Duplicate state transitions where only one is expected

[ricmua]

During normal operation of the task, duplicate state transitions are being observed where only one is expected. The testing infrastructure ~is~~ was not set up to detect this.

An illustration of the repeated state from the console:

[INFO] [main-2]: process started with pid [NNNNNN]
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: start_block
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: intertrial
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: timeout
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: to_move_a
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: move_a
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: move_a
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: timeout
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: failure
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: timeout
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: end_trial
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: intertrial
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: intertrial
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: timeout
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: to_move_a
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: move_a
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: State: move_a
[node-1] [INFO] [NNNNNNNNNN.NNNNNNNNN] [machine.delay_out_center_task]: Event: timeout

Likely, this is actually an issue with the delay_out_center_task package. However, the testing infrastructure for this package is currently superior. The following code can be used to illustrate the issue:

from delay_out_center_task import Environment, Machine, Model
def timeout(self, *args, **kwargs):
    print('Timeout')
    self.cancel_timeout()
Model.timeout = timeout
environment = Environment()
model = Model(environment=environment)
machine = Machine(model=model)
model.trigger('start_block')
model.trigger('timeout')

Created the [ricmua/duplicate_state_transition_issue](https://github.com/ricmua/ros_delay_out_center_task/tree/ricmua/duplicate_state_transition_issue) branch to illustrate and deal with this issue.

[ricmua]

The issue is that automatic transitions are being triggered during the on_enter methods of the state machine model. Since the machine never actually enters the associated state, it is reported as the subsequent state. For example, the on_enter_trial_setup method triggers a transition to move_a (via self.to_move_a()), and so the trial_setup state is erroneously reported as move_a.

Two obvious options:

• Leave the code as it is, and just account for the flawed reporting.
• Replace automatic transitions with (nearly instantaneous) timeouts. Add the timeout transitions to the state machine.