Testing

Main paths

Control Flow

1. Start -> Spinning -> Ready
2. Start -> Spinning -> Error
3. Start -> Error
4. Start -> Ready

Data Flow

• run operation (none | some) x timeout (any | default)
• ready (no data | data)
• error (no data | data)

Strategy

We will mix control and data flow in our test sequences.

Oracle

1. Start -> Spinning -> Ready
   • No run:
     • emit timer | render fallback | render main
   • run:
     • run task | emit timer | render fallback | render main
2. Start -> Spinning -> Error
   • No run:
     • emit timer | render fallback | render error
   • run:
     • run task | emit timer | render fallback | render error
3. Start -> Ready
   • No run:
     • emit timer | render main
   • run:
     • run task | emit timer | render main
4. Start -> Error
   • No run:
     • emit timer | render main
   • run:
     • run task | emit timer | render main

Properties

Abstracting null commands:

• emit timer is always there
• run task command is always before emit timer when present
• render main is last when present
• render error is last when present
• the render fallback command when present is always immediately after emit timer

Edge cases

Basically not starting with start, receiving unexpected events etc.

• make some myself: take the oracle and intersperse event there, I know where the null will be from the oracle
• make some random one