent benefit.
Besides kernel users potentially creating multiple contexts, this change may also be useful for tests creating multiple setups.
ACKs for top commit:
stickies-v:
re-ACK 93fb0e7897000072ea790a91816aea876718ad27
maflcko:
ACK 93fb0e7897000072ea790a91816aea876718ad27 👝
tdb3:
re ACK 93fb0e7897000072ea790a91816aea876718ad27
Tree-SHA512: c8418c23b34883b9b6af2b93c48760a931c246c9190fae372fb808f573408d332f53ca43b9c783eef561c4a6681e2fb63f215c939b40a87d597c0518dabea22a
|
|
a6efc7e16ed23377705a0a6a3445cab0acfd7ccc test: fix intermittent failures in feature_proxy.py (Martin Zumsande)
Pull request description:
Fixes #29871
If addnode connections are made with v2transport and the peer immediately disconnects us, reconnections with v1 are scheduled. This could interfere with later checks depending on timing. Avoid this by using `v2transport=False` in the addnode rpc - this test isn't about the message layer anyway, so running it with v2 would add no value.
ACKs for top commit:
maflcko:
ACK a6efc7e16ed23377705a0a6a3445cab0acfd7ccc
tdb3:
cr re ACK a6efc7e16ed23377705a0a6a3445cab0acfd7ccc
Tree-SHA512: 39353a392e75e4c6257d971ceecb65fb76ec6d3b121a087869831c24b767a18f57e2ae2968da445c7fa731cb03053c90df37dd2cd6e86f786ad4121bc68ca235
|
|
ec5e294e4b830766dcc4a80add0613d3705c1794 test: fix constructor of msg_tx (Martin Zumsande)
Pull request description:
In python, if the default value is a mutable object (here: a class) it is shared over all instances, so that one instance being changed would affect others to be changed as well.
This was the source of #30543, and possibly various other intermittent bugs in the functional tests, see
https://github.com/bitcoin/bitcoin/issues/29621#issuecomment-1999298224.
Fixes #30543
Fixes #29621
Fixes #25128
ACKs for top commit:
sipa:
utACK ec5e294e4b830766dcc4a80add0613d3705c1794. I believe some linters even warn about doing this.
maflcko:
ACK ec5e294e4b830766dcc4a80add0613d3705c1794
vasild:
ACK ec5e294e4b830766dcc4a80add0613d3705c1794 :heart:
theStack:
ACK ec5e294e4b830766dcc4a80add0613d3705c1794
Tree-SHA512: a6204fb1a326de3f9aa965f345fd658f6a4dcf78731db25cc905ff6eb8d4eeb65d14cc316305eebd89387aec8748c57c3a4f4ca62408f8e5ee53f535b88b1411
|
|
|
|
|
|
zmq) in OpenBSD build docs
903def1ffd2903615de509d5869cadf5315f04e2 doc: mention optional dependencies (qrencode, zmq) in OpenBSD build docs (Sebastian Falbesoner)
Pull request description:
The wording is taken from the FreeBSD build docs.
Tested on OpenBSD 7.5. See the following links for the package names:
- https://openbsd.app/?search=libqrencode
- https://openbsd.app/?search=zeromq
Thanks to hebasto for noticing that this was missing.
ACKs for top commit:
maflcko:
review ACK 903def1ffd2903615de509d5869cadf5315f04e2
hebasto:
ACK 903def1ffd2903615de509d5869cadf5315f04e2, I can successfully build with the `libqrencode` and `zeromq` packages on my OpenBSD 7.5 installation.
Tree-SHA512: 955e4892948a7703627d304a41a774f7cca0e4c672bdfa0edf531587d6970444aa49195b0f6f531ce375c8e7c2af6bbfa1a12e0612ae7a65f3e454fb17958672
|
|
efc26826400762207cde9f23802cfe75a737963c
6ee000e56f7ceba82791bc9f6fca59328a2eb0ad guix: bump time-machine to efc26826400762207cde9f23802cfe75a737963c (fanquake)
cbeb2c20e1f8076a9789436816ccb72b59a471c3 guix: patch /gnu/store paths out of winpthreads (fanquake)
Pull request description:
Needed for https://github.com/bitcoin/bitcoin/issues/30210. This doesn't switch runtimes, because upstream is
still configured to use the old runtime. See:
https://git.savannah.gnu.org/cgit/guix.git/commit/?id=17188be0f723e00377b21b767f5447d7938a116e.
git-mimimal `2.45.1` -> `2.45.2`
Kernel Headers `6.1.92` -> `6.1.100`
LLVM `18.1.6` -> `18.1.8`
mingw-w64 `11.0.1` -> `12.0.0`
NSIS `3.09` -> `3.10`
patch `2.7.6` -> `2.7.6-0.f144b35`
ACKs for top commit:
TheCharlatan:
ACK 6ee000e56f7ceba82791bc9f6fca59328a2eb0ad
Tree-SHA512: f4f99d16dd8cab5b2b7c5d82111af86de20e1669cc3b4054d72ab4a67b2956757df170f0df28c96d18653c1c9d2ebdd0ecc441005726a20cd963d98513b4a851
|
|
This removes the default value, because there should not be a use-case
to fall back to a an empty leaf_script by default. (If there was, it
could trivially be added back)
|
|
In python, if the default value is a mutable object (here: a class)
its shared over all instances, so that one instance being changed
would affect others to be changed as well.
This was likely the source of various intermittent bugs in the
functional tests.
|
|
The wording is taken from the FreeBSD build docs.
See the following links for the package names:
- https://openbsd.app/?search=libqrencode
- https://openbsd.app/?search=zeromq
Thanks to hebasto for noticing that this was missing.
|
|
|
|
|
|
|
|
|
|
|
|
Only standard when non-nested.
|
|
These outputs are called anchors, and allow
key-less anchor spends which are vsize-minimized
versus keyed anchors which require larger outputs
when creating and inputs when spending.
|
|
41a1a8615dd48fdd9811b9824c49ceb934c6375e gui: Hide peers details (@RandyMcMillan)
Pull request description:
Add a close (X) button to the Peers Detail panel.
Reuse the same icon used in the Console Tab.
The close button deselects the peer highlighted
in the PeerTableView and hides the detail panel.
fixes #485
Co-authored-by: @w0xlt <w0xlt@users.noreply.github.com>
ACKs for top commit:
pablomartin4btc:
re ACK 41a1a8615dd48fdd9811b9824c49ceb934c6375e
hebasto:
ACK 41a1a8615dd48fdd9811b9824c49ceb934c6375e, tested on Ubuntu 23.10.
Tree-SHA512: fc692891eec61bd1e6878f2433b478de3c69bf0b3ce3471f2faafda6f63d371e2cc125ae8290fd2ac3e4d8659031b79d85665318cfc5a9481e967ef99d245f9c
|
|
The globals setup by the function calls when creating a new kernel
context only need to be setup once. Calling them multiple times may be
wasteful and has no apparent benefit.
Besides kernel users potentially creating multiple contexts, this change
may also be useful for tests creating multiple setups.
Co-authored-by: stickies-v <stickies-v@protonmail.com>
|
|
The diff is generated by executing `make -C src translate`.
|
|
|
|
The diff is generated by executing the `update-translations.py` script.
Removals of translation files have been discarded.
|
|
If addnode connections are made with v2transport and the peer immediately disconnects us, reconnections
with v1 are scheduled. This could interfere with later checks depending on timing. Avoid this by using
`v2transport=False` in the addnode rpc - this test isn't about the message layer anyway, so running it
with v2 would add no value.
|
|
The discussion of libbitcoinkernel has moved from #24303 to #27587,
this commit updates the documentation accordingly.
|
|
bda537f7c484d61a55a72c5856f2d4420664db8f depends: remove ENV unsetting for darwin (fanquake)
1807760f094b600fdebe4ebd0018e4fd5d506ce9 guix: improve ENV unsetting for macOS (fanquake)
0b2aeee21d24645394d70086e992408948952952 depends: patch explicit -lm usage out of Qt tools (fanquake)
Pull request description:
Now that we use the native compiler, and have fixed Qt, and these vars
are (almost) unset in Guix, we can remove the unsetting from our compiler
command here.
I couldn't manage to make a darwin-clang-cross only exclusion of `-lm` work properly
for Qt, so opted for just removing the explicit link entirely. I do not think this should have
any other unwanted side-effects.
Fixes #21552.
ACKs for top commit:
TheCharlatan:
ACK bda537f7c484d61a55a72c5856f2d4420664db8f
Tree-SHA512: 97a2d85de7d4b1d65717ecb521399ecba5f53863b8aef21af62ede5ceee59ee1a9392663da3a3852cad1b6d8b420dd4b0b5f0eea38d30a81785d8b2718620b5f
|
|
This resolves the compiler warning about potential uninitialized
use of window_tx_count introduced in fa2dada.
|
|
93ee17c1d610f811e86fcf17db00df24f7de8c1b ci: enable berkley db on test each commit job (Max Edwards)
Pull request description:
As the "test each commit" job installs `libdb++-dev` it looks like it was intended that it would compile with Berkeley DB support.
This PR enables it.
CI run log with the change: https://github.com/m3dwards/bitcoin/actions/runs/10142921800/job/28043223197?pr=1
ACKs for top commit:
maflcko:
ACK 93ee17c1d610f811e86fcf17db00df24f7de8c1b
danielabrozzoni:
ACK 93ee17c1d610f811e86fcf17db00df24f7de8c1b
hebasto:
ACK 93ee17c1d610f811e86fcf17db00df24f7de8c1b.
Tree-SHA512: b6c2a7cea104a84221814fd121fd4fef2d7c0c3717f8c3fe35ec3c42e850e72085e9e6407b61ca1f9e86571346fa33a2cdd924210b26155956835011775320b2
|
|
|