There was one more case where we have saved the stack trace into the _remoteStackTraceString
field in the exception when the exception was passing from reflection invoked code
to the caller of that code. While there is no visible difference in the Exception.ToString,
a SOS test was failing due to that. And it is not necessary to save the stack trace
there.

I have thought about the cases when we really need the stack trace saved into the
_remoteStackTraceString and I believe that actually the only case is when an existing
exception is thrown again in managed code. So I have removed the option to save the stack
trace from all the variants of the DispatchManagedException.