opencode-feishu-plugin 0.2.0 → 0.2.2

package/README.en.md

@@ -414,12 +414,14 @@ OpenCode **evicts idle Locations**, which unloads plugins and closes the Feishu
 
 Both dispose the plugin (closing the Feishu WS). **After that, no request means no recovery — the bot stays silent permanently** (see issues [#51343](https://github.com/anomalyco/opencode/issues/51343), [#48691](https://github.com/anomalyco/opencode/issues/48691), [#51828](https://github.com/anomalyco/opencode/issues/51828); the TTL has no config knob).
 
-The plugin ships a two-channel keep-alive (default every 20 min):
+The plugin ships two built-in layers (both on by default, **no external script required**):
 
-1. **session-scoped `GET /api/session/{id}`** → `locations.get()` renews the LayerMap entry; if it was already evicted, this request **re-creates the Location** (plugin reloaded, WS reconnected);
-2. **create + immediately delete a probe session** → the `session.created` event renews `LocationActivity`.
+1. **Gateway keep-alive** (every 20 min, `keepaliveIntervalMs`): ① a session-scoped `GET /api/session/{id}` → `locations.get()` renews the LayerMap entry, and **re-creates the Location** if it was evicted; ② create + GET + delete a probe session → renews `LocationActivity` (via the `session.created` event).
+2. **Process-wide gateway watchdog** (same interval, exactly one timer per process): every Location's plugin instance registers it, and it performs a session-scoped GET against the gateway Location. Result:
+   - if the gateway instance was evicted, it is revived automatically as long as **any other Location** still has a loaded instance (e.g. you have a TUI/Web open in another project);
+   - after a **service restart**, the first use of any Location starts the watchdog, which fires an immediate probe after ~3s and brings the gateway back up.
 
-> **External safety net**: a plugin cannot revive itself once evicted (its timer dies with it), and it is not loaded after a **service restart** either. Add a cron / systemd timer that performs one session-scoped GET every 15–30 minutes (see `scripts/keepalive-feishu.sh`) to also cover restarts and long sleeps.
+> **No external script / cron / systemd setup is required**: both layers run inside the plugin process. The only case neither can cover is "the opencode process is fully down and nothing is used for a long time" — no plugin can run then; the watchdog restores the gateway as soon as opencode is used again. Set `keepalive: false` to disable all keep-alive.
 
 ---
 
@@ -471,7 +473,7 @@ otherwise (per session preset) → ask ─────────────
 | No plugin logs | Plugin stderr is discarded in service mode; set `logFile: true` and read `<configDir>/plugins/feishu.log` |
 | Main chat replies with a hint card | Expected: the main chat is management-only. Use `/new` and work inside a topic; set `threadRouting: false` to revert |
 | Session looks stuck and messages only queue | The watchdog auto-interrupts it after `staleExecutionMs` (default 5 min) and cancels the queue, then sends a notice card; you can also tap the card's "⏹ force stop" or send `/stop` |
-| **Bot goes completely silent after ~1 hour idle** (no "long connection started" in the log) | OpenCode evicted the idle Location (hardcoded 60-min TTL). The built-in keep-alive is on by default; if it still happens (service restart / long sleep / keep-alive disabled), one session-scoped request re-creates it: `opencode api get /api/session/{id}`, or set up the external cron described under "Location keep-alive" |
+| **Bot goes completely silent after ~1 hour idle** (no "long connection started" in the log) | OpenCode evicted the idle Location (hardcoded 60-min TTL). The built-in keep-alive + process-wide watchdog are on by default and restore it automatically; you can also force it manually with one session-scoped request: `opencode api get /api/session/{id}`. If it never recovers, check that `keepalive` is not set to `false` |
 | Switched `/model` but older messages still show the old model | Expected: a switch only affects **subsequent** replies; history keeps each message's model. The receipt / run-card footer / `/current` all show the read-back truth |
 
 ---

package/README.md

@@ -293,12 +293,17 @@ opencode 会**回收空闲的 location**，这会连带卸载插件、关闭飞
 
 两者都会让插件被 dispose（飞书长连接关闭）。**此后若该 location 再无请求，插件不会自行恢复 → 机器人永久沉默**（官方 issue：[#51343](https://github.com/anomalyco/opencode/issues/51343)、[#51891→#48691](https://github.com/anomalyco/opencode/issues/48691)、[#51828](https://github.com/anomalyco/opencode/issues/51828)；TTL 无配置项）。
 
-插件内置双通道保活（默认每 20 分钟）：
+插件内置两道防线（都是默认开启，**无需任何外部脚本**）：
 
-1. **会话级 `GET /api/session/{id}`** → `locations.get()` 续期 LayerMap；若已被回收，**该请求会重建 location**（插件重新加载、长连接重连）；
-2. **创建 + 立即删除探针会话** → `session.created` 事件续期 `LocationActivity`。
+1. **网关保活**（每 20 分钟，`keepaliveIntervalMs`）：① 会话级 `GET /api/session/{id}` → `locations.get()` 续期 LayerMap，若已被回收则**重建 location**；② 创建 + GET + 删除一个探针会话 → 续期 `LocationActivity`（`session.created` 事件）。
+2. **进程级网关看门狗**（同样每 20 分钟，每进程仅一个定时器）：**任意** location 的插件实例都会登记，周期性对网关 location 做会话级 GET。效果：
+   - 网关实例即使已被回收，只要进程里还有**别的** location 存活（例如你在别的项目里开了 TUI/Web），网关会被自动救活；
+   - **服务重启后**，你第一次使用任意 location 时看门狗即启动（并在约 3 秒后立即探测一次），网关随之上线；
+   - 首次探测带 3 秒延迟，重启后恢复很快。
 
-> **外部兜底建议**：插件自身被回收后无法自救（定时器随插件销毁），且**服务重启后**也不会自动加载。可再加一条 cron / systemd timer 每 15–30 分钟做一次「会话级 GET」作为唤起器（`keepalive-feishu.sh` 示例见仓库 `scripts/`），覆盖重启 / 长时间休眠场景。
+> **不需要任何外部脚本 / cron / systemd 配置**：以上两道防线都在插件进程内完成。
+> 唯一无法覆盖的是「opencode 进程整个挂掉且长时间无人使用」——此时任何插件都无从执行；
+> 重新使用 opencode 时会由看门狗自动恢复。`keepalive: false` 可关闭全部保活。
 
 ---
 
@@ -317,7 +322,7 @@ opencode 会**回收空闲的 location**，这会连带卸载插件、关闭飞
 | 卡片内容被截断 | 飞书卡片上限约 30KB，超长会话丢弃卡片上最旧块（完整内容仍在会话里） |
 | 建会话后没看到话题 | 表单卡会改写为「✅ 已创建 · …」并附手动创建话题指引 |
 | 会话像卡死、发消息只排队 | 看门狗默认 5 分钟后自动中断；也可点「⏹ 强制停止」或发 `/stop` |
-| **空闲约 1 小时后机器人完全失联**（日志无「长连接已启动」） | opencode 回收了空闲 location（60 分钟硬编码 TTL）。内置保活默认开启；若仍发生（服务重启 / 长休眠 / 保活被关），用 `opencode api get /api/session/{id}` 任一本地会话即可唤起，或按上方「位置保活」配置外部 cron |
+| **空闲约 1 小时后机器人完全失联**（日志无「长连接已启动」） | opencode 回收了空闲 location（60 分钟硬编码 TTL）。内置保活 + 进程级看门狗默认开启，会自动恢复；也可手动用 `opencode api get /api/session/{id}`（任一本地会话）立即唤起。若长期不恢复，确认 `keepalive` 未被设为 `false` |
 | 看不到插件日志 | 服务模式下 stderr 被丢弃，设 `logFile: true` |
 | 切了 `/model` 但历史还是旧模型 | 预期行为：切换只影响后续回复，历史消息保留各自当时的模型 |
 

package/dist/index.js

@@ -147137,6 +147137,15 @@ async function touchLocationOverHttp(directory, deps) {
     deps.log.debug("\u4FDD\u6D3B\u63A2\u9488\u521B\u5EFA\u5F02\u5E38", { directory, error: errorMessage(err2) });
   }
   if (sessionID) {
+    try {
+      const getRes = await doFetch(`${service3.url}/api/session/${encodeURIComponent(sessionID)}`, {
+        method: "GET",
+        headers,
+        signal: signal()
+      });
+      touched = touched || getRes.ok;
+    } catch {
+    }
     try {
       await doFetch(`${service3.url}/api/session/${encodeURIComponent(sessionID)}`, {
         method: "DELETE",
@@ -147149,6 +147158,44 @@ async function touchLocationOverHttp(directory, deps) {
   if (touched) deps.log.debug("\u4FDD\u6D3B\u5FC3\u8DF3\u5DF2\u53D1\u9001", { directory, probe: sessionID ?? "(none)" });
   return touched;
 }
+var WATCHDOG_SLOT = /* @__PURE__ */ Symbol.for("opencode-feishu-v2/gateway-watchdog");
+function ensureGatewayWatchdog(input) {
+  const g = globalThis;
+  const existing = g[WATCHDOG_SLOT];
+  if (existing) {
+    if (input.authoritative && existing.target !== input.directory) {
+      existing.target = input.directory;
+      input.log.debug("\u7F51\u5173\u770B\u95E8\u72D7\u76EE\u6807\u66F4\u65B0\u4E3A\u7F51\u5173 location", { directory: input.directory });
+    }
+    return false;
+  }
+  const setIntervalImpl = input.setIntervalImpl ?? setInterval;
+  input.clearIntervalImpl ?? clearInterval;
+  const setTimeoutImpl = input.setTimeoutImpl ?? setTimeout;
+  input.clearTimeoutImpl ?? clearTimeout;
+  const probe = input.probe ?? ((dir) => touchLocationOverHttp(dir, { log: input.log }));
+  const slot = {
+    target: input.directory,
+    timer: void 0,
+    initial: void 0
+  };
+  const run2 = () => {
+    void probe(slot.target).catch((err2) => {
+      input.log.debug("\u7F51\u5173\u770B\u95E8\u72D7\u63A2\u6D4B\u5F02\u5E38", { error: errorMessage(err2) });
+    });
+  };
+  const delay3 = input.immediateDelayMs ?? 3e3;
+  if (delay3 > 0) slot.initial = setTimeoutImpl(run2, delay3);
+  slot.timer = setIntervalImpl(run2, input.intervalMs);
+  slot.timer.unref?.();
+  slot.initial?.unref?.();
+  g[WATCHDOG_SLOT] = slot;
+  input.log.info("\u7F51\u5173\u770B\u95E8\u72D7\u5DF2\u542F\u52A8\uFF08\u8FDB\u7A0B\u7EA7\uFF09", {
+    directory: input.directory,
+    intervalMs: input.intervalMs
+  });
+  return true;
+}
 function startKeepalive(deps) {
   const setIntervalImpl = deps.setIntervalImpl ?? setInterval;
   const clearIntervalImpl = deps.clearIntervalImpl ?? clearInterval;
@@ -150917,24 +150964,37 @@ var index_default = plugin_exports.define({
       hasAppSecret: hasSecret(config.appSecret)
     });
     const here = ctx.location?.directory;
+    let startedWatchdog = false;
+    if (config.keepalive) {
+      const watchdogTarget = config.gatewayLocation ?? here;
+      if (watchdogTarget) {
+        startedWatchdog = ensureGatewayWatchdog({
+          log: log2,
+          directory: watchdogTarget,
+          intervalMs: config.keepaliveIntervalMs
+        });
+      }
+    }
     if (config.gatewayLocation) {
       const matched = Boolean(here) && isUnder(here, config.gatewayLocation);
+      const keepSink = matched || startedWatchdog;
       trackGatewayLocationSeen({
         here: here ?? "(unknown)",
         expected: config.gatewayLocation,
         matched,
         warn: (message) => log2.warn(message),
-        ...matched ? {} : { onSettled: () => logSink?.close() }
+        ...keepSink ? {} : { onSettled: () => logSink?.close() }
       });
       if (!matched) {
         log2.debug("\u8DF3\u8FC7\u975E\u7F51\u5173 location", { here, expected: config.gatewayLocation });
+        if (!keepSink) logSink?.close();
         return async () => {
         };
       }
     }
     if (!acquireProcessGuard()) {
       log2.debug("\u68C0\u6D4B\u5230\u540C\u8FDB\u7A0B\u91CD\u590D setup\uFF0C\u8DF3\u8FC7\u542F\u52A8\uFF08\u4EC5\u9996\u4E2A\u5B9E\u4F8B\u751F\u6548\uFF09");
-      logSink?.close();
+      if (!startedWatchdog) logSink?.close();
       return async () => {
       };
     }
@@ -151487,6 +151547,15 @@ async function start(ctx, config, log2, logSink) {
   }
   const here = ctx.location?.directory;
   const keepaliveDirectory = here ?? config.gatewayLocation;
+  if (config.keepalive && keepaliveDirectory) {
+    ensureGatewayWatchdog({
+      log: log2,
+      directory: keepaliveDirectory,
+      intervalMs: config.keepaliveIntervalMs,
+      authoritative: true,
+      immediateDelayMs: 0
+    });
+  }
   const stopKeepalive = config.keepalive && keepaliveDirectory ? startKeepalive({
     log: log2,
     directory: keepaliveDirectory,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "opencode-feishu-plugin",
-  "version": "0.2.0",
+  "version": "0.2.2",
   "description": "Feishu/Lark plugin for OpenCode V2 — single-user topics as sessions, in-card permission approvals, streaming cards. Minimal permissions, no public endpoint.",
   "type": "module",
   "main": "index.js",