labath added a comment. I see two problems with this patch.
- it makes shutting down lldb safely impossible (because you never know if there are any unfinished tasks running). We already have problems with shutting down (grep for "intentionally leaked"), but this is guaranteed to make the problem worse. Using the global thread pool would make things better, as then one could flush the pool in SBDebugger::Terminate, or something similar - there seems to be a race between new debuggers/targets being added, and their existing instances being notified. I'm not 100% sure what will happen in this case, but it seems very easy to miss some debuggers (or targets) that were added after we took the "snapshot" of the current state (There's also the (obvious) problem of any breakpoints that are set on code within those modules becoming effective at an unpredictable future data, but I am assuming you are aware of that.) CHANGES SINCE LAST ACTION https://reviews.llvm.org/D131328/new/ https://reviews.llvm.org/D131328 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
