From ac6687181e2e43a5e3359da7bbc4c8b7355bbd5c Mon Sep 17 00:00:00 2001
From: yilsem <yilsempazarlama@gmail.com>
Date: Sat, 12 Sep 2026 11:52:12 +0300
Subject: [PATCH] fix(windows): probe compute host liveness without sending
 signals

---
 tests/tui_gateway/test_compute_host_phase1.py | 21 +++++++++++++++++++
 tui_gateway/host_supervisor.py                | 16 +++++---------
 2 files changed, 26 insertions(+), 11 deletions(-)

diff --git a/tests/tui_gateway/test_compute_host_phase1.py b/tests/tui_gateway/test_compute_host_phase1.py
index edf78cf..f713eb7 100644
--- a/tests/tui_gateway/test_compute_host_phase1.py
+++ b/tests/tui_gateway/test_compute_host_phase1.py
@@ -15,6 +15,7 @@ from tui_gateway.host_supervisor import (
     MUTATOR_ROUTE_TABLE,
     HostSupervisor,
     append_log_record,
+    _pid_alive,
 )
 
 
@@ -164,6 +165,26 @@ def test_supervisor_startup_reconcile_pid_reuse_guard(tmp_path, monkeypatch):
     assert not registry.exists()
 
 
+def test_pid_probe_does_not_terminate_a_live_process():
+    # Use the base interpreter, not the Windows venv launcher/worker chain.
+    child = subprocess.Popen(
+        [getattr(sys, "_base_executable", sys.executable), "-c",
+         "import sys; sys.stdin.readline(); print('still-alive', flush=True)"],
+        stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE,
+        creationflags=getattr(subprocess, "CREATE_NO_WINDOW", 0),
+    )
+    try:
+        assert _pid_alive(child.pid)
+        output, errors = child.communicate(input=b"ping\n", timeout=10)
+        assert child.returncode == 0, errors.decode("utf-8", errors="replace")
+        assert output.strip() == b"still-alive", "A liveness query terminated the process"
+    finally:
+        if child.poll() is None:
+            child.terminate()
+        child.wait(timeout=10)
+    assert not _pid_alive(child.pid)
+
+
 def _make_compress_host_session(events: list) -> dict:
     class _Agent:
         model = "host-model"
diff --git a/tui_gateway/host_supervisor.py b/tui_gateway/host_supervisor.py
index ca50e3a..db644d3 100644
--- a/tui_gateway/host_supervisor.py
+++ b/tui_gateway/host_supervisor.py
@@ -22,6 +22,8 @@ from collections.abc import Callable
 from pathlib import Path
 from typing import Any
 
+import psutil
+
 from hermes_constants import get_hermes_home
 from tools.environments.local import hermes_subprocess_env
 
@@ -112,17 +114,9 @@ def _default_registry_path() -> Path:
 
 
 def _pid_alive(pid: int) -> bool:
-    if pid <= 0:
-        return False
-    try:
-        os.kill(pid, 0)
-        return True
-    except ProcessLookupError:
-        return False
-    except PermissionError:
-        return True
-    except Exception:
-        return False
+    # Signal zero is not a portable liveness query (CTRL_C_EVENT on Windows).
+    # pid_exists also distinguishes exited Windows processes with open handles.
+    return pid > 0 and psutil.pid_exists(pid)
 
 
 def _pid_command(pid: int) -> str: