Post #18710, there isn’t much left using boost::thread_group, so should just be able to replace it with the standard library. This also removes the last use of boost::thread_interrupted.
After this change, last piece of Boost Thread we’d be using is boost::shared_mutex. See the commentary here as to why it may be non-trivial to swap that for std::shared_mutex in the near future.
Closes #17307
155@@ -155,7 +156,7 @@ static std::unique_ptr<ECCVerifyHandle> globalVerifyHandle;
156
157 static std::thread g_load_block;
158
159-static boost::thread_group threadGroup;
160+static std::thread g_scheduler;
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.
Reviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.
This also removes the last use of
boost::thread_interrupted.
Concept ACK on it.
This also removes the last use of boost::thread_interrupted.
Concept ACK thank youuu
1341@@ -1342,7 +1342,7 @@ bool AppInitMain(const util::Ref& context, NodeContext& node, interfaces::BlockA
1342 node.scheduler = MakeUnique<CScheduler>();
1343
1344 // Start the lightweight task scheduler thread
1345- threadGroup.create_thread([&] { TraceThread("scheduler", [&] { node.scheduler->serviceQueue(); }); });
1346+ g_scheduler = std::thread(&TraceThread<std::function<void()>>, "scheduler", [&] { node.scheduler->serviceQueue(); });
I’m probably missing some context so feel free to ignore: can this continue to use a lambda rather than std::function for possibly less compile time / run time overhead? This compiles and the tests pass for me locally.
0- g_scheduler = std::thread(&TraceThread<std::function<void()>>, "scheduler", [&] { node.scheduler->serviceQueue(); });
1+ g_scheduler = std::thread([&] { TraceThread("scheduler", [&] { node.scheduler->serviceQueue(); }); });
CScheduler is sort of written / documented as if it will have multiple threads running serviceQueue(), the only place we actually do that is in the scheduler tests.
35@@ -35,6 +36,8 @@ class CScheduler
36 CScheduler();
37 ~CScheduler();
38
39+ std::thread m_service_queue;
m_service_queue is a bit of a weird name for a data structure that’s not a queue. m_service_thread maybe?
m_service_thread.
This is a straightforward change.
Code review ACK 4b45d7857c06c908d001d6b792fb85066f9ea60f
Code review re-ACK dc8be12510c2fd5a809d9a82d2c14b464b5e5a3f
review ACK dc8be12510c2fd5a809d9a82d2c14b464b5e5a3f 🔁
Signature:
0-----BEGIN PGP SIGNED MESSAGE-----
1Hash: SHA512
2
3review ACK dc8be12510c2fd5a809d9a82d2c14b464b5e5a3f 🔁
4-----BEGIN PGP SIGNATURE-----
5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16=mxeP
17-----END PGP SIGNATURE-----
Timestamp of file with hash d4a9b161264d0b7adbc481cbca5ceed62333c7b8ce658c5ce2fbf697aa94a943 -