memory-eternal 0.10.39 → 0.10.41

package/lib/watchdog.js

@@ -351,6 +351,8 @@ export async function stopWatchdogs({ env = process.env, port, timeoutMs = 5000,
     for (const p of ports) {
       const found = await probeListener(p)
       if (found && found.pid) portOwners.push({ port: p, pid: Number(found.pid), command: String(found.command || '') })
+      // 「未知」要留痕（issue #45 后续）：静默等于把人引回「端口空闲」的老误判
+      else if (found && found.unknown) warnings.push(`端口 ${p} 的占用者无法判定（进程枚举超时/不可用）—— 未知，不代表空闲`)
     }
     if (portOwners.length && forcePortOccupant) {
       for (const occ of portOwners) {
@@ -576,6 +578,46 @@ export async function probe(port, timeoutMs = 1000) {
  *
  * @returns {Promise<string>} 取不到返回空串
  */
+/**
+ * 内部：向某端口发一条 GET 并解析 JSON。**每次都销毁 socket**（issue #45 后续第 4 条）。
+ * @returns {Promise<{status:number, data:any, error:string}>}
+ */
+function requestJson(port, route, timeoutMs = 1200) {
+  return new Promise((resolve) => {
+    const req = http.request({ host: '127.0.0.1', port, path: route, method: 'GET', timeout: timeoutMs }, (res) => {
+      const chunks = []
+      res.on('data', (c) => chunks.push(c))
+      res.on('end', () => {
+        try { resolve({ status: res.statusCode || 0, data: JSON.parse(Buffer.concat(chunks).toString('utf8')), error: '' }) }
+        catch { resolve({ status: res.statusCode || 0, data: null, error: 'BAD_JSON' }) }
+      })
+      res.on('error', (e) => { try { res.destroy() } catch {} resolve({ status: 0, data: null, error: 'RES_ERROR:' + String((e && e.code) || e) }) })
+    })
+    req.on('error', (e) => resolve({ status: 0, data: null, error: String((e && e.code) || e) }))
+    req.on('timeout', () => { try { req.destroy() } catch {} resolve({ status: 0, data: null, error: 'TIMEOUT' }) })
+    req.end()
+  })
+}
+
+/**
+ * 「端口上的版本」+ **失败原因**（issue #45 后续第 3 条：`超时` 与 `没有服务` 必须能分开报）。
+ * 纯诊断用；返回的 version 为空时，timedOut 告诉你这是超时还是根本没人应答。
+ * @returns {Promise<{version:string, timedOut:boolean, lastError:string}>}
+ */
+export async function probeServedVersionDetailed(port, timeoutMs = 1200, { tries = 3 } = {}) {
+  let lastError = ''
+  for (let i = 0; i < Math.max(1, tries); i++) {
+    const a = await requestJson(port, `${API_PREFIX}/version-check`, timeoutMs)
+    if (a.data && a.data.ok === true && typeof a.data.loaded === 'string' && a.data.loaded) return { version: String(a.data.loaded), timedOut: false, lastError: '' }
+    if (a.error) lastError = a.error
+    const b = await requestJson(port, `${API_PREFIX}/overview`, timeoutMs)
+    if (b.data && typeof b.data.version === 'string' && b.data.version) return { version: String(b.data.version), timedOut: false, lastError: '' }
+    if (b.error) lastError = b.error
+    if (i < tries - 1) await new Promise((r) => setTimeout(r, 150 * (i + 1)))
+  }
+  return { version: '', timedOut: /TIMEOUT/i.test(lastError), lastError }
+}
+
 export async function probeServedVersion(port, timeoutMs = 1200) {
   // issue #31/#33 第②层：`/overview` 会碰数据库，库只读/损坏时它 500 —— 于是「端口上明明活着、
   // 版本却拿不到」：verify 报「（无响应）」、status 甚至误判成「v0.10.4 之前的旧版」，进而触发
@@ -603,7 +645,7 @@ export async function probeServedVersion(port, timeoutMs = 1200) {
 async function probeViaVersionCheck(port, timeoutMs = 1200) {
   return new Promise((resolve) => {
     const req = http.request({ host: '127.0.0.1', port, path: `${API_PREFIX}/version-check`, method: 'GET', timeout: timeoutMs }, (res) => {
-      if (res.statusCode !== 200) return resolve('')
+      if (res.statusCode !== 200) { try { res.resume(); res.destroy() } catch { /* 已关闭 */ } return resolve('') }
       const chunks = []
       res.on('data', (c) => chunks.push(c))
       res.on('end', () => {
@@ -917,7 +959,8 @@ export async function clearResidentForTakeover({ port = 7999, env = process.env,
   const findListener = deps.findListener || findPortListener
   // 枚举本身要能被时间预算约束（否则 8s 预算被一次 13.6s 的 PowerShell 查询吃光，决策基于过期数据）
   const listenerOpts = { timeoutMs: Math.max(500, Math.min(2500, Number(timeoutMs) || 8000)) }
-  const out = { stopped: [], webStopped: [], portStopped: [], warnings: [], failed: [], portFree: false }
+  // portFree 只是兼容布尔；**判定看 portState**（issue #45 后续：超时 ≠ 空闲）。
+  const out = { stopped: [], webStopped: [], portStopped: [], warnings: [], failed: [], portFree: false, portState: 'unknown' }
   const rawKill = deps.kill || ((pid) => { try { process.kill(pid, 'SIGTERM'); return true } catch { return false } })
   // 记下「已经判定收不掉」的 pid：立刻再来一次只会同样失败，白折腾还可能误导用户以为是权限抖动
   const failedKills = new Set()
@@ -949,7 +992,14 @@ export async function clearResidentForTakeover({ port = 7999, env = process.env,
   }
   if (left && left.pid) {
     out.portFree = false
+    out.portState = 'occupied'
     out.warnings.push(`端口 ${p} 仍被 pid ${left.pid} 占用（${left.command || '命令行未知'}）`)
+  } else if (left && left.unknown) {
+    // 枚举器超时/不可用 —— 这是「未知」，不是「空闲」：中止接管（不杀、不漂移、不动锁）。
+    out.portFree = false
+    out.portState = 'unknown'
+    out.unknown = true
+    out.warnings.push(`端口 ${p} 的占用者**无法判定**（进程枚举超时/不可用）→ 已中止接管：既不按「端口空闲」继续，也不把 web 挪到 port+1`)
   } else {
     // 枚举不到 ≠ 端口是空的（issue #30 现场）：受限环境/被拒的进程查询会让 findListener 返回 null，
     // 旧实现据此认为"端口已空出" → 新 web 绑不上 7999 → 退让或失败 → 自检「（无响应）」+ 锁被清空。
@@ -959,9 +1009,11 @@ export async function clearResidentForTakeover({ port = 7999, env = process.env,
     try { listening = Boolean(await probe(p)) } catch { listening = false }
     if (listening) {
       out.portFree = false
+      out.portState = 'occupied'
       out.warnings.push(`端口 ${p} 仍在监听，但占用者无法识别（进程枚举被拒/不可用）—— 已中止接管，且未改动任何锁登记`)
     } else {
       out.portFree = true
+      out.portState = 'free'
     }
   }
   return out
@@ -1001,6 +1053,31 @@ export function parseLsofFields(text) {
  * @param {{exec?: (cmd:string,args:string[])=>Promise<string>}} [opts]
  * @returns {Promise<{pid:number, command:string}|null>}
  */
+/**
+ * 纯函数：从 `netstat -ano` 输出里找出**监听指定端口**的 pid（issue #45 后续实测）。
+ *
+ * 为什么它是 Windows 上的首选：报告者实测 `Get-NetTCPConnection` 在本机要 **4.1–61.4 秒**
+ * （TCP 表本身很大，它要先枚举整表再过滤），而 `netstat -ano` 全量只要 **1.8 秒**，且直接给 pid。
+ * netstat 不连端口、不枚举两次 —— 对「端口上那个进程把 backlog 打满、连不上」的现场
+ * （正是 0.10.36 自探递归留下的状态）也能给出正确结论。
+ *
+ * @returns {number} pid，没找到返回 0
+ */
+export function parseNetstatListener(text, port) {
+  const want = String(Number(port) || 0)
+  if (!want || want === '0') return 0
+  for (const line of String(text || '').split(/\r?\n/)) {
+    if (!/LISTENING/i.test(line)) continue
+    // 形如： TCP    127.0.0.1:7999    0.0.0.0:0    LISTENING    10924
+    //       TCP    [::1]:7999        [::]:0       LISTENING    10924
+    const m = /^\s*TCP\s+\S*:(\d+)\s+\S*:\d+\s+LISTENING\s+(\d+)\s*$/i.exec(line.trim())
+    if (!m) continue
+    if (String(Number(m[1])) !== want) continue
+    return Number(m[2]) || 0
+  }
+  return 0
+}
+
 export async function findPortListener(port, { exec, platform = process.platform, timeoutMs = 4000 } = {}) {
   const p = Number(port)
   if (!Number.isFinite(p) || p <= 0) return null
@@ -1008,6 +1085,24 @@ export async function findPortListener(port, { exec, platform = process.platform
   const execOpts = { timeoutMs: Math.max(200, Number(timeoutMs) || 4000) }
   try {
     if (platform === 'win32') {
+      // ① netstat -ano 优先（issue #45 后续）：不连端口，1.8s 量级给出监听 pid。
+      //    PowerShell 的 Get-NetTCPConnection 在 TCP 表大的机器上实测 4.1–61.4s ——
+      //    一旦超时，旧实现返回 null，调用方把它当「端口空闲」，于是新看门狗去绑一个**还被旧 web 占着**的端口，
+      //    只能漂到 port+1，旧 web 永远活着、修复永远轮不到执行（报告者的现场）。
+      let netOut = null
+      try { netOut = await runExec('netstat', ['-ano'], { timeoutMs: Math.max(3000, execOpts.timeoutMs) }) } catch { netOut = null }
+      const netPid = netOut == null ? 0 : parseNetstatListener(netOut, p)
+      if (netPid) {
+        let cmd = ''
+        try {
+          cmd = String((await runExec('powershell', ['-NoProfile', '-Command', `(Get-CimInstance Win32_Process -Filter "ProcessId=${netPid}" | Select-Object -ExpandProperty CommandLine)`], execOpts)) || '').trim()
+        } catch { cmd = '' }
+        return { pid: netPid, command: cmd }
+      }
+      // netOut 为 null = 命令没跑起来/被杀超时 → **未知**，绝不能当成「端口空闲」
+      // 空输出同样不可信：netstat 正常时至少会有一堆行；空 = 命令没跑起来/被超时杀掉（defaultExec 超时返回 ''）
+      if (netOut == null || String(netOut).trim() === '') return { pid: 0, command: '', unknown: true }
+      // ② 兜底：老路径（netstat 明确没有该端口的 LISTEN 行时，用 PowerShell 再确认一次，兼容旧系统）
       const out = await runExec('powershell', ['-NoProfile', '-Command', `Get-NetTCPConnection -LocalPort ${p} -State Listen -ErrorAction SilentlyContinue | Select-Object -ExpandProperty OwningProcess`], execOpts)
       const pid = Number(String(out || '').split(/\r?\n/).map((s) => s.trim()).find((s) => /^\d+$/.test(s))) || 0
       if (!pid) return null
@@ -1400,7 +1495,7 @@ if (process.argv[1] && /[\\/]watchdog\.js$/.test(process.argv[1])) {
       const claim = acquireWatchdogLock({ port, vault: vaultRoot || '', pkgVersion: currentPkgVersion(), force: true })
       if (claim.acquired) log('已登记本次替换尝试（端口仍被占用，看门狗未启动）：pid ' + process.pid)
       else log('⚠ 未能登记本次替换（锁不可写？）：' + (claim.file || ''))
-      const why = `${summary} —— 端口没空出来，已中止接管（不退让到 port+1，避免又留一个孤儿 web）`
+      const why = `${summary} —— ${cleared.portState === 'unknown' ? '占用者无法判定（枚举超时/不可用），按未知处理' : '端口没空出来'}，已中止接管（不退让到 port+1，避免又留一个孤儿 web）`
       await noteReplace('fail', why)
       process.exit(1)
     }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "memory-eternal",
-  "version": "0.10.39",
+  "version": "0.10.41",
   "description": "记忆核心（Memory Eternal）：自研的 DeepSeek Harness 记忆插件，不移植任何既有记忆框架——对话结束后自动沉淀知识卡到本地 Markdown Vault（自研去重、自研 CJK 检索、可 git 管理），设置页提供图形化知识库（统计 / 搜索 / 知识图谱 + 侧边栏一键弹窗），Agent 通过 memory_recall 工具按需召回历史上下文。零人工干预。",
   "license": "MIT",
   "repository": {
@@ -64,7 +64,7 @@
     "prepublishOnly": "node build.mjs",
     "release": "node scripts/release.mjs",
     "release:dry": "node scripts/release.mjs --dry-run",
-    "test": "node build.mjs && node tests/vault.test.mjs && node tests/capture.test.mjs && node tests/capture-route.test.mjs && node tests/stall.test.mjs && node tests/vault-path.test.mjs && node tests/vault-resolve.test.mjs && node tests/minhash.test.mjs && node tests/settings-save.test.mjs && node tests/config-coverage.test.mjs && node tests/config-sync.test.mjs && node tests/config-merge.test.mjs && node tests/sse.test.mjs && node tests/api.test.mjs && node tests/import-roundtrip.test.mjs && node tests/graph-perf.test.mjs && node tests/settings-compat.test.mjs && node tests/node-bin.test.mjs && node tests/watchdog-lock.test.mjs && node tests/markdown.test.mjs && node tests/feedback.test.mjs && node tests/graph-lod.test.mjs && node tests/ssr.test.mjs && node tests/web-assets.test.mjs && node tests/manifests.test.mjs && node tests/two-store.test.mjs && node tests/noise-gate.test.mjs && node tests/e2e-recall.test.mjs && node tests/issue-fixes.test.mjs && node tests/issue-fixes-2.test.mjs && node tests/issue-fixes-3.test.mjs && node tests/watchdog-orphans.test.mjs && node tests/watchdog-restart.test.mjs && node tests/readme-i18n.test.mjs && node tests/issue-fixes-4.test.mjs && node tests/host-drain.test.mjs && node tests/e2e-host-drain.mjs && node tests/issue-fixes-5.test.mjs && node tests/issue-fixes-6.test.mjs && node tests/capture-json-repair.test.mjs && node tests/csrf-origin.test.mjs && node tests/watchdog-orphan-web.test.mjs && node tests/config-validate.test.mjs && node tests/issue-fixes-8.test.mjs && node tests/issue-fixes-9.test.mjs && node tests/release-verify.test.mjs && node tests/refine.test.mjs && node tests/db-restore.test.mjs && node tests/align.test.mjs && node tests/tags-normalize.test.mjs && node tests/resident-state.test.mjs && node tests/llm-route.test.mjs && node tests/host-managed.test.mjs && node tests/replace-claim.test.mjs && node tests/version-check-self.test.mjs && node tests/setup-codex.test.mjs"
+    "test": "node build.mjs && node tests/vault.test.mjs && node tests/capture.test.mjs && node tests/capture-route.test.mjs && node tests/stall.test.mjs && node tests/vault-path.test.mjs && node tests/vault-resolve.test.mjs && node tests/minhash.test.mjs && node tests/settings-save.test.mjs && node tests/config-coverage.test.mjs && node tests/config-sync.test.mjs && node tests/config-merge.test.mjs && node tests/sse.test.mjs && node tests/api.test.mjs && node tests/import-roundtrip.test.mjs && node tests/graph-perf.test.mjs && node tests/settings-compat.test.mjs && node tests/node-bin.test.mjs && node tests/watchdog-lock.test.mjs && node tests/markdown.test.mjs && node tests/feedback.test.mjs && node tests/graph-lod.test.mjs && node tests/ssr.test.mjs && node tests/web-assets.test.mjs && node tests/manifests.test.mjs && node tests/two-store.test.mjs && node tests/noise-gate.test.mjs && node tests/e2e-recall.test.mjs && node tests/issue-fixes.test.mjs && node tests/issue-fixes-2.test.mjs && node tests/issue-fixes-3.test.mjs && node tests/watchdog-orphans.test.mjs && node tests/watchdog-restart.test.mjs && node tests/readme-i18n.test.mjs && node tests/issue-fixes-4.test.mjs && node tests/host-drain.test.mjs && node tests/e2e-host-drain.mjs && node tests/issue-fixes-5.test.mjs && node tests/issue-fixes-6.test.mjs && node tests/capture-json-repair.test.mjs && node tests/csrf-origin.test.mjs && node tests/watchdog-orphan-web.test.mjs && node tests/config-validate.test.mjs && node tests/issue-fixes-8.test.mjs && node tests/issue-fixes-9.test.mjs && node tests/release-verify.test.mjs && node tests/refine.test.mjs && node tests/db-restore.test.mjs && node tests/align.test.mjs && node tests/tags-normalize.test.mjs && node tests/resident-state.test.mjs && node tests/llm-route.test.mjs && node tests/host-managed.test.mjs && node tests/replace-claim.test.mjs && node tests/version-check-self.test.mjs && node tests/setup-codex.test.mjs && node tests/listener-netstat.test.mjs && node tests/issue-46-overlay.test.mjs && node tests/graph-initial-zoom.test.mjs"
   },
   "dependencies": {
     "@deepseek-ai/schemastery": "^3.18.1"