emrg: TUI daemon spawn throttle — complete the anti-storm fix

Agent.md

@@ -93,7 +93,7 @@ Community needs voiced in HN agent-UI discussions map directly to EMRG's design:
 pkill -f "emrg.server"; rm -f ~/.emrg/emrgd.port; python -m emrg
 ```
 
-Python: `uv run pytest tests/ -v` (650) — import check: `uv run python -c "from emrg.client.app import run_client"`
+Python: `uv run pytest tests/ -v` (652) — import check: `uv run python -c "from emrg.client.app import run_client"`
 GUI: `cd emrg/gui && npm test` (101: 27 daemon_client + 22 app-commands + 27 renderer smoke + 15 i18n + 7 integration + 3 commands) — syntax: `node --check main.js preload.js daemon_client.js renderer/js/*.js`
 CI: `uv run pytest` + GUI tests + **actionlint workflow lint** (`rhysd/actionlint@v1.7.12` gate, #444 — workflow 解析错误在 PR CI 即失败，如 `if:` secrets 上下文)
 Re-trigger: `scripts/re-trigger-ci.sh [branch]` (workflow_dispatch, #527 — 替代空 commit 重触发：Actions outage 会整段丢弃 push 事件，dispatch 走 API 路径不受影响)

README.cn.md

@@ -274,7 +274,7 @@ EMRG 不只是追赶——它自己追上来。
 git clone https://github.com/argszero/emrg.git
 cd emrg
 uv sync # 安装依赖
-uv run pytest tests/ -v # 跑测试（当前 650 项）
+uv run pytest tests/ -v # 跑测试（当前 652 项）
 uv run python -m emrg # 启动 TUI
 # CI 含 actionlint workflow 门禁（#444）：workflow 解析错误在 PR 即失败
 

README.md

@@ -273,7 +273,7 @@ EMRG doesn't just keep up — it catches up on its own.
 git clone https://github.com/argszero/emrg.git
 cd emrg
 uv sync # install deps
-uv run pytest tests/ -v # run tests (currently 650 items)
+uv run pytest tests/ -v # run tests (currently 652 items)
 uv run python -m emrg # launch TUI
 # CI includes actionlint workflow gate (#444): workflow parse errors fail PR CI
 

emrg/client/app.py

@@ -310,6 +310,9 @@ async def _reconnect():
  # close stale connection
  try: await conn.close()
  except Exception: pass
+ # Rant 2026-08-09T13:16:36 ⑤: spawn 节流命中后提示宿主手动启动
+ # （否则每 1s 静默重试 spawn 一台新 daemon，Windows 上即弹窗风暴）。
+ _throttle_warned = False
  while True:
  try:
  await asyncio.sleep(1)
@@ -320,6 +323,13 @@ async def _reconnect():
  status.update(center=server_id or "emrg")
  term.render()
  return
+ except RuntimeError as e:
+ if "failed to start after" in str(e) and not _throttle_warned:
+ _throttle_warned = True
+ chat.add("system", f"⚠ {e}")
+ status.update(center="daemon down — run 'emrg server'")
+ term.render()
+ continue
  except Exception:
  continue
 

emrg/client/daemon_manager.py

@@ -70,9 +70,25 @@ def is_running() -> bool:
  return is_server_running_sync()
 
 
+# Rant 2026-08-09T13:16:36 ⑤（防风暴总闸）：daemon 启动失败时不得无限重拉——
+# TUI app.py _reconnect 循环每 1s 调 ensure_connected → start_daemon 会每 1s
+# spawn 一个新 daemon 进程（Windows 上每个 spawn 都是 cmd 窗口来源）。单个
+# "连接生命周期"内最多 _MAX_SPAWN_ATTEMPTS 次 spawn，超限抛错提示宿主手动
+# `emrg server`；成功连接后归零。
+_MAX_SPAWN_ATTEMPTS = 3
+_spawn_attempts = 0
+
+
 async def start_daemon() -> subprocess.Popen:
  """Start emrgd in the background and wait until it accepts connections."""
- logger.info("starting emrgd daemon...")
+ global _spawn_attempts
+ if _spawn_attempts >= _MAX_SPAWN_ATTEMPTS:
+ raise RuntimeError(
+ f"daemon failed to start after {_MAX_SPAWN_ATTEMPTS} attempts — "
+ "please run 'emrg server' manually and check emrgd.log"
+ )
+ _spawn_attempts += 1
+ logger.info("starting emrgd daemon (attempt %d/%d)...", _spawn_attempts, _MAX_SPAWN_ATTEMPTS)
  cleanup_server()
  proc = await asyncio.create_subprocess_exec(
  sys.executable, "-m", "emrg.server",
@@ -187,11 +203,15 @@ async def ensure_connected() -> "DaemonConnection":
 
  内部改名：check_and_restart_if_stale / is_running / start_daemon。
  """
+ global _spawn_attempts
  await check_and_restart_if_stale()
  if not is_running():
  cleanup_server()
  await start_daemon()
- return DaemonConnection(await connect_to_server())
+ conn = DaemonConnection(await connect_to_server())
+ # 连接生命周期成功 → spawn 节流计数归零（对照 GUI daemon_client.js auth_ok）
+ _spawn_attempts = 0
+ return conn
 
 
 # ── 协议客户端封装 ─────────────────────────────────────────────────────

tests/test_daemon_manager.py

@@ -24,6 +24,14 @@
 from emrg.client import daemon_manager
 
 
+@pytest.fixture(autouse=True)
+def _reset_spawn_attempts():
+ """每个测试前重置模块级 spawn 节流计数（跨测试状态不泄漏）。"""
+ daemon_manager._spawn_attempts = 0
+ yield
+ daemon_manager._spawn_attempts = 0
+
+
 class FakeWS:
  """Minimal websockets-like fake: send/recv/close."""
 
@@ -233,6 +241,54 @@ async def _run():
  asyncio.run(_run())
 
 
+# ── Spawn throttle (rant 2026-08-09T13:16:36 ⑤) ─────────────
+# TUI _reconnect 循环每 1s 调 ensure_connected → start_daemon 会每 1s spawn 一台
+# 新 daemon（Windows 每个 spawn 都是 cmd 窗口来源）。单个连接生命周期内最多
+# _MAX_SPAWN_ATTEMPTS 次，超限抛节流错误；成功连接后归零。
+
+class TestSpawnThrottle:
+ @patch("emrg.client.daemon_manager.is_running", return_value=False)
+ @patch("emrg.client.daemon_manager.cleanup_server")
+ @patch("emrg.client.daemon_manager.asyncio.create_subprocess_exec",
+ new_callable=AsyncMock)
+ def test_start_daemon_throttles_after_max_attempts(self, mock_spawn,
+ mock_cleanup, mock_is_running):
+ proc = MagicMock(pid=1234)
+ mock_spawn.return_value = proc
+
+ async def _run():
+ # 前 3 次：真正 spawn（is_running 恒 False → 超时抛错）
+ for _ in range(3):
+ with pytest.raises(RuntimeError, match="failed to start"):
+ await daemon_manager.start_daemon()
+ assert mock_spawn.await_count == 3
+ # 第 4 次：不再 spawn，直接抛节流错误（提示手动 emrg server）
+ with pytest.raises(RuntimeError, match="after 3 attempts"):
+ await daemon_manager.start_daemon()
+ assert mock_spawn.await_count == 3, "超过上限后不得再 spawn（防弹窗/重试风暴）"
+ assert daemon_manager._spawn_attempts == 3
+
+ asyncio.run(_run())
+
+ @patch("emrg.client.daemon_manager.is_running", return_value=False)
+ @patch("emrg.client.daemon_manager.check_and_restart_if_stale",
+ new_callable=AsyncMock)
+ @patch("emrg.client.daemon_manager.start_daemon", new_callable=AsyncMock)
+ @patch("emrg.client.daemon_manager.cleanup_server")
+ @patch("emrg.client.daemon_manager.connect_to_server", new_callable=AsyncMock)
+ def test_spawn_attempts_reset_on_success(self, mock_connect, mock_cleanup,
+ mock_start, mock_check, mock_running):
+ daemon_manager._spawn_attempts = 2 # 模拟已有失败
+ ws = FakeWS([json.dumps({"type": "auth_ok"})])
+ mock_connect.return_value = ws
+
+ async def _run():
+ await daemon_manager.ensure_connected()
+ assert daemon_manager._spawn_attempts == 0, "成功连接后节流计数必须归零"
+
+ asyncio.run(_run())
+
+
 # ── DaemonConnection ─────────────────────────────────────────
 
 class TestDaemonConnection: