Since C++11, the STL has it's own threading implementation that's covering all our use cases; the only thing missing is a shared mutex implementation.
The change is mostly straightforward except for the io context cancellation in the first commit, but it improves compilation times by ~4% and reduces peak memory usage by ~17%.
Since C++11, the STL has it's own threading implementation that's covering all our use cases; the only thing missing is a shared mutex implementation.
The change is mostly straightforward except for the io context cancellation in the first commit, but it improves compilation times by ~4% and reduces peak memory usage by ~17%.