some times you may have a test with lots of complicated setup and it might hit a breakpoint you are looking for many times before it hits it in the situation you want. instead of having to continue untill you get to the point you want it would be nice if there was a db.trap() or trap() shell helper that sends SIGTRAP to a process you are debugging. that way you can wait until you get to the situation you want and then set your other breakpoint in c++ code.