plan: one question, one answer � a flow verdict and the state it reads (#1518) #1614
No reviewers
Labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
jaylouisw/packetforge!1614
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "plan/1116-0610"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Forward plan only, no code.
docs/plans/flow-determinism.md.What I measured on
bfc4f2ea#1518 claims the first
openFlowon a fresh topology always failsE_UNREACHABLE. I could not reproduce it: 13 shapes (plain L2, two switches with VLAN 10, routed, router-on-a-stick, switch SVI, DHCP client, UDP, LAG, serialize/fromJSON round trip, three identical rebuilds) all returnedok: trueon their first flow.What does reproduce, on the same board, asked the same question:
pingsayspeer port shut (d2 Fa0/2) (dropped by d3);openFlowsaysE_UNREACHABLE / no response from 10.0.0.80:80while its own timeline shows the shut port._flowFailure()(simulator.js:3764) does not use the ranked_lastDiscard()(simulator.js:2922) that #1431/#1445 gave ping.dev.arpSeenAtis empty and everymacTableentry isatMs: null, because_nowMs()(simulator.js:4275) returns null whenrunAtMsis Infinity � the default. So the 300 s / 4 h TTLs that docs/SIMULATION.md:69-74 advertises never fire.Plan
Five steps, one PR each: a determinism suite (S), a flow that reports the ranked reason (S), dating what was learned (M, needs Jay on which clock),
openFlow({atMs})(S, needs Jay on the API field), and � only if a divergence survives � moving run state onto the Simulator (L, would move docs/SIMULATION.md:66 with it).Closes #1518 via steps 1-3. Deliberately out: the
/resettable-clear omission, server-exposure work, ping wording already in flight.