jingham added a comment. Getting watchpoints to propagate to threads that are created after the watchpoint has been set is one of the trickier parts of watchpoint propagation. On macOS, (where we don't get new thread creation notification) we rely on the kernel to propagate the control register settings to the thread contexts of newly created threads - a feature nobody but the debugger uses so we need to make sure it doesn't regress. On systems that don't have kernel cooperation you have to make sure that you catch new thread events and by hand set the control registers, again something we need to make sure we keep doing. If there's no other test that this works, then we should keep testing that in this test. It is the most fragile part of the interaction between watchpoints and threads.
OTOH making this a debug info independent test seems correct. https://reviews.llvm.org/D43857 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
