CB-114: readiness-gate timeout + presence cleanup (delegated review findings)

An off-sub worker's review of CB-113 (delegated through the bridge) surfaced two real gaps in the readiness gate: - A worker that herdr reports idle but whose Claude never connects the bridge MCP (crashed during boot, or wedged on a startup prompt) left its message queued forever: ready.test() never passed, the target was polled indefinitely, and the caller's future never completed (async waiter hung for the full 30-min window). Injector now counts injectable-but-not-ready samples and, after a ~60s grace (READINESS_GRACE_POLLS, deliberately longer than the UNKNOWN stall grace since a first boot is slower than an in-turn blip), fails the queued messages, fires onTurnFailed so blocking/async waiters resolve WORKER_FAILED, and reclaims the target. Mirrors the CB-109 UNKNOWN-stall path. - WorkerPresence.forget had no caller, so a worker's readiness lingered past its life. Injector now clears presence via a forget callback on drop() (pane crash) and on the readiness timeout. +3 InjectorTest cases (never-ready failure, ready-within-grace delivery, drop clears presence). 105 tests green.
2026-07-16 05:01:58 +02:00
parent 37f7ad4185
commit 2fb46f670f
3 changed files with 107 additions and 1 deletions
@@ -308,6 +308,58 @@ class InjectorTest {
        assertEquals(List.of(T), cap.failed, "a delivery that vanishes before pickup still fails its send");
    }

+    // ~60s of idle-but-not-ready at the 250ms prod poll interval; enough to trip the readiness grace.
+    private static final int READINESS_SAMPLES = 245;
+
+    @Test
+    void failsAQueuedMessageWhoseWorkerNeverBecomesReady() {
+        // CB-114: herdr keeps reporting the worker idle, but its Claude never connects the bridge MCP,
+        // so the readiness gate never opens. The message must not be held (and the target polled)
+        // forever — after the grace it fails, the caller unblocks via the worker-failure path, the
+        // target is reclaimed, and the never-set presence is cleared.
+        Captor cap = new Captor();
+        List<String> forgotten = new ArrayList<>();
+        Injector inj = new Injector(new AgentControl(herdr), cap, _ -> false, forgotten::add);
+        CompletableFuture<Void> f = inj.enqueue(T, "task");
+
+        for (int i = 0; i < READINESS_SAMPLES; i++) inj.onStatus(T, AgentStatus.IDLE);
+
+        assertEquals(List.of(), sent(), "a never-ready worker is never delivered to");
+        assertTrue(f.isCompletedExceptionally(), "the caller's future fails instead of hanging forever");
+        assertEquals(List.of(T), cap.failed, "the awaiting send resolves through the worker-failure path");
+        assertEquals(List.of(), cap.completed, "a never-ready worker is a failure, not a completion");
+        assertEquals(List.of(T), forgotten, "the never-ready worker's presence is cleared");
+        assertTrue(inj.activeTargets().isEmpty(), "the target is reclaimed, not polled forever");
+    }
+
+    @Test
+    void aWorkerThatBecomesReadyWithinTheGraceIsDeliveredNormally() {
+        // The readiness grace must not fail a worker that is merely slow to boot: once it becomes
+        // available before the grace elapses, delivery proceeds as usual (the counter resets).
+        Set<String> ready = new java.util.HashSet<>();
+        Injector inj = new Injector(new AgentControl(herdr), TurnListener.NOOP, ready::contains, _ -> {
+        });
+        inj.enqueue(T, "task");
+
+        for (int i = 0; i < 100; i++) inj.onStatus(T, AgentStatus.IDLE); // still booting, well under grace
+        assertEquals(List.of(), sent());
+
+        ready.add(T); // MCP connects before the grace elapses
+        inj.onStatus(T, AgentStatus.IDLE);
+        assertEquals(List.of("task"), sent(), "a worker that connects within the grace is delivered to");
+    }
+
+    @Test
+    void dropClearsWorkerPresence() {
+        // CB-114 (finding #1): a vanished worker's readiness must be forgotten so a stale entry cannot
+        // linger past the worker's life (WorkerPresence.forget had no caller before this).
+        List<String> forgotten = new ArrayList<>();
+        Injector inj = new Injector(new AgentControl(herdr), TurnListener.NOOP, _ -> true, forgotten::add);
+        inj.enqueue(T, "orphan");
+        inj.drop(T, new HerdrException("worker gone", "pane_not_found", null));
+        assertEquals(List.of(T), forgotten, "drop clears the gone worker's presence");
+    }
+
    @Test
    void pollerDeliversToAnIdleWorker() throws Exception {
        // End-to-end through the poller: idle worker → message delivered without manual onStatus.