Skip to content

[PR] docs: point the media-split digest comments at #6356, not the unrelated #6351 (post-merge supplement to #6357) - #6536

Open
ranxianglei wants to merge 1 commit into
masterfrom
2026-11-13_media-split-comment-ref
Open

ranxianglei wants to merge 1 commit into
masterfrom
2026-11-13_media-split-comment-ref

Conversation

@ranxianglei

Copy link
Copy Markdown
Owner

Post-merge supplement to #6357 (landed via batch #6464 as 47f0b96).

What

The batch cherry-pick landed the fix commit but dropped the follow-up docs commit 828aed7, so the two comments it introduced still cite #6351 — an unrelated open PR (fold before pi's cancelled auto-compaction) — instead of #6356, the bug they describe. A reader following the comment ref lands on the wrong PR.

This restores that exact two-line change, cherry-picked with -x from 828aed7:

  • src/storage/shadow.ts: doc comment above messageSignature
  • tests/media-split-digest-parity.test.ts: header comment

Why it matters

Comments only — no code change, no behavior change. Pure reference hygiene so the landed fix points at its own issue.

Verification

Closes nothing — the source issue #6356 is already closed as completed; this only completes the landing of the author's own commit that the batch dropped.

@ranxianglei ranxianglei added documentation Improvements or additions to documentation p3 scope/storage labels Nov 13, 2026
@github-actions

github-actions Bot commented Nov 13, 2026 •

Copy link
Copy Markdown
Contributor

📦 Built Package Artifact

Branch: 2026-11-13_media-split-comment-ref (d7edf18)

Latest build — install this exact version

npm install -g billion-context@0.0.9567-pr.6536.9e62d7f

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
d7edf18 billion-context@0.0.9567-pr.6536.9e62d7f build
731eacf billion-context@0.0.9547-pr.6536.f943459 build

Download artifact

  1. Download the artifact from the Actions run
  2. Extract the tarball and install:
tar xzf billion-context-pr6536.tgz
npm install -g package

Install instructions are refreshed on each push.

#6351

The two comments introduced by this PR cite #6351, which is an unrelated
open PR (fold before pi's cancelled auto-compaction). The bug they
describe is #6356, which this PR closes.

(cherry picked from commit 828aed7)
@ranxianglei
ranxianglei force-pushed the 2026-11-13_media-split-comment-ref branch from 731eacf to d7edf18 Compare November 13, 2026 13:43
@ranxianglei

Copy link
Copy Markdown
Owner Author

🤖 Powered by ework · qwen3.8-27b

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

This is a [PR] thread — PR #6536 already exists with the correct diff, so I verified it rather than re-doing the work, and I fixed its one red CI shard.

What I verified

CI — the original head (731eacf) had one red shard: test (ubuntu-latest, 24, 1) failing on tests/codex-mcp-shell.test.ts:125 "resolution priority: repo layout wins over plugin-dir self" (expected repo-layout, got ''). That test is a different file and subsystem from the 2-line comment change, and the files it exercises were untouched by every master commit between the branch's base and current master — so it's a load flake, not a regression from this PR.

Action taken — per §7.2 I rebased the branch onto current master (a007974) and pushed (new head d7edf1896b). The diff is still exactly the 2 comment lines; the rebase is conflict-free (master's only in-range change, #6480 pure-SQLite default, doesn't touch either file). This triggers a fresh CI run on the current base.

Status — PR #6536 is open, mergeable, now on current master, labels documentation / p3 / scope/storage already set. Ready to merge once the fresh CI goes green. #6536

八问(中文):

  1. 是什么问题:这是 [PR] fix(storage): hash the split media form on both verification legs, so media-bearing tail slots can converge (Closes #6356) #6357 合入后的补漏——批量 cherry-pick([PR] chore(batch): land 8 open fragmented fixes (ReineLan ×4, PatelVro ×2, dimas623 ×2) in one history-preserving batch #6464)落了修复 commit 但丢了两行注释的 follow-up,导致注释仍指向无关的 [PR] feat: fold before pi's cancelled auto-compaction (POST /__bili/plugin/compress) #6351,而非它们描述的 bug [Bug][storage] Sampled verifier hashes split media against hydrated media, so any media-bearing tail slot drifts forever and the session full-ingests every save | 采样校验器把 split 媒体与 hydrated 媒体对比 → 含媒体槽位永久漂移、每次保存全量 ingest #6356。
  2. 什么原因导致的:批量 cherry-pick 只挑了修复 commit,漏掉了作者随后的纯注释 follow-up commit 828aed7。
  3. 影响范围与严重程度:仅注释引用错误,无代码/行为影响;读者顺着注释 ref 会点到错的 PR。体验瑕疵级(p3)。
  4. 怎么修复的:恢复那两行注释改动(从 828aed7 cherry-pick),把 #6351 改回 #6356;并把分支 rebase 到当前 master 触发全新 CI。
  5. 是否有回归问题:无。纯注释不触碰任何执行路径;原 head 唯一红项是无关的 codex-mcp-shell flake(不同文件、不同子系统,且该文件在分支基线到当前 master 间零改动),非本 PR 引入。
  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(等全新 CI 绿)。

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

Labels

documentation Improvements or additions to documentation p3 scope/storage

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant