Anybody know why the Cascade class has a method startState(self) that overrides the SM class variable startState? And why this is required for a Feedback machine that uses a Cascade machine? I find if I remove the assignment of the Cascade's startState method to the Feedback instance's startState variable - the Feedback machine won't work (for a simple counter - INCR-DELAY). WHY?

Question

Anybody know why the Cascade class has a method startState(self) that overrides the SM class variable startState?  And why this is required for a Feedback machine that uses a Cascade machine?  I find if I remove the assignment of the Cascade's startState method to the Feedback instance's startState variable - the Feedback machine won't work (for a simple counter - INCR->DELAY). WHY?

anonymous · Answer

def __init__(self, sm):
        self.m = sm
        #self.startState = self.m.startState
        print 'Feedback startState = ' + str(self.m.startState)

Above is the Feedback's __init__ method. If I comment out the line as in the above, what does that change?  Specifically when sm is a Cascade machine.

anonymous · Answer

ignore the print statement that was just my debugging
(actually what was printed was a memory address to the location of the method Cascade.startState)