[Aboysky]

Fixes #7435

Since 155dee1 ("refactor(authority): retire unmigrated operational read fallbacks"),
todo list on a Goal without initialized canonical authority fails closed with
canonical_authority_required. This test still read the source Markdown first and
initialized the authority only afterwards, so all four parameterizations
(monitor_changed/todo_done × file/sqlite) failed at the first read with
ok: false.

Seed the fixture's canonical authority before the first read via the existing
_initialize_fixture_authority helper (the same pattern the sibling tests in
tests/control_plane/test_quota_settlement_cli.py already use), and drop the
now-redundant post-read initialize_canonical_authority call — the helper builds
the same soft_claim shadow projection from the same source state. Product
behavior is unchanged: the four parameterizations go from failing at the first
todo list to passing, and the rest of the test body (quota guard, resume-when +
successor update, unsettled_host_turn_recovery replay, refresh-state recovery,
idempotent replay) passes unchanged.

Other tests that still read before initializing authority (the failures tracked in
#7433/#7434/#7436) likely need the same pattern; keeping that out of this PR.

Verification (on main f7a1ad7, Node 24.19.0):

Check	Result
uv run --extra test pytest tests/test_quota_bound_wait_recovery.py -q	4 passed
uv run --extra test ruff check tests/test_quota_bound_wait_recovery.py	pass
uv run --extra test mypy (repo config; files covers loopx/** only)	pass, no issues in 17 source files
uv run --extra test loopx canary premerge --from-git-diff --git-diff-base upstream/main	pass (diff checks, RFC bilingual gate, py_compile, 3 canaries)
git diff --check	clean

Not run: the full pytest suite beyond this file. Running mypy directly on the test
file reports a pre-existing duplicate-module error that also occurs on unmodified
main (tests/ has no __init__.py), hence the config-implied run above.

AI-drafted under my direction, reviewed and validated by me.

[mergify]

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

Choose the remote for the base repository, not an out-of-date fork.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: actor_kind=model_agent; model=gpt-6.1-sol; provider=OpenAI; declaration_source=runtime_reported; reasoning_effort=max

动机 / Motivation

维护者在检查等待恢复回归时，会运行四个“等待解除方式 × 本地存储”组合。旧基线在尚未建立规范任务存储时先读任务，四个场景都在准备阶段失败，因而测不到原回合恢复；提案先完成夹具初始化，再运行原来的恢复断言。四个场景在旧基线全部失败，在该 head 全部通过；当前 main 也全部通过，因为同一初始化修复已经合入。本次只审查测试夹具，不修改产品恢复、调度或扣额语义，也不以用例通过证明长期收益。

这里的“规范任务存储”指测试真正调用的 File/SQLite authority，不能靠先读 Markdown 回填来绕过已结束的读取兼容。历史修复有意义；当前需要先核对主干已交付的结果。

改动思路 / Approach

提案把初始化放在全部任务写完之后、第一次 todo list 之前，直接复用已有 _initialize_fixture_authority。旧流程先调用被禁止的未迁移读取，后面的初始化根本到不了；新流程先建规范存储，原恢复断言得以执行。这是合理的测试维护方式，没有恢复旧 fallback、引入第二 authority 或改产品选择规则。

但“不再修改”这一更小方案已经能达到相同结果：当前 target 的 既有修复 使用相同 canonical fixture owner，在读取前初始化，而且去掉了多余的首个 todo list。该提交的时间为 2026-10-27 15:06:24Z，早于本 PR 创建时间 16:15:36Z。当前 main 的四个组合独立运行全部通过；因此应先对齐已有交付，而不能把旧基线红、提案绿直接解释成当前必要增量。

具体改动 / Changes

精确 head 485326ebf65890d533ae9b1dcdf58d7a88c37613，相对不可变 base f7a1ad73fec0c9058fe9feecfdff66438be37476 只有 tests/test_quota_bound_wait_recovery.py 一条路径，+5/-19；五条新增包括 helper import、说明注释及初始化调用，删除旧手工投影初始化和不用的 imports。完整下游断言未改变。

规范依据为 Issue #7435 及 不可变旧测试。spec_revision=f7a1ad73fec0c9058fe9feecfdff66438be37476 指被报告失败的固定测试来源；以下 criterion_id 是按原问题和仓库当前交付规则映射，并非新编运行时协议。

criterion_id	处置
four recovery parameterizations	implemented：head 的 monitor_changed/todo_done × File/SQLite 四例全部通过。
retain original recovery obligations	implemented：原 task/Turn、validator、claim、一次结算与新回合选择断言完整保留。
repair a current target gap	not_met：当前 main 已在 00f6be3e4f86294608e857acc994a1b54195ec98 完成相同修复，独立四例全部通过。

关键内容讲解

head 初始化位置 在最后一段 fixture authoring 后立即建立 canonical authority；共享 helper 从完整 source Todos 初始化相应后端。后面的测试仍验证：等待解除时先恢复原回合与原任务，合法收尾之后重放不重复扣额，下一新回合才选择替代任务。附近 settlement 测试没有替代这四种等待组合的全部断言，因此应保留它们；本评审并不建议删除或弱化覆盖。

对主干的风险 / Risks

[P2] 先与主干已合入的同一修复核对增量。 这不是产品运行时 bug，也不是仅凭 merge conflict 提出的 blocker。当前 main 文件 已在第一次 canonical read 前初始化。实际相同命令 uv run --extra test python -m pytest -q tests/test_quota_bound_wait_recovery.py 的结果为：旧 base 四败、head 四过、当前 main 四过。完整比较仅剩 setup/import 写法差异，原恢复断言一致。请 rebase 后保留主干有效 setup；若仍有不同且可复现的维护问题，说明该增量，否则关闭为已被取代即可，不必另造后继任务。

PASS：干净 exact-head 工作树的标准 premerge 为 5 direct + 3 selected 全过、零 manual hold，Ruff 通过。预检与质量回执均仅覆盖上述一条已审查路径。CQR 物理规范回执在首次公开结论前完整读取，质量检查 PASS。没有查询或等待远端 CI；全仓、Windows 和安装态未新验。没有发现断言削弱或产品正确性回归。

我的整体评价 / Verdict

当前判断为 off_goal：历史问题和历史修法都成立，但全 head 尚未证明相对实际 target 的新增成果。long_horizon 为 preserved，已有恢复和一次结算覆盖保留；user_experience 为 not_applicable，只有测试准备变化，不涉及实际 UI/CLI 展示，visual_review 亦 not_applicable、相关 rendered 未验范围为 none。唯一必要处置是与已合入版本核对并保留有用覆盖；不要求修复无关 runtime、改预算或补长期模型试验。仓库经验在此用于核对“可通过”与“当前需要”的差别，不作为自动否定小 PR 的规则。

English verdict: REQUEST_CHANGES at 485326e: the fixture repair is valid against its old base, but current main already contains the equivalent fix and passes all four cases. Reconcile the current-target delta or close the superseded proposal; no product correctness regression was found.

[Aboysky]

Closing — this was fixed on main in the meantime by #7502 (00f6be3), which seeds the canonical fixture authority before the first read the same way. Verified on current main (787f15e): all four parameterizations of test_bound_wait_recovers_original_turn_before_selecting_successor pass (4 passed in 10.87s, Node 24).