An off-chain append with no hook in between would shift Pi live positions
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 4/5
- 预计耗时
- 3-5 天
- 新手友好度
- 25/100
- Issue 类型
- 缺陷
- 描述清晰度
- 基本清楚
- 活跃度
- 活跃
- 技术栈
- javascript
- 领域
- backend
调研方向
Read packages/pi-extension/index.js:138-165, especially the parent-chain walk in capture(), and inspect the pi_tui_cli_capture step 5 live-versus-backfill position comparison. First verify on a real Pi client whether an off-chain append can occur without the listed hooks firing; done means the comparison shows matching positions across navigation and compaction, or a real trace demonstrates the uncovered case.
由索引模型根据 Issue 内容生成。
描述
Source: #1874 at 5ab56bcaa3e41a8fcb6a095eb83e9b6d76c3f38e
Location: packages/pi-extension/index.js:138-165
Severity: minor
Evidence
The parent-chain walk in capture() counts only ancestors of the new head, so an entry appended off the current chain is never counted by appended(), shifting later positions by the same mechanism as finding 17. The stronger form of this claim, that ordinary branch navigation breaks, was probed and rejected in round 2: with session_before_tree and session_tree firing, live and recovery agree because checkpoint() resyncs by delivering the appended tail.
Why deferral is safe
The shift requires Pi to append off-chain while firing none of turn_end, agent_end, agent_settled, session_compact, session_before_tree, or session_tree, and there is no evidence it ever does. That is a speculative risk about undocumented client behavior, not an observed defect, and the same acceptance-gate step 5 live-vs-backfill position comparison covers it on a real client. Confirmed the walk is unchanged at head 5ab56bca.
Acceptance condition
The pi_tui_cli_capture step 5 comparison on a real Pi shows live and backfill positions agree across navigation and compaction; if a real client ever demonstrates an uncovered off-chain append, promote this to a defect with that trace.
Review preference; backlog only. A human may explicitly delegate it with neutral:fix.
- 主要语言
- JavaScript
- 星标
- 11
- 派生
- 0
- 平均合并
- 2 小时 25 分钟
- 30 天内合并 PR
- 311
环境准备
这个项目没有提供开发容器、Dockerfile 或贡献指南，环境需要你自己搭建：先看它的 README，通用步骤见我们的新手贡献指南。
从这里开始
- 先读完整个 Issue，再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库，在一个分支上完成修改。
- 提交 Pull Request，并在描述里引用这个 Issue 编号。
hyparam/hypaware 的其他 Issue
-
Windows: mutation guard rethrows EPERM from a lock dir being deleted, so the ingest-seq reserve skips its retry可能已有人在做 @HypForge 今天认领。 未关闭
难度 2/5 1-3 小时 新手友好度 77/100
hyparam/hypaware#2886 · 2 条评论 ·
维护者通常 1 天内回复
-
docs: ollama.md says --source ollama adds to an existing install, but without --force it refuses未关闭
难度 2/5 1 小时以内 新手友好度 62/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 84/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 68/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 76/100
维护者通常 1 天内回复
相似的 Issue
-
Administration: required plugin configuration field stays invalid after restoring its saved value未关闭needs-triage Platform(Default)
难度 2/5 1-3 小时 新手友好度 74/100
维护者通常 1 天内回复
-
bug
难度 2/5 1-3 小时 新手友好度 82/100
nearform/ag-grid-url-sync#314 · 1 个 reaction ·
维护者通常 1 天内回复
-
chore
难度 2/5 1-3 小时 新手友好度 74/100
idean3885/claude-ops-agent#910 ·
维护者通常 1 天内回复
-
S: triage
难度 2/5 1-3 小时 新手友好度 75/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 72/100
oskarrough/agent-ide#29 · 1 条评论 ·
维护者通常 1 天内回复