test(subagent): exercise native Windows Claude shim

2026-08-10 13:34:10 +08:00
parent aee58c73e0
commit 9f6ec0ead6
5 changed files with 16 additions and 39 deletions
@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-08-04-claude-code-and-codex-subagent-backends.md
-2026-08-04-claude-code-and-codex-subagent-backends.md: 783a8a6537d2b05ab99f1962bda88787e2e4e938
-2026-08-04-claude-code-and-codex-subagent-backends.zh.md: f331dc698001e24df62733d219f28b59d2358fac
+2026-08-04-claude-code-and-codex-subagent-backends.md: eb4c4ab0116cdf1e8b9e6dd53655035a78afdfa0
+2026-08-04-claude-code-and-codex-subagent-backends.zh.md: c14b7977a267c2b7325e4a4382593c8f30dddea4
@@ -79,10 +79,6 @@ The project owner's distribution authorization is scoped to the official `@anthr

 **A model-visible product selector.** Product availability and authentication are deployment facts. Two fixed tools keep each schema and provider binding explicit and avoid adding dynamic selection state to the common service.

-**Global product enable settings and a product-specific Web page.** Those controls make Codex and Claude Code exceptions to the Agent Preset composition that already owns one agent's tool set, and one process-wide choice cannot represent two sessions using different presets. The host always supplies the providers; the preset alone decides which fixed tools its agent receives.
-
-**One shipped preset per product combination.** Four preset variants encode a two-boolean choice in preset identities and multiply every future standard-preset change. Independent ordinary rows express the same result in the user's copied preset without adding a roster taxonomy.
-
 **Product doubles as required evidence.** Doubles cover exhaustive private protocol branches but do not prove package exports, official distributions, authentication, or real process behavior. Required evidence drives each official product against a loopback model fixture.

 **Plugin-managed login, product home, models, settings, or permissions.** Those choices would create another authority beside each product's native configuration and enlarge a one-shot provider into account management. The providers expose only an explicit environment overlay and teardown grace; unattended interaction fails closed.
@@ -79,10 +79,6 @@ Claude Code 证据锁定 Agent SDK 0.3.220 及与其查询协议兼容的原生

 **面向模型的产品选择器。** 产品可用性和身份验证属于部署事实。两个固定工具使各自的 schema 与提供方绑定保持明确，也避免在通用服务中添加动态选择状态。

-**全局产品启用设置与产品专属 Web 页面。** 这类控制会让 Codex 与 Claude Code 成为 Agent Preset 组装的例外，而后者本就拥有单个 agent 的工具集；一个进程级选择也无法表达两个会话使用不同 preset。宿主始终提供 provider，只有 preset 决定其 agent 获得哪些固定工具。
-
-**为每种产品组合交付一份 preset。** 四个 preset 变体把两个布尔选择编码成 preset 身份，并让未来每次标准 preset 修改都要同步多份副本。用户复制的 preset 中两条独立普通行已经能表达同一结果，无需新增名单分类。
-
 **以产品替身作为强制证据。** 替身可以穷尽覆盖私有协议分支，但无法证明包导出、官方发行版、身份验证或真实进程行为。强制证据会驱动每个官方产品连接回环模型 fixture（测试前置数据）。

 **由插件管理登录、产品主目录、模型、设置或权限。** 这些选择会在每个产品的原生配置之外建立另一套权威来源，并将一次性提供方扩张为账户管理功能。提供方只公开显式环境覆盖项和清理宽限期；无人值守交互会以默认拒绝方式失败。
@@ -132,24 +132,6 @@ describe('the shipped Web composition', () => {
    expect(ctx.agentPresets.defaultId).toBe('standard')
  })

-  it('keeps product providers on the host while shipped presets leave their tools disabled', async () => {
-    expect(ctx.subagents.list()).toEqual(expect.arrayContaining([
-      'spawn', 'fork', 'codex', 'claude-code',
-    ]))
-
-    const handle = await ctx.agents.create({
-      sessionId: SessionId('preset-products-disabled'),
-      setup: agentCtx => ctx.agentPresets.mount(agentCtx, 'standard').then(() => undefined),
-    })
-    try {
-      expect(toolNames(ctx, handle.agent)).not.toEqual(expect.arrayContaining([
-        'subagent_codex', 'subagent_claude_code',
-      ]))
-    } finally {
-      await handle.dispose()
-    }
-  })
-
  it('composes the full agent from `standard`', async () => {
    const handle = await ctx.agents.create({
      sessionId: SessionId('preset-standard'),
@@ -1,6 +1,5 @@
 import { execFile } from 'node:child_process'
 import {
-  copyFileSync,
  mkdirSync,
  mkdtempSync,
  readFileSync,
@@ -116,14 +115,17 @@ async function realHarness(behavior: MessagesBehavior): Promise<{
  const workspace = join(root, 'workspace')
  const claudeConfig = join(root, 'claude-config')
  const xdgConfig = join(root, 'xdg')
-  const nativeBin = join(root, 'native-bin')
+  const nativeBin = join(root, 'native bin')
  mkdirSync(workspace)
  mkdirSync(claudeConfig)
  mkdirSync(xdgConfig)
  mkdirSync(nativeBin)
-  const executable = join(nativeBin, process.platform === 'win32' ? 'claude.exe' : 'claude')
-  if (process.platform === 'win32') copyFileSync(claudeBin, executable)
-  else symlinkSync(claudeBin, executable)
+  const executable = join(nativeBin, process.platform === 'win32' ? 'claude.cmd' : 'claude')
+  if (process.platform === 'win32') {
+    writeFileSync(executable, `@echo off\r\n"${claudeBin}" %*\r\n`)
+  } else {
+    symlinkSync(claudeBin, executable)
+  }
  writeFileSync(
    join(claudeConfig, 'settings.json'),
    `${JSON.stringify({ model: settingsModel }, null, 2)}\n`,
@@ -207,7 +209,7 @@ describe('real Claude Agent SDK 0.3.220 and Claude Code 2.1.220', {
    expect(sdkPackage.version).toBe('0.3.220')
    expect(sdkPackage.claudeCodeVersion).toBe('2.1.220')
    expect(sdkPackage.optionalDependencies[platformPackage]).toBe('0.3.220')
-    const version = await execFileAsync(harness.executable, ['--version'], {
+    const version = await execFileAsync(process.platform === 'win32' ? claudeBin : harness.executable, ['--version'], {
      env: { ...process.env, ...harness.env },
    })
    expect(version.stdout.trim()).toBe('2.1.220 (Claude Code)')
@@ -224,11 +226,12 @@ describe('real Claude Agent SDK 0.3.220 and Claude Code 2.1.220', {
        message.type === 'system' && message.subtype === 'init',
    )
    expect(initMessage?.claude_code_version).toBe('2.1.220')
-    const spawnedExecutable = harness.spawnSpecs[0]?.argv[0]
-    expect(spawnedExecutable).toBeDefined()
-    if (spawnedExecutable !== undefined) {
-      expect(process.platform === 'win32' ? spawnedExecutable.toLowerCase() : spawnedExecutable)
-        .toBe(process.platform === 'win32' ? harness.executable.toLowerCase() : harness.executable)
+    if (process.platform === 'win32') {
+      expect(harness.spawnSpecs[0]?.argv.slice(0, 5)).toEqual([
+        'cmd.exe', '/d', '/s', '/c', harness.executable,
+      ])
+    } else {
+      expect(harness.spawnSpecs[0]?.argv[0]).toBe(harness.executable)
    }

    expect(fixture.requests).toHaveLength(1)