Genarrative/apps/ai-game-creator-shell/scripts/start-dev-stack.mjs

import { spawn } from 'node:child_process';
import { existsSync, readdirSync, readFileSync } from 'node:fs';
import http from 'node:http';
import net from 'node:net';
import { resolve } from 'node:path';
import { fileURLToPath } from 'node:url';

import {
  agcVitePortEnvKey,
  readAgcDevEndpoint,
  resolveAgcDevEndpoint,
  withAgcDevEndpointEnv,
} from './dev-port.mjs';

const appRoot = fileURLToPath(new URL('..', import.meta.url));
const repoRoot = resolve(appRoot, '../..');
const devStackStatePath = resolve(repoRoot, '.app/dev-stack.json');
const defaultApiTarget =
  process.env.RUST_SERVER_TARGET || 'http://127.0.0.1:8082';
const backendDatabase = 'genarrative-game-creator-dev';
const backendSpacetimeDataDir = resolve(
  repoRoot,
  'server-rs/.spacetimedb/ai-game-creator/data',
);
const npm = process.platform === 'win32' ? 'npm.cmd' : 'npm';
const childLifecycles = new WeakMap();

function readJson(path) {
  if (!existsSync(path)) {
    return null;
  }
  try {
    return JSON.parse(readFileSync(path, 'utf8'));
  } catch {
    return null;
  }
}

function httpGetText(url, timeout = 1000) {
  return new Promise((resolveRequest) => {
    const request = http.get(url, { timeout }, (response) => {
      let body = '';
      response.setEncoding('utf8');
      response.on('data', (chunk) => {
        if (body.length < 4096) {
          body += chunk;
        }
      });
      response.on('end', () => {
        resolveRequest({
          statusCode: response.statusCode ?? 0,
          body,
        });
      });
    });
    request.on('timeout', () => {
      request.destroy();
      resolveRequest(null);
    });
    request.on('error', () => resolveRequest(null));
  });
}

async function isHttpReady(url) {
  const response = await httpGetText(url);
  return Boolean(
    response && response.statusCode >= 200 && response.statusCode < 300,
  );
}

function resolveBackendTargetsFromState(
  state,
  {
    requireAgcBackend = false,
    expectedDatabase = backendDatabase,
    expectedSpacetimeDataDir = backendSpacetimeDataDir,
    fallbackApiTarget = defaultApiTarget,
  } = {},
) {
  const apiServer = state?.services?.['api-server'];
  const spacetime = state?.services?.spacetime;
  const isActive = (service) =>
    service && ['running', 'reused', 'starting'].includes(service.status ?? '');
  const database = typeof state?.database === 'string' ? state.database : '';
  const spacetimeDataDir =
    typeof state?.spacetimeDataDir === 'string'
      ? resolve(state.spacetimeDataDir)
      : '';
  const hasMatchingDatabase = database === expectedDatabase;
  const hasMatchingDataDir =
    Boolean(spacetimeDataDir) &&
    spacetimeDataDir === resolve(expectedSpacetimeDataDir);
  const hasMatchingBackend = hasMatchingDatabase && hasMatchingDataDir;
  const canReuseState = !requireAgcBackend || hasMatchingBackend;
  const apiUrl =
    canReuseState && isActive(apiServer) && apiServer.url
      ? apiServer.url
      : requireAgcBackend
        ? ''
        : fallbackApiTarget;
  const spacetimeUrl =
    canReuseState && isActive(spacetime) && spacetime.url
      ? spacetime.url
      : requireAgcBackend
        ? ''
        : 'http://127.0.0.1:3101';
  return {
    apiUrl,
    spacetimeUrl,
    database,
    spacetimeDataDir,
    hasMatchingDatabase,
    hasMatchingDataDir,
    hasMatchingBackend,
  };
}

function readBackendTargets({ requireAgcBackend = false } = {}) {
  return resolveBackendTargetsFromState(readJson(devStackStatePath), {
    requireAgcBackend,
  });
}

async function isBackendReady() {
  const { apiUrl, spacetimeUrl, hasMatchingBackend } = readBackendTargets({
    requireAgcBackend: true,
  });
  return (
    hasMatchingBackend &&
    Boolean(apiUrl) &&
    Boolean(spacetimeUrl) &&
    (await isHttpReady(`${apiUrl}/healthz`)) &&
    (await isHttpReady(`${spacetimeUrl}/v1/ping`))
  );
}

async function readExistingViteServer(endpoint = readAgcDevEndpoint()) {
  return httpGetText(endpoint.url);
}

function isVitePortListening(endpoint = readAgcDevEndpoint()) {
  return new Promise((resolveRequest) => {
    const socket = net.connect({ host: endpoint.host, port: endpoint.port });
    socket.once('connect', () => {
      socket.destroy();
      resolveRequest(true);
    });
    socket.once('error', () => resolveRequest(false));
    socket.setTimeout(1000, () => {
      socket.destroy();
      resolveRequest(true);
    });
  });
}

function isAiGameCreatorServer(response) {
  return (
    response &&
    response.statusCode >= 200 &&
    response.statusCode < 500 &&
    response.body.includes('<title>陶泥儿</title>') &&
    response.body.includes('/src/main.tsx')
  );
}

async function readExistingViteMarker(endpoint = readAgcDevEndpoint()) {
  const response = await httpGetText(endpoint.markerUrl, 2000);
  if (!response || response.statusCode !== 200) {
    return null;
  }
  try {
    return JSON.parse(response.body);
  } catch {
    return null;
  }
}

async function preflightExistingVite({
  endpoint = readAgcDevEndpoint(),
  readServer = readExistingViteServer,
  portListening = isVitePortListening,
  readMarker = readExistingViteMarker,
} = {}) {
  const existing = await readServer(endpoint);
  if (!existing) {
    if (await portListening(endpoint)) {
      throw new Error(
        `${endpoint.url} is already in use by a non-HTTP or unrecognized server. Stop it before starting Tauri dev.`,
      );
    }
    return { status: 'available', apiTarget: '' };
  }

  if (!isAiGameCreatorServer(existing)) {
    throw new Error(
      `${endpoint.url} is already in use by another server. Stop it before starting Tauri dev.`,
    );
  }

  const marker = await readMarker(endpoint);
  const markerApiTarget =
    marker?.schemaVersion === 1 &&
    marker?.app === 'ai-game-creator-shell' &&
    typeof marker?.apiTarget === 'string'
      ? marker.apiTarget
      : '';
  const actualTarget = markerApiTarget || 'unknown';
  throw new Error(
    `${endpoint.url} is already running with API target ${actualTarget}. Its owning worktree cannot be proven, so it will not be reused. Stop that Vite dev server before starting Tauri dev.`,
  );
}

function spawnChild(command, args, options, spawnImpl = spawn) {
  const isPosix = process.platform !== 'win32';
  const useShell = options.shell ?? !isPosix;
  const child = spawnImpl(command, args, {
    ...options,
    shell: useShell,
    // POSIX 下让每个长驻服务拥有独立进程组，退出时可以连同 npm、
    // Node、Cargo 及其子进程一起清理，避免残留订阅任务持续刷日志。
    detached: isPosix,
    stdio: 'inherit',
  });
  const lifecycle = {
    failure: null,
    promise: null,
    // detached 子进程在 POSIX 下以自身 PID 作为 PGID。leader 退出后
    // child.pid 仍是清理其后代的唯一稳定句柄，必须随生命周期保留。
    processGroupId: isPosix && Number.isInteger(child.pid) ? child.pid : null,
  };
  lifecycle.promise = new Promise((resolveLifecycle) => {
    child.once('error', (error) => {
      lifecycle.failure = { type: 'error', error };
      resolveLifecycle(lifecycle.failure);
    });
    child.once('exit', (code, signal) => {
      if (!lifecycle.failure) {
        lifecycle.failure = { type: 'exit', code, signal };
      }
      resolveLifecycle(lifecycle.failure);
    });
  });
  childLifecycles.set(child, lifecycle);
  return child;
}

function readChildFailure(child) {
  return childLifecycles.get(child)?.failure ?? null;
}

function waitForChildTermination(child) {
  const lifecycle = childLifecycles.get(child);
  if (!lifecycle) {
    return Promise.resolve({
      type: 'error',
      error: new Error('子进程未注册生命周期监听'),
    });
  }
  return lifecycle.promise;
}

function formatChildFailure(failure) {
  if (failure?.type === 'error') {
    return failure.error instanceof Error
      ? failure.error.message
      : String(failure.error);
  }
  return failure?.signal
    ? `signal=${failure.signal}`
    : `code=${failure?.code ?? 0}`;
}

function stopChild(child, signal = 'SIGTERM') {
  if (!child) {
    return;
  }

  if (process.platform !== 'win32') {
    const processGroupId = childLifecycles.get(child)?.processGroupId;
    if (Number.isInteger(processGroupId)) {
      try {
        process.kill(-processGroupId, signal);
        return;
      } catch (error) {
        if (error?.code === 'ESRCH') {
          return;
        }
        // leader 尚存活时保留 direct child fallback；leader 已退出则仍以
        // 负 PGID kill 的失败为准，不能误以为 descendants 已清理。
        if (child.exitCode != null || child.signalCode != null) {
          return;
        }
      }
    }
  }

  if (child.exitCode != null || child.signalCode != null) {
    return;
  }
  try {
    child.kill(signal);
  } catch {
    // ignore cleanup races
  }
}

function readLinuxProcessGroupAlive(
  processGroupId,
  { readdirImpl = readdirSync, readFileImpl = readFileSync } = {},
) {
  let processIds;
  try {
    processIds = readdirImpl('/proc');
  } catch {
    return null;
  }

  for (const processId of processIds) {
    if (!/^\d+$/.test(processId)) {
      continue;
    }
    let stat;
    try {
      stat = readFileImpl(`/proc/${processId}/stat`, 'utf8');
    } catch {
      continue;
    }
    const commandEnd = stat.lastIndexOf(') ');
    if (commandEnd < 0) {
      continue;
    }
    const [state, , processGroup] = stat
      .slice(commandEnd + 2)
      .trim()
      .split(/\s+/);
    if (
      Number(processGroup) === processGroupId &&
      state !== 'Z' &&
      state !== 'X'
    ) {
      return true;
    }
  }
  return false;
}

function isProcessGroupAlive(
  processGroupId,
  {
    platform = process.platform,
    killImpl = process.kill,
    readLinuxGroupAlive = readLinuxProcessGroupAlive,
  } = {},
) {
  if (!Number.isInteger(processGroupId)) {
    return false;
  }
  try {
    killImpl(-processGroupId, 0);
  } catch (error) {
    return error?.code !== 'ESRCH';
  }
  if (platform === 'linux') {
    const linuxGroupAlive = readLinuxGroupAlive(processGroupId);
    if (typeof linuxGroupAlive === 'boolean') {
      return linuxGroupAlive;
    }
  }
  return true;
}

async function waitUntil(check, timeoutMs, pollIntervalMs = 25) {
  const deadline = Date.now() + timeoutMs;
  while (Date.now() < deadline) {
    if (await check()) {
      return true;
    }
    await new Promise((resolveWait) => setTimeout(resolveWait, pollIntervalMs));
  }
  return check();
}

function runWindowsTaskkill(
  processId,
  { spawnImpl = spawn, timeoutMs = 5000 } = {},
) {
  return new Promise((resolveRequest) => {
    const taskkill = spawnImpl(
      'taskkill.exe',
      ['/PID', String(processId), '/T', '/F'],
      {
        shell: false,
        stdio: 'ignore',
        windowsHide: true,
      },
    );
    let settled = false;
    const timeout = setTimeout(() => {
      try {
        taskkill.kill('SIGKILL');
      } catch {
        // ignore taskkill timeout races
      }
      finish({ timedOut: true, code: null, error: null });
    }, timeoutMs);
    const finish = (result) => {
      if (settled) {
        return;
      }
      settled = true;
      clearTimeout(timeout);
      resolveRequest(result);
    };
    taskkill.once('error', (error) =>
      finish({ timedOut: false, code: null, error }),
    );
    taskkill.once('exit', (code) =>
      finish({ timedOut: false, code: code ?? 0, error: null }),
    );
  });
}

async function terminateChildTree(
  child,
  {
    platform = process.platform,
    gracefulTimeoutMs = 2500,
    forceTimeoutMs = 2000,
    killImpl = process.kill,
    taskkillImpl = runWindowsTaskkill,
  } = {},
) {
  if (!child) {
    return { stopped: true, forced: false };
  }

  if (platform === 'win32') {
    if (!Number.isInteger(child.pid)) {
      stopChild(child, 'SIGTERM');
      return { stopped: true, forced: false };
    }
    const result = await taskkillImpl(child.pid);
    return {
      stopped:
        !result?.timedOut &&
        !result?.error &&
        [0, 128].includes(result?.code ?? 0),
      forced: true,
      result,
    };
  }

  const processGroupId = childLifecycles.get(child)?.processGroupId;
  if (!Number.isInteger(processGroupId)) {
    stopChild(child, 'SIGTERM');
    const lifecycle = childLifecycles.get(child);
    if (lifecycle) {
      await Promise.race([
        lifecycle.promise,
        new Promise((resolveWait) =>
          setTimeout(resolveWait, gracefulTimeoutMs),
        ),
      ]);
    }
    if (child.exitCode == null && child.signalCode == null) {
      stopChild(child, 'SIGKILL');
      return { stopped: false, forced: true };
    }
    return { stopped: true, forced: false };
  }

  stopChild(child, 'SIGTERM');
  if (
    await waitUntil(
      () => !isProcessGroupAlive(processGroupId, { platform, killImpl }),
      gracefulTimeoutMs,
    )
  ) {
    return { stopped: true, forced: false };
  }

  try {
    killImpl(-processGroupId, 'SIGKILL');
  } catch (error) {
    if (error?.code !== 'ESRCH') {
      return { stopped: false, forced: true, error };
    }
  }
  const stopped = await waitUntil(
    () => !isProcessGroupAlive(processGroupId, { platform, killImpl }),
    forceTimeoutMs,
  );
  return { stopped, forced: true };
}

async function waitForBackendReady(backendChild, timeoutMs = 600_000) {
  const startedAt = Date.now();
  while (Date.now() - startedAt < timeoutMs) {
    if (await isBackendReady()) {
      return readBackendTargets();
    }
    const failure = readChildFailure(backendChild);
    if (failure) {
      throw new Error(`配套后端启动失败: ${formatChildFailure(failure)}`);
    }
    await Promise.race([
      new Promise((resolveWait) => setTimeout(resolveWait, 1000)),
      waitForChildTermination(backendChild),
    ]);
  }
  throw new Error('等待配套后端和数据库启动超时');
}

async function ensureBackend({
  onBackendChild = () => {},
  checkBackendReady = isBackendReady,
  resolveTargets = readBackendTargets,
  spawnBackend = () =>
    spawnChild(
      npm,
      [
        '--prefix',
        '../..',
        'run',
        'agc:backend',
        '--',
        '--database',
        backendDatabase,
        '--spacetime-data-dir',
        backendSpacetimeDataDir,
        '--no-interactive',
      ],
      { cwd: appRoot },
    ),
  waitUntilReady = waitForBackendReady,
} = {}) {
  if (await checkBackendReady()) {
    const targets = resolveTargets();
    console.log(`[ai-game-creator-shell] reuse backend ${targets.apiUrl}`);
    return { backendChild: null, targets };
  }

  console.log('[ai-game-creator-shell] starting backend stack');
  const backendChild = spawnBackend();
  try {
    onBackendChild(backendChild);
    const targets = await waitUntilReady(backendChild);
    console.log(`[ai-game-creator-shell] backend ready ${targets.apiUrl}`);
    return { backendChild, targets };
  } catch (error) {
    stopChild(backendChild);
    throw error;
  }
}

async function startVite(apiTarget, endpoint = readAgcDevEndpoint()) {
  const { apiUrl } = readBackendTargets();
  if (apiUrl !== apiTarget) {
    throw new Error(
      `dev-stack state API target ${apiUrl} does not match paired backend ${apiTarget}.`,
    );
  }

  const existing = await readExistingViteServer(endpoint);
  if (existing) {
    if (isAiGameCreatorServer(existing)) {
      throw new Error(
        `${endpoint.url} is already running and cannot be safely reused. Stop it before starting Tauri dev.`,
      );
    }
    throw new Error(
      `${endpoint.url} is already in use by another server. Stop it before starting Tauri dev.`,
    );
  }

  return spawnChild(
    npm,
    [
      '--prefix',
      '../..',
      'exec',
      'vite',
      '--',
      '--config',
      'vite.config.ts',
      '--port',
      String(endpoint.port),
    ],
    { cwd: appRoot, env: withAgcDevEndpointEnv(endpoint) },
  );
}

async function main() {
  let backendChild = null;
  let viteChild = null;
  let shutdownSignal = '';
  const signalHandlers = new Map();

  for (const signal of ['SIGINT', 'SIGTERM']) {
    const handler = () => {
      shutdownSignal = signal;
      stopChild(viteChild, signal);
      stopChild(backendChild, signal);
    };
    signalHandlers.set(signal, handler);
    process.on(signal, handler);
  }

  try {
    const endpoint = await resolveAgcDevEndpoint({ strictConfigured: true });
    process.env[agcVitePortEnvKey] = String(endpoint.port);
    await preflightExistingVite({ endpoint });
    const backend = await ensureBackend({
      onBackendChild(child) {
        backendChild = child;
        if (shutdownSignal) {
          stopChild(child, shutdownSignal);
        }
      },
    });
    backendChild = backend.backendChild;
    if (shutdownSignal) {
      throw new Error(`启动期收到 ${shutdownSignal}，已停止配套后端`);
    }

    viteChild = await startVite(backend.targets.apiUrl, endpoint);
    if (shutdownSignal) {
      stopChild(viteChild, shutdownSignal);
      throw new Error(`启动期收到 ${shutdownSignal}，已停止前端服务`);
    }

    const children = [backendChild, viteChild].filter(Boolean);
    if (children.length === 0) {
      return 0;
    }

    const failure = await Promise.race(
      children.map((child) => waitForChildTermination(child)),
    );
    stopChild(viteChild);
    stopChild(backendChild);
    return failure.type === 'error' || failure.signal ? 1 : (failure.code ?? 0);
  } catch (error) {
    stopChild(viteChild);
    stopChild(backendChild);
    console.error(
      `[ai-game-creator-shell] ${error instanceof Error ? error.message : String(error)}`,
    );
    return 1;
  } finally {
    await Promise.all([
      terminateChildTree(viteChild),
      terminateChildTree(backendChild),
    ]);
    for (const [signal, handler] of signalHandlers) {
      process.off(signal, handler);
    }
  }
}

function isDirectModuleExecution() {
  return Boolean(
    process.argv[1] &&
      resolve(process.argv[1]) === fileURLToPath(import.meta.url),
  );
}

export {
  ensureBackend,
  formatChildFailure,
  isDirectModuleExecution,
  isProcessGroupAlive,
  preflightExistingVite,
  readChildFailure,
  readLinuxProcessGroupAlive,
  resolveBackendTargetsFromState,
  runWindowsTaskkill,
  spawnChild,
  stopChild,
  terminateChildTree,
  waitForBackendReady,
  waitForChildTermination,
};

if (isDirectModuleExecution()) {
  process.exitCode = await main();
}