-
Notifications
You must be signed in to change notification settings - Fork 3
plat 318
| Coordination | Value |
|---|---|
| Assigned agent | Codex |
| Ticket state | Implemented and regression-tested; release verification pending |
| Last synchronized | 2026-09-12 |
The Plan view waited on plan.json and then step_config.json. Shared cache-promise consumers used a different loading/error path, and late results from a prior workflow could update the newly selected workflow's state.
Plan/config reads now run concurrently, with 20-second per-read timeouts. Both fresh and shared loads clear their loading state; cached results clear a previous spinner. Results are applied only while their workspace remains current. Optional step-config absence retains existing behavior. A DOM hook test covers concurrent fetch, shared requests, workflow switching and late-response isolation.
During investigation RTS load was 0.86 with about 12 GB available memory; a local workspace API read of rts-latency plan.json returned HTTP 200 in 3 ms. These are point-in-time diagnostics, not a full public browser performance benchmark. Type checking, targeted lint and the loader regression passed. A production browser timing comparison remains unverified.
The latest follow-up above is included in RTS release 3a37a1c-20260912143119
(app 3a37a1c75). Source revisions and all three services were verified after
activation. The configured-environment Linux sandbox regression passed.
Earlier pending-deployment notes are superseded; feature-specific live acceptance
limits remain as documented. No user accounts/sharing or notification recipients
were changed during verification.
Auto-synced from docs/ on main. Edit there, not here.