[JasonBuildAI]

Restores the Release Artifacts packaged-browser gate: four recent Release Artifacts runs fail the research-route-reading scenario and skip upload-release and publish-pypi, so maintainers cannot produce release artifacts from current main. This is a follow-up to #8003.

Goal And Delivered Outcome

• Outcome basis / optional anchor: Follow-up to merged fix(workspace): ground research reading in real benchmark state #8003; its second sequential research-analysis read exposed a single-resolver loading gate in the packaged-browser fixture.
• Goal/source and gap: examples/personal-workspace-browser/research-route-reading.mjs creates a pending promise for each loading-state research-analysis GET but releases only the last resolver. With fix(workspace): ground research reading in real benchmark state #8003's second original-pack read, one GET stays pending, the analysis button remains disabled, and the scenario times out.
• Observable before → after, with the validation row that proves it: Before, the scenario timed out after 30s waiting for .research-analysis-entry-actions > button:not(:disabled); after, the packaged research-route-reading scenario completes in 132.8s and reports personal-workspace-browser-smoke (packaged): ok.
• Issue/task and intended base: Fixes [Bug]: Release Artifacts browser smoke hangs after #8003 adds a second research-analysis read #8032; intended base is upstream loopx-project/loopx:main.

Author Declaration

• Written by: model_agent — OpenAI Codex, under user direction; no human review has occurred.

Implemented against

• Specification and revision: No separate written specification; the linked issue and the reproduced Release Artifacts failure are the basis.
• Criteria:

Criterion (spec clause)	Disposition	Symbol / path	Test or command
Loading state keeps analysis reads pending until explicitly released	implemented	examples/personal-workspace-browser/research-route-reading.mjs:77-101	packaged research-route-reading smoke
Both sequential research-analysis reads resume after release	implemented	examples/personal-workspace-browser/research-route-reading.mjs:80,101,107	packaged research-route-reading smoke
Existing language, width, route-state, and layout assertions still run	implemented	examples/personal-workspace-browser/research-route-reading.mjs:96-108	packaged research-route-reading smoke
Release Artifacts build job no longer hangs on this scenario	implemented	same test path	Local packaged smoke; remote Release Artifacts run remains post-merge verification

• Self-check before submission: Read the issue/PR templates, docs/development/good-smokes.md, the predecessor fix(workspace): ground research reading in real benchmark state #8003 diff, and the production readAnalysisSnapshot path. Verified the failing selector and before/after behavior. Deliberately left production code, fixture data, and unrelated browser scenarios unchanged.

Scope And Continuation

• Completed scope and remaining work: The packaged-browser loading gate now creates one releasable gate per loading case, so every request admitted while loading resumes together. No production behavior or fixture data changes.
• Slice boundary / successor: Complete within this scope; no successor is needed. Remote release qualification is expected after merge, not a separate code task.

Validation

• Tested revision: 958c17f0654fc2e916360dbaf7c75290cd11c442
• Run state: finished
• Input classes: synthetic

Check kind	Result	Public-safe evidence / limitation
static	passed	git diff --check; node --check examples/personal-workspace-browser/research-route-reading.mjs
unit	passed	PYTHONUTF8=1 uv run --extra test python -m pytest -q tests/presentation/test_research_history_fixture.py tests/presentation/test_explore_history.py → 13 passed. An initial run under the Windows GBK default failed two fixture decodes before this UTF-8 rerun; those failures are environment-codec related, not this diff.
integration	passed	PYTHONUTF8=1 uv run --extra test loopx canary premerge --from-git-diff --git-diff-base upstream/main → status: passed, merge_gate_passed: true. One advisory inherited failure remained: control-plane-maintainability-ratchet-smoke.py did not mention changed files.
real_entrypoint	passed	Packaged dashboard bundle built with npm run build:chat; research-route-reading reported scenario-end=research-route-reading duration_ms=132841 and personal-workspace-browser-smoke (packaged): ok.
regression_parity	passed	Pre-fix packaged reproduction timed out after 39.3s on the same scenario; public CI shows the same 30s locator timeout in four Release Artifacts runs. The patched head completes the full matrix.

• Coverage and gaps: The local packaged smoke exercises the full research route-reading matrix (two languages, three widths, seven route states) against the built frontend and synthetic fixture. No remote Release Artifacts run exists for the branch head yet; the post-merge run is the remaining qualification. Non-packaged CI coverage is unchanged.

Frontend / Visual Evidence

• UI impact: none — the diff changes only a browser-smoke route gate; no product rendering or interaction changes.
• Before: N/A
• After: N/A
• First viewport and preview: N/A
• States and viewports shown: N/A
• Source data: synthetic (the existing route-reading browser fixture)
• Attention review: N/A; no user-visible surface changed.
• Craft review: N/A; no user-visible surface changed.

Type of Change

• Bug fix
• New feature
• Breaking change
• Refactoring (no functional changes)
• Documentation update
• Test update

LoopX Area

• Control plane (goals, todos, quota, scheduler, registry, runtime)
• Benchmark boundary (adapters, runners, verifiers, scoring, evidence)
• Capability or extension (providers, adapters, skills)
• Public docs or presentation surface (README, protocols, dashboard)
• Build, packaging, installer, or CI
• Host or runtime integration

Technical Direction

• Direction / acceptance reference, when applicable: N/A

Shared-authority RFC fixture impact

• Production-scale fixture schema: N/A
• Semantic dimensions changed, or reviewed no-impact rationale: N/A; synthetic browser test fixture only.
• Provider conformance arms run: N/A
• Read-only legacy/file/PostgreSQL three-arm rehearsal (required for promotion, runtime-routing, or compatibility-projection changes): N/A

Boundary Checklist

• Neither the diff nor this PR body/comments/attachments disclose private state, credentials, raw traces or verifier output, internal links, or local machine paths (including .loopx/, .codex/goals/, and live ACTIVE_GOAL_STATE.md).
• I did not duplicate maintainer-owned benchmark work unless a maintainer split out a public issue for it.
• I kept the change scoped to the linked issue/task.
• Existing English/Chinese RFC pairs (including ledger entries) both have a diff, and the local premerge check passes, or no paired RFC changed.
• I completed the visual evidence section for UI changes, or marked UI impact none.
• Every commit includes a DCO Signed-off-by trailer (git commit -s).

AI disclosure: this change was drafted with AI assistance at the user's direction; no human review has occurred.

[mergify]

This pull request has merge conflicts with main and cannot be merged
until they are resolved. Please rebase or merge the base branch, @JasonBuildAI.

Choose the remote for the base repository, not an out-of-date fork.
For a fork clone, first inspect git remote -v; upstream must point
to https://github.com/loopx-project/loopx.git. If it is absent, add it
with git remote add upstream https://github.com/loopx-project/loopx.git.
Then run:

git fetch upstream
git rebase upstream/main
# Resolve each conflict, git add the resolved files, then git rebase --continue.
git push --force-with-lease origin HEAD

For a same-repository clone whose origin points to
https://github.com/loopx-project/loopx.git, use origin instead of
upstream for fetch/rebase. If you prefer merging the base, use
git merge <base-remote>/main and push normally.

Keep the DCO Signed-off-by trailer on every commit when you rebase.
https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[loopx-agent]

Reviewer: model_agent; gpt-6.1-sol; OpenAI; runtime_reported; reasoning_effort=max.

动机

未发现阻塞性问题。评审对象为 958c17f0654fc2e916360dbaf7c75290cd11c442，比较基线为 ed2e4591309014861472cf72fbf55fcbb0b28b20。

• 受影响调用者：维护 Personal Workspace Research 路由浏览器回归的贡献者，以及依赖该 loading 断言识别真实回归的维护者。
• 前后场景：带 language 的读取完成后，页面还会读取 original pack；基线只释放第一个 Promise，第二次读取继续挂起，按钮一直禁用。head 让同一 loading case 的两次读取共享一个 Promise，一次 release 后按钮可用；下一 case 重新阻塞。
• 可观察结果：修复既有 loading fixture 的单次释放语义，保留释放前禁用、释放后可用、只读拒绝和逐 case 重置断言。
• 非目标：不改变产品请求、布局或权限；不声称完整 42-case 原生矩阵、发布产物或远端 CI 已验证，也不处理分支冲突。

验收依据为 #8032 的 Expected behavior，不可变代码基线为 ed2e4591309014861472cf72fbf55fcbb0b28b20，并对照基线 good-smokes 约束与实际产品两次读取流程。此 PR 交付 loading fixture 的完整小修复；不把它等同于所有 release 检查已通过。

改动思路

同一个 loading case 应有一个可释放边界。把 Promise 放在 case 级，两个请求共享等待；release 后到达的第二个读取也会立即通过，下一 loading case 再建立新边界。默认 Promise.resolve() 保持其他 state 不阻塞，现有非 GET 拒绝规则不变。

这只需要一文件 +3/-2。重复调用 resolver 仍依赖请求到达顺序；resolver 数组和 released flag 更复杂。生产层保留现有语言读取与 original-pack fallback，不需要为测试夹具改产品请求或权限。

具体改动

关键代码讲解

1. loadingGate 与请求 callback：loadingGate 默认已完成，只有 loading state 的 GET 等待它，避免每个请求重建 Promise 并覆盖唯一的 releaseLoading。
2. 逐 case 重置：每个 loading case 都创建新的 Promise/resolver。因此第一次释放不会让后面的 loading 检查失去阻塞。
3. 真实产品消费者：语言版本缺失时还会取 original pack。这段生产代码未改，解释了原来的单请求等待为什么不成立。

独立验证：

• 执行实际 base/head 源码中的 callback 与 case reset：基线顺序第二次读取挂起，并发读取丢失 resolver；head 顺序、并发、迟到读取都只需一次 release，下一 case 重新阻塞，POST 仍被拒绝。
• 故意移除 head 的 case reset：首个 loading/read 立即通过，独立的释放前禁用/重置 oracle 失败；证明没有只看“最终完成”的空洞绿灯。
• 真实 packaged 页面 + 原生只读 history backend + 合成 analysis gate：基线 release 后第二个 GET pending、按钮仍禁用；head 在 zh-CN/en × 1440/768/390 六种 loading case 中均有两次 GET fulfilled、writes=0、按钮由禁用变为可用。五个主操作控件的间隔、容纳、最小高度和顶部对齐检查通过。两边产品 bundle 相同，因为 diff 只改 fixture。
• 相关原生 history/fixture 测试 13 passed，node --check 与 npm run build:chat 通过；quick premerge 的 diff/RFC/semantic-vocabulary/public-boundary gate passed，精确范围 CQR 通过。build 的大 chunk 提示保留，未放宽预算。

对主干的风险

最强风险是已释放 Promise 泄漏到下一 case，使 loading 断言失效；逐 case 重置和 no-reset 反例覆盖了该风险。原有七种 state、语言/宽度、只读和操作行断言没有删改。复用当前 smoke，不增加近似副本；同作者 JasonBuildAI 的 #8031 为 UTF-8 修复，与本 PR 有不同消费者和缺陷。#8055 的作者是 keoyyo85，不纳入同作者扫描；本次有界扫描未发现足以建议限制贡献者的重复堆积证据。

维护性 ratchet 实际返回 1，涉及 terminal.py、interaction_contract.py、pr_review.py、slash_command_install.py 的四项 module-budget 诊断。原生 runner 标为 advisory_inherited_failure，均不在此 diff；没有独立 base 同命令归因，保留历史原因未确认，不声称修复或扩大预算。当前 loading 不变量有上述单独通过的证据。

首次私有 browser proxy 因 Origin/Host 不匹配返回 asset 403，发生在 analysis 请求之前；确认原因并仅规范自身 origin 后继续，失败记录保留，未把它算作 PR 回归或成功 case。analysis payload 是合成的；完整原生 7 × 2 × 3 = 42-case 矩阵、发布/安装产物和远端 CI 未执行或查询。不能把六个 loading case 扩写为完整矩阵通过。当前 packet 已报告分支冲突；评审通过不解决冲突，改 head 后需重新确认。

我的整体评价

同意此 exact head：修复有真实两次读取消费者，且释放前、释放后、下一次与相反方向写请求均有独立证据。生产操作步骤、布局、权限和额外交互轮数保持不变；测试可以完成并保持重复 case 的有效性，不外推长期效率收益。

视觉评审不适用：此 diff 只改 examples/personal-workspace-browser/research-route-reading.mjs 的测试等待逻辑，没有 rendered presentation 变化；实际 packaged loading 操作行做了自动检查。未验证范围：完整 42-case 原生矩阵、发布/安装应用及美学意义的整页评审。没有未解决的阻塞性 finding；分支冲突、合并策略和授权单独保留，本评审不授权合并。

English verdict: APPROVE — one per-case Promise releases both research reads at 958c17f0654fc2e916360dbaf7c75290cd11c442, while loading/reset and read-only assertions remain meaningful. The full native matrix and release qualification are unverified; merge conflicts remain a separate hold.

[huangruiteng]

This fix is superseded by #8037, whose merge includes f04ec308f6 on main.

At PR head 958c17f065, loadingGate is created once per loading case and both research-analysis GETs share it. The landed main implementation does the same through loadingRead; comparing f04ec308f6 with 958c17f065 leaves only the local variable name and an explanatory comment. There is no behavioral delta left to merge.

#8037 also records the packaged research-route-reading validation across Chinese/English and 1440/768/390px, including all seven lifecycle states. I am closing this conflicting duplicate rather than creating a no-op merge or treating the old-head approval as qualification for a changed head.