fix: replace stale codex threads

2026-06-02 12:36:34 +08:00
parent 7475ea0116
commit e40585fd66
4 changed files with 29 additions and 15 deletions
@@ -105,18 +105,19 @@ const selfTest: SelfTestCase = async (context) => {
      env: { CODEX_HOME: context.codexHome, AGENTRUN_FAKE_CODEX_MODE: "resume-no-rollout" },
      oneShot: true,
    }) as JsonRecord;
-    assert.equal(staleThreadResult.terminalStatus, "failed", "stale thread resume must fail instead of creating a replacement thread");
-    assert.equal(staleThreadResult.failureKind, "thread-resume-failed", "stale thread resume failure kind");
+    assert.equal(staleThreadResult.terminalStatus, "completed", "stale thread resume should create a replacement thread and continue the turn");
+    assert.equal(staleThreadResult.failureKind, null, "replacement stale thread turn should not expose thread-resume-failed as terminal failure");
    const staleEnvelope = await client.get(`/api/v1/runs/${staleThread.runId}/commands/${staleThread.commandId}/result`) as JsonRecord;
-    assert.equal(staleEnvelope.terminalStatus, "failed");
-    assert.equal(staleEnvelope.failureKind, "thread-resume-failed");
-    assert.equal(staleEnvelope.completed, false);
-    assert.equal((staleEnvelope.sessionRef as JsonRecord).threadId, "thread_missing_rollout");
+    assert.equal(staleEnvelope.terminalStatus, "completed");
+    assert.equal(staleEnvelope.failureKind, null);
+    assert.equal(staleEnvelope.completed, true);
+    assert.equal((staleEnvelope.sessionRef as JsonRecord).threadId, "thread_selftest_1");
    const staleEvents = await client.get(`/api/v1/runs/${staleThread.runId}/events?afterSeq=0&limit=100`) as { items?: Array<{ type: string; payload: unknown }> };
    const stalePhases = (staleEvents.items ?? []).filter((event) => event.type === "backend_status").map((event) => String(eventPayload(event).phase ?? ""));
-    assert.equal(stalePhases.some((phase) => phase.startsWith("thread/start")), false, "stale resume must not start a replacement thread");
+    assert.equal(staleEvents.items?.some((event) => event.type === "backend_status" && eventPayload(event).phase === "thread/resume:non-resumable"), true, "stale resume must be visible as non-resumable before replacement");
+    assert.equal(stalePhases.some((phase) => phase === "thread/replacement-start:completed"), true, "stale resume must start a replacement thread");
    assert.equal(staleEvents.items?.some((event) => event.type === "backend_status" && eventPayload(event).phase === "thread/resume:completed"), false, "stale resume must not be reported as a successful resume");
-    assert.equal(staleEvents.items?.some((event) => event.type === "error" && eventPayload(event).failureKind === "thread-resume-failed"), true, "stale resume failure must be visible as an error event");
+    assert.equal(staleEvents.items?.some((event) => event.type === "error" && eventPayload(event).failureKind === "thread-resume-failed"), false, "replacement stale resume must not surface terminal thread-resume-failed error");
    assertNoSecretLeak({ staleThreadResult, staleEnvelope, staleEvents });

    const live = await createRunWithCommand(client, context, "hello live events", "selftest-live-tool-events", 15_000);
@@ -185,7 +186,7 @@ const selfTest: SelfTestCase = async (context) => {
    await runSecretFailureCase({ client, managerUrl: server.baseUrl, context });
    await runSpawnFailureCase({ client, managerUrl: server.baseUrl, context });

-    return { name: "codex-stdio", tests: ["runner-lease-heartbeat", "codex-stdio-fake-turn", "codex-stdio-projected-writable-home", "codex-stdio-deepseek-profile-fake-turn", "codex-stdio-minimax-m3-profile-fake-turn", "codex-stdio-deepseek-missing-secret-no-fallback", "codex-stdio-minimax-m3-missing-secret-no-fallback", "codex-stdio-config-model-authoritative", "codex-stdio-explicit-model-forwarded", "codex-stdio-final-agent-message-only", "codex-stdio-stale-thread-resume-fails", "codex-stdio-live-tool-events", "codex-stdio-noisy-reasoning-suppression", "codex-stdio-missing-turn-result", "codex-stdio-provider-auth-failed", "codex-stdio-provider-rate-limited", "codex-stdio-provider-invalid-tool-call", "codex-stdio-provider-503-rpc-error", "codex-stdio-provider-503-terminal", "codex-stdio-provider-503-retry-event", "codex-stdio-invalid-json", "codex-stdio-timeout", "codex-stdio-secret-unavailable", "codex-stdio-spawn-failure"] };
+    return { name: "codex-stdio", tests: ["runner-lease-heartbeat", "codex-stdio-fake-turn", "codex-stdio-projected-writable-home", "codex-stdio-deepseek-profile-fake-turn", "codex-stdio-minimax-m3-profile-fake-turn", "codex-stdio-deepseek-missing-secret-no-fallback", "codex-stdio-minimax-m3-missing-secret-no-fallback", "codex-stdio-config-model-authoritative", "codex-stdio-explicit-model-forwarded", "codex-stdio-final-agent-message-only", "codex-stdio-stale-thread-replacement", "codex-stdio-live-tool-events", "codex-stdio-noisy-reasoning-suppression", "codex-stdio-missing-turn-result", "codex-stdio-provider-auth-failed", "codex-stdio-provider-rate-limited", "codex-stdio-provider-invalid-tool-call", "codex-stdio-provider-503-rpc-error", "codex-stdio-provider-503-terminal", "codex-stdio-provider-503-retry-event", "codex-stdio-invalid-json", "codex-stdio-timeout", "codex-stdio-secret-unavailable", "codex-stdio-spawn-failure"] };
  } finally {
    await new Promise<void>((resolve) => server.server.close(() => resolve()));
  }