Suppose that user performs a step over code that contains a breakpoint.
In this case the step is interrupted with the breakpoint, the code suspends on the breakpoint and the user can decide to either continue debugging from the breakpoint, or do continue to finish the step.
That works fine if there is no code evaluation before the continue. If there is, the step is resubmitted from the current (breakpoint) location. This is unexpected since the original step would not be completed and a new local step is performed after continue instead.
We need to assure that the step is not resubmitted from the breakpoint location and the original step is completed as expected.