Skip to content

fix(#2929): fold claim window covers the post-fold cold tail - #2934

Merged
ranxianglei merged 17 commits into
masterfrom
2026-10-13_fold-claim-cold-tail
Oct 12, 2026
Merged

ranxianglei merged 17 commits into
masterfrom
2026-10-13_fold-claim-cold-tail

Conversation

@ranxianglei

@ranxianglei ranxianglei commented Oct 12, 2026 •

Copy link
Copy Markdown
Owner

What

The fold claim window is the post-fold cold tail, not just the first sample. Mirrors the #1847 model-switch cold-window discipline, applied to folds.

Why (#2929)

Issue #2929's log showed 7 cache-seam warnings (~284K tokens re-billed on the chip), each landing 25–66s after a degenerate head fold (foldPoint=m00001/m00009, shrink 81–100%). Mechanism:

  1. A degenerate head fold rewrites the whole prefix → the next request can only hit the floor.
  2. The fold's claim window consumed only the FIRST sample after it — the proxy re-request (bili re-request view).
  3. The client's echo re-bill a minute later (mid-history byte divergence, cold hit, no switch/restart) had no fold attribution → fell into the unattributed ttlRepay residual and tripped the cache-seam detector with "no fold/switch/restart attribution".
  4. bili's own observability already knew (degenerate-fold warn + "cache ceiling ≥0%" prediction) — the seam detector just didn't credit it.

Change

  • kernel buildCacheReport: a fold claims its first sample unconditionally, then COLD continuations (hit < FOLD_CLAIM_WARM_HIT_PCT = 85) up to FOLD_CLAIM_COLD_ROUNDS = 4; a warm line (≥85%) retires every open window. Constants exported for host parity. kernel 0.0.119 → 0.0.120.
  • host cache-ledger: same state machine via sparse claims/retired fields on the persisted LedgerFold (legacy ledgers without the fields keep first-claim semantics).
  • web trajectory chart: the gap-cause heuristic mirrors the window (uses line.foldSeq when present; fold-between-times fallback for pre-codex 命中率问题分析 #2929 ledgers), so the chart's comp-coloring matches the ledger's comp bucket.
  • i18n (zh+en): the residual bucket is honestly labeled "未归因 / unattributed" (upstream TTL/eviction/unknown) everywhere it was called "TTL过期/TTL repay"; the seam tip now mentions post-fold echo re-bills being fold-owned.

Effect on the #2929 numbers: the echo re-bills move from "unattributed TTL" (284K chip) into 压缩回补 (comp repay) where they belong; seams no longer fire for this shape (the a211 sessions had 10 folds but only 1 seam-shaped echo — now claimed).

Testing

  • kernel: cache-report.test.ts +3 (cold-tail full chain: re-request + echo + recovery all foldSeq=1, warm retire, later blip falls to unattributed; COLD_ROUNDS bound; warm-first single-claim) — kernel suite 1281/1281.
  • host: new tests/issue2929-fold-claim-window.test.ts 5/5 (echo claimed + no seam + cr bucket; warm retire; bound; sparse-field legacy; constants parity) — ledger battery (cache-ledger + 4 switch families + seam-detector + 1547/1891/1843 + provenance + compress-observability + 2929) 122/122.
  • web suites 80/80, tsc --noEmit clean, npm run build OK.

Refs #2929 (accounting/attribution half; the growth-threshold tuning it also surfaces is separate).


kernel-change: #2929 — buildCacheReport must apply the same post-fold COLD TAIL claim window as the host's incremental cache-ledger (first following sample claimed unconditionally, then cold continuations hit<85% up to FOLD_CLAIM_COLD_ROUNDS=4; any warm line ≥85% retires every open window — the #1847 model-switch discipline mirrored for folds). The batch report is built inside the kernel, so the host cannot express this without duplicating report logic; FOLD_CLAIM_WARM_HIT_PCT/FOLD_CLAIM_COLD_ROUNDS are exported from the kernel as the single source of truth for host parity.

Storage-render golden regeneration (review follow-up)

tests/golden/storage-render/{anthropic,responses,google}.json regenerated via node --import tsx scripts/update-storage-render-goldens.ts. The persisted LedgerFold now carries the sparse claims/retired fields this PR introduces, so the goldens' metadata section digests drift BY DESIGN — only that section moved (state/blockContents/stats unchanged; chat.json untouched, no cache ledger on that wire). A/B dump of the full persisted session (master f8f7d40 vs this head) confirms the only changed bytes are the cache-ledger subtree: the fold claims 4 samples and retires, the echo re-bill lines move from unattributed ttlRepay (foldSeq: null) into comp repay (foldSeq: 1), and seamSuspects drops 2 → 0 in the fixed scenario. This is the intended #2929 attribution change pinned into the snapshot, not incidental drift.
The follow-up commit 518ff78e1 (provider-side miss split by idle gap) added the new providerTtl{Misses,Missed} ledger fields and re-refreshed ALL four goldens (a59b9e2ec) so the snapshots pin both attribution changes; legacy ledgers migrate the new agg fields to 0 via the existing getCacheLedger guard.

@github-actions

github-actions Bot commented Oct 12, 2026 •

Copy link
Copy Markdown
Contributor

📦 Built Package Artifact

Branch: 2026-10-13_fold-claim-cold-tail (e76b35f)

Option A — Install this exact build (recommended)

npm install -g billion-context@0.0.3312-pr.2934.27b87e8

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

Option B — Download artifact

  1. Download the artifact from the Actions run
  2. Extract the tarball and install:
tar xzf billion-context-pr2934.tgz
npm install -g package

Install instructions are refreshed on each push.

@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 收到,开始 review。PR #2934 分支 2026-10-13_fold-claim-cold-tail(5a35f14)已拉取核对中。

初步 CI 状态:大部分绿,两处红——kernel-guard(PR 正文缺 kernel-change: 声明,触了 kernel/)与 lock-sync,另有 2 个 e2e job 仍在跑。我会先读完整 diff 做回归审查,再处理这两处门禁(正文补声明属流程修复,lockfile 若确为漏更会直接修)。完整结论按八问体例随后回复。

ranxianglei pushed a commit that referenced this pull request Oct 12, 2026
ranxianglei pushed a commit that referenced this pull request Oct 12, 2026
@ranxianglei
ranxianglei force-pushed the 2026-10-13_fold-claim-cold-tail branch from 7ffa5b1 to c019591 Compare October 12, 2026 13:56
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 Review 完成,修复已作为增量 commit 推到分支(c01959161),CI 守卫全绿,等测试矩阵出结果后给结构化结论。

审查结论(方向):根因修对了——#2929 的 seam 告警是「折叠后冷尾回显重计费无归因」掉进 ttlRepay 残差,本 PR 在产生方(kernel buildCacheReport + host cache-ledger 状态机)把 claim window 扩到 post-fold 冷尾,不是症状层拦截。kernel/host/web 三层状态机逐行核对一致,decomposeSample 未动,legacy ledger 经 sparse 字段保持首 claim 语义。

审查中发现并已修复(独立标注 commit):

  1. 7fe7e0ee3 — root package-lock.json 缺 kernel 0.0.120 同步(lock-sync gate 失败原因),1 行再生成;
  2. fa8fa8b6f — kernel/package-lock.json 在 master 上本就 stale(0.0.114 vs pkg 0.0.119),随本次 bump 一并再生到 0.0.120,独立 commit;
  3. ff5e0aba6 — web gap-cause 启发式两处缺陷:①对每条 foldSeq 行都重新武装 foldTailClaims=4(ledger 只在 opener 武装一次)→ 持续冷段会多染 ~4 个样本;②warm 行的 early-return 跳过了 counter 清零 → warm-first 场景把 warm 之后的 TTL 抖动误染成 comp。已改为仅 opener 路径武装、warm 无条件清零,并用 5 场景 A/B 脚本对照旧行为验证(codex 命中率问题分析 #2929 echo+warm 不变、COLD_ROUNDS 边界精确、legacy 恰好 4、warm-first 修正、unattributed opener 对齐 ledger)。纯图表着色,不动数据。
  4. PR body 补 kernel-change: 声明(kernel-guard gate 要求;内容:为何 host 侧无法替代——claim window 是 kernel cache-report 与 host ledger 的双侧契约,常量必须单一来源)。
  5. 并行会话中途推上的 2 个 i18n 术语 commit(b47cebe/c0195b1 系列)已 cherry-pick 保留,零冲突。

另:分支已 rebase 到当前 master(f8f7d40,#2923),无文件交集。本地定向测试全绿(host 42/0、kernel cache-report 25/0、web-i18n 10/0),全量矩阵交给 CI。

ework-agent and others added 9 commits October 12, 2026 22:31
A degenerate head fold (81-100% coverage) rewrote the whole prefix, but the
claim window only consumed the FIRST sample after it - the proxy re-request.
The client's echo re-bill ~25-66s later carried no foldSeq, fell into the
unattributed ttlRepay residual, and tripped cache-seam with "no fold
attribution". Every one of #2929's seven seams was exactly this shape.

Mirror the #1847 switch discipline for folds:
- kernel buildCacheReport: a fold claims its first sample unconditionally,
  then COLD continuations (hit < FOLD_CLAIM_WARM_HIT_PCT=85) up to
  FOLD_CLAIM_COLD_ROUNDS=4; a warm line retires every open window.
- cache-ledger recordCacheSample: same state machine on the sparse
  LedgerFold.claims/retired fields (legacy persisted ledgers keep the
  first-claim behavior).
- web trajectory chart: cause heuristic mirrors the window (foldSeq-first,
  fold-between-times fallback for pre-#2929 ledgers).
- i18n: the residual bucket is honestly "unattributed" (upstream
  TTL/eviction/unknown), not "TTL expiry"; seam tip mentions post-fold
  echo re-bills.

kernel 0.0.119 -> 0.0.120.
The same comp bucket appeared as 压缩折损/压缩/压缩后重新建缓存/压缩复付/压缩回补
and the residual bucket as 未归因前缀失效/未归因/未归因回补 across the overview,
session list, detail card, miss-split line, chart tooltips and economics copy.

Canonical zh terms now: 新内容 / 压缩回补 / 未归因回补(上游TTL/淘汰/未知)/ 残差;
en mirrors: new / comp repay / unattributed repay / residual.

Also fixes the attribution-section note, which claimed the seam/provider/
rewind/prompt/projection partitions are "not counted in TOTAL above" — they
are refinements OF the unattributed repay bucket and are already included in
it (abort stays an overlay counted in no bucket).
…normal

The provider-side arm (byte-stable / append-shaped miss with no fold or
switch attribution) conflated two very different situations: a miss
seconds after the previous request (abnormal — upstream eviction or
relay rotation) and the same miss after a 20-minute idle (normal — the
upstream prefix cache simply aged out). The web page labeled both
"上游侧失效", which reads as a failure when it is expected cost.

- cache-ledger: bookProvider() splits by idle gap
  (PROVIDER_TTL_GAP_MS = 10 min, mirroring the chart's GAP_MS) into
  agg.providerTtlMissed (normal) vs agg.providerSideMissed (abnormal);
  seam.providerTtl{count,missed} rides the ledger + /acp-cache report
  (⏱ UPSTREAM TTL EXPIRY — NORMAL section).
- web: attribution card gains the ⏱ 上游 TTL 过期(正常) mini (green);
  ▲ 上游侧失效 re-labeled (异常) with a short-gap qualifier tip.
- web: the headline hit-rate subline is the three-bucket split only —
  the seam/provider/rewind/abort/prompt/projection chips duplicated the
  归因 · 未解释残差 card below and mixed taxonomy levels (#2929).
- taxonomy rename: 折叠后重读重建 → 折叠后重建 / post-fold re-read →
  post-fold rebuild.
- tests: gap-split case (short-gap → providerSide, long-idle →
  providerTtl) with a pinned Date.now (settleUsageReport stamps it
  internally); dead chip keys removed from both locales.
…s an inline breakdown

The economics card and the separate attribution card read as two parallel
grids, hiding that seam/provider/TTL/rewind/abort are subsets of the
unattributed-repay bucket. Render them as an inline breakdown inside the
未归因回补 mini (zero rows skipped, residual row only when >8192) and keep
the standalone card only for the seam event list. Rows use literal t()
calls so the #1024 dead-key scan still sees every tip key.
@ranxianglei
ranxianglei force-pushed the 2026-10-13_fold-claim-cold-tail branch from bfc0289 to 41eeefa Compare October 12, 2026 14:32
ework-agent and others added 5 commits October 12, 2026 22:39
…o duplicate grids

The previous pass nested a line list inside the unattributed KPI tile and
kept a second attribution grid with a duplicate seam mini + event list
(the events table already lives under the trajectory chart). Now the four
KPI tiles stay clean, the bucket composition renders as one subordinate
strip directly beneath them (containment is the layout), the redundant
attribution section and attr_title/attr_note keys are gone, and the
stale en 'Post-fold re-read' label is renamed to 'Post-fold rebuild'.
…ution

The chart used to re-derive miss causes with its own gap heuristic, so the
paint and the economics card told two different stories, and the legend
asserted a guess ('likely upstream TTL/eviction'). detectSeam now stamps
each settled line with the arm it booked (prov: ttl | side | rewind; seam
already existed), the report exposes seam/prov per line, and the chart
paints those fields first — the gap heuristic remains only as the legacy
fallback for old ledgers. The unattributed family renders as one bordered
legend group (⊇) with only the sub-causes that actually occur, matching
the economics composition strip bucket-for-bucket.
… mini()

Hand-rolled the tile with class="mini", which is not a CSS class in this
page (the helper emits class="stat"), so the tile rendered as bare text
between its boxed neighbours and the row looked broken. One-liner: call
mini() like its three siblings, which also gives the tile the "?" qmark
with the bucket tooltip.
…ion stamps

detectSeam now stamps each settled line booked into a provider-side arm
(prov: ttl | side | rewind); the golden fixture's 8 side-arm lines carry
the new sparse field. A/B persisted-state dump (b896a09 vs this head)
shows exactly those 8 fields and nothing else.
…ds bind to their bars

Requested after visual review of the rendered page:
- The 'of which' breakdown (seam / TTL / provider / rewind / rest) was a caption
  floating under the KPI row — it is now part of the unattributed-repay tile
  itself: label · value · then the parts, top to bottom. Containment is the
  layout, not an annotation beside it. The dead ce_ttl_tip qmark key goes away.
- The trajectory chart only labelled samples 1/5/10 and each miss-cause band
  floated free of any bar, so band-to-bar mapping required detective work.
  Now every sample gets a seq label (thinned only when dense) and each band
  drops a thin same-colour spine from its floor to the baseline, tying it to
  the exact bar column it annotates.
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

[bot] 🏷 Review complete — PR #2934 ready to merge. CI fully green on head c20078d43 (rebased onto master da15214e6), all 32 checks passing. This supersedes my earlier interim status comment.

结构化八问

  1. 是什么问题:degenerate head fold 重写整个前缀后,客户端 echo re-bill(mid-history 字节分歧、冷命中、无切换/重启)没有折叠归因通道——落进未归因 TTL 残差并误触发 cache-seam 检测器(codex 命中率问题分析 #2929 的 7 条告警、~284K tokens 被记成"上游失效")。
  2. 什么原因导致的:fold claim window 只消费 fold 后的第一个样本(proxy re-request 视图),后续冷尾样本无任何归因路径。是归因状态机窗口过窄(机制缺失),不是检测阈值问题。
  3. 影响范围与严重程度:所有"degenerate fold 后跟冷尾 echo"的会话(codex 命中率问题分析 #2929 chip 实例)。后果是 seam 误报 + 计费归因错桶(压缩回补记成了 TTL 过期)——纯 observability/记账层错误,无数据丢失、无 wire 影响。
  4. 怎么修复的:fold claim window 覆盖 post-fold 冷尾——fold 无条件认领第一个样本,之后对冷样本(hit<85%)继续认领至多 4 轮;任何 warm line(≥85%)retire 全部 open window(证明 provider 已重缓存)。kernel buildCacheReport 与 host cache-ledger 实现同一状态机,常量 FOLD_CLAIM_WARM_HIT_PCT=85 / FOLD_CLAIM_COLD_ROUNDS=4 由 kernel 导出作为 host parity 单一事实源;web 图表启发式镜像同一窗口。修在归因决策点(producer),不是症状层遮罩——同一根因换形状出现时修复依然成立。
  5. 是否有回归问题:逐项清点 diff 触碰路径的既有行为——
    • 归因桶:echo re-bill 从 ttlRepay(未归因) → compRepay(压缩回补):有意变更(本 PR 目的),A/B dump 证实 agg tr 180973→142495 全部来自该迁移;
    • seam detector 不再对 post-fold 冷尾 shape 触发(A/B:seamSuspects 2→0);
    • legacy ledger(无 claims/retired 字段)保持 first-claim 语义——向后兼容已测;
    • consumedFoldSeq 变为 write-only(保留 schema 兼容,无害死状态);
    • i18n "TTL过期/TTL repay" → "未归因/unattributed":诚实标注修正,键名不变;
    • 并行会话 follow-up(aba017fe5)把 provider-side miss 按 idle gap 拆成正常 TTL(>10min,⏱ 绿)vs 异常(▲):新 taxonomy,migration guard 给 legacy 默认 0;我验证了其 prevSampleAt 索引无 off-by-one(lines.push src/cache-ledger.ts:969 先于 detectSeam :1264);
    • 并行会话 6f4cd2380(trajectory bands)让 detectSeam 在每个 booked 样本上盖 prov: ttl|side|rewind 戳(与 agg 记账同一分支内设置,一致性由构造保证),report 按行暴露 seam/prov,图表优先读权威字段、gap 启发式降级为纯 legacy fallback(仅当 foldSeq/seam/prov 三者皆无)——我逐行核过,且该设计取代了我此前对启发式的修补(修补仍保留在 fold/comp 段,与其注释一致);
    • storage-render goldens ×4 wire 两轮重新生成(claims/retired 字段一轮、prov 戳一轮):A/B state dump(worktree @ b896a09 vs head)证明第二轮唯一变化 = 8 个新 sparse prov:"side" 字段(lines[3..8,13,14]),其余逐字节不变;第一轮 A/B 证明变化仅在 metadata cache-ledger 子树(fold {claims:4,retired:true}、echo 行 cr/foldSeq、seamSuspects、agg 汇总),state/blockContents/stats 不变。
      结论:无意回归未发现。
  6. 是否破坏缓存:不涉及。prefix cache 策略、重试、去重键、增量游标均未触碰;全部改动在事后归因统计层。
  7. 是否改了设计方向:不涉及。这是把 cache-ledger 归因账本:同一残差重复计入多个重叠桶、switch flag 只比紧邻上一条、未归因残差占绝对主导、unknown-cache 样本被画成 0% 平台 #1847 model-switch cold-window 纪律扩展到 folds(issue 中 owner 已如此定框),无新边界/数据流。边界情况已在 PR body 记录:长 idle 跨缝最多认领 4 轮 TTL blip(与 cache-ledger 归因账本:同一残差重复计入多个重叠桶、switch flag 只比紧邻上一条、未归因残差占绝对主导、unknown-cache 样本被画成 0% 平台 #1847 已接受 tradeoff 一致);legacy ledger 重启升级后获得一次 fresh first-claim(有界,warm line 出现即自愈)。
  8. 是否引入了新的配置:不涉及。无新 env/config/CLI flag;两个 kernel 常量是代码常量(导出仅为 host parity),非用户可调旋钮。

Review 过程说明(diff 之外做了什么)

  • 原 issue 方向审核:codex 命中率问题分析 #2929 机制分析准确(7 条 seam 各落在 degenerate fold 后 25–66s;bili 自身观测早已知道 degenerate fold + "cache ceiling ≥0%" 预测,只是 seam detector 没给它 credit)——方案打到核心。
  • 复现/查证:A/B state dump(.tmp/state-pr.json vs .tmp/state-base.json,driveWire 全量对比)确认只有预期归因字节变化;本地定向测试最终 head 全绿(host battery 58/58;kernel cache-report 套件在对应 head 全绿),远程 CI 最终 head 全矩阵 32/32。
  • 修掉的门禁失败 ×3:PR body 补 kernel-change: 声明(kernel-guard);root package-lock 同步 kernel 0.0.120(lock-sync);goldens 重新生成(test matrix 因此才第一次真正跑到)。原始作者看不到 golden 失败是因为 guards 先挂了整个 test matrix。
  • 修掉的真缺陷 ×1(web 图表启发式,独立 commit):旧代码在每个 foldSeq-stamped 行都重新 arm 冷尾计数器 → 新 ledger 上图表比 ledger 多染 ~4 个样本的 comp 色;且 warm-first 场景 early return 跳过 warm reset,把 warm 之后的 TTL blip 误染成 comp。现改为只在 opener 行 arm + warm 恒 reset;用 5 组场景对照旧行为验证(含 codex 命中率问题分析 #2929 原例、持续冷尾、legacy ledger、warm-first、未归因 opener)。后续被并行会话的权威字段设计取代(见 ⑤),该修补保留在 fold/comp 段仍正确。
  • 补做并行会话漏掉的 golden 再生成 ×2 轮:第一轮(claims/retired 字段)与第二轮(6f4cd2380 的 prov 戳——该 commit 改了 cache-ledger.ts 但没跑 golden 测试,本地必挂 ×3 wire)均为独立标注 commit,各附 A/B dump 证据。
  • 多次 rebase 保持分支新鲜:master 前进两次(fix: 释放会话派生缓存与诊断请求底层字符串 #2923 → fix: fork 子会话 nudge 全盲死区 — 收据播种 sizing 基线(0.1.195 卡死复现) #2936/fix: surface DSH no-output-ceiling degraded budget on user-visible surfaces (#2904) #2922)+ 并行会话推送数次,全部零冲突合入;并行会话的 9 个 follow-up commit(i18n 术语统一 ×2、providerTtl 正常/异常拆分及其 golden 刷新、attribution minis 内联进未归因桶、归因展示条去重清理、trajectory bands 权威归因、KPI tile mini() 一行修复、composition 移入未归因 tile + 图表 band 绑定到所属 bar)全部原样保留,每个都单独过审(除 trajectory bands 触碰 cache-ledger.ts 外均 web-only)。

合并判定

可合——但按 Kernel Boundary 规则 3,kernel/** 变更永远 human-review tier,不 auto-merge,合并由你执行。一个流程项请确认:Kernel Boundary 规则 1 要求 kernel 变更实施前有 owner 批准的 dedicated issue——本 PR 中 owner 亲自撰写的 #2929/#2934 issue 本身就是该 kernel 变更的精确规格(buildCacheReport 窗口 + 导出常量 + 0.0.119→0.0.120),我将其视为批准记录;若你认为还需要单独的 kernel issue 请说一声,我补建。

ework-agent added 3 commits October 12, 2026 23:54
…loor

`Math.ceil(24 / Math.max(barW, 1))` clamped sub-pixel bar widths UP to
1px, so sessions with thousands of samples (barW≈0.36) always got
tStep=24 → 100+ overlapping vertical labels. The demo case (N=10,
barW≈99 > 1) never exercised the clamp, which is how it shipped.

Drop the inner max (the ternary guarantees barW>0 for N≥2, N=1 → 1):
dense charts now hold the intended ~24px horizontal pitch
(N=2494 → tStep=68 → 37 labels).
Show a one-line small-text hint on the 折叠后重建 tile whenever the value
is non-zero: lowering compress.preserveRecentMessages (the protected
recent-message count) reduces how much recent content each fold drags
back into context — minimum recommended value 1. Bilingual zh/en,
conditional so zero-rebuild sessions stay clean.
▲上游侧失效 reused #bf8700 (the 折叠后重建 amber) while 其余 rendered as
#656d76 dim — two muddy earth tones that read as one color at chip size.
Align the provider-side chip with its chart band (#f0883e) and lighten
the rest-of-dust chip to neutral #9aa1a8, so every bucket now has a
distinct hue: red seam / green TTL / orange provider-side / slate rewind
/ gold abort / light-gray rest.
@ranxianglei
ranxianglei merged commit aeb39a7 into master Oct 12, 2026
32 checks passed
@github-actions

Copy link
Copy Markdown
Contributor

📦 Merged Build Published

This merge to master was published as 0.0.339-master.aeb39a7 (rolling master tag).

npm install -g billion-context@0.0.339-master.aeb39a7  # this exact build
npm install -g billion-context@master  # always the newest merged build

npm view billion-context@master prints the current version at any time.

ranxianglei pushed a commit to YunzhouLi-hub/billion-context that referenced this pull request Oct 13, 2026
The chart legend's 'rest' swatch shared the parent group's purple
(#8250df == det.cause_unattrib), so 其余 was indistinguishable from
未归因回补 itself. Point CAUSE_COLOR.rest and the two standalone
legend swatches at #9aa1a8 — the same light gray the economics card
chips already use (ranxianglei#2934/e76b35f9), keeping card<->chart consistent.
@github-actions github-actions Bot mentioned this pull request Oct 13, 2026
@ranxianglei
ranxianglei deleted the 2026-10-13_fold-claim-cold-tail branch October 15, 2026 05:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant