Skip to content

[PR] fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone - #5006

Open
ranxianglei wants to merge 13 commits into
masterfrom
2026-10-28_doom-gate-billable-caliber
Open

ranxianglei wants to merge 13 commits into
masterfrom
2026-10-28_doom-gate-billable-caliber

Conversation

@ranxianglei

Copy link
Copy Markdown
Owner

Related: #5000 — the measurement facet. This PR does NOT close #5000 (see Out of scope).

What

The three doomed gates — egress fuse (src/server/relay.ts), preflight fail-fast probe (#2313), post-compress recheck (src/server.ts) — priced ONLY core message bytes against the bar max(2MB, window×7.5 B/token). But preflight's trigger prices messages PLUS system/tools wire overhead (estimateWireOverhead, #470/#829). A payload whose billable total exceeds every model window could therefore clear the floor on the residual caliber. Field evidence from #5000 (dsh 0.1.208): a 26,903-message session with only 1.44 MiB of core text produced 55.2 MiB request bodies that every gate read as "small".

Fix

Every gate site now prices billable content = core message text + system/tools wire overhead, both measured from the ACTUAL outbound body at that site (whatever prepare really transmitted — injected tools, hoisted system, dropped reasoning all reflected automatically):

  • src/server/budget.ts: extraction factored into exported wireOverheadTexts(protocol, parsed) — single source of truth for BOTH calibers. estimateWireOverhead (tokens; trigger/clamp channel, track: plugin-mode proxy-side emergency truncation (last-resort backstop, deferred from #467) #470) computes its original piece-sum over those texts (output identical — pinned by a no-drift test plus the existing tests/wire-overhead.test.ts); new estimateWireOverheadBytes prices the same pieces in UTF-8 bytes.
  • src/server/relay.ts (egress fuse), src/server.ts (fail-fast + post-compress recheck): floor mass = estimateCoreMessagesUpperBytes(msgs) + estimateWireOverheadBytes(protocol, body); the operator-facing 413 messages now carry the component breakdown (~N B message text + ~M B system/tools overhead) and the remedy names the manifest class.

Deliberately NOT counted: the JSON envelope (transport syntax no tokenizer bills — counting it would break the bytes÷7.5 > tokens ⇒ doomed invariant into false positives on dense-but-fitting payloads) and signature-only thinking mass (a derived token projection, not measured bytes — already priced by the calibrated trigger channel). Media stays out of the byte floor as before (separate learned image-token channel).

Behavior change disclosure (old → new)

Tests

New tests/issue5000-doom-floor-overhead.test.ts:

Local targeted run: 34/34 green across issue5000-*, issue4079-*, issue2847-*, issue2490-*, wire-overhead; typecheck clean. Full-suite verdict comes from CI on this head.

Out of scope (tracked separately — decision request filed in-thread on #5000)

Config surface: none. Kernel: untouched.

…t alone (#5000)

The three doomed gates (egress fuse, preflight fail-fast, post-compress
recheck) priced only core message bytes while preflight's trigger priced
messages PLUS system/tools wire overhead — a payload whose billable total
exceeds every window could clear the floor on the residual caliber. Field
evidence (dsh 0.1.208): a 26,903-message session with 1.44 MiB of core text
sent 55.2 MiB request bodies the gate never saw.

All three sites now price core message text + system/tools wire overhead,
both measured from the ACTUAL outbound body via the shared extraction in
wireOverheadTexts (single source for both calibers — token and byte — so
they can never drift). The JSON envelope stays out of the floor: it is
transport syntax no tokenizer bills, and counting it would false-block
dense-but-fitting payloads (#3617 both-ways pinned by tests B/D).
@ranxianglei ranxianglei changed the title fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone [PR] fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone Oct 27, 2026
@github-actions

github-actions Bot commented Oct 27, 2026 •

Copy link
Copy Markdown
Contributor

📦 Built Package Artifact

Branch: 2026-10-28_doom-gate-billable-caliber (938cdec)

Latest build — install this exact version

npm install -g billion-context@0.0.7190-pr.5006.470a0ae

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
938cdec billion-context@0.0.7190-pr.5006.470a0ae build
cd2928a billion-context@0.0.7062-pr.5006.82c82f2 build
b92db90 billion-context@0.0.7035-pr.5006.977503d build
7d0c673 billion-context@0.0.6882-pr.5006.e78aa49 build
8508caf billion-context@0.0.6613-pr.5006.d03175b build
f543809 billion-context@0.0.6587-pr.5006.1177140 build
36b693a billion-context@0.0.6574-pr.5006.34763f4 build
d65fe00 billion-context@0.0.6561-pr.5006.5b1d7ec build
0e5b046 billion-context@0.0.6558-pr.5006.6ab207b build

Download artifact

  1. Download the artifact from the Actions run
  2. Extract the tarball and install:
tar xzf billion-context-pr5006.tgz
npm install -g package

Install instructions are refreshed on each push.

@ranxianglei ranxianglei added bug Something isn't working dsh p1 scope/wiring labels Oct 27, 2026
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

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

独立审核完成。结论:可合(待一个已知 flake 分片重跑转绿;merge 归人)。 审核基于 rebase 后的 head d65fe007(merge-guard 已把 master 9b6ec2345 合入分支;git range-diff 确认 PR patch 与原 head 0e5b046f9 逐字节相同,即我对原 diff 的逐行审读对新 head 完全成立)。查重结论见首条回复:不重复——是 #5000「立即开修」线的落地,不 Closes #5000;PR 声称的 decision request 已核实存在(#5007,p1,来源标记齐备,两条目各列 A/B/C+推荐)。

结构化八问(独立产出)

1. 是什么问题:三道判死闸(egress fuse / preflight fail-fast / 压缩后再检)的字节地板只量 core message text,对 system/tools manifest(MB 级常见)完全失明;而 preflight 触发器侧一直是全口径(含 overhead)。同一份 payload,触发器按全口径判超窗、守卫按残口径判能装——超限请求可以穿过全部三道闸打到上游吃 413(#5000 的 55.2 MiB 实例)。

2. 什么原因导致的:#2490 把地板不变量(bytes÷7.5>tokens⇒doomed)钉在 Σ|m.text| 上;#470/#829 后来给 trigger/clamp 通道补了 system/tools 项(estimateWireOverhead),但没有回头对齐三个 doom 闸点——计量口径分裂是历史遗留,不是本次引入。

3. 影响范围与严重程度:所有走 preflight/egress 的车道(dsh/opencode/pi/codex 同等)。触发条件:manifest 质量把可计费总量推过 bar、但残差文本低于 bar。修复前严重度 p1 类(该形态下会话每请求被上游拒收);修复后仅按披露方向变严,误杀方向有双向测试钉住。

4. 怎么修复的(机制层):三个闸点统一改为「core message text + system/tools wire overhead」,两项都从该点的真实出站体测(relay fuse 用捕获的 wire body、fail-fast 用 prepared.body、再检用 outbound.body——注入工具/hoist 后的 system/dropped reasoning 自动反映,不猜);提取逻辑收敛为单一来源 wireOverheadTexts,token/byte 双口径共用,结构上防漂移;bar(max(2MB, window×7.5)、window 下限 1M)一字未动;envelope 与 thinking 签名质量刻意不计(理由成立:envelope 是不计费的传输语法,计入会把不变量打穿成假阳性)。

5. 是否有回归问题:逐项独立核验过——

6. 是否破坏缓存:不涉及——prefix cache/cache ref/增量游标/去重键零触碰,改动只在三个拒收点的判决量与文案。

7. 是否改了设计方向:无架构转向——把闸的计量对齐到它自己的 trigger 口径(单一提取源)。判决模型的深面(count 墙、fused→unknown fall-through)正确隔离到 #5007 等拍板,没有夹带进本 PR。

8. 是否引入了新的配置:无。零新 env/config/CLI 字段。

独立验证证据(非采信 PR 自述)

两个非阻塞观察(记录在案,不动分支)

  1. 新增的 no-drift 单测(issue5000-doom-floor-overhead.test.ts:112)结构上是自指断言(用同一提取源重算 piece-sum 再比函数本身),真正的 legacy 锚点是既有 wire-overhead.test.ts 的精确值 pin——锚点存在且通过,故不构成缺陷,只提示该单测的实际保护力有限。
  2. 第三个闸点(post-compress recheck,src/server.ts:4021)没有专属 e2e;它与另两个已测闸点共用同一求和式,接线(outbound.protocol/outbound.body)我已逐行目验,风险可接受。

判定:无阻塞发现,可合。 唯一前置动作:人工重跑 flake 分片(或任意方式重触发 CI)转绿后 merge。merge 归人。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

(ai: vllm-qwen/qwen3.8-27b) Second sync: merged origin/master (new wave: #4774/#4352 claude subagent split, #4985/#4999 pi inline streaming — zero file overlap with this PR) and pushed 36b693a79; clean merge, typecheck clean, 34/34 targeted doom-gate tests green locally; branch now contains origin/master (no longer behind), CI re-running on the new head with 0 failures so far — full-suite verdict from CI on this head.

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

(ai: vllm-qwen/qwen3.8-27b) Third sync: merged origin/master (new wave: #4485/#4484 kernel merge-coverage fix + kernel 0.0.169 bump, #4862 invariant-judge egress pairing, #4945 openai usage regressions, #4989 split-brain graceful stop — zero file overlap with this PR; root lockfile delta is only the kernel file-link version) and pushed f54380924; clean merge, npm ci + typecheck clean (kernel rebuilt at 0.0.169), 34/34 targeted doom-gate tests green locally; branch now contains origin/master (no longer behind), CI re-running on the new head — full-suite verdict from CI on this head.

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

(ai: vllm-qwen/qwen3.8-27b) Fourth sync: merged origin/master (new wave: #5016/#4982 draining-twin port claim fix — touches src/server.ts but hunks are disjoint from this PR's doom-gate edits: ~L97/~L1392-1562 vs ~L3547+/~L4239) and pushed 8508cafde; clean merge, typecheck clean, 34/34 targeted doom-gate tests green locally; branch now contains origin/master (no longer behind), full-suite verdict from CI on this head.

ci-check and others added 8 commits October 30, 2026 01:51
…uires (#4599)

The #5140 recovery test proves forkSnapshot reconstructs a non-retained
parent's raw tail from the UNIFIED index — but it never pinned its own
storage mode. Under the CI storage baseline (preload-storage-baseline.mjs
sets BILI_STORAGE_UNIFIED=legacy for the broad corpus) no unified rows
exist, so the reconstruction has no source and forkSnapshot fails closed
with 409 SNAPSHOT_UNAVAILABLE. Deterministic, not a flake: reproduces 100%
on master and any merged branch with BILI_STORAGE_UNIFIED=legacy, passes
without it (default shadow). Pin full like the sibling persist tests and
restore the ambient value afterwards.
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b) Eighth sync: merged origin/master (#5188 sidecar-recovery wave — src/persist.ts / src/plugin.ts / src/storage/shadow.ts + new issue5140/issue5188 suites; zero overlap with this PR's doom-gate files, and the wave's doc-file deltas were already byte-identical to what the prior ci-fixes merge brought in, so the master merge produced no delta there) and re-merged sentinel origin/ci-fixes for its newest commit (67f8e8940 test pin for the cold raw-tail storage mode) — both merges clean, pushed 938cdec12; branch now contains all of origin/master AND origin/ci-fixes (no longer behind either); no manifest changes in the window (npm ci only), typecheck clean, 34/34 targeted doom-gate tests green locally; full-suite verdict from CI on this head.

@ranxianglei

Copy link
Copy Markdown
Owner Author

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

Superseded by #5932 (merged 2026-11-08, fe963d7) — same mechanism (doom gates priced only m.text), different approach: #5932 meters the actual outbound body via estimateRawBodyTokens (the #554 CJK-aware leaf walk, image keys excluded, text-only doom, image overflow routed to the #496/#1800 billing arbitration) at all three forwarding gates, plus wires the #1839 forward-once arm, fixes the handle.ts metadata double-billing, and adds the 32MiB decoded-admission band (#5794) — broader than this PR.

Where this PR diverged — adding system/tools bytes to the preflight fail-fast probe floor — #5932 deliberately keeps the #2490 byte floor at that site (the probe's 413 is the designed #2313 refusal evidence; the #1001 shape must stay able to bet one request). That is the more correct call; this PR would have regressed that contract.

Recommend closing this PR as superseded (agents cannot close). Nothing to salvage: its bidirectional e2e cases and wording pins are bound to its own implementation, and #5932 carries its own regression coverage (issue4079 / request-body-budget / issue4185 / side-request-isolation image_url pin). See the #5000 thread for the full verification.

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

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

CI lane ran. I verified your supersession independently and did not rebase or push — the conflict resolution turned out to be moot, not mechanical.

What I checked this run (live sources only):

Why I stopped rather than resolving: those conflicts are a design choice, not a textual lag — resolving them means picking which implementation wins, which is the decision you already made in-thread (#5932's measured-outbound meter; this PR's overhead-byte divergence "would have regressed that contract"). A mechanical CI lane shouldn't re-make an already-decided trade-off; forcing the rebase would either resurrect the rejected design or empty the PR. Remote branch left unchanged at 938cdec12e.

On red-CI: per the mirror cache (the sanctioned source), the cached checks are stale (all queued, fetched days ago) and mergeable_state: blocked — moot once the PR is superseded and won't merge, so there's no CI fix to apply.

Per your recommendation this should be closed as superseded by #5932. I can't close it (agents can't), and I applied no label because there's no open decision left — you've already decided. Nothing further for me on this branch unless you want it kept.


结构化八问(中文,供人类 reviewer)

  1. 是什么问题:PR [PR] fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone #5006 与 base(master)冲突(merge state dirty),CI 车道被要求 rebase 解冲突;但本轮核实到该 PR 已被已合并的 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 取代——同一 doom-gate 字节地板机制,[PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 改用「实测出站体」计量,取代本 PR 的「核心文本 + system/tools 开销字节」口径。
  2. 什么原因导致的:两个 PR 改同一处代码(relay.ts 的 egress fuse 判定、server.ts、issue4079 测试)。[PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 先合入 master,本 PR 分支停在旧基线(merge-base 44067a5bc7),把 master 合进来必然在相同行上撞出设计级冲突——不是单纯文本落后。
  3. 影响范围与严重程度:仅影响本 PR 自身(一个待关闭的分支)。无用户面/线上影响——本 PR 从未发布,[PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 已在 master 生效。严重度:不涉及可用性,属「无需再处理」。
  4. 怎么修复的:未做 rebase / 未 push。理由:3 处冲突全是「两种实现二选一」的设计决策,而该取舍 owner 已在 thread 拍板([PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 胜出,本 PR 口径会回归 dsh压缩与bili压缩冲突 #2490/billion-context 卡死问题 —— 上报材料 usage获取不到导致上下文估计失误 #2313 的 probe 拒绝契约)。机械 CI 车道不应替 owner 重做已决定的取舍;强推只会复活被否方案或把 PR 掏空。远端分支保持原样 938cdec12e。
  5. 是否有回归问题:我没改动仓库任何代码,零回归风险。反向看,若强行按本 PR 口径合入反而引入回归(把 probe fail-fast 的字节地板从 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 保留的口径改成叠加 system/tools 开销)——正是你指出的点。
  6. 是否破坏缓存:不涉及。未触碰任何缓存/重试/游标/去重逻辑。
  7. 是否改了设计方向:没有。维持 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 已确立的方向(实测出站体计量),不转向本 PR 的方向。
  8. 是否引入了新的配置:不涉及。无任何新 env/config/默认值。

收工判定:本 PR 已被 #5932 取代,冲突解决与 CI 修复均无意义,建议由 owner 直接以「superseded by #5932」关闭(agent 无权关闭)。我这边无后续动作,除非你希望保留该分支。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

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

Re-verified live before acting (mirror cache, fetched today): head 938cdec12e has exactly two red shards — test (windows-latest, 22, 1) and test (windows-latest, 22, 2). Everything else is green: all 6 ubuntu shards (22+24), the other 4 windows-22 shards, static-gates (both OS), and mysql-contract. The remaining lanes are env-gated skips, not failures. So the two reds are real, not a stale event.

I did not fix or push. Reason: last run I established — and you confirmed in-thread — that #5006 is superseded by #5932 (merged, fe963d7f5c), which handles the same doom-gate byte-floor mechanism via measured outbound body. These two windows shards are running this PR's head, i.e. the discarded byte-overhead approach. Green-ing a branch you've assessed as "nothing to salvage / close as superseded" is wasted churn on a decision site, and each fix attempt costs a full push→CI→verify cycle on a branch slated for closure.

Pattern note (I did not pull the job logs, since it wouldn't change the call): only 2 of 12 windows shards fail while all ubuntu and 4 other windows shards pass — that reads as environmental/flaky or a narrow platform-specific quirk of this PR's code, not a broad break. Happy to pull the logs and triage properly if you decide to keep the branch alive.

So the single action left is yours: close #5006 as superseded by #5932 (agents can't close). I applied no label — there's no open design decision (you've already decided supersession) and the lane's tools/add-label.sh isn't present in this workdir anyway. If you'd rather keep it open, say so and I'll dig into the two windows failures specifically.


结构化八问(中文,供人类 reviewer)

  1. 是什么问题:车道报告 PR [PR] fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone #5006 有 2 个 windows 测试分片红(test (windows-latest, 22, 1) / (22, 2)),要求在本分支修复。本轮先按「状态新鲜度」纪律用镜像缓存核对了实时状态,确认非陈旧事件。
  2. 什么原因导致的:这两个分片确实红,但它们跑的是本 PR 的 head 938cdec12e——即已被 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 取代、即将关闭的「字节开销」旧口径实现。红是「被弃方案自身的平台相关/环境性失败」,不是需要抢救的活代码缺陷。
  3. 影响范围与严重程度:仅本 PR 分支自身;无用户面/线上影响——[PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 已在 master 生效且合入时 CI 通过。其余 10 个测试分片 + static-gates + mysql-contract 全绿。严重度:不涉及可用性,属「待关闭分支上的无关红」。
  4. 怎么处理的:未修、未 push。理由:PR 已被已合并的 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 取代、owner 已评估「无可挽救 / 建议以 superseded 关闭」;给一个将被关闭的分支刷绿是决策现场的无效 churn,且每次尝试都要一轮 push→CI→验证的舰队算力开销。远端分支保持原样 938cdec12e。
  5. 是否有回归问题:我未改动任何代码,零回归风险。(反向提示:若强行把本 PR 口径合入反而会引入回归——把 probe fail-fast 字节地板改成叠加 system/tools 开销,即你指出的 dsh压缩与bili压缩冲突 #2490/billion-context 卡死问题 —— 上报材料 usage获取不到导致上下文估计失误 #2313 契约回归。)
  6. 是否破坏缓存:不涉及。未触碰缓存/重试/游标/去重逻辑。
  7. 是否改了设计方向:没有。维持 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 的方向(实测出站体计量)。
  8. 是否引入了新的配置:不涉及。无新 env/config/默认值。

收工判定:实时状态已核实(2 个 windows 分片红、其余全绿);因该 PR 已被 #5932 取代,修复这些红无意义,建议由 owner 直接以「superseded by #5932」关闭(agent 无权关闭)。若你决定保留该分支,我再针对这两个 windows 失败拉日志定位。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

1 similar comment
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)

State unchanged this run (9th lane wake). Verified live: PR #5006 head 938cdec12e (unchanged, via ls-remote refs/pull/5006/head), master has since advanced to e0d63bf8c85d, PR still open + mergeable_state: dirty.

Nothing has moved since the owner's 2026-11-08 note ("Superseded by #5932 — recommend closing, nothing to salvage"). The conflict is the supersession: this branch merged cleanly with master until #5932 (2026-11-08) rewrote the same doom-gate byte-floor code this PR rewrites, so every conflicting hunk is "this PR's discarded byte-overhead approach vs #5932's measured-outbound approach." There is no mechanical rebase — resolving the conflict is a design decision the owner already made against this PR.

Decision needed (owner-only; I cannot close):

I'll stop re-posting this analysis on each automated re-trigger and re-engage only on your A/B call or a new head sha. (The tools/add-label.sh the lane references for the needs-decision label is not present in this workdir, so I can't tag it — noted, not blocking.)

结构化八问(中文)

  1. 是什么问题:[PR] fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone #5006 与 master 冲突、CI 车道反复触发;本轮确认状态未变。
  2. 什么原因导致的:[PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 合入后改写本 PR 同一段 doom-gate 代码,冲突即「取代」本身,无机械 rebase。
  3. 影响范围与严重程度:仅本 PR 分支自身,无用户面/线上影响。
  4. 怎么处理的:未 rebase、未 push;等待 owner 拍板 A(关闭)/B(保留再 rebase)。
  5. 是否有回归问题:未改任何代码,零回归。
  6. 是否破坏缓存:不涉及。
  7. 是否改了设计方向:没有,维持 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 方向。
  8. 是否引入新配置:不涉及。

收工判定:状态稳定、阻塞于 owner 决策(owner 最后表态为建议关闭);不再逐次重复,等待 A 或 B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

10th lane wake — state unchanged, verified live this run:

Action taken this run: applied the needs-decision label to the PR. (The lane's tools/add-label.sh is absent from this workdir, so I used the mirror label API directly; all prior labels preserved — bug, dsh, p1, pr, scope/wiring.)

The conflict is the #5932 supersession itself — no mechanical rebase exists. Decision needed (owner-only; I cannot close):

I'm parking here and will not re-post the full analysis on each automated re-trigger — I'll re-engage only on your A/B call or a new head sha.

结构化八问(中文)

  1. 是什么问题:[PR] fix: doom-gate byte floor prices billable content (core text + system/tools), not core text alone #5006 与 master 冲突、CI 车道反复触发;本轮确认状态未变。
  2. 什么原因导致的:[PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 合入后改写本 PR 同一段 doom-gate 代码,冲突即「取代」本身,无机械 rebase。
  3. 影响范围与严重程度:仅本 PR 分支自身,无用户面/线上影响。
  4. 怎么处理的:本轮打上 needs-decision 标签(add-label.sh 缺失,走镜像标签 API);未 rebase、未 push;等 owner 拍板 A/B。
  5. 是否有回归问题:未改任何代码,零回归。
  6. 是否破坏缓存:不涉及。
  7. 是否改了设计方向:没有,维持 [PR] Over-window requests no longer sent raw upstream: three forwarding gates metered by actual outbound body size + 32 MiB c | 超窗请求不再裸发上游:三道转发闸按真实出站体计量 + 大请求 32MiB 成本带准入 (#5000 #4543 #5885 #5794) #5932 方向。
  8. 是否引入新配置:不涉及。

收工判定:状态稳定、已打 needs-decision 标签、阻塞于 owner 决策;parking,等待 A 或 B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

11th wake — no change. Verified live: head 938cdec12e (unchanged), master 301e1a5459a7, no owner response since floor 38 (2026-11-08, "recommend closing"). needs-decision label already applied.

Still blocked on your call: A (recommended) close #5006 as superseded by #5932 / B keep + rebase (no-op branch). I cannot close it myself. Parking — no further action until you decide or a new head sha lands.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;未改代码零回归;parking 等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

1 similar comment
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

12th wake — no change. Head 938cdec12e (unchanged), master 77fe0b1d63f9, no owner response since floor 38. needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op). Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;零回归;等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

1 similar comment
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

13th wake — no change. Head 938cdec12e (unchanged), master a259e0b8fabf, no owner response since floor 38. needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op). Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;零回归;等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

14th wake — no change. Verified live: head 938cdec12e (unchanged, 14th consecutive check), master 548e396022b9, no owner response since floor 38 (2026-11-08, "recommend closing"). needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op branch). I cannot close it myself. Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;未改代码零回归;等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

15th wake — no change. Verified live: head 938cdec12e (unchanged, 15th consecutive check), master a446a3c1c817, no owner response since floor 38 (2026-11-08, "recommend closing"). needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op branch). I cannot close it myself. Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;未改代码零回归;等 A/B。

@ranxianglei ranxianglei mentioned this pull request Nov 19, 2026
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

16th wake — no change. Verified live: head 938cdec12e (unchanged, 16th consecutive check), master 6453feeeb842, no owner response since floor 38 (2026-11-08, "recommend closing"). needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op branch). I cannot close it myself. Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;未改代码零回归;等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

17th wake — no change. Verified live: head 938cdec12e (unchanged, 17th consecutive check), master 6453feeeb842 (unchanged), no owner response since floor 38 (2026-11-08, "recommend closing"). needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op branch). I cannot close it myself. Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;未改代码零回归;等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 (ai: vllm-qwen/qwen3.8-27b)
🔶 需要决策

18th wake — no change. Verified live: head 938cdec12e (unchanged, 18th consecutive check), master b10c1a0d4ee5, no owner response since floor 38 (2026-11-08, "recommend closing"). needs-decision label applied.

Blocked on your call: A (rec) close as superseded by #5932 / B rebase (no-op branch). I cannot close it myself. Parking.

八问(简):状态未变;冲突 = #5932 取代本身,无机械 rebase;未改代码零回归;等 A/B。

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[ework] auto-fix stalled: test (windows-latest, 22, 1), test (windows-latest, 22, 2) — stopping for human review

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

1 participant