95975dd08d sync: detect double lock from the same thread (Vasil Dimov)
4df6567e4c sync: make EnterCritical() & push_lock() type safe (Vasil Dimov)
Pull request description:
Double lock of the same (non-recursive) mutex from the same thread would produce an undefined behavior. Detect this from `DEBUG_LOCKORDER` and react similarly to the deadlock detection.
This came up during discussion in another, related PR: https://github.com/bitcoin/bitcoin/pull/19238#discussion_r442394521.
ACKs for top commit:
laanwj:
code review ACK 95975dd08d
hebasto:
re-ACK 95975dd08d
Tree-SHA512: 375c62db7819e348bfaecc3bd82a7907fcd8f5af24f7d637ac82f3f16789da9fc127dbd0e37158a08e0dcbba01a55c6635caf1d8e9e827cf5a3747f7690a498e
fa18e7cbc5 This change to the appveyor CI config for msvc builds reverses a change introduced in #19960. It re-applies a setting to inform vcpkg to only build release vesions of the dependencies rather than the default of debug and release. (Aaron Clauson)
Pull request description:
This change to the appveyor CI config for msvc builds reverses a change introduced in #19960. It re-applies a setting to inform vcpkg to only build release versions of the dependencies rather than the default of debug and release.
It had been expected that the vcpkg manifest mechanism introduced in #19960 would do this automatically but it turns out not to be the case.
ACKs for top commit:
MarcoFalke:
ACK fa18e7cbc5 if green
hebasto:
ACK fa18e7cbc5, AppVeyor build takes 5 minutes less.
Tree-SHA512: 427e7e78190c20e0d85dad9b29beed2b6fa13f99c6bc72bcc1839dfb51237a7cc785ab707b4f851c527c1bb0d3e7ebad9e640969e19d29778584bbaeec75cecf
e3e7446305 Add lifetimebound to attributes for general-purpose usage (Cory Fields)
1d58cc7cb0 span: add lifetimebound attribute (Cory Fields)
62733fee87 span: (almost) match std::span's constructor behavior (Cory Fields)
Pull request description:
Replaces #19382 with a different approach. See [this comment](https://github.com/bitcoin/bitcoin/pull/19382#discussion_r446332852) for the reasoning behind the switch.
--
Description from #19382:
See [here](http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2018/p0936r0.pdf) for more detail on lifetimebound.
This is implemented using preprocesor macros rather than configure checks in order to keep span.h self-contained.
The ```[[clang::lifetimebound]]``` syntax was chosen over ```__attribute__((lifetimebound))``` because the former is more flexible and works to guard ```this``` as well as function parameters, and also because at least for now, it's available only in clang.
There are currently no violations in our codebase, but this can easily be tested by inserting one like this somewhere and compiling with a modern clang:
```c++
Span<const int> bad(std::vector<int>{1,2,3});
```
The result:
> warning: temporary whose address is used as value of local variable 'bad' will be destroyed at the end of the full-expression [-Wdangling]
Span<const int> bad(std::vector<int>{1,2,3});
```
ACKs for top commit:
sipa:
ACK e3e7446305
ajtowns:
ACK e3e7446305 (drive by; only a quick skim of code and some basic sanity checks)
MarcoFalke:
review ACK e3e7446305🔗
jonatack:
ACK e3e7446 change since last review is adding `[[clang::lifetimebound]]` as `LIFETIMEBOUND` to src/attributes.h as suggested in https://github.com/bitcoin/bitcoin/pull/19387#issuecomment-650752959.
Tree-SHA512: 05a3440ee595ef0e8d693a2820b360707695c016a68e15df47c20cd8d053646cc6c8cca8addd7db40e72b3fce208879a41c8102ba7ae9223e4366e5de1175211
3eb6f8b2e6 wallet (not for backport): improve upgradewallet error messages (Jon Atack)
ca8cd893bb wallet: fix and improve upgradewallet error responses (Jon Atack)
99d56e3571 wallet: fix and improve upgradewallet result responses (Jon Atack)
2498b04ce8 Don't upgrade to HD split if it is already supported (Andrew Chow)
c46c18b788 wallet: refactor GetClosestWalletFeature() (Jon Atack)
Pull request description:
This follows up on #18836 and #20282 to fix and improve the as-yet unreleased `upgradewallet` feature and also implement review follow-up in https://github.com/bitcoin/bitcoin/pull/18836#discussion_r519328607.
This PR fixes 4 upgradewallet issues:
- this bug: https://github.com/bitcoin/bitcoin/pull/20403#discussion_r526063920
- it returns nothing in the absence of an RPC error, which isn't reassuring for users
- it returns the same thing both in the case of a successful upgrade and when no upgrade took place
- the error message object is currently dead code
This PR fixes the above and provides:
...user feedback to not silently return without upgrading
```
{
"wallet_name": "disable private keys",
"previous_version": 169900,
"current_version": 169900,
"result": "Already at latest version. Wallet version unchanged."
}
```
...better feedback after successfully upgrading
```
{
"wallet_name": "watch-only",
"previous_version": 159900,
"current_version": 169900,
"result": "Wallet upgraded successfully from version 159900 to version 169900."
}
```
...helpful error responses
```
{
"wallet_name": "blank",
"previous_version": 169900,
"current_version": 169900,
"error": "Cannot downgrade wallet from version 169900 to version 159900. Wallet version unchanged."
}
{
"wallet_name": "blank",
"previous_version": 130000,
"current_version": 130000,
"error": "Cannot upgrade a non HD split wallet from version 130000 to version 169899 without upgrading to support pre-split keypool. Please use version 169900 or no version specified."
}
```
updated help:
```
upgradewallet ( version )
Upgrade the wallet. Upgrades to the latest version if no version number is specified.
New keys may be generated and a new wallet backup will need to be made.
Arguments:
1. version (numeric, optional, default=169900) The version number to upgrade to. Default is the latest wallet version.
Result:
{ (json object)
"wallet_name" : "str", (string) Name of wallet this operation was performed on
"previous_version" : n, (numeric) Version of wallet before this operation
"current_version" : n, (numeric) Version of wallet after this operation
"result" : "str", (string, optional) Description of result, if no error
"error" : "str" (string, optional) Error message (if there is one)
}
```
ACKs for top commit:
achow101:
ACK 3eb6f8b
MarcoFalke:
review ACK 3eb6f8b2e6 🛡
Tree-SHA512: b767314069e26b5933b123acfea6aa40708507f504bdb22884da020a4ca1332af38a7072b061e36281533af9f4e236d94d3c129daf6fe5b55241127537038eed
e95aaefe25 build: Avoid secp256k1.h include from system (Niklas Gögge)
Pull request description:
While building i ran into an error because i had a version of `secp256k1.h` under `/usr/local/include` that was incompatible with the secp256k1 code in the repository. This caused a problem because `$(BOOST_CPPFLAGS)` contained `-I/usr/local/include` and the include paths are searched by the compiler in order from left to right, so in the end `$(BITCOIN_INCLUDES)` contained `-I/usr/local/include` before `-I$(srcdir)/secp256k1/include` which caused the compiler to find `secp256k1.h` under `/usr/local/include`.
Looking at git blame i am wondering how this has not happened to anyone else in several years: cb89e18845/src/Makefile.am (L25)
I am on macOS 10.15.
ACKs for top commit:
laanwj:
Code review ACK e95aaefe25
hebasto:
ACK e95aaefe25, tested on macOS 11 Big Sur by adding `#error` into `/usr/local/include/secp256k1.h`.
Tree-SHA512: 1f0b395725936c179ab60dee3582ec7b21e2f9c0f1895e160d84a487cf0db16d0c7aa47d05800e0aded31685b4362056cac9b9ecca1bb8c308a4c5a810e8dc1d
fa69c2c784 wallet: Do not treat default constructed types as None-type (MarcoFalke)
fac4e136fa refactor: Change pointer to reference because it can not be null (MarcoFalke)
Pull request description:
Equating `0==None` and `""==None` is confusing, unneeded and undocumented
ACKs for top commit:
jonatack:
ACK fa69c2c784
achow101:
ACK fa69c2c784
Sjors:
tACK fa69c2c784 modulo `unset`
Tree-SHA512: c4c8d0ad80c6697621d356a9545caf28ca2facc82bb2fa8e70eceb52372d25f0685237c73688c4b01da0e75d213c77c0d45011a8bdfe81ea783d85f045786dac
8f7d1b39ef Fix QPainter non-determinism on macOS (Andrew Chow)
Pull request description:
Aplies a patch to Qt that fixes the non-determinism by modifying Qt. The source of the non-determinism is how LLVM 8 optimizes qt_intersect_spans when compiling. The particular optimization that seems to be causing the problems is that a temp variable is being added for spans->y. For some reason, when it does this, it chooses different instructions to use when making that variable. We bypass this problem by patching qt_intersect_spans to always make and use this local variable.
Potential alternative to #20436 and #20440
ACKs for top commit:
hebasto:
re-ACK 8f7d1b39ef ~for merging into the 0.21 branch, but [not into the master](https://github.com/bitcoin/bitcoin/pull/20454) branch.~
fanquake:
ACK 8f7d1b39ef
Tree-SHA512: b0d00a77643554021736524fb64611462ef2ec849a220543c12d99edb0f52f2e8128d2cc61fa82176b7e13b294574774a92d6b649badf8b7630c6d6a7e70ce10
05c1095388 test: Add testing of ParseInt/ParseUInt edge cases with leading +/-/0:s (practicalswift)
Pull request description:
Add testing of `ParseInt`/`ParseUInt` edge cases with leading `+`/`-`/`0`:s.
Context: While working on #20457 and #20452 I noticed some edge cases which our unit tests are currently not covering.
ACKs for top commit:
MarcoFalke:
review ACK 05c1095388
laanwj:
Code review ACK 05c1095388
jonatack:
ACK 05c1095388
promag:
Code review ACK 05c1095388.
Tree-SHA512: bdfb94d8fa0293512dbba89907cb6dd0f8b1418d878267dd6d49c8c397a0e5b9714441345565d41a6a909a1cda052ef7cccece822f355ff604fcf85f2dc8136f
b1f59d55d9 RPC/Wallet: unloadwallet: Clarify docs/error when both the RPC endpoint and wallet_name parameter specify a wallet (Luke Dashjr)
Pull request description:
Just documentation clarifications from #20448
ACKs for top commit:
MarcoFalke:
review ACK b1f59d55d9
jonatack:
re-ACK b1f59d55d9 per `git diff e8303a0 b1f59d5`
Tree-SHA512: ac068b0aa7ceed49496367fdd9425b59dbba18b56e89b26afc22a6c8ece51f0b92a169cacd55740b1cadab2b32f4f8e8700e609066ab7e59d3b53c7891da585e
0918eb49d5 doc: Document current boost dependency as 1.71.0 (Wladimir J. van der Laan)
Pull request description:
This was forgotten in #19764.
ACKs for top commit:
practicalswift:
ACK 0918eb49d5
fanquake:
ACK 0918eb49d5
Tree-SHA512: bd4a39b96b95adeb725767b283f4cf04d9f0d6ac352e7dc67f88cf575b00a24c6d3f4bf51fe362e0c89aeebb6c7e8e9add9f9f17e843121efd30f8edef6128bc
f190343c96 depends: boost: Specify cflags+compileflags (Carl Dong)
b2328b7989 depends: boost: Remove unnecessary _archiver_ (Carl Dong)
ab9e047cc2 depends: boost: Cleanup toolset selection (Carl Dong)
86002e7e90 depends: boost: Cleanup architecture/address-model (Carl Dong)
d7048fa73f depends: boost: Disable all compression (Carl Dong)
9cf2ee54d3 depends: boost: Split into non-/native packages (Carl Dong)
a57b498560 depends: boost: Bump to 1.71.0 (Carl Dong)
800655ff31 depends: boost: Refer to version in URL (Carl Dong)
Pull request description:
This PR improves the robustness of our boost package in depends, most notably:
1. Bumps boost from `1.70.0` to `1.71.0`, because `1.71.0`:
1. Removes the need to patch out the unused variable.
f8462a6d27/depends/packages/boost.mk (L36)
Upstream boost patched it out in d20b64cf37, which was first included in the `1.71.0` release
2. Comes packaged with a version of `b2` which allows us to override its `CXX` and `CXXFLAGS`. Previously, choosing a toolset while building `b2` such as `clang` or `gcc` would force `b2`'s build system to invoke the compiler as a bare, hardcoded `clang` or `gcc`. However, our `depends` build system often want to customize this behaviour, adding extra flags or invoking the compiler by an alternate name. So this is useful.
1. Commit where `CXX` was introduced: 374f96516a
2. Commit where `CXXFLAGS` was introduced: 5d49abc1f2
2. The boost package is now split into `native_b2` and `boost`, better representing what actually happens.
- In our `depends` build system, we have a distinction between `native` packages and non-`native` packages. The output of `native` packages are meant to be used on the machine that's performing the build, and the output of non-`native` packages are meant to be used on/for the machine that will ultimately be running bitcoin. Previously, `boost` existed in `depends` as a non-`native` package, but that's partly inaccurate because the `./bootstrap.sh` invocation in its `$(package)_config_cmds` stage actually produced a binary called `b2`, which is run on the machine that's performing the build. This means that `b2` is a `native` package which is being built in an environment set up for the non-`native` package `boost`. This reveals a hidden unintended behavior in our `depends` build system: for linux->darwin cross builds, we use `gcc` for `native` packages, and `clang` for non-`native` packages. But `b2` was actually being built using `clang`, since it was being built in an environment set up for non-`native` packages.
theuni you might be interested in taking a look
ACKs for top commit:
laanwj:
Concept and code review ACK f190343c96
Tree-SHA512: f8b728a34da4f0a9a985a819a5762f2fc2689ea24c7eba1d24d26dfbd4c59f202227c699b0a4069dab10b6329cf9f4c6dd95082685776ee43dd5f7b659acdef1
fa7a4385d0 ci: Fix doc typos in .cirrus.yml (MarcoFalke)
fa73674738 ci: Run i686 centos ci config on cirrus (MarcoFalke)
fa1f949a4d ci: Run nowallet ci config on cirrus (MarcoFalke)
Pull request description:
Travis CI Org is shutting down, so move the configs to cirrus ci
ACKs for top commit:
practicalswift:
ACK fa7a4385d0: patch looks correct
Tree-SHA512: 1b7125c7f0d2288931fb8c5e90345891d5f7c1f00c4af136afbeb36bd2836f72920a1877dacc7be787e2c8d84de2a733c1ca71931e5c101b222c1fea588619b3
Aplies a patch to Qt that fixes the non-determinism by modifying Qt. The
source of the non-determinism is how LLVM 8 optimizes qt_intersect_spans
when compiling. The particular optimization that seems to be causing the
problems is that a temp variable is being added for spans->y. For some
reason, when it does this, it chooses different instructions to use when
making that variable. We bypass this problem by patching
qt_intersect_spans to always make and use this local variable.
b87caf10b5 test: add is_bdb_compiled helper (Sjors Provoost)
Pull request description:
Followup for #20202, needed by #16546.
Allow the functional test suite to skip tests that require BDB, as well as introduce specific logic to handle whether BDB support is enabled or not. It follows the same pattern as `skip_if_no_sqlite` and `is_sqlite_compiled`.
ACKs for top commit:
laanwj:
Code review ACK b87caf10b5
Tree-SHA512: e84fb22e017b28f0f75d17e5368fcba22a893484b31b12082cfe9354e6fbd566daf34b3b82f7deb7205b2061c9c61538e402df000e2f05428affae6dbea05c5e
a52ecc936a build: set minimum supported macOS to 10.14 (fanquake)
Pull request description:
This is a requirement for C++17 support. See my comments [here](https://github.com/bitcoin/bitcoin/issues/16684#issuecomment-643722538):
> You cannot use std::get with std::variant on macOS < 10.14, because Apples libc++ doesn't support the std::bad_variant_access exception. [Relevant comment](https://github.com/bitcoin/bitcoin/pull/19183#discussion_r439794318) in #19183.
> While we could work around this in our own code, using std::get_if, this would still be a problem for 3rd-party dependencies.
> I've been testing Qt 5.15LTS (we'll have to enable C++17 in qt, and may upgrade to a newer version at the same time), and you can't enable -std c++17, while targeting a macOS deployment version < 10.14, configuring will fail. They are making use of std::get with std::variant throughout their cocoa code.
We would have to had to have bumped to at least 10.13 in any case, as Qt 5.15 (#19716) [requires 10.13+](https://doc.qt.io/qt-5/supported-platforms.html).
ACKs for top commit:
hebasto:
ACK a52ecc936a, I have reviewed the code and it looks OK, I agree it can be merged.
Tree-SHA512: f669b2fc777aeea1e9afdbbc7bd9afe3997418211db6ba53c934cae0e62a9b999603da539518c229f34961d275c9e2f315c7b022cf5fb97bd201a69c85d470cc
68c2ef13e9 Fix version string in Windows and Mac installers (Andrew Chow)
Pull request description:
Apparently NSIS requires a 4 digit version number, and #20223 dropped the 4th digit. So this adds a 4th 0 digit so that building the Windows installer doesn't fail. Also fixes a typo in that version string that was also present in a plist file.
ACKs for top commit:
fanquake:
ACK 68c2ef13e9
laanwj:
Code review ACK 68c2ef13e9
hebasto:
Approach ACK 68c2ef13e9, tested on Linux Mint 20 (x86_64, nsis 3.05-2).
Tree-SHA512: 845560ff176eae8081096426790c928a773fa75d366f42a2a4631c1be2ae9234d7a5b72854ccfaa7fa1a32002b937ca393b12168ffacf9a5e3e311a76725483a
d52f502b1e Fix mock SQLiteDatabases (Andrew Chow)
99309ab3e9 Allow disabling BDB in configure with --without-bdb (Andrew Chow)
ee47f11f73 GUI: Force descriptor wallets when BDB is not compiled (Andrew Chow)
71e40b33bd RPC: Require descriptors=True for createwallet when BDB is not compiled (Andrew Chow)
6ebc41bf9c Enforce salvage is only for BDB wallets (Andrew Chow)
a58b719cf7 Do not compile BDB things when USE_BDB is defined (Andrew Chow)
b33af48210 Include wallet/bdb.h where it is actually being used (Andrew Chow)
Pull request description:
Adds a `--without-bdb` option to `configure` which disables the compilation of the BDB stuff. Legacy wallets will not be created when BDB is not compiled. A legacy-sqlite wallet can be loaded, but we will not create them.
Based on #20156 to resolve the situation where both `--without-sqlite` and `--without-bdb` are provided. In that case, the wallet is disabled and `--disable-wallet` is effectively set.
ACKs for top commit:
laanwj:
Code review ACK d52f502b1e
Tree-SHA512: 5a92ba7a542acc2e27003e9d4e5940e0d02d5c1f110db06cdcab831372bfd83e8d89c269caff31dd5bff062c1cf5f04683becff12bd23a33be731676f346553d
fabecce719 net: Treat raw message bytes as uint8_t (MarcoFalke)
Pull request description:
Using `uint8_t` from the beginning when messages are `recv`ed has two style benefits:
* The signedness is clear from reading the code, as it does not depend on the architecture
* When passing the bytes on, the need for static signedness casts is dropped, making the code a bit less verbose and more coherent
ACKs for top commit:
laanwj:
Code review ACK fabecce719
theStack:
Code Review ACK fabecce719
jonatack:
Tested ACK fabecce719
Tree-SHA512: e6d9803c78633fde3304faf592afa961ff9462a7912d1da97a24720265274aa10ab4168d71b6ec2756b7448dd42585321afee0e5c889e705be778ce9a330d145
e1f2553e11 build: remove global_init_link_order from mac qt qmake.conf (fanquake)
498fa16bea build: document preprocessing steps in qt package (fanquake)
bd5d9336d9 build: don't copy Info.plist.* into mkspec for macOS qt build (fanquake)
bfd7e33b4b build: remove plugin_no_soname from mac qt qmake.conf (fanquake)
fdde4c7ce6 build: pass XCODE_VERSION through to qt macOS cross compile conf (fanquake)
49473ef211 build: convert "echo" usage into a patch in qt package (fanquake)
Pull request description:
Follow up on removing `sed` usage in #19761. Also nice to revisit & cleanup before 5.15.x.
ACKs for top commit:
laanwj:
Code review ACK e1f2553e11
Tree-SHA512: 4e6489d877aaa300f69e091d7117136da49611bd80afd45adfbd7ddeb5b3c9c76fb0f87a3249cbe63ba93129df56281fd4a9389daadc852211325c5ca9ac6567
9f08780dd7 Use the correct incremental fee constant in bumpfee help (Jon Atack)
3f1e10b2b1 Update feeRate (BTC/kvB) to fee_rate (sat/vB) in wallet_bumpfee (Jon Atack)
1b3d700928 Allow zero-fee fundrawtxn and walletcreatefundedpsbt calls (Jon Atack)
Pull request description:
- Fixes https://github.com/bitcoin/bitcoin/pull/20305/files#r525406176. A check to raise an error on zero-fee txns was mistakenly extended in a0d4957 from the bumpfee and send{toaddress, many} RPCs to also include fundrawtransaction and walletcreatefundedpsbt. This commit re-overrides zero fee rate checking for these two RPCs, not only for the feeRate (BTC/kvB) arg to return to previous behavior, but also for the new fee_rate (sat/vB) arg. Negative fee rates will still raise "amount out of range" by the MoneyRange check in src/bitcoin-tx.cpp::AmountFromValue.
- Updates a wallet bumpfee test from feeRate (BTC/kvB) to fee_rate (sat/vB)
- Fixes https://github.com/bitcoin/bitcoin/pull/20305/files#r525405363 to use the correct incremental fee rate constant in the bumpfee help (thanks Marco Falke for the catch) and rectifies "1.000 sat/vB sat/vB" in the help to "1.000 sat/vB"
ACKs for top commit:
MarcoFalke:
review ACK 9f08780dd7🐾
promag:
Code review ACK 9f08780dd7.
Xekyo:
Code review reACK 9f08780dd7.
Tree-SHA512: 413dfb4f23ebaf3d2ef210dd04610a843272e64eabba428699f5de4d646a86ac4911dab66b5e2f5ebea53b76e4be8347ef40824c1592c750d5eaa12579d3cdf6
8f7b930475 Drop the leading 0 from the version number (Andrew Chow)
Pull request description:
Removes the leading 0 from the version number. The minor version, which we had been using as the major version, is now the major version. The revision, which we had been using as the minor version, is now the minor version. The revision number is dropped. The build number is promoted to being part of the version number. This also avoids issues where it was accidentally not included in the version number.
The CLIENT_VERSION remains the same format as previous as previously, as the Major version was 0 so it never actually got included in it.
The user agent string formatter is updated to follow this new versioning.
***
Honestly I'm just tired of all of the people asking for "1.0" that maybe this'll shut them up. Skip the whole 1.0 thing and go straight to version 22.0!
Also, this means that the terminology we commonly use lines up with how the variables are named. So major versions are actually bumping the major version number, etc.
ACKs for top commit:
jnewbery:
Code review ACK 8f7b930475
MarcoFalke:
review ACK 8f7b930475🎻
Tree-SHA512: b5c3fae14d4c0a9c0ab3b1db7c949ecc0ac3537646306b13d98dd0efc17c489cdd16d43f0a24aaa28e9c4a92ea360500e05480a335b03f9fb308010cdd93a436
2a55a0ed30 Squashed 'src/univalue/' changes from 98261b1e7b..98fadc0909 (MarcoFalke)
Pull request description:
Just a minor bugfix: Currently we don't push booleans into arrays, but if we did they'd be pushed as integers.
Can be tested by reverting the diff in `include/` and observing a test failure.
ACKs for top commit:
laanwj:
ACK fa17eef627
practicalswift:
cr ACK fa17eef627
Tree-SHA512: d87ca5be6769b4cd0c9b9e319973bc0c4f2b7121779f9554e11f34a4edb0013997e875c7edb7bc6eb9309ff5c13379d22f436cd4fb9e6e68df6f0aee29fed914
