Prototype test script for TTINTEL-239 #21
Conversation
Signed-off-by: David Spence <dspence@redhat.com>
| r'.*$', | ||
| )) | ||
|
|
||
| Parsed = namedtuple('Parsed', ('timestamp', 'interface', 'terror', 'state')) |
|
|
||
| def handle_S0_in_TESTING_INITIALIZING(fsm, arg): | ||
| """Handle event S0 in state /TESTING/INITIALIZING""" | ||
| fsm.state = STATE_TESTING_INITIALIZING |
There was a problem hiding this comment.
I'm not loving the assignments in these functions that have no lasting effect. On first reading it me wonder if there was a side effect of the first assignment that makes it necessary, though I can't see anything to back that up.
What's the purpose of the assignments that set fsm.state to the value it already has?
There was a problem hiding this comment.
The state assignments are output on each "state boundary crossing" for an external transition. An external transition is one where the FSM leaves the current state before entering the target state. The state boundaries crossed are all those up to (but not including) the common ancestor of current and target states. With regard to "assigning fsm.state to the value it already has", that will be due to an external transition back into the same state.
Currently I expect these to be optimized away (by Python itself). I've left them in the generated output as they document the boundary crossings, allowing visual inspection that the intended path is taken.
I agree they're ugly. Getting rid of them is on my TODO list.
|
Replying to Jose's comment... for this prototype, the purpose was twofold: 1) to make it easy for me to identify events so I could draw vertical lines in the plot; 2) to stimulate conversation around separation of concerns, how we combine code to produce higher order tests. |
@crwr45 @spenced Propose to close this PR since this work will be moved to https://github.com/redhat-partner-solutions/vse-sync-pp |
No description provided.