修复 AGC Ctrl+C 残留上个工作树后端导致切换工作树复用旧后端 (#315)

closes #314 ## 现象 `npm run agc` 按 Ctrl+C 后有概率残留上个工作树的 `api-server.exe` / SpacetimeDB，切换 worktree 再启动时 AGC 复用旧后端，改过数据库 / schema 的工作树会串库。 ## 根因 1. Windows 下长驻服务都经 Node `shell: true` 的 `cmd.exe /d /s /c` 包装层启动，Ctrl+C 先杀包装层（`0xC000013A`）；`dev.mjs` 的 `stopProcess` 见到直接子进程已退出就 return，`taskkill /PID <已退出 PID> /T /F` 也只会失败，深处的 `cargo → api-server.exe` 无人清理。 2. 按根 PID 遍历依赖快照里的父子链，中间层先消失时链断，只能拿到根 PID。 3. 复用判据只看 `.app/dev-stack.json` status 与 `/healthz`、`/readyz`、`/v1/ping`，不校验端口上的进程属于哪个工作树，残留后端照样被判为健康并复用。 ## 改动 - 新增 `scripts/dev-windows-process.mjs`：按根 PID 遍历 + 按身份匹配（`server-rs/target/debug/api-server.exe` 绝对路径、SpacetimeDB `--data-dir`）两条独立清理路径，带 1s 快照缓存避免清理被拖慢。 - `scripts/dev.mjs`：直接子进程已退出时仍按记录 PID 清理后代；退出时按身份兜底清扫本工作树后端（复用他人 standalone 时跳过）；启动前清理旧 api-server 保留 `Wait-Process` 语义，避免 `failed to remove file`。 - `apps/ai-game-creator-shell/scripts/start-dev-stack.mjs`：复用前校验端口监听进程归属，无法证明归属就不复用、改为启动本工作树后端并允许端口漂移；信号与 `finally` 各兜底清扫一次；`taskkill` 失败时降级按 PID 遍历；等待就绪时输出归属校验失败原因，避免静默超时。探测不可用时退化为旧行为，不阻断本地启动。 - 测试与文档：新增 `scripts/dev-windows-process.test.ts`、扩充 AGC 复用门禁用例；同步 `docs/project-memory/shared-memory/pitfalls.md` 与本地开发运维文档。 ## 验证 - 伪造 `api-server.exe` 进程：按身份精确命中并杀掉（`matched=[17284] stopped=[17284]`）。 - 3 个真实监听进程下归属判定：`owned` / `api-server-owner-mismatch` / `spacetime-owner-mismatch` 均正确。 - `npx vitest run scripts/dev.test.ts scripts/dev-windows-process.test.ts scripts/dev-stack-port-utils.test.ts apps/ai-game-creator-shell/tests/...`：119 passed（唯一失败为 Windows 文件权限用例，已确认在合并基线 `origin/master` 上同样失败）。 - `node --check`、`eslint --max-warnings 0`、`prettier --check`、`npm run check:encoding`、`git diff --check` 全部通过。 ## 备注 Rust 侧 `api-server` 的 `with_graceful_shutdown` 没有超时上限，是「有概率」的来源之一；本次只在 Node 侧收口，是否给优雅退出加 deadline 可另行评估。 Reviewed-on: https://git.genarrative.world/git/GenarrativeAI/Genarrative/pulls/315 Co-authored-by: Suzumiya <suzmii@qq.com> Co-committed-by: Suzumiya <suzmii@qq.com>
2026-09-09 20:08:04 +08:00
parent 7c49e7e51c
commit a1b9b24891
7 changed files with 992 additions and 103 deletions
@@ -0,0 +1,180 @@
+import { describe, expect, test } from 'vitest';
+
+import {
+  normalizeWindowsPath,
+  parseWindowsProcessSnapshot,
+  selectProcessTreeIds,
+  selectWorktreeOwnedProcessIds,
+} from './dev-windows-process.mjs';
+
+function processEntry(overrides) {
+  return {
+    ProcessId: 1,
+    ParentProcessId: 0,
+    Name: 'node.exe',
+    ExecutablePath: null,
+    CommandLine: null,
+    ...overrides,
+  };
+}
+
+describe('Windows 进程快照解析', () => {
+  test('单个进程对象也归一为数组', () => {
+    const single = parseWindowsProcessSnapshot(
+      JSON.stringify({ ProcessId: 42, ParentProcessId: 1 }),
+    );
+    expect(single).toHaveLength(1);
+    expect(single[0].ProcessId).toBe(42);
+  });
+
+  test('空输出和非法 JSON 返回空数组', () => {
+    expect(parseWindowsProcessSnapshot('')).toEqual([]);
+    expect(parseWindowsProcessSnapshot('null')).toEqual([]);
+    expect(parseWindowsProcessSnapshot('not json')).toEqual([]);
+  });
+
+  test('路径归一化去掉 \\\\?\\ 前缀、尾部分隔符并忽略大小写', () => {
+    expect(
+      normalizeWindowsPath('\\\\?\\C:\\Repo\\target\\debug\\api-server.exe'),
+    ).toBe('c:\\repo\\target\\debug\\api-server.exe');
+    expect(normalizeWindowsPath('C:\\Repo\\data\\')).toBe('c:\\repo\\data');
+    expect(normalizeWindowsPath(null)).toBe('');
+  });
+});
+
+describe('按根 PID 遍历进程树', () => {
+  test('中间层进程已从快照消失时无法再到达更深的后代', () => {
+    // cmd(10) -> wrapper(20) -> api-server(30)。Ctrl+C 先杀掉 10 和 20，
+    // 快照里只剩 30（ParentProcessId 仍指向已消失的 20），父链断开后
+    // 按根 PID 遍历只能拿到根自己，这正是后端被漏杀的原因。
+    const snapshot = [
+      processEntry({
+        ProcessId: 30,
+        ParentProcessId: 20,
+        Name: 'api-server.exe',
+      }),
+    ];
+    expect(selectProcessTreeIds(snapshot, 10)).toEqual([10]);
+  });
+
+  test('根已退出但中间层仍在快照里时仍可收全后代', () => {
+    const snapshot = [
+      processEntry({ ProcessId: 20, ParentProcessId: 10, Name: 'cargo.exe' }),
+      processEntry({
+        ProcessId: 30,
+        ParentProcessId: 20,
+        Name: 'api-server.exe',
+      }),
+    ];
+    expect(selectProcessTreeIds(snapshot, 10).sort((a, b) => a - b)).toEqual([
+      10, 20, 30,
+    ]);
+  });
+
+  test('父链完整时能收全后代', () => {
+    const snapshot = [
+      processEntry({ ProcessId: 10, ParentProcessId: 1, Name: 'cmd.exe' }),
+      processEntry({ ProcessId: 20, ParentProcessId: 10, Name: 'cargo.exe' }),
+      processEntry({
+        ProcessId: 30,
+        ParentProcessId: 20,
+        Name: 'api-server.exe',
+      }),
+      processEntry({ ProcessId: 40, ParentProcessId: 1, Name: 'other.exe' }),
+    ];
+    expect(selectProcessTreeIds(snapshot, 10).sort((a, b) => a - b)).toEqual([
+      10, 20, 30,
+    ]);
+  });
+});
+
+describe('按身份匹配本工作树后端进程', () => {
+  const apiServerExePath = 'C:\\Repo\\server-rs\\target\\debug\\api-server.exe';
+  const spacetimeDataDir =
+    'C:\\Repo\\server-rs\\.spacetimedb\\ai-game-creator\\data';
+
+  test('只收本工作树的 api-server.exe 与同一 data-dir 的 SpacetimeDB', () => {
+    const snapshot = [
+      processEntry({
+        ProcessId: 100,
+        Name: 'api-server.exe',
+        ExecutablePath: apiServerExePath,
+        CommandLine: '"server-rs\\target\\debug\\api-server.exe"',
+      }),
+      processEntry({
+        ProcessId: 101,
+        Name: 'api-server.exe',
+        ExecutablePath: 'C:\\Other\\server-rs\\target\\debug\\api-server.exe',
+      }),
+      processEntry({
+        ProcessId: 102,
+        Name: 'spacetimedb-standalone.exe',
+        ExecutablePath:
+          'C:\\Users\\me\\SpacetimeDB\\spacetimedb-standalone.exe',
+        CommandLine: `spacetimedb-standalone.exe start --data-dir ${spacetimeDataDir} --listen-addr 127.0.0.1:3101`,
+      }),
+      processEntry({
+        ProcessId: 103,
+        Name: 'spacetimedb-standalone.exe',
+        CommandLine:
+          'spacetimedb-standalone.exe start --data-dir C:\\Other\\data --listen-addr 127.0.0.1:3101',
+      }),
+      processEntry({
+        ProcessId: 104,
+        Name: 'node.exe',
+        CommandLine: `node dev.mjs --spacetime-data-dir ${spacetimeDataDir}`,
+      }),
+    ];
+
+    expect(
+      selectWorktreeOwnedProcessIds(snapshot, {
+        apiServerExePath,
+        spacetimeDataDir,
+      }).sort((a, b) => a - b),
+    ).toEqual([100, 102]);
+  });
+
+  test('只给 api-server 路径时不会误伤 SpacetimeDB', () => {
+    const snapshot = [
+      processEntry({
+        ProcessId: 100,
+        Name: 'api-server.exe',
+        ExecutablePath: apiServerExePath,
+      }),
+      processEntry({
+        ProcessId: 102,
+        Name: 'spacetimedb-standalone.exe',
+        CommandLine: `--data-dir ${spacetimeDataDir}`,
+      }),
+    ];
+
+    expect(
+      selectWorktreeOwnedProcessIds(snapshot, { apiServerExePath }),
+    ).toEqual([100]);
+  });
+
+  test('排除自身进程且路径大小写不敏感', () => {
+    const snapshot = [
+      processEntry({
+        ProcessId: 200,
+        Name: 'api-server.exe',
+        ExecutablePath: apiServerExePath.toUpperCase(),
+      }),
+    ];
+
+    expect(
+      selectWorktreeOwnedProcessIds(snapshot, {
+        apiServerExePath,
+        selfPid: 200,
+      }),
+    ).toEqual([]);
+    expect(
+      selectWorktreeOwnedProcessIds(snapshot, { apiServerExePath }),
+    ).toEqual([200]);
+  });
+
+  test('没有可匹配身份时返回空数组', () => {
+    const snapshot = [processEntry({ ProcessId: 100, Name: 'api-server.exe' })];
+    expect(selectWorktreeOwnedProcessIds(snapshot, {})).toEqual([]);
+  });
+});