feat: 解码 worker 化、Playwright E2E 测试体系与文件夹对比（0.6.0）

2026-08-18 14:40:44 +08:00
parent 2b43b5e676
commit be619c2377
24 changed files with 2030 additions and 151 deletions
@@ -1,9 +1,11 @@
-import { join, dirname } from 'path'
+import { join, dirname, sep } from 'path'
 import { app, shell, BrowserWindow, Menu, ipcMain, dialog, clipboard, nativeTheme, screen } from 'electron'
+import { Worker } from 'worker_threads'
 import { electronApp, optimizer, is } from '@electron-toolkit/utils'
-import iconv from 'iconv-lite'
 import fs from 'fs'
 import { parseState, clampBounds, type AppState } from './windowState'
+import { decodeText, type DecodeResult } from './decode'
+import { scanFolders } from './folderScan'

 /** 应用状态持久化文件（窗口 bounds + 上次文件目录），位于 userData 目录 */
 const stateFile = (): string => join(app.getPath('userData'), 'window-state.json')
@@ -88,48 +90,88 @@ function createWindow(): void {
 /** 单侧文件大小上限（10MB）：超限直接拦截，避免超大文件卡死界面 */
 const MAX_FILE_SIZE = 10 * 1024 * 1024

-/**
- * 缓冲区二进制启发式判定：出现 NUL 字节立即判定；
- * 否则统计前 8KB 内不可见控制字符占比（> 5% 视为二进制）。
- */
-function looksBinary(buf: Buffer): boolean {
-  const len = Math.min(buf.length, 8000)
-  if (len === 0) return false
-  let suspicious = 0
-  for (let i = 0; i < len; i++) {
-    const b = buf[i]
-    if (b === 0) return true
-    if (b < 0x09 || (b > 0x0d && b < 0x20)) suspicious++
-  }
-  return suspicious / len > 0.05
-}
+/** 解码快路径阈值（256KB）：小缓冲同步解码（毫秒级），大缓冲派发 worker 后台解码 */
+const DECODE_FAST_PATH_BYTES = 256 * 1024

 /**
- * 探测文本编码：优先 BOM，其次严格 UTF-8，
- * 失败则回退 GBK 解码，返回 { text, encoding, binary }。
+ * 解码 worker（worker_threads）单例管理。
+ * 大文件解码（GBK 纯 JS 解码可达数百毫秒）移入后台线程，避免阻塞主进程事件循环；
+ * 崩溃时拒绝在途任务并销毁实例（下次请求重建），调用方回退主进程同步解码。
 */
-function decodeText(buf: Buffer): { text: string; encoding: string; binary: boolean } {
-  if (buf.length >= 3 && buf[0] === 0xef && buf[1] === 0xbb && buf[2] === 0xbf) {
-    return { text: buf.subarray(3).toString('utf8'), encoding: 'UTF-8 (BOM)', binary: false }
-  }
-  if (buf.length >= 2 && buf[0] === 0xfe && buf[1] === 0xff) {
-    return { text: iconv.decode(buf.subarray(2), 'utf16-be'), encoding: 'UTF-16 BE', binary: false }
-  }
-  if (buf.length >= 2 && buf[0] === 0xff && buf[1] === 0xfe) {
-    return { text: buf.subarray(2).toString('utf16le'), encoding: 'UTF-16 LE', binary: false }
-  }
-  // 严格 UTF-8 探测
+let decodeWorker: Worker | null = null
+let decodeJobSeq = 0
+const pendingDecode = new Map<
+  number,
+  { resolve: (r: DecodeResult) => void; reject: (e: Error) => void }
+>()
+
+/** 解码 worker 脚本路径：打包后位于 asar.unpacked（Electron 的 asar 补丁不覆盖 worker 线程） */
+function decodeWorkerPath(): string {
+  const p = join(__dirname, 'decodeWorker.js')
+  return p.includes(`app.asar${sep}`) ? p.replace(`app.asar${sep}`, `app.asar.unpacked${sep}`) : p
+}
+
+function getDecodeWorker(): Worker {
+  if (decodeWorker) return decodeWorker
+  const w = new Worker(decodeWorkerPath())
+  w.on('message', (msg: { jobId: number; result: DecodeResult }) => {
+    pendingDecode.get(msg.jobId)?.resolve(msg.result)
+    pendingDecode.delete(msg.jobId)
+  })
+  w.on('error', (err: Error) => {
+    for (const { reject } of pendingDecode.values()) reject(err)
+    pendingDecode.clear()
+    const dead = decodeWorker
+    decodeWorker = null
+    dead?.terminate().catch(() => {})
+  })
+  // 意外退出兜底：拒绝全部在途任务（error 分支已处理时此处为 no-op），下次请求重建
+  w.on('exit', () => {
+    for (const { reject } of pendingDecode.values()) reject(new Error('decode worker exited'))
+    pendingDecode.clear()
+    if (decodeWorker === w) decodeWorker = null
+  })
+  decodeWorker = w
+  return w
+}
+
+/** 大缓冲派发 worker 解码；worker 创建/运行失败时回退主进程同步解码（行为不降级） */
+async function decodeMaybeWorker(buf: Buffer): Promise<DecodeResult> {
+  if (buf.length < DECODE_FAST_PATH_BYTES) return decodeText(buf)
+  const jobId = ++decodeJobSeq
+  return new Promise<DecodeResult>((resolve, reject) => {
+    let w: Worker
+    try {
+      w = getDecodeWorker()
+    } catch (e) {
+      reject(e instanceof Error ? e : new Error('decode worker unavailable'))
+      return
+    }
+    pendingDecode.set(jobId, { resolve, reject })
+    w.postMessage({ jobId, buffer: new Uint8Array(buf.buffer, buf.byteOffset, buf.byteLength) })
+  }).catch(() => decodeText(buf))
+}
+
+/** 按路径读取并解码文本文件（file:open 与 file:read-by-path 共用；10MB 上限） */
+async function readTextFile(
+  filePath: string
+): Promise<
+  | { path: string; name: string; text: string; encoding: string; binary: boolean }
+  | { error: 'too-large' | 'read-failed'; name: string; size?: number }
+> {
+  const name = filePath.split(/[\\/]/).pop() ?? filePath
+  let buf: Buffer
  try {
-    const decoder = new TextDecoder('utf-8', { fatal: true })
-    const text = decoder.decode(buf)
-    return { text, encoding: 'UTF-8', binary: false }
+    const st = await fs.promises.stat(filePath)
+    if (st.size > MAX_FILE_SIZE) {
+      return { error: 'too-large', name, size: st.size }
+    }
+    buf = await fs.promises.readFile(filePath)
  } catch {
-    const text = iconv.decode(buf, 'gbk')
-    // GBK 覆盖面广，二进制内容也能解出“文字”；用字节启发式 + 替换符占比双保险
-    const bad = (text.match(/\uFFFD/g) ?? []).length
-    const binary = looksBinary(buf) || (text.length > 0 && bad / text.length > 0.05)
-    return { text, encoding: 'GBK', binary }
+    return { error: 'read-failed', name }
  }
+  const { text, encoding, binary } = await decodeMaybeWorker(buf)
+  return { path: filePath, name, text, encoding, binary }
 }

 /** IPC: 打开文件对话框并读取内容 */
@@ -151,24 +193,12 @@ ipcMain.handle('file:open', async (_event, side: 'left' | 'right' | null) => {
  // 记录所在目录供下次对话框起始定位（与窗口 bounds 同存一个状态文件）
  appState.lastDir = dirname(filePath)
  saveState()
-  const name = filePath.split(/[\\/]/).pop() ?? filePath
-  let buf: Buffer
-  try {
-    const stat = await fs.promises.stat(filePath)
-    if (stat.size > MAX_FILE_SIZE) {
-      return { error: 'too-large', name, size: stat.size }
-    }
-    buf = await fs.promises.readFile(filePath)
-  } catch {
-    return { error: 'read-failed', name }
-  }
-  const { text, encoding, binary } = decodeText(buf)
-  return { path: filePath, name, text, encoding, binary }
+  return readTextFile(filePath)
 })

-/** IPC: 解码拖拽传入的原始字节（复用编码探测逻辑） */
+/** IPC: 解码拖拽传入的原始字节（复用编码探测逻辑，大缓冲走 worker） */
 ipcMain.handle('file:decode-buffer', async (_event, buffer: ArrayBuffer) => {
-  return decodeText(Buffer.from(buffer))
+  return decodeMaybeWorker(Buffer.from(buffer))
 })

 /** IPC: 保存差异报告（系统保存对话框 + 写入内容） */
@@ -200,6 +230,35 @@ ipcMain.handle('clipboard:write', (_event, text: string) => {
  return true
 })

+/** IPC: 选择文件夹（文件夹对比入口；记忆上次目录） */
+ipcMain.handle('folder:pick', async () => {
+  const result = await dialog.showOpenDialog({
+    title: '选择要对比的文件夹',
+    defaultPath: appState.lastDir,
+    properties: ['openDirectory']
+  })
+  if (result.canceled || result.filePaths.length === 0) return null
+  const dirPath = result.filePaths[0]
+  appState.lastDir = dirPath
+  saveState()
+  return dirPath
+})
+
+/** IPC: 扫描对比两个文件夹（递归对齐 + 字节级判定，异步 fs 不阻塞主进程） */
+ipcMain.handle('folder:scan', async (_event, leftDir: string, rightDir: string) => {
+  try {
+    return await scanFolders(leftDir, rightDir)
+  } catch {
+    // 目录不存在/不可读等：转为错误标记，渲染端提示
+    return { error: 'scan-failed' as const }
+  }
+})
+
+/** IPC: 按绝对路径读取文本文件（文件夹对比双击进入单文件对比；与 file:open 同一读取管线） */
+ipcMain.handle('file:read-by-path', async (_event, filePath: string) => {
+  return readTextFile(filePath)
+})
+
 app.whenReady().then(() => {
  electronApp.setAppUserModelId('com.metonateam.difflens')