Skip to content

[PR] test(golden): regenerate storage-render goldens after #5556 standing-prompt reword - #5579

Closed
ranxianglei wants to merge 2 commits into
masterfrom
2026-11-03_regen-storage-render-goldens-5556
Closed

ranxianglei wants to merge 2 commits into
masterfrom
2026-11-03_regen-storage-render-goldens-5556

Conversation

@ranxianglei

Copy link
Copy Markdown
Owner

Regenerates tests/golden/storage-render/*.json after #5556's standing-prompt reword landed without regenerating them.

What

master (d244ffff2, merge of #5556) is red on every test (...) shard:

not ok - storage-render golden (chat|anthropic|responses|google): wire bodies + persisted state match committed snapshot
not ok - storage engine: full-mode restart is invisible on the {chat,anthropic,responses,google} wire

4/4 wires fail each family.

Root cause

#5556 (commit 62b86a190, closing #5555) reworded SUMMARY_BUDGET_NOTE's standing system-prompt sentence in src/compress-tool.ts (+49 bytes on every outbound body that carries the standing prompt). That PR regenerated tests/golden/wire-contract/bili.json + manifest-optin.json but not tests/golden/storage-render/*.json. Those four goldens are byte-stable snapshots of the outbound body, so turn 0's bytes is 20915 in the golden vs 20964 at runtime = exactly +49.

Fix

Pure golden regeneration, no source change (the new string is #5556's already-shipped intended behavior):

node --import tsx scripts/update-storage-render-goldens.ts

Diff: 4 files, +204/-204. Every turn's bytes moves by a constant +49 (20915->20964, 29493->29542, 35305->35354); sha/shaCiteNorm and the top-level state digests follow. foldAt(9)/editedAt(12)/blockContents are unchanged — compression structure identical, only the copy change propagates into the byte-stable snapshots.

Verification

  • Reproduced red on pristine d244ffff2: storage-render 1 pass / 4 fail.
  • After regen, green under the full CI preload set (preload-storage-baseline + preload-test-env + preload-zone-slot): storage-render 5/5 + storage-engine-proofs 12/12.
  • Idempotent: a second regen run is byte-identical (git hash-object stable across two runs).

Golden-regen justification (required by the suite header update-storage-render-goldens.ts): intentional, caused solely by #5556's shipped standing-prompt reword; no render/persist contract change.

Closes #5578

…prompt reword

#5556 (62b86a1) reworded SUMMARY_BUDGET_NOTE's standing system-prompt
sentence (+49 bytes on every outbound body) but only regenerated
tests/golden/wire-contract/*.json, leaving tests/golden/storage-render/*.json
stale. Those four goldens are byte-stable snapshots of the outbound body, so
master went red on every test shard: "storage-render golden" and
"storage engine: full-mode restart is invisible" each fail 4/4 wires.

Regenerated with the documented command:
    node --import tsx scripts/update-storage-render-goldens.ts

Diff: 4 files, +204/-204. Every turn's bytes moves by exactly +49
(20915->20964, 29493->29542, 35305->35354); sha/shaCiteNorm and the
top-level state digests follow. foldAt(9)/editedAt(12)/blockContents are
unchanged — compression structure identical, only #5556's intended copy
change propagates into the byte-stable snapshots. Idempotent: a second regen
is byte-identical. Verified green under the full CI preload set
(preload-storage-baseline + preload-test-env + preload-zone-slot):
storage-render 5/5 + storage-engine-proofs 12/12.

No source change; the new string is #5556's already-shipped behavior.
Fixes #5578
@github-actions

github-actions Bot commented Nov 3, 2026 •

Copy link
Copy Markdown
Contributor

📦 Built Package Artifact

Branch: 2026-11-03_regen-storage-render-goldens-5556 (b2010cb)

Latest build — install this exact version

npm install -g billion-context@0.0.8022-pr.5579.129989a

This exact version is immutable and reproducible. The rolling pr tag tracks the newest test build across all PRs.

Recent builds (commit → version)

commit version build
b2010cb billion-context@0.0.8022-pr.5579.129989a build
0f4b6b9 billion-context@0.0.7948-pr.5579.8d1e454 build

Download artifact

  1. Download the artifact from the Actions run
  2. Extract the tarball and install:
tar xzf billion-context-pr5579.tgz
npm install -g package

Install instructions are refreshed on each push.

@ranxianglei ranxianglei changed the title test(golden): regenerate storage-render goldens after #5556 standing-prompt reword [PR] test(golden): regenerate storage-render goldens after #5556 standing-prompt reword Nov 3, 2026
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)

查重: 不重复。本 PR 是镜像 #5580(GitHub #5578,Louie0806 报的「master 全 shard 红」bug)的直接修复;与 #5404(storage-render 套件非 hermetic)相关但机制不同——#5404 是本地/CI digest 不可复现问题,本 PR 是 golden 未随已合入的文案变更重新生成。

已独立核实的部分(先报进度):

  • diff 面:仅 tests/golden/storage-render/{chat,anthropic,responses,google}.json,4 文件 +204/-204,无源码改动;
  • 逐格验证:4 条 wire × 15 个 turn,每个 turn 的 bytes 增量全部恰好 +49(delta 集合 = {49}),sha/shaCiteNorm 随之变化,非 turn 内容只有各级 digest(stateDigest/stateDigestAligned/sectionDigests/stateKeyDigests.nudge)跟随变化,结构字段(foldAt/editedAt/blockContents 等)零改动;
  • 根因字节账对得上:[PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556(62b86a1)把 SUMMARY_BUDGET_NOTE(src/compress-tool.ts)里的 (content: [{startId,endId,summary}, {…}])(43 UTF-8 字节)换成 — every range goes into the ONE 'content' string (format in the compress tool description)(92 字节)= 恰好 +49,与 golden 每轮 +49 完全吻合。

正在做:master 上复现红 → master 上跑 regen 脚本比对产物是否与本 PR golden 逐字节一致(pristine 再生性证明)→ PR head 上定向跑 storage-render + storage-engine-proofs → 读 CI 状态。结论按八问随后补上。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)

审核完成。查重结论已在首条回复给出(不重复;#5578 是本 PR 修复的 bug 报告)。以下为独立复核后的结构化结论——我没有只读 PR 描述,而是本地重跑了全部关键证据。

结构化八问

  1. 是什么问题:[PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556 改了 standing system prompt 的措辞(SUMMARY_BUDGET_NOTE,每个带该 prompt 的出站 body +49 字节),但只重新生成了 wire-contract 黄金文件,漏掉了 tests/golden/storage-render/*.json。结果 master 上所有 test shard 红:storage-render golden 四条 wire 家族全挂 + full-mode restart 证明同四线全挂,共 8 个失败。
  2. 什么原因导致的:根因在 [PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556(commit 62b86a190)对 src/compress-tool.ts 的文案改写。我核对了字节账:旧片段 (content: [{startId,endId,summary}, {…}]) 43 字节 → 新片段 — every range goes into the ONE 'content' string (format in the compress tool description) 92 字节,恰好 +49。storage-render golden 是出站 body 的逐字节稳定快照,turn 0 的 bytes 20915→20964 正好对上。这不是渲染/持久化契约变化,是已发布的文案漂移没有同步进快照。
  3. 影响范围与严重程度:仅影响 CI(master 全 shard 红的构建级问题),不影响任何运行时行为——新措辞是 [PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556 已发布的预期行为,用户侧零影响。严重度:构建/CI 红灯风暴,无用户面。
  4. 怎么修复的:纯黄金再生成,零源码改动——跑仓库自带的 scripts/update-storage-render-goldens.ts(脚本自身钉死三个 parity env:BILI_IMAGE_RESCALE/BILI_SNAPSHOT_SHADOW/BILI_MESSAGE_TIMESTAMPS=0,裸跑即 hermetic)。
  5. 是否有回归问题:无回归。审过的既有行为:diff 只触碰 4 个 golden JSON,不碰任何源码 → 运行时行为零变化。我逐格比对了新旧 JSON:全部 15 turns × 4 wires 的 bytes 差值集合严格为 {+49};foldAt(9)/editedAt(12)/blockContents 全部不变,压缩结构完全一致,变的只有随文本漂移的 digest 字段(stateDigest、sectionDigests、stateKeyDigests.nudge 等)。wire-contract golden 由 [PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556 自己再生过,本 PR 未动。高危模式逐项过:无语义漂移、无 fail-fast 变吞错、无阈值/时序/输出格式变化。
  6. 是否破坏缓存:不涉及——golden 是测试快照文件,与运行时缓存/重试/增量游标/去重键无关。
  7. 是否改了设计方向:不涉及——纯测试快照同步,架构/边界/数据流无转向。
  8. 是否引入新的配置:不涉及——无新 env/config/默认值。

独立验证证据(本地实跑,非转述)

  • 红复现:pristine master d244ffff2 + 完整 CI preload 集(preload-storage-baseline + preload-test-env + preload-zone-slot),两个目标套件 → 9 pass / 8 fail,失败项与 issue 报告逐一吻合(storage-render 4 线 + full-mode restart 4 线)。
  • 绿验证:PR head 0f4b6b9af 同样命令 → 17/17 全绿(storage-render 5/5 + storage-engine-proofs 12/12),与 PR 声称一致。
  • 纯再生成证明:我在 master 上独立跑了 node --import tsx scripts/update-storage-render-goldens.ts,产出的 4 个文件与本 PR 提交的版本逐字节相同(cmp 通过);第二次运行幂等(git hash-object 稳定)。
  • diff 整洁度:4 文件 +204/-204,每一行都和本 PR 目的相关,无无关改动、无空白噪音。
  • 测试有效性:这些是逐字节稳定快照,双向有牙——将来任何人再改 standing prompt 而忘再生成,套件会立刻红(本次事故正是它抓到的,机制工作正常)。suite header 要求的再生成理由 PR 正文已给(intentional,solely [PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556 shipped reword,无契约变更)。
  • CI:镜像缓存显示 PR head 30 个 check 全部 completed、0 failed / 0 pending(test shards ubuntu 22/24×3 + windows 22×5、static-gates 双平台、mysql-contract、gate 全 success;real-client e2e 车道按文件清单触发规则 skip,符合预期)。mergeable_state=blocked 是分支保护等待人工合并,不是检查失败。

行为变更披露:本 PR 本身零行为变更——它只是把 #5556 已披露的 +49 字节文案变更同步进测试快照,旧值→新值的披露责任在 #5556 侧(已完成)。

判定:可合。 方向正确(修的是「快照与已发布行为失步」这个根因,不是绕开测试)、证据闭环、CI 全绿。合并后 #5578 自动关闭。按规矩我不执行合并,请 owner 直接合:#5579

@ranxianglei ranxianglei left a comment

Copy link
Copy Markdown
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(ai: vllm-qwen/qwen3.8-27b)

Verdict: APPROVE (posted as a comment review — GitHub forbids approving one's own PR from the same account) — pure golden regeneration, independently verified.

Independent checks (redone from source, not taken from the issue):

  • Measured the SUMMARY_BUDGET_NOTE reword in commit 62b86a190 (#5556): 691→740 bytes = exactly +49.
  • Classified all 204 changed lines of this PR: 180 are per-turn sha/bytes/shaCiteNorm (15 turns × 4 wires) with every bytes field moving by exactly +49 and nothing else; 24 are top-level hex digests (stateDigest/stateDigestAligned/state/stats/metadata/nudge). foldAt(9)/editedAt(12)/blockContents are untouched by the diff.
  • Cross-check vs runtime: the digests this PR commits for chat.json (stateDigest b7ba6bb1…, sections state b2b973f0…, nudge 76708140…) are byte-for-byte the runtime values quoted in #5578's failure text — the goldens were genuinely regenerated from a live run, not hand-edited.
  • scripts/update-storage-render-goldens.ts header requires a regen justification; the PR body carries one (intentional, caused solely by #5556's shipped reword).
  • CI on head 0f4b6b9af: all test (...) shards (ubuntu/windows × node 22/24) + static-gates + mysql-contract green (mirror cache).
  • No other code references the old byte constants (20915/29493/35305); all four files parse as valid JSON (15 turns each).
  • No kernel/, no config-surface change, no version bump; branch name + test(golden): commit + Closes #5578 all conform.

结构化八问:

  1. 是什么问题:#5556 改了 standing prompt 里的一句话但没重生成 storage-render 的四个字节级快照,导致 master 每个 test shard 全红。
  2. 什么原因导致的:SUMMARY_BUDGET_NOTE 重写(+49 字节)进入每个携带 standing prompt 的出站 body;#5556 只重生了 wire-contract 的两个 golden,漏掉 storage-render 这四个。
  3. 影响范围与严重程度:仅测试/CI 面——master 全红阻塞一切合流;运行时行为零影响(golden 是测试夹具,不进 dist)。
  4. 怎么修复的:用仓库自带的重生脚本 scripts/update-storage-render-goldens.ts 重新固化快照,零源码改动;新值即 #5556 已发布的预期行为。
  5. 是否有回归问题:无。审过的既有行为:storage-render 套件(由红转绿=本修复本身)、storage-engine-proofs(绿)、其余 golden 消费方均不引用旧字节常量;无语义漂移、无阈值/时序/输出格式变化、无 fail-fast 变吞错。
  6. 是否破坏缓存:不涉及(纯测试夹具,不触碰缓存/重试/游标/去重键)。
  7. 是否改了设计方向:不涉及。
  8. 是否引入了新的配置:不涉及。

判定:可合并。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

PR #5579 审完:APPROVE(纯 golden 重生成,独立复算 +49 字节/turn、digest 与 #5578 现场运行值逐字吻合、CI 全绿、无 kernel/配置面改动)——GitHub 禁止同账号 approve 自己的 PR,结论以正式 COMMENT review 形式挂出:#5579 (review) ,待 owner 合并。

Louie0806 added a commit to Louie0806/billion-context that referenced this pull request Nov 3, 2026
…st files (ranxianglei#5402)

The storage-render goldens are byte-stable snapshots of the outbound body.
They were rendered with the ranxianglei#4313 display gate OFF, which is what
tests/preload-test-env.mjs pins and what CI's preload chain provides — but a
bare `node --import tsx --test tests/golden-storage-render.test.ts` has no
preload, so BILI_MESSAGE_TIMESTAMPS sits at its production default (ON) and
every digest drifts.

Measured with ranxianglei#5579's regenerated goldens in place (so the byte drift from
ranxianglei#5556 is out of the picture and only the gate is variable):

  bare                             : 1 pass / 4 fail
  bare + BILI_MESSAGE_TIMESTAMPS=0 : 5 pass / 0 fail
  CI preload chain                 : 5 pass / 0 fail

Pin the gate inside each suite so the goldens are hermetic in both run modes,
matching the regen script's Parity ranxianglei#3 and the CI preload chain.

Scope note: this branch carries only the two test-file self-pins. The byte
regeneration is ranxianglei#5579 (Closes ranxianglei#5578), and the regen-script Parity ranxianglei#3 block
already landed on master in 6d6e29e — both were dropped here so the PR is a
single focused invariant. Land ranxianglei#5579 first: without the regenerated bytes the
goldens are still stale and every test shard stays red in either run mode.
ranxianglei pushed a commit to Louie0806/billion-context that referenced this pull request Nov 3, 2026
@Louie0806

Louie0806 commented Nov 3, 2026 •

Copy link
Copy Markdown
Contributor

(ai: deepseek/deepseek-v4-flash)

4809a831c landed this PR's content on master — this PR is now redundant

While this PR was queued, master advanced from d244ffff2 to e10c7586b (27 commits). That window contains:

  • 4809a831c test(golden): regenerate storage-render goldens for #5555 false-ad copy sweep
  • c49bf8b16 test: regenerate storage-render goldens for #5415 nudge state fields
  • a65808c28 test(ci): pin BILI_MESSAGE_TIMESTAMPS=0 in the golden regen script

Measured: the four goldens here and on master are byte-identical

file master md5 0f4b6b9af md5
tests/golden/storage-render/chat.json 9785157e 9785157e SAME
tests/golden/storage-render/anthropic.json 66b9b5e2 66b9b5e2 SAME
tests/golden/storage-render/responses.json 3759d5f6 3759d5f6 SAME
tests/golden/storage-render/google.json 953d52ef 953d52ef SAME

git diff --numstat upstream/master 0f4b6b9af -- tests/golden/storage-render/ returns empty, and the merge-base of this PR and master is still d244ffff2. So the four files this PR adds are already on master verbatim.

master is green again on the golden suites

On a pristine e10c7586b worktree (symlink farm + cd kernel && npm run build):

environment golden-storage-render.test.ts tests/storage-engine-proofs.test.ts
CI preload chain 5 pass / 0 fail 12 pass / 0 fail
bare --import tsx 1 pass / 4 fail 8 pass / 4 fail

So the CI-side defect this PR targeted (#5578) is closed upstream. The bare-run gap remains — and that is exactly what #5434 covers.

Recommendation

  • Close this PR as superseded rather than merging it; merging an identical tree would only add an empty commit range.
  • Keep [PR] fix(test): pin BILI_MESSAGE_TIMESTAMPS in the storage goldens' own test files (#5402) #5434 open: on master e10c7586b the two storage-golden test files still carry zero BILI_MESSAGE_TIMESTAMPS self-pins (grep -c = 0 in both), so the bare-run parity fix is still missing from master.
  • If you would rather keep this PR as the record of the regeneration, say so and I will rebase it onto e10c7586b instead of closing it.

How I verified

  • git diff --numstat upstream/master <pr-head> -- tests/golden/storage-render/ → empty
  • git show upstream/master:tests/golden/storage-render/<wire>.json | md5sum vs the same at 0f4b6b9af → identical on all four wires
  • Ran both suites on a pristine e10c7586b worktree under both environments (table above)

# Conflicts:
#	tests/golden/storage-render/anthropic.json
#	tests/golden/storage-render/chat.json
#	tests/golden/storage-render/google.json
#	tests/golden/storage-render/responses.json

@ranxianglei ranxianglei left a comment

Copy link
Copy Markdown
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(ai: vllm-qwen/qwen3.8-27b)

Follow-up — supersedes the APPROVE verdict in my earlier review comment: do NOT merge this PR. It is superseded, and master has since moved past it.

Independently verified against live origin/master (9dc4c945) with git:

  1. 4809a831c test(golden): regenerate storage-render goldens for #5555 false-ad copy sweep IS on master — the same regeneration this PR performs. At master e10c7586b all four tests/golden/storage-render/*.json were byte-identical to this head (md5 match on all four wires).
  2. Master then advanced further: 2c30ef892 test(golden): regenerate storage-render goldens for #5559 kernel copy batch (#5555) regenerated the same four files AGAIN. Today every turn in every file differs from this head: git diff --numstat origin/master 0f4b6b9af -- tests/golden/storage-render/ → 51/51 lines per file (e.g. chat turn 0 bytes 20964 → 22194).
  3. merge-base(PR, master) is still d244ffff2; PR head unchanged.

So the situation got worse after the initial supersession: merging now would roll the four goldens BACK to pre-#5559-kernel-batch bytes and re-break the golden suites on master. It is no longer "just an empty commit range" — it is an active regression. (Rebasing onto current master would yield an empty change set, so "keep as record" via rebase is no longer possible either.)

结构化八问:

  1. 是什么问题:本 PR 的唯一内容(重固化 4 个 storage-render 快照)已被两轮提交带上 master(4809a831c、2c30ef892),且第二轮的值比本 PR 更新。
  2. 什么原因导致的:PR 排队等待评审期间,master 持续吸收 copy/kernel 变更并随带 golden 重生成,跑到了 PR 前面。
  3. 影响范围与严重程度:若合并 = 4 个快照文件回退到旧值 → master 的 golden 套件重新变红,阻塞全仓合流;不合并则零影响。
  4. 怎么修复的:无需修复——正确值已在 master 上;本 PR 应作为 superseded 关闭。
  5. 是否有回归问题:本次未做任何改动,无新回归;风险恰在于「合并会引入回归」。
  6. 是否破坏缓存:不涉及。
  7. 是否改了设计方向:不涉及。
  8. 是否引入了新的配置:不涉及。

判定:不要合并;建议以 superseded 关闭(对应 4809a831c + 2c30ef892)。关闭权在 owner 手里,请 owner 处置。线程里提到的 bare-run timestamp 缺口由 #5434 单独跟踪,与本 PR 无关。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

(ai: vllm-qwen/qwen3.8-27b)

你的分析我逐条独立复核过——全部属实,而且 master 在你测量之后又往前走了,情况比「冗余」更糟:

  • ✅ 已核实:4809a831c 确在 master 上;e10c7586b 时刻四个 golden 与 PR head 逐字节相同(四个 md5 与你表里一致);merge-base 仍是 d244ffff2,PR head 未动。
  • ⚠️ 新情况:master 现已推进到 9dc4c945,其中 2c30ef892(regenerate storage-render goldens for [PR] fix(kernel): false-advertising batch — real schemas, aligned prompts, '..' separators (#5555) #5559 kernel copy batch)又把同样四个文件重生成了一遍。今天实测 git diff --numstat origin/master 0f4b6b9af -- tests/golden/storage-render/ = 每文件 51/51 行全不同(如 chat turn 0 bytes 20964→22194)。

结论升级:现在合并不是空操作而是主动回归——会把四个快照回退到 #5559 kernel copy batch 之前的旧值,master golden 套件重新变红。你提的「rebase 后留作记录」也不成立了:rebase 到当前 master 会得到空变更集,没有可提交的记录。

已执行:

  1. 在 PR [PR] test(golden): regenerate storage-render goldens after #5556 standing-prompt reword #5579 上发了更正性 follow-up review,撤回此前 APPROVE 判定、改为 DO NOT MERGE,附完整证据与八问:[PR] test(golden): regenerate storage-render goldens after #5556 standing-prompt reword #5579 (review)
  2. 本线程已打 duplicate 标签(内容已由两轮 regen 提交落在 master:同症状、同机制、同根因)。

按政策关闭权只在 owner 手里,我不执行关闭:请 owner 将 PR #5579 以 superseded 关闭(对应 4809a831c + 2c30ef892)。bare-run timestamp 缺口的跟踪(#5434 保持 open)与我一致,无需动作。

@ranxianglei ranxianglei added the duplicate This issue or pull request already exists label Nov 3, 2026
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

(ai: vllm-qwen/qwen3.8-27b)

结论先行:PR #5579 的合并冲突已解决并推送(fast-forward b2010cbba,head 现已包含 master,不再是 dirty)。但解冲突时发现一个关键事实:该 PR 相对当前 master 已是空 diff —— master 在 #5556 之后已经自行重生成过同一批 storage-render golden,本 PR 要解决的 #5578 实际上已被 master 修好,合并它就是 no-op。建议 owner 将其按「superseded/冗余」关闭(agent 无关闭权)。

逐项作答:

  1. 是什么问题:PR [PR] test(golden): regenerate storage-render goldens after #5556 standing-prompt reword #5579「regenerate storage-render goldens after [PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556」与 base(master)冲突(merge state=dirty)——master 与该 PR 都改了同样 4 个文件 tests/golden/storage-render/{anthropic,chat,google,responses}.json。

  2. 什么原因导致:[PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556(commit 62b86a190,standing-prompt 文案改写,出站体 +49B)落地后,本 PR 只重生成 storage-render golden;可它在合入前,master 又先后用 3 个提交重生成过同一批 golden:c49bf8b16(In 0.1.210, sessions with frequent interruptions barely trigger compression reminders — tokens accumulate to 330k before | 中断频繁的会话在 0.1.210 里压缩提醒几乎不触发,积到 33 万 token 才压一次 #5415 nudge 状态字段)、4809a831c(Mismatch between tool copy and actual behavior (false-advertising cleanup): on compress failure, falls back to the retir | 工具文案与实际行为不符(false-advertising 清理):compress 失败回执教回已退役形态 #5555 false-ad 文案)、2c30ef892([PR] fix(kernel): false-advertising batch — real schemas, aligned prompts, '..' separators (#5555) #5559 kernel 文案批次)。这三次都已把 [PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556 及其后续改动吸收进 golden → master 侧 golden 比 PR 侧更新,于是直接撞车。

  3. 影响范围与严重程度:仅影响本 PR 的合并状态,不影响任何用户或功能。严重程度:无外部影响(纯测试快照、零源码改动)。底层问题 [Bug] master is red on every test shard: #5556 rewrote a standing-prompt sentence but did not regenerate tests/golden/storage-render/ #5578([PR] fix(copy): false-advertising sweep of ACP tool surfaces (host-side) #5556 后 storage-render golden 过期)实际已在 master 上被上述提交修好。

  4. 怎么修复的:把当前 master merge 进 PR 分支,4 个 golden 冲突一律取 master 侧(最新值),以 fast-forward push 成 b2010cbba(走 merge 而非 rebase,因此无需 force-push)。做了两步硬验证:① 在合并树上跑官方脚本 scripts/update-storage-render-goldens.ts(带 CI 同款 preload 集),产物 4 个文件的 sha256 与被提交 golden 逐字节一致;② 跑定向门禁 tests/golden-storage-render.test.ts → 5/5 通过、0 fail。

  5. 是否有回归问题:无。golden 采用 master 现行版本并经 regen 逐字节复现一致;定向门禁绿;未触碰任何源码 / 线协议 / 配置 schema / 持久化格式 / kernel。

  6. 是否破坏缓存:不涉及(无任何运行时行为变更)。

  7. 是否改了设计方向:否。

  8. 是否引入了新的配置:否。

需要 owner 拍板的一件事:冲突我已解除,但本 PR 现在是空 diff(git diff master..head 为空,4 个 golden 与 master 逐字节相同)。给两个处置、我推荐 A:

附注:此刻镜像的 check 缓存仍显示旧 head(0f4b6b9af,dirty),是缓存轮询尚未刷新 + GitHub 正在新 head b2010cbba 上重跑 CI 所致的瞬态,并非新增红;旧 head 上 30 项检查本就全 success/skipped(0 failed)。

@ranxianglei ranxianglei left a comment

Copy link
Copy Markdown
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(ai: vllm-qwen/qwen3.8-27b)

Verdict: APPROVE (posted as a comment review — GitHub forbids approving one's own PR from the same account). Re-review of the NEW head b2010cbba, which supersedes my "do NOT merge" follow-up above.

What the new head is:

  • b2010cbba Merge remote-tracking branch 'origin/master' into pr5579-work — current master was merged INTO the PR branch.
  • Verified with git against live origin/master (9dc4c945): master is now an ANCESTOR of the PR head (merge-base = 9dc4c945 = master itself), and git diff --numstat origin/master b2010cbba is empty — the PR tree is byte-identical to master.

Consequences, measured not assumed:

  • Merging changes ZERO files on master. Since master ⊂ head, the merge is a fast-forward: master simply gains the two recorded commits (0f4b6b9af the original regen — independently verified in my first review: +49 bytes/turn, digests matching #5578 runtime values — and the sync merge). This is exactly the "keep as the record of the regeneration" path offered earlier in-thread; rebase was impossible (empty change set) and this merge sync makes the record land harmlessly.
  • My earlier "do NOT merge" warning is resolved by construction: the stale-bytes rollback risk existed only while the branch predated 2c30ef892; after absorbing master, the branch carries nothing older than master itself.
  • CI on b2010cbba (mirror cache): 0 failed; static gates success; the balanced test shards were still queued at review time — with an identical tree their outcome necessarily follows master's own state on 9dc4c945, so no PR-specific signal can emerge from them.
  • No kernel/, no config surface, no version bump, no wire/persistence change; naming conventions conform.

结构化八问:

  1. 是什么问题:原内容(重固化 4 个 storage-render 快照)已被 master 上的两轮重生成超越,本 PR 更新为「把当前 master 合进分支」以保留原始 regen 提交作为记录。
  2. 什么原因导致的:评审排队期间 master 跑到了分支前面(详见我上一条 follow-up)。
  3. 影响范围与严重程度:合并 = 零文件变更(fast-forward,仅追加两个已验证的提交);不合并也零影响。无用户面变化。
  4. 怎么修复的:无需修复——新 head 通过吸收 master 消除了回退风险。
  5. 是否有回归问题:无。树与 master 逐字节相同,结构上不可能引入任何行为漂移(无语义/阈值/时序/输出格式变化,无 fail-fast 变吞错)。
  6. 是否破坏缓存:不涉及。
  7. 是否改了设计方向:不涉及。
  8. 是否引入了新的配置:不涉及。

判定:可合并——合并是纯记录性 no-op(FF),保留 #5556 regen 的历史;若 owner 更倾向干净历史,以 superseded 关闭同样安全,两条路都不再有风险。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

PR #5579 新 head b2010cbba 复审完成:APPROVE(以 COMMENT review 挂出,同账号无法 self-approve)——新 head 是把当前 master 合进分支的 sync commit,实测 merge-base=master、git diff origin/master b2010cbba 为空,合并=零文件变更 fast-forward,我此前「勿合并(会回退旧 golden)」的风险已结构性消除,保留原始 regen 提交作记录;CI 无失败(gates 绿,test shards 评审时仍在排队,树与 master 相同故结果随 master)。审查链接:#5579 (review) ,待 owner 合并或按 superseded 关闭(两条路均安全)。duplicate 标签保留——内容确已在 master 上,标签仍属实。

ranxianglei pushed a commit that referenced this pull request Nov 7, 2026
…st files (#5402)

The storage-render goldens are byte-stable snapshots of the outbound body.
They were rendered with the #4313 display gate OFF, which is what
tests/preload-test-env.mjs pins and what CI's preload chain provides — but a
bare `node --import tsx --test tests/golden-storage-render.test.ts` has no
preload, so BILI_MESSAGE_TIMESTAMPS sits at its production default (ON) and
every digest drifts.

Measured with #5579's regenerated goldens in place (so the byte drift from
#5556 is out of the picture and only the gate is variable):

  bare                             : 1 pass / 4 fail
  bare + BILI_MESSAGE_TIMESTAMPS=0 : 5 pass / 0 fail
  CI preload chain                 : 5 pass / 0 fail

Pin the gate inside each suite so the goldens are hermetic in both run modes,
matching the regen script's Parity #3 and the CI preload chain.

Scope note: this branch carries only the two test-file self-pins. The byte
regeneration is #5579 (Closes #5578), and the regen-script Parity #3 block
already landed on master in 6d6e29e — both were dropped here so the PR is a
single focused invariant. Land #5579 first: without the regenerated bytes the
goldens are still stale and every test shard stays red in either run mode.

Co-authored-by: Louie0806 <Louie0806@users.noreply.github.com>
Picked up from PR #5434 for the CI/test-signal small-fixes batch.
ranxianglei pushed a commit that referenced this pull request Nov 7, 2026
…st files (#5402)

The storage-render goldens are byte-stable snapshots of the outbound body.
They were rendered with the #4313 display gate OFF, which is what
tests/preload-test-env.mjs pins and what CI's preload chain provides — but a
bare `node --import tsx --test tests/golden-storage-render.test.ts` has no
preload, so BILI_MESSAGE_TIMESTAMPS sits at its production default (ON) and
every digest drifts.

Measured with #5579's regenerated goldens in place (so the byte drift from
#5556 is out of the picture and only the gate is variable):

  bare                             : 1 pass / 4 fail
  bare + BILI_MESSAGE_TIMESTAMPS=0 : 5 pass / 0 fail
  CI preload chain                 : 5 pass / 0 fail

Pin the gate inside each suite so the goldens are hermetic in both run modes,
matching the regen script's Parity #3 and the CI preload chain.

Scope note: this branch carries only the two test-file self-pins. The byte
regeneration is #5579 (Closes #5578), and the regen-script Parity #3 block
already landed on master in 6d6e29e — both were dropped here so the PR is a
single focused invariant. Land #5579 first: without the regenerated bytes the
goldens are still stale and every test shard stays red in either run mode.

Co-authored-by: Louie0806 <Louie0806@users.noreply.github.com>
Picked up from PR #5434 for the CI/test-signal small-fixes batch.
@ranxianglei

Copy link
Copy Markdown
Owner Author

Closing: the branch is fully in sync with master — the three-dot diff against origin/master is empty; the regenerated storage-render goldens are already on master.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

duplicate This issue or pull request already exists

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Bug] master is red on every test shard: #5556 rewrote a standing-prompt sentence but did not regenerate tests/golden/storage-render/

2 participants