b223c3c21e test: Add functional test for symlinked blocks directory (laanwj)
ddb75c2e87 test: Add fs_tests/create_directories unit test (Hennadii Stepanov)
1f46b6e46e util: Work around libstdc++ create_directories issue (laanwj)
Pull request description:
Work around libstdc++ issue [PR101510] with create_directories where the leaf already exists as a symlink. Fixes#24257, introduced by the switch to `std::filesystem`. It is meant to be more thorough than #24266, which worked around one instance of the problem.
The issue was [fixed upstream](https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=124eaa50e0a34f5f89572c1aa812c50979da58fc), but unfortunately we'll have to carry a fix for it for a while.
This introduces a function `fs::create_directories` which wraps
`std::filesystem::create_directories`. This allows easiliy reverting the
workaround when it is no longer necessary.
ACKs for top commit:
jonatack:
re-ACK b223c3c21e per `git range-diff df08250 67019cd b223c3c`
hebasto:
re-ACK b223c3c21e
w0xlt:
re-ACK b223c3c
vasild:
ACK b223c3c21e
Tree-SHA512: 028321717c8b10d16185c3711b35da6b05fb7aa31cee1c8c7e754e92bf5a0b02719a3785cd0f6f8bf052b3bd759f644af212320672baabc9e44e0b93ba464abc
eb8b22d517 block_connected: re-use previous GetTimeMicros (William Casarin)
80e1c55687 block_connected: don't serialize block hash twice (William Casarin)
Pull request description:
In the validation:block_connected tracepoint, we call block->GetHash(), which
ends up calling CBlockHeader::GetHash(), executing around 8000 serialization
instructions. We don't need to do this extra work, because block->GetHash() is
already called further up in the function. Let's save that value as a local
variable and re-use it in our tracepoint so there is no unnecessary tracepoint
overhead.
Shave off an extra 100 or so instructions from the validation:block_connected
tracepoint by reusing a nearby GetTimeMicros(). This brings the tracepoint down
to 54 instructions. Still high, but much better than the previous ~154 and
8000 instructions which it was originally.
Signed-off-by: William Casarin <jb55@jb55.com>
ACKs for top commit:
0xB10C:
ACK eb8b22d517
laanwj:
Code review ACK eb8b22d517
theStack:
re-ACK eb8b22d517
Tree-SHA512: 92ae585e487554e0f73042a8abaa239f630502c1d198e010bd7c1de252d882bccb627bbf0e4faec09c1253e782b145bcf153f9fee78cdb8456188044a96f8267
Work around libstdc++ issue [PR101510] with create_directories where the
leaf already exists as a symlink. Fixes#24257, introduced by the switch
to `std::filesystem`. It is meant to be more thorough than #24266, which
only worked around one instance of the problem.
The issue was fixed upstream in
https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=124eaa50e0a34f5f89572c1aa812c50979da58fc,
but unfortunately we'll have to carry a fix for it for a while.
This introduces a function `fs::create_directories` which wraps
`std::filesystem::create_directories`. This allows easiliy reverting the
workaround when it is no longer necessary.
dc01cbc538 test: Add fs_tests/rename unit test (Hennadii Stepanov)
d4999d40b9 util: Revert back MoveFileExW call for MinGW-w64 (Hennadii Stepanov)
Pull request description:
Unfortunately, bitcoin/bitcoin#24308 introduced a [regression](https://github.com/bitcoin/bitcoin/pull/24308#issuecomment-1037259386) for mingw builds.
The root of the problem is a broken implementation of [`std::filesystem::rename`](https://en.cppreference.com/w/cpp/filesystem/rename). In particular, the expected behavior
> If `old_p` is a non-directory file, then `new_p` must be ... existing non-directory file: `new_p` _is first deleted_...
fails with the "File exists" error.
This PR reverts back the `MoveFileExW` call, and adds the [suggested](https://github.com/bitcoin/bitcoin/pull/24308#pullrequestreview-878832906) unit test.
ACKs for top commit:
vasild:
ACK dc01cbc538
Tree-SHA512: c8e5a98844cfa32bec0ad67a1aaa58fe2efd0c5474d3e83490211985b110f83245758a742dcaa0a933a192ab66a7f11807e0c53ae69260b7dd02fc99f6d03849
f485a07454 Add missing thread safety lock assertions in validation.h (Jon Atack)
37af8a20cf Add missing thread safety lock assertions in validation.cpp (Jon Atack)
Pull request description:
A number of functions in validation.{h,cpp} have a thread safety lock annotation in the declaration but are missing the corresponding run-time lock assertion in the definition.
ACKs for top commit:
hebasto:
re-ACK f485a07454, only suggested change since my [previous](https://github.com/bitcoin/bitcoin/pull/24177#pullrequestreview-877810465) review.
vasild:
ACK f485a07454
Tree-SHA512: c86c0c0e8fe6ec7ae9ed9890f1dd7d042aa482ecf99feb6679a670aa004f6e9a99f7bc047205a34968fab7f1f841898c59b48c3ed6245c166e3b5abbf0867445
bfcd60f5d5 test: activate all index types in feature_init.py (Martin Zumsande)
0243907fae index: Don't commit without valid m_best_block_index (Martin Zumsande)
Pull request description:
When an index thread receives an interrupt during init before it got to index anything (so `m_best_block_index == nullptr` still), it will still try to commit previous "work" before stopping the thread. That means that `BaseIndex::CommitInternal()` calls `GetLocator(nullptr)`, which returns an locator to the tip ([code](06b6369766/src/chain.cpp (L31-L32))), and saves it to the index DB.
On the next startup, this locator will be read and it will be assumed that we have successfully synced the index to the tip, when in reality we have indexed nothing.
In the case of coinstatsindex, this would lead to a shutdown of bitcoind without any indication what went wrong. For the other indexes, there would be no immediate shutdown, but the index would be corrupt.
This PR fixes this by not committing when `m_best_block_index==nullptr`, and it also adds an error log message to the silent coinstatsindex shutdown path.
This is another small bug found by `feature_init.py` - the second commit enables blockfilterindex and coinstatsindex for this test, enabling coinstatsindex without the first commit would have led to frequent failures.
ACKs for top commit:
fjahr:
reACK bfcd60f5d5
shaavan:
reACK bfcd60f5d5
Tree-SHA512: 8e2bac0fc40cde209518a9e59b597ae0a5a875a2a90898673987c91733718d40e528dada942bf552b58bc021bf46e59da2d0cc5a61045f48f9bae2b1baf6033b
f11dad22a5 test: refactor: remove unneeded bytes<->hex conversions in `byte_to_base58` (Sebastian Falbesoner)
Pull request description:
It seems like the only reason for using hex strings in this method was to have a convenient way to convert to an integer from the input data interpreted as big-endian. In Python3 we have `int.from_bytes(..., 'big')` for that purpose, hence there is no need for that anymore and we can simply operate on bytes only.
ACKs for top commit:
laanwj:
Code review ACK f11dad22a5
Tree-SHA512: 9b1563010066ca74d85139c3b9259e9a5bb49e1f141c30b6506a0445afddb2bde7fd421fdd917dc516956e66f93610e2c21d720817640daee8f57f803be76ee4
faef344f84 Print enable_fuzz_binary in configure (MarcoFalke)
Pull request description:
A *disabled* `enable_fuzz` on current master does *not* mean the the fuzz binary is not compiled. This is confusion, so fix it.
* `enable_fuzz` toggles compilation flags for fuzzing and disables all other target. There is no need to print this in the configure result, because the compilation flags are already printed. Also, all other targets are already printed as `no`.
* `enable_fuzz_binary` does what it says it does and is currently not printed. So print it.
ACKs for top commit:
hebasto:
ACK faef344f84, tested on Linux Mint 20.2 (x86_64):
Tree-SHA512: 9b02b05c4b9c5fc92cf3487497392690303c36eace5e217f18b4349f059b5a23a7c0e0d030fb6fa7bbad83e927576a5e81c00099164f9ed8e185c0969dc17689
fa455975e5 util: Add missing unlinkat to syscall sandbox (MarcoFalke)
Pull request description:
This will be needed for g++-12 (after libstdc++6 12-20220206).
Steps to reproduce:
```
gdb --args ./src/bitcoind -sandbox=log-and-abort -regtest
./src/bitcoin-cli -regtest -named createwallet wallet_name=a descriptors=false
./src/bitcoin-cli -regtest stop
```
BT:
```
Thread 1 "b-shutoff" received signal SIGSYS, Bad system call.
0x00007ffff79564f7 in unlinkat () from /lib/x86_64-linux-gnu/libc.so.6
(gdb) bt
#0 0x00007ffff79564f7 in unlinkat () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007ffff7cc7335 in ?? () from /usr/lib/x86_64-linux-gnu/libstdc++.so.6
#2 0x00007ffff7cc94e3 in std::filesystem::remove_all(std::filesystem::__cxx11::path const&) () from /usr/lib/x86_64-linux-gnu/libstdc++.so.6
#3 0x00005555559d4918 in wallet::BerkeleyEnvironment::Flush (this=0x7fffc4005160, fShutdown=<optimized out>) at /usr/include/c++/12/bits/fs_path.h:595
#4 0x000055555592c058 in wallet::StopWallets (context=...) at /usr/include/c++/12/bits/shared_ptr_base.h:1665
#5 0x00005555556617ca in Shutdown (node=...) at ./src/init.cpp:293
#6 0x000055555563ada6 in AppInit (argv=<optimized out>, argc=<optimized out>, node=...) at ./src/bitcoind.cpp:249
#7 main (argc=<optimized out>, argv=<optimized out>) at ./src/bitcoind.cpp:273
ACKs for top commit:
laanwj:
Code review ACK fa455975e5
Tree-SHA512: e80a38828f8656040954c9befa2d1c9d5170e204dc09c61031633349897f51ccd85cc5c99a089c4726d7f5237875cd9ed3fa8ef864cd6c1c8a2b8250b392d57f
64645fa3e0 Release process: fix broken link to Guix building docs (Jeremy Rand)
Pull request description:
Not 100% sure whether this link was always broken or if the Guix docs renamed the heading at some point. Either way, seems good to fix it.
ACKs for top commit:
fanquake:
ACK 64645fa3e0
Tree-SHA512: 4932059fe583c0d27c70febf8f4dd8cffd3e15567359c5429d2691e221afc6da319bf43ebcd264ae0f98302e1eeb67ffd763d3d7d06ab1633913555ee7461643
460fa8e0d9 test: remove `import socket` in test_ipv6_local (brunoerg)
Pull request description:
Since this module (`socket`) is imported at the top of file, there is no need to import it again within the function.
ACKs for top commit:
MarcoFalke:
cr ACK 460fa8e0d9
Tree-SHA512: 031c17a776dedaa21b3ec6458ca822304e76a5a3f4494406e6b7b04f08cc2abefcfe742c462b60c9b3e2fee3cd110a69ed5ad413357886dc7b823abc916ea40e
f730bd7d58 scripted-diff: Rename ShowModalDialogAndDeleteOnClose (Hennadii Stepanov)
5d7666b151 qt: Revert 7fa91e8312 partially (Hennadii Stepanov)
89c277a6fc qt: Delay shutdown while a modal dialog is active (Hennadii Stepanov)
8c0eb80f41 qt: Disable tray icon menu when a modal dialog is active (Hennadii Stepanov)
92427354dd qt, refactor: Use local QAction instances for the tray icon menu (Hennadii Stepanov)
58e16035c1 qt, refactor: Drop BitcoinGUI::{send,receive}CoinsMenuAction members (Hennadii Stepanov)
fd667e73cd qt: Make show_hide_action dependent on the main window actual state (Hennadii Stepanov)
ee151d0327 qt: Drop BitcoinGUI::toggleHideAction member (Hennadii Stepanov)
78189daac8 qt, refactor: Fill up trayIconMenu before connections (Hennadii Stepanov)
66afa286e5 qt, refactor: Replace BitcoinGUI::trayIconActivated with a lambda (Hennadii Stepanov)
c3ca8364b2 qt, refactor: Replace BitcoinGUI::macosDockIconActivated with a lambda (Hennadii Stepanov)
Pull request description:
As pointed in bitcoin/bitcoin#23790 a regression in wallet unlock was introduced in bitcoin-core/gui#336 when a synchronous `AskPassphraseDialog` has been replaced with an asynchronous one.
This PR reverts a call back to a synchronous mode.
To make synchronous dialogs behave nice during shutdown some additional changes were made.
Please note that disabling the tray icon menu when a modal dialog is active is useful itself as on master (4ad59042b3) it is possible to switch to the "Receive" tab while the GUI is waiting for a password for the "Send" tab:
This is confusing and must be avoided.
Fixesbitcoin/bitcoin#23790.
ACKs for top commit:
prayank23:
tACK f730bd7d58
Tree-SHA512: 2b68275754190e4a9831b96e882d3c5b005e03909aeb6f2c5846da07199bb3efbb74ce87a9d25bb139f643c43d377a2051b221d553281fa5aefdd3181a58077f
This comment described how the constructor of CFeeRate was previously
indirectly used to parse fee rate arguments from RPCs. The command line
input was actually in sat/vB but due to the use of AmountFromValue() it
got converted to BTC/vB. In the constructor this could be rectified by
creating a CFeeRate from that given value (in BTC/vB) and COIN as the
transaction size, turning the input effectively to sat/vB. Since this
usage pattern was removed from the codebase some months ago, the comment
is now obsolete.
Also:
• Use vsize and vbyte instead of size and byte
abc057c603 build: Add Boost.Process usage check (Hennadii Stepanov)
Pull request description:
This PR adds a check that Boost.Process can be used without linking any libraries (header-only).
Disable the functionality if that is not the case.
Fixesbitcoin/bitcoin#24314.
ACKs for top commit:
fanquake:
ACK abc057c603
Tree-SHA512: ed2a32b1f751ec6f88cc7220766edd4cdab93c1d0515c892aa3094ee8d5b13ef569830d6e7a7a00c0197b117585dc526d00d943cc99a1f8c8a66ac4e20fe2061
bcd36e14f0 build: correct depends FreeBSD C{XX}FLAGS (fanquake)
7b06ffce9c build: add NetBSD support to depends (fanquake)
Pull request description:
Similar to #23948. Doesn't build the Qt package; I haven't looked at doing that yet, but have an assumption that it's going to fail out of the box similar to the FreeBSD build.
Guix Build:
```bash
```
ACKs for top commit:
theuni:
ACK bcd36e14f0
Tree-SHA512: 9a0946cefbcb9a92dd730b885463f3213e304c8d4b39fea8d831fc013a73d2ef998ca84e384bf45a01fa1449cf5a35eaffaa5b57a9062c2cdda34312d33ec3fc
