pi-crew 0.11.41 → 0.11.43

package/CHANGELOG.md

@@ -2,6 +2,66 @@
 
 ## [Unreleased]
 
+## [0.11.43] — fix wave 15: render trusts task heartbeats + teardown sweep batch 1 (2026-10-25)
+
+Wave 13 fixed the watchdog's frozen-evidence hung verdicts by adding task heartbeats, but the render layer still
+judged liveness from `lastActivityAt` alone — which freezes by construction during one long tool call. A run
+executing a 20-min test suite displayed as "stale"/"orphaned" on the dashboard while its heartbeat was fresh.
+Meanwhile the bare-`rmSync` inventory from the same deep review counted ~1,400 sites across 470 files; this wave
+ships the render half and the first sweep batch (SPEC `fixwave15-display-heartbeat-sweep1-2026-10-25`).
+
+**Added — `hasFreshTaskHeartbeatEvidence(tasks, now, ttl)`.** Pure predicate export from
+`src/runtime/process-status.ts` (shared predicates untouched per GAP-4 purity). The render layer — run dashboard
+and agents-jobs-browser — now treats a run with a fresh task heartbeat (`lastSeenAt` within the 5-min TTL,
+`alive !== false`) as active and suppresses the "stale"/"orphaned" labels, closing the render half of wave 13's
+liveness work. Old behavior pinned: no heartbeat / expired / `alive === false` ⇒ stale label stands.
+
+**Fixed — teardown races in batch 1 of the bare-`rmSync` sweep.** 336 sites across 84 files in `test/unit/state/**`
++ `test/unit/prompt/**` migrated to `cleanupStateRoot`/`removeDirWithRetry` (drain-then-remove over
+retry-racing the async event-log writer). Intentional residuals kept: the `drain-event-log-writers` in-body
+plain-`rmSync` assertion (it pins the exact race the API exists to fix), single-file deletes, and child-source
+strings.
+
+**Battery (PASS, adjudicated).** targeted ×3 (h1 heartbeat-evidence 11/11, h2 run-dashboard 4/4, h3
+agents-jobs-browser 5/5 incl. 2 new heartbeat tests), sweep clusters 1,895 test-runs (4 load-timing fail-events,
+all solo ×3 green), `test:critical` 124/124, full unit 10,081 pass / 1 adjudicated flake / 4 skipped, integration
+144/0/4, static gates 0, dist MATCH (1,893,087 B).
+
+**Docs.** Battery report
+`docs/real-test/reports/real-test-2026-10-25-wave15-render-heartbeat-sweep1-battery.md`. Residual noted:
+health-monitor is not wired to the new predicate (its scan loop consumes tasks internally, not display).
+Backlog: sweep batches 2a (415 sites/124 files), 2b (816/210), 3 (88/33).
+
+## [0.11.42] — fix wave 14: event-log drain API + zombie-scanner seam (2026-10-25)
+
+Wave 12d papered over the recurring `ENOTEMPTY` teardown race with `removeDirWithRetry`; the root cause stayed:
+production fires **66 event writes across 23 files** without awaiting them (`void appendEventAsync` ×23,
+`appendEventFireAndForget` ×43), and each queued write re-creates the run dir before appending — so teardown that
+`rmSync`s the state dir mid-flight races the async writer and loses. There was no way for a test to await one
+run's writers: the queue drain was private and `beforeExit`-only (SPEC `fixwave14-drain-api-zombie-seam-2026-10-25`).
+
+**Added — `drainEventLogWriters(eventsPath?)`.** Public per-run drain API for the event-log async writer queues
+(allSettled tail await + per-path buffer flush, never throws). Lets tests await in-flight writers before removing
+a run state dir instead of retry-racing them — the race repro (drain → plain `rmSync`, no retry) is now a pinned
+test.
+
+**Added — `cleanupStateRoot(dir, eventsPath?)` test helper.** Best-effort drain then `removeDirWithRetry(settleMs: 0)`.
+
+**Added — `ZombieScanOptions` seam on `scanZombieSubagents`** (`platform`/`getuid`/`readProcFile`/`listProcPids`/
+`pidAlive`/`getParentPid`, defaults byte-exact). **7 previously linux-only tests now run on every OS** via
+synthetic `/proc` fixtures (+2 new seam tests; 2 real-`/proc` default-path tests stay linux-gated).
+
+**Fixed — teardown races by adopting drain-then-remove in pilot tests.** `team-tool-steer-clamp` (retry-only →
+drain), `dwf-setresult` (12 bare `rmSync`), `event-log-batch`/`event-log-async`, plus dedup of two local
+`removeDirWithRetry` variants (`subagent-steer-live`, `subagent-tools-integration`).
+
+**Battery (PASS).** targeted ×3 (drain 4/4, zombie 26/26 skip=0, 6 pilot files), `test:critical` 124/124, full
+unit 10,075 pass / 1 ambient-adjudicated / 4 skipped, integration 144/0/4, static gates 0, dist MATCH.
+
+**Docs.** Battery report `docs/real-test/reports/real-test-2026-10-25-wave14-drain-zombie-battery.md`; closed
+investigation (no code change): the "hung" alert for an already-cancelled run was delivery timing — `cancelled`
+is already terminal in both the emit and delivery gates.
+
 ## [0.11.41] — fix wave 13: watchdog liveness — false "appears hung" alerts on healthy direct runs (2026-10-25)
 
 Five false alarms landed in one working session (2026-10-24/25), all the same shape: healthy direct single-agent