Optimize DebuggerScope objects

Today, we create `DebuggerScope` objects during byte code generation and it stores metadata information that is consumed by debugger and memory profiler. However (at least today) there is no way to get put `Chakracore` into debug mode when it is started under non-debug mode E.g. If `node-chakracore` is running under non-debugger mode, it can never go back to debugger mode. Vice versa if it is in debugger mode using `--debug` flag, we can know before hand and make use of this information whether to populate `DebuggerScope` objects or not.

For memory profilers, in future, when we add Jsrt APIs, we can evaluate if we need to pre-populate this meta data during bytecode generation when chakracore is running in non-debugger mode.

This issue is to investigate what can be done to optimize the working set because of `DebuggerScope` objects.