You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/ONECOMPUTER-LIVE-E2E.md
+2Lines changed: 2 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -62,6 +62,8 @@ This is an integration blocker, not a reason to weaken the ONEVibe boundary:
62
62
63
63
Until (1) and (2) are implemented, a caller can cancel before it receives an ID, leaving the provider unable to participate in reliable automatic cleanup. That is unacceptable for production ephemeral-workspace guarantees.
64
64
65
+
The ONEVibe runner now persists the provider-returned sandbox ID/state immediately and emits typed state-transition evidence while polling. This client-side prerequisite is covered by tests; the live gate remains the Azure deployment of the corresponding provider repair and its fresh provenance response.
66
+
65
67
## Required successful proof
66
68
67
69
Run the following after the provider lifecycle change:
expect(store.listEvents(task.id).filter((event)=>event.label==='ONEComputer sandbox state observed').map((event)=>event.payload.state)).toEqual(['creating','started'])
0 commit comments