jeffreytan81 wrote: @jimingham, sounds good. I will move the logic into ValueObject.
> TryAllThreads is true I did not see any client code setting this to true, I assume this happens because `m_try_others` is true by default? Maybe we should set `m_try_others`'s default value to false to avoid side effect like this? It sucks that we have to use timeout to guess and detect potential deadlock and resume all threads. On Windows, I remember I implemented the deadlock-free expression evaluation for VS debugger much accurately because Windows has an API to detect mutex block chain. With the APIs, I can even find out which dependency thread to resume to resolve the deadlock. It does not cover all synchronization situations though. https://github.com/llvm/llvm-project/pull/108414 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
