useAgentStream · runId lifecycle
The test mocked a state the machine can’t reach
reachable states of
useAgentStream
idle
runId: null
running
runId: set
applyComplete() keeps prev.runId
complete
runId: still set
carried forward
no transition produces this
the state the test mocked
{ status:
"complete"
, runId:
null
}
The mocked fixture sits outside the reachable state space.
applyComplete()
carries
runId
forward into
complete
, so
complete
+
runId: null
is a pair no transition produces.