## Description
Pull request was inspired by the need to debug lock problems when
working on https://github.com/dashpay/dash/pull/5352.
As far as I'm aware, only macOS has `-Werror=thread-safety` as part of
its default `CXXFLAGS` despite the capability being present on Linux as
well. This PR introduces thread safety checks for that into our thread
sanitizer build.
Additionally, since we're using Clang, something that on first glimpse,
appears to be something that `stacktraces.cpp` isn't happy with, due to
`-Wl,-wrap` being available only on GCC, that no longer seems to be the
case, since the version of Clang with comes with `focal`, its `lld`
_does_ have support for `-wrap` (see [man page for `lld` on
`focal`](https://manpages.ubuntu.com/manpages/focal/en/man1/lld.1.html)).
The current `stable` version of Clang/LLVM is 15, at the time of this
pull request (see https://apt.llvm.org/) but `focal` ships with an older
version, requiring us to use the official LLVM APT repository. I feel we
should be testing with recent compilers alongside the ones shipped by
LTS distributions.
Certain bugs are only made apparent when testing on rolling release
distros or distros that have faster update cycles, like Fedora (see
https://github.com/dashpay/dash/pull/5295 for an illustration of that),
which ship with more recent compilers. Until we overhaul our CI systems
to test using those distros directly (our current infrastructure is
centered around using a "development image" with an LTS distro as the
base), this is the best we can do.
A similar pull request testing against the latest GCC stable will be
welcome as that is currently outside the scope of this PR as the changes
made were to make sure that builds were operating as expected on
Clang/LLVM 15.
---------
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
* style: use clang-tidy style named parameters
* refactor: make IsTimeOutOfBounds testable by having current time be a parameter
* style: use x-> not (*x).
* refactor: make SelectCoinsGroupedByAddresses return a vector, remove out param
previous semantics was return false if the vecTally vector was empty. Now we just let the caller check if it is empty or not
* refactor: fix some sign-compare warnings
* refactor: consistently pre-declare stuff as struct / class inline with underlying type
* refactor: don't return const bool
* refactor: use ref to string
* refactor: use = default for CompactTallyItem
* refactor: adjust "initialization" ordering
* refactor: adjust how we handle negatives in GetProjectedMNPayees, use std::min
* refactor: don't bind a reference to a temporary value
* refactor: use a ref
* refactor: ensure attempt in SelectMemberForRecovery is non-negative.
* refactor: remove unused this capture
* refactor: fix numerous sign-compare warnings
* refactor: more consistently use size_t, use empty()
* src/evo/evodb.cpp:57:29: warning: Assert statement calls a function which may have desired side effects: 'IsClean'. [assertWithSideEffect]
* src/llmq/quorums.cpp:635:37: note: Null pointer dereference
src/llmq/quorums.cpp:635:37: warning: Either the condition 'pFrom==nullptr' is redundant or there is possible null pointer dereference: pFrom. [nullPointerRedundantCheck]
src/llmq/quorums.cpp:636:81: note: Assuming that condition 'pFrom==nullptr' is not redundant
* fix a bunch of cppcheck warnings
* cppcheck: run on many more files. Enable all checks except a few ignored ones.
ignored
```
"Consider using std::transform algorithm instead of a raw loop."
"Consider using std::accumulate algorithm instead of a raw loop."
```
* ci: build specific version of cppcheck instead of install from apt
* ci: use cppcheck 2.4, remove commented out line, fix symlink
cppcheck 2.6 is latest, however causes issues
```
src/spork.cpp:135:51: warning: Analysis failed. If the code is valid then please report this failure. [cppcheckError]
```
cppcheck 2.5 appears to get into an infinite loop
* no need to check presence before insertion
* use if-init, remove redundant check
* remove redundant check
* don't remove cmake? fix macOs depends build?
* cppcheck: one per line, alphabetize
* remove duplicate cmake install
* scripted-diff: Replace #include "" with #include <> (ryanofsky)
-BEGIN VERIFY SCRIPT-
for f in \
src/*.cpp \
src/*.h \
src/bench/*.cpp \
src/bench/*.h \
src/compat/*.cpp \
src/compat/*.h \
src/consensus/*.cpp \
src/consensus/*.h \
src/crypto/*.cpp \
src/crypto/*.h \
src/crypto/ctaes/*.h \
src/policy/*.cpp \
src/policy/*.h \
src/primitives/*.cpp \
src/primitives/*.h \
src/qt/*.cpp \
src/qt/*.h \
src/qt/test/*.cpp \
src/qt/test/*.h \
src/rpc/*.cpp \
src/rpc/*.h \
src/script/*.cpp \
src/script/*.h \
src/support/*.cpp \
src/support/*.h \
src/support/allocators/*.h \
src/test/*.cpp \
src/test/*.h \
src/wallet/*.cpp \
src/wallet/*.h \
src/wallet/test/*.cpp \
src/wallet/test/*.h \
src/zmq/*.cpp \
src/zmq/*.h
do
base=${f%/*}/ relbase=${base#src/} sed -i "s:#include \"\(.*\)\"\(.*\):if test -e \$base'\\1'; then echo \"#include <\"\$relbase\"\\1>\\2\"; else echo \"#include <\\1>\\2\"; fi:e" $f
done
-END VERIFY SCRIPT-
Signed-off-by: Pasta <pasta@dashboost.org>
* scripted-diff: Replace #include "" with #include <> (Dash Specific)
-BEGIN VERIFY SCRIPT-
for f in \
src/bls/*.cpp \
src/bls/*.h \
src/evo/*.cpp \
src/evo/*.h \
src/governance/*.cpp \
src/governance/*.h \
src/llmq/*.cpp \
src/llmq/*.h \
src/masternode/*.cpp \
src/masternode/*.h \
src/privatesend/*.cpp \
src/privatesend/*.h
do
base=${f%/*}/ relbase=${base#src/} sed -i "s:#include \"\(.*\)\"\(.*\):if test -e \$base'\\1'; then echo \"#include <\"\$relbase\"\\1>\\2\"; else echo \"#include <\\1>\\2\"; fi:e" $f
done
-END VERIFY SCRIPT-
Signed-off-by: Pasta <pasta@dashboost.org>
* build: Remove -I for everything but project root
Remove -I from build system for everything but the project root,
and built-in dependencies.
Signed-off-by: Pasta <pasta@dashboost.org>
# Conflicts:
# src/Makefile.test.include
* qt: refactor: Use absolute include paths in .ui files
* qt: refactor: Changes to make include paths absolute
This makes all include paths in the GUI absolute.
Many changes are involved as every single source file in
src/qt/ assumes to be able to use relative includes.
Signed-off-by: Pasta <pasta@dashboost.org>
# Conflicts:
# src/qt/dash.cpp
# src/qt/optionsmodel.cpp
# src/qt/test/rpcnestedtests.cpp
* test: refactor: Use absolute include paths for test data files
* Recommend #include<> syntax in developer notes
* refactor: Include obj/build.h instead of build.h
* END BACKPORT #11651 Remove trailing whitespace causing travis failure
* fix backport 11651
Signed-off-by: Pasta <pasta@dashboost.org>
* More of 11651
* fix blockchain.cpp
Signed-off-by: pasta <pasta@dashboost.org>
* Add missing "qt/" in includes
* Add missing "test/" in includes
* Fix trailing whitespaces
Co-authored-by: Wladimir J. van der Laan <laanwj@gmail.com>
Co-authored-by: Russell Yanofsky <russ@yanofsky.org>
Co-authored-by: MeshCollider <dobsonsa68@gmail.com>
Co-authored-by: UdjinM6 <UdjinM6@users.noreply.github.com>
Not returning a value from a function with non-void return type
is undefined behaviour in C++ (except for main() function).
Fortunately, this was causing a crash for multiwallet functional test
when building with GCC 9.2 with default optimisation flags for building
Fedora packages, so I was able to investigate, find this bug and fix it.
But even if it was not causing a crash and just returning some random
garbage in eax register, it would lead to incorrect behaviour:
non-zero return from a handler causes dl_iterate_phdr() to stop iteration.
Signed-off-by: Oleg Girko <ol@infoserver.lv>
* Remove use of -rdynamic
This causes check-symbols to fail horribly and also turned out to be not
required when using libbacktrace. It was only required when using
"backtrace()" from "<execinfo.h>"
* Remove spurious ], from configure.ac
* Add -DENABLE_STACKTRACES=1 to CMakeLists.txt
* Remove unused method my_backtrace_simple_callback
* Use fs::path().filename() instead of basename()
* Add static g_exeFileName and g_exeFileBaseName
* Use .exe.dbg file when available
* Use uint64_t instead of uintptr_t
* Implement GetBaseAddress() for unix and win32
* Implement unified crash_info and use it everywhere before printing crash info
* Print a serialized version of crash_info when there is no debug info
* Implement "-printcrashinfo" command line option
* Compile stacktrace support unconditionally and only make crash hooks conditional
This also renames the --enable-stacktraces option to --enable-crash-hooks
* Enable crash hooks in win/linux Gitian builds
* Try to load .debug file on MacOS and enable crash hooks for osx Gitian builds
* Check for dsymutil and if it needs --flat
* Create .debug files in osx Gitian build
* Handle review comments
* Also print crash description when no stacktrace is available
* Unconditionally add -g1 debug information
Instead of making it dependent on "--enable-crash-hooks". We will need the
debug info every time now, even in release builds.
* Put MacOS debug info into dSYM symbols instead of plain .debug files
* Implement MacOS specific GetBaseAddress
* Add libbacktrace to depends
This is currently only useful to extract symbols. It fails to gather
stacktraces when compiled with MinGW, so we can only use it to get symbol
information from a stack trace which we gathered outside of libbacktrace.
* Add -mbig-obj to CXXFLAGS for MinGW builds
* Implement stacktraces for C++ exceptions
This is a hack and should only be used for debugging. It works by wrapping
the C++ ABI __wrap___cxa_allocate_exception. The wrapper records a backtrace
and stores it in a global map. Later the stacktrace can be retrieved with
GetExceptionStacktraceStr.
This commit also adds handlers to pretty print uncaught exceptions and
signals.
* Use GetPrettyExceptionStr for all unhandled exceptions
* Use --enable-stacktraces in CI for linux32/linux64
* Register exception translators to pretty print exceptions in unit tests
* Catch and print python exceptions when stopping nodes
Otherwise the code at the bottom is never executed when nodes crash,
leading to no output of debug.log files on Travis.
* Remove now unneeded/unused TestCrash methods
