refactor(cli): remove unused EnumSelector component

[qqqys]

What this PR does

Removes a UI selector component that is no longer referenced anywhere in the codebase, together with the test and snapshot that existed solely to cover it. The component renders a left-right value picker for enum-typed settings, but no screen or component mounts it; its only callers were its own unit test and snapshot. Nothing else changes.

Why it's needed

Dead surface is a cost: every contributor who reads it has to work out what wires it up, and the answer here is "nothing". The component arrived with an upstream sync and was never referenced in this repository's history — no production code, no string-keyed registry entry, no dynamic import path. Its dedicated test and snapshot exist only to cover the component itself, so they go with it.

Reviewer Test Plan

How to verify

This is a pure deletion of code nothing references. Verify the absence of consumers rather than behavior:

1. rg -n '\bEnumSelector\b' packages integrations integration-tests scripts .github docs-site — expect zero hits after this PR (the only three before it were the component, its test, and its snapshot).
2. npm run build && npm run bundle && npm run typecheck — all green locally with this change (a type-only break from a removed export would surface here; typecheck does not run in CI).
3. npm run lint:ci — green locally with this change.

No user-visible behavior changes, so no before/after TUI evidence applies.

Evidence (Before & After)

N/A — non-UI change (dead-code deletion; nothing rendered this component).

Tested on

OS	Status
🍏 macOS	⚠️
🪟 Windows	⚠️
🐧 Linux	✅

Environment (optional)

Local npm run build && npm run bundle && npm run typecheck and npm run lint:ci in a clean worktree off origin/main.

Risk & Scope

• Main risk or tradeoff: essentially none — the component has zero production consumers, verified by full-corpus search including string keys, dynamic-import shapes, build graphs, and the test corpus; if a consumer were ever found, reverting the single commit restores everything.
• Not validated / out of scope: macOS/Windows builds (no platform-specific code involved); no integration test run (no CLI behavior is touched by removing unmounted code).
• Breaking changes / migration notes: none; nothing exported from any package surface was consumed externally.

Linked Issues

Part of the dead-surface cleanup tracked in #10000.

Machine details (find-simplifications)

• id / class: enum-selector — class 2, dead component with its test and snapshot.
• Surface removed: packages/cli/src/ui/components/shared/EnumSelector.tsx (87 lines), EnumSelector.test.tsx (152 lines), __snapshots__/EnumSelector.test.tsx.snap (9 lines) — 248 lines deleted, 0 added (git diff --numstat).
• Every consumer found and its kind: full-corpus grep (all extensions, including tests, snapshots, docs, .github, .husky, .vscode, patches, root build/lint manifests, tracked .qwen files) resolves to exactly three paths — the component itself (declaration), its test (test-only), and its snapshot (snapshot). Zero production consumers; zero string-keyed references (checked string-keys, build-graph, generated, vi-mock, mirrors, dyn-import, cli-flags, assets checklist rows); no integration-test reference (second pass without test exclusions); no design-doc or plan mentions it.
• Proof steps run (survey protocol §3): ledger check (no tombstone), recency (introduced via the 2025-10-23 Gemini-CLI v0.8.2 sync per GitHub full-history path dating — local clone is shallow; no commit in the repo's history ever referenced the symbol), published-surface escape (territory is packages/cli/src — landable, not packages/core/channels/SDK/acp-bridge), full-corpus grep, own-file check (no in-file caller beyond the declaration), hidden-consumer checklist, test-only check, unwire history (no binding text ever existed — it was never wired), design-doc ownership (none).
• Re-verified at land time against fresh origin/main (4a281f2efc): same three paths, no commits touching the symbol since the survey, branch base freshly fetched.
• Verification: npm run build && npm run bundle && npm run typecheck green; npm run lint:ci green; targeted vitest skipped by design (the deletion carries the surface's only test; corpus re-grep on the landing checkout returned zero surviving references). Self-audit: two clean passes.
• No collateral: no eslint.legacy-filenames.mjs entry (.tsx is outside the kebab-case rule), no i18n locale keys (component took display strings via props), no docs rows.

中文说明

本 PR 做了什么

删除一个在代码库中已无任何引用的 UI 选择器组件，以及仅为覆盖它而存在的测试和快照。该组件渲染一个用于枚举类设置项的左右切换选择器，但没有任何界面或组件挂载它；它唯一的调用方就是它自己的单元测试和快照。除此之外没有任何其他改动。

为什么需要

死代码是一种成本：每个读到它的人都得弄清是什么在用它，而这里的答案是“没有任何东西”。该组件随一次上游同步进入本仓库，在整个历史中从未被引用——没有生产代码、没有字符串键注册项、没有动态导入路径。它专属的测试和快照只为覆盖组件本身而存在，因此一并删除。

评审者测试计划

如何验证

这是一次对无任何引用代码的纯删除。请验证消费者的缺失，而不是行为：

1. rg -n '\bEnumSelector\b' packages integrations integration-tests scripts .github docs-site —— 本 PR 之后应无任何命中（之前仅有的三处即组件本身、其测试与其快照）。
2. npm run build && npm run bundle && npm run typecheck —— 本地在包含本改动的情况下全部通过（删除导出可能引发的纯类型问题会在这里暴露；typecheck 不在 CI 中运行）。
3. npm run lint:ci —— 本地在包含本改动的情况下通过。

由于没有用户可见的行为变化，无需前后对比的 TUI 证据。

证据（改动前后）

N/A —— 非 UI 改动（死代码删除；该组件从未被渲染）。

测试环境

OS	状态
🍏 macOS	⚠️ 未测试
🪟 Windows	⚠️ 未测试
🐧 Linux	✅ 已测试

环境（可选）

在基于 origin/main 的干净 worktree 中本地运行 npm run build && npm run bundle && npm run typecheck 与 npm run lint:ci。

风险与范围

• 主要风险或权衡：基本没有——该组件零生产消费者，已通过全语料库搜索（含字符串键、动态导入形态、构建图与测试语料）验证；万一未来发现消费者，回滚这一个提交即可完整恢复。
• 未验证 / 超出范围：macOS/Windows 构建（不涉及平台相关代码）；未运行集成测试（删除未挂载的代码不触及任何 CLI 行为）。
• 破坏性变更 / 迁移说明：无；没有任何从包对外表面导出的内容被外部消费。

关联 Issue

属于 #10000 跟踪的死代码清理工作。

机器细节见上方英文 “Machine details (find-simplifications)” 折叠块：候选编号 enum-selector（第 2 类：死组件连同其测试与快照），共删除 248 行、新增 0 行；全语料库搜索证明消费者恰为组件自身、其测试与其快照三处；调查协议各步骤及落地前的重新核验均已执行；构建、打包、类型检查与 lint:ci 全部通过；因删除本身带走了该表面唯一的测试，按流程跳过定向 vitest，改以落地检出上的语料库复扫（零残留引用）兜底；无允许列表条目、无 i18n 键、无文档行等连带清理项。

[qwen-code-ci-bot]

✅ Qwen Triage finished — view run. See the stage comments in this thread for the result.

✅ Qwen Triage 已完成 —— 查看运行。结果见本线程中的各阶段评论。

[qwen-code-ci-bot]

Thanks for the PR!

Template looks good ✓

Problem: Verified, not theoretical. I grepped main independently: EnumSelector appears in exactly three code files — the component, its test, and its snapshot. No production consumer, no string-key variants, no barrel re-export. (The only other mentions are in the find-simplifications skill docs, which use this very deletion as a worked example.) The dead surface is real.

Direction: Aligned — this is subtraction, not addition; removing a component nothing mounts is a pure maintenance win. No CHANGELOG direction signal needed for deleting unreferenced code.

Size: Not applicable — no core paths touched (all three files are under packages/cli/src/ui/components/shared/). 0 additions, 248 deletions.

Approach: The scope is exactly right. The PR deletes precisely the three files needed and nothing else, and correctly leaves SettingEnumOption in settingsSchema.ts alone — it is still used by the schema itself and by serve/routes/workspace-settings.ts. No drive-by changes, no unrelated churn.

Risk: No elevated risk signals — no high-risk paths matched. One note for transparency: the author @qqqys has write access and is a /packages/core/ code owner, but this is still a fork + refactor-title PR, so the approval guardrail applies regardless — I will not auto-approve; a maintainer signs off in Stage 3.

Moving on to code review. 🔍

中文说明

感谢贡献！

模板完整 ✓

问题：已核实，非理论问题。我在 main 上独立检索：EnumSelector 恰好出现在三个代码文件中——组件本身、其测试与其快照。无生产消费者、无字符串键变体、无 barrel 再导出。（其余提及仅出现在 find-simplifications 技能文档中，该文档正以本次删除作为示例。）死代码表面真实存在。

方向：对齐——这是减法而非加法；删除无人挂载的组件是纯维护收益。删除无引用代码无需 CHANGELOG 方向信号。

规模：不适用——未触及核心路径（三个文件均在 packages/cli/src/ui/components/shared/ 下）。新增 0 行，删除 248 行。

方案：范围恰好。PR 精确删除所需的三个文件，别无其他，并且正确地保留了 settingsSchema.ts 中的 SettingEnumOption——它仍被 schema 本身与 serve/routes/workspace-settings.ts 使用。无顺手改动，无无关变更。

风险：无升级风险信号——未命中高风险路径。一点透明说明：作者 @qqqys 拥有 write 权限且是 /packages/core/ 的 code owner，但这仍是 fork + refactor 标题的 PR，审批护栏依然适用——不会自动批准；由维护者在 Stage 3 签核。

进入代码审查 🔍

— Qwen Code · qwen3.8-max

_{Reviewed at db1b9cc6764c6e656110d893a73162a819aa8aee · re-run with @qwen-code /triage}

[qwen-code-ci-bot]

Code review

My independent approach to "remove an unused component" would be: prove zero consumers (exact name, case/kebab/snake variants, barrel re-exports, dynamic-import shapes), then delete the component together with its colocated test and snapshot, and let CI confirm nothing breaks. This PR does exactly that.

Going through the diff against main:

• Deletion-only, as advertised — 0 additions, 248 deletions; the diff carries nothing but the three file removals.
• Zero-consumer claim holds — I re-ran the search independently: EnumSelector appears only in the three deleted files. No string-key variants, no barrel export, no dynamic import, no reference in integration-tests/, docs/, .github/, or scripts/. The only surviving mentions are in the find-simplifications skill docs, which use this deletion as a worked example — docs, not consumers.
• No collateral left behind — SettingEnumOption (imported by the deleted files) correctly stays: it is still used by settingsSchema.ts and serve/routes/workspace-settings.ts. No eslint.legacy-filenames.mjs entry exists for it (.tsx is outside the kebab-case rule), and the component had no i18n keys — it received display strings via props. Verified both.

No critical blockers, no convention violations. Nothing to fix.

Test evidence

Unattended CI run — I never build or execute PR code; the evidence below is the PR's own CI, fetched via the API for commit db1b9cc6764c6e656110d893a73162a819aa8aee. The main unit suite (Test (ubuntu-latest, Node 22.x)) is still running; macOS/Windows test and integration test jobs show skipped, consistent with gating behind the Linux leg. The table below is updated in place by the finalize job once CI settles.

Not verified / not applicable: no user-visible behavior changes, so there is no TUI surface to drive and no behavioural claim for a sandboxed lane to settle. The author reports local build && bundle && typecheck and lint:ci green — that is the author's claim, not independently re-run here; for a pure deletion the stronger evidence is the zero-consumer verification above, since a removed export can only break something that imports it, and nothing does.

Final CI results for db1b9cc (auto-updated by the triage finalize job after CI completed):

Check	Conclusion
Test (ubuntu-latest, Node 22.x)	❌ failure
Classify PR	✅ success
Dependency CVE audit	✅ success
Desktop Shell (ubuntu-22.04)	✅ success
Desktop Shell (windows-2022)	✅ success
route	✅ success
Secret scan (TruffleHog)	✅ success
web-shell E2E Smoke (ubuntu-latest, Node 22.x)	✅ success

_{One row per check name (latest run); skipped checks omitted; failures sort first. / 每个检查名一行（取最新一次运行），省略 skipped，失败项排在最前。}

中文说明

代码审查

我对"删除未使用组件"的独立方案是：先证明零消费者（精确名、大小写/kebab/snake 变体、barrel 再导出、动态导入形态），再连同其同目录测试与快照一并删除，由 CI 兜底确认。本 PR 的做法与此完全一致。

对照 main 逐条核对：

• 确为纯删除——新增 0 行、删除 248 行，diff 中除三个文件的移除外别无他物。
• 零消费者成立——我独立复跑了检索：EnumSelector 仅出现在被删的三个文件中。无字符串键变体、无 barrel 导出、无动态导入，integration-tests/、docs/、.github/、scripts/ 中均无引用。唯一残留的提及在 find-simplifications 技能文档里，且是以本次删除作为示例——是文档，不是消费者。
• 无遗留连带项——被删文件导入的 SettingEnumOption 被正确保留：它仍被 settingsSchema.ts 与 serve/routes/workspace-settings.ts 使用。eslint.legacy-filenames.mjs 中无对应条目（.tsx 不在 kebab-case 规则范围内），组件也不含 i18n 键——显示文案均经 props 传入。两者均已核实。

无阻塞性问题，无规范违背。无需修改。

测试证据

无人值守 CI 运行——不会构建或执行 PR 代码；以上证据来自 PR 自身的 CI，经 API 拉取（提交 db1b9cc6764c6e656110d893a73162a819aa8aee）。主单测套件（Test (ubuntu-latest, Node 22.x)）仍在运行；macOS/Windows 测试与集成测试作业显示 skipped，与"以 Linux 通道为前置"的门控一致。上表将由 finalize 作业在 CI 收敛后就地更新。

未验证 / 不适用：无用户可见行为变化，故无 TUI 表面可驱动，也无行为性主张需要沙箱通道裁决。作者自述本地 build && bundle && typecheck 与 lint:ci 通过——此为作者声明，未在此独立复跑；对纯删除而言，更强的证据是上文的零消费者核验——被移除的导出只会破坏导入它的事物，而没有任何事物导入它。

— Qwen Code · qwen3.8-max

_{Reviewed at db1b9cc6764c6e656110d893a73162a819aa8aee · re-run with @qwen-code /triage}

[qwen-code-ci-bot]

Confidence: 3/5 — clean review, but the fork-refactor guardrail needs a maintainer's sign-off.

Stepping back: this is about as clean as a deletion PR gets. The problem is verified (I independently confirmed zero consumers on main), the diff is exactly the minimal set — component, test, snapshot, nothing else — and it correctly leaves SettingEnumOption in place since other code still uses it. No blockers, no drive-by changes, no hygiene complaints. In six months nobody will curse this PR; they'll just not notice the dead weight is gone, which is the point.

Why not approve then? Pure policy, not doubt. The approval guardrail is mechanical: a cross-repository PR with a refactor title is never auto-approved, because that exact combination once got auto-approved and merged wrongly. This PR is isCrossRepository: true and its title starts with refactor, so the guardrail fires regardless of how clean the review is — and no approve-on-green marker is emitted either (CI on db1b9cc6 is also still running: the Linux unit leg in flight).

Context that should make the sign-off quick: the author @qqqys is a maintainer with write access and a /packages/core/ code owner, submitting from a fork — so this is a trusted author caught by a deliberately blunt rule, not an unknown contributor. A human maintainer can sign off on sight once CI lands green.

⏸️ Deferring to @wenshao @pomelo-nwu — the fork-refactor approval guardrail blocks auto-approval; review itself is clean (see Stages 1–2), needs a human maintainer's sign-off.

中文说明

置信度：3/5 —— 审查干净，但 fork-refactor 护栏需要维护者签核。

退一步看：这是一个删除类 PR 能做到的最干净的程度。问题已核实（我在 main 上独立确认零消费者），diff 恰好是最小集合——组件、测试、快照，别无其他——并且正确地保留了 SettingEnumOption，因为其他代码仍在使用它。无阻塞项、无顺手改动、无规范问题。六个月后不会有人骂这个 PR；大家只会注意到死代码不见了——这正是目的所在。

那为何不批准？纯粹是政策，而非疑虑。审批护栏是机械的：跨仓库（fork）且标题为 refactor 类型的 PR 永不自动批准，因为这一组合曾经被错误地自动批准并合并。本 PR isCrossRepository: true 且标题以 refactor 开头，护栏必然触发，与审查多干净无关——同样不会发出 approve-on-green 标记（db1b9cc6 上的 CI 也仍在运行：Linux 单测通道进行中）。

让签核可以快速完成的背景：作者 @qqqys 是拥有 write 权限的维护者、/packages/core/ 的 code owner，从 fork 提交——这是被一条刻意"一刀切"的规则拦下的可信作者，而非陌生贡献者。CI 变绿后，人类维护者看一眼即可签核。

⏸️ 转交 @wenshao @pomelo-nwu —— fork-refactor 审批护栏阻止自动批准；审查本身干净（见 Stage 1–2），需要人类维护者签核。

— Qwen Code · qwen3.8-max

_{Reviewed at db1b9cc6764c6e656110d893a73162a819aa8aee · re-run with @qwen-code /triage}

[qwen-code-ci-bot]

⚠️ Downgraded from Approve to Comment: CI failing: Test (ubuntu-latest, Node 22.x). Reviewed.

中文说明

⚠️ 已从批准降级为评论：CI failing: Test (ubuntu-latest, Node 22.x)。 已审查。

— qwen3.8-max via Qwen Code /review (v0.22.0)

[qqqys]

@qwen-code /takeover

[qwen-code-dev-bot]

🤝 Takeover engaged: the autofix loop now manages this PR — it will address new review feedback and resolve base conflicts until the label is removed or the round cap is reached. This is a fork PR, so the first round comes from the next scheduled scan (usually within minutes). Remove the autofix/takeover label (or comment @qwen-code /takeover stop) to release.

中文说明

🤝 已接管：autofix 循环现在管理此 PR —— 将持续处理新的评审反馈与 base 冲突，直到移除标签或达到轮次上限。本 PR 来自 fork，首轮处理将由下一次定时扫描执行（通常几分钟内）。移除 autofix/takeover 标签（或评论 @qwen-code /takeover stop）即可释放。

[qwen-code-dev-bot]

🔀 Base updated: red check(s) [Test (ubuntu-latest, Node 22.x)] pass on current main — merged current main via update-branch; CI will re-run.

中文说明

🔀 已更新 base：红色检查 [Test (ubuntu-latest, Node 22.x)] 在当前 main 上通过 —— 已通过 update-branch 合入当前 main，CI 将重新运行。

[qwen-code-ci-bot]

No issues found. LGTM! ✅

中文说明

未发现问题。LGTM！✅

— qwen3.8-max via Qwen Code /review (v0.22.0)

[qwen-code-dev-bot]

✅ AutoFix round 1 finished — view run. See this round's report below.

中文说明

✅ AutoFix 第 1 轮已完成 —— 查看运行。本轮报告见下方。

[qwen-code-dev-bot]

🤖 Reviewed the latest feedback — no changes needed. Why, point by point: · 已审阅最新反馈——无需改动。逐点说明原因如下：

Autofix review round — no action needed

Feedback triaged: 1 item — review [rv:5017090095]: "⚠️ Downgraded from Approve to Comment: CI failing: Test (ubuntu-latest, Node 22.x)".

Decision: no code change. The CI-failure claim was true when the review ran but is now refuted by current evidence; there is nothing to fix in this PR.

Evidence

• The downgrade was recorded while reviewing commit db1b9cc6 (the single substantive commit of this PR), at which point the Test (ubuntu-latest, Node 22.x) check was red.
• The current PR head (88c182dd82, the merge of main) has a green CI: the latest Qwen Code CI run (32833676259) shows Test (ubuntu-latest, Node 22.x) = SUCCESS (completed 2026-08-25T10:11:01Z). This round's feedback lists no entries under "Failed checks" or "Still-red checks". The earlier red state was a transient failure cleared by the re-run; no code defect in this PR is implicated.
• Cross-checked locally on this branch to rule out a real regression:
  • npm run build — passed
  • npm run typecheck — passed
  • npm run lint — passed
  • packages/cli full Vitest suite — 23996 passed, 33 failed. Every one of the 33 failures is in a file unrelated to this PR (which only deletes the unreferenced EnumSelector component, its test, and its snapshot; a repo-wide search finds zero remaining references to EnumSelector). All 33 are artifacts of the verification sandbox container this round ran in: the SANDBOX/QWEN_HOME environment diverts settings/config/docs code paths (e.g. the Footer renders a docker | sandbox pill absent from the golden snapshots), and the filesystem sandbox rejects mkdtemp outside the workspace. Re-running representative files (Footer.test.tsx, docsCommand.test.ts, cdCommand.test.ts) with the sandbox environment variables removed turns them green except the one test that needs to write outside the workspace, which only passes on a real CI runner. CI on the same OS/Node is green, consistent with this diagnosis.

Outcome

No commit was made and no files were changed. No further action is required from the autofix loop on this finding; the next CI evaluation of the PR head should report the check as passing.

中文说明

Autofix 审查轮次 — 无需任何操作

处理的反馈： 1 条 — 审查 [rv:5017090095]："⚠️ 已从批准降级为评论：CI failing: Test (ubuntu-latest, Node 22.x)"。

决定：不修改代码。 该 CI 失败的指控在审查发生时属实，但已被当前证据推翻；本 PR 没有任何需要修复的内容。

证据

• 降级记录于审查提交 db1b9cc6（本 PR 唯一的实质性提交）时，当时 Test (ubuntu-latest, Node 22.x) 检查为红色（失败）。
• 当前 PR 头部（88c182dd82，即合并 main 后的提交）CI 已为绿色：最新的 Qwen Code CI 运行（32833676259）显示 Test (ubuntu-latest, Node 22.x) = SUCCESS（完成于 2026-08-25T10:11:01Z）。本轮反馈的 "Failed checks" 与 "Still-red checks" 部分均为空。此前的红色状态是瞬态失败，已随重新运行消除；与本 PR 的代码缺陷无关。
• 在本分支上进行了本地交叉验证，以排除真实回归：
  • npm run build — 通过
  • npm run typecheck — 通过
  • npm run lint — 通过
  • packages/cli 完整 Vitest 套件 — 23996 通过，33 失败。这 33 个失败全部位于与本 PR 无关的文件中（本 PR 仅删除了无任何引用的 EnumSelector 组件、其测试及快照；全仓库搜索确认已无任何 EnumSelector 残留引用）。这 33 个失败全部是本轮所处的验证沙箱容器的环境假象：SANDBOX/QWEN_HOME 环境变量改变了 settings/config/docs 的代码路径（例如 Footer 渲染出金色快照中不存在的 docker | 沙箱标识），且文件系统沙箱拒绝在工作区之外执行 mkdtemp。移除沙箱环境变量后重新运行代表性文件（Footer.test.tsx、docsCommand.test.ts、cdCommand.test.ts）全部转绿，仅余一个需要在工作区外写入的测试失败，该测试只在真实 CI runner 上通过。相同操作系统/Node 版本下的 CI 为绿色，与此诊断一致。

结果

未创建任何提交，未修改任何文件。autofix 循环无需对该指控采取进一步操作；下一次对 PR 头部的 CI 评估应显示该检查已通过。

Base-conflict check · 基分支冲突检查: no conflict with main. · 与 main 无冲突。

---

🧠 Handled by Qwen Code · model/模型 qwen3.8-max

[qwen-code-ci-bot]

Released in v0.22.2.