Floor headroom can hide a new barrier at a collapsed identity until the floor is re-armed
维护者通常 1 天内回复
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 35/100
- Issue 类型
- 功能
- 描述清晰度
- 基本清楚
- 活跃度
- 冷清
- 技术栈
- playwright, typescript
调研方向
Read the ground truth section on the evidence floor and inspect the engine's headroom reporting and floor prune behavior. The issue proposes stronger element keys, a non-flapping automatic re-arm design, and possibly per-identity counts in receipts; begin by tracing how collapsed identities and observed counts are represented. Done means a design addresses the documented false green path without flapping on dynamic row counts.
由索引模型根据 Issue 内容生成。
描述
What
The evidence floor records, for each collapsed identity (screen, rule, element key), the number of barriers it accepted. When fewer barriers are present than the floor recorded, the difference is headroom. A genuinely new barrier that arrives at that identity fills the headroom, is counted as carried, and does not block until usabl floor prune re-arms the floor.
Why it matters
This is a false green path in the verdict authority. It was proven against the built engine: floor count 2, one barrier fixed, a new barrier arrives, observed count 2, verdict verified with a receipt.
What ships now
The engine discloses headroom on every run where it exists, on the summary line and under the recorded debt group, and floor prune lowers counts to what was observed so re-arming is one command. Blocking on headroom was rejected because collapsed identities on live pages track row counts (icon buttons per table row reached 15 on one screen), and blocking would flap on an unchanged codebase.
Direction
- Stronger element keys so fewer identities collapse.
- A design for automatic re-arm that does not flap on dynamic row counts, for example re-arming only when the same identity has been observed below its count on consecutive runs against an unchanged tree.
- Consider recording per-identity observed counts in the receipt so a reviewer can see the trend.
The residual is documented in the ground truth under the evidence floor.
- 主要语言
- TypeScript
- 星标
- 0
- 派生
- 0
- 平均合并
- 1 小时 22 分钟
- 30 天内合并 PR
- 19
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 没有贡献指南
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。
usabl-dev/usabl 的其他 Issue
-
docsBaseUrl is not checked for credentials可能已有人在做 @eparenti 今天认领。 未关闭area: setup bug security
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
-
area: setup bug
难度 2/5 1-3 小时 新手友好度 78/100
维护者通常 1 天内回复
-
area: playwright bug
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
-
area: engine bug
难度 2/5 1-3 小时 新手友好度 86/100
维护者通常 1 天内回复
-
gate() does not check draft enums, so a misspelled evidence class never gates可能已有人在做 @eparenti 于 3 天前认领。 未关闭area: engine bug security
难度 2/5 1-3 小时 新手友好度 72/100
维护者通常 1 天内回复
相似的 Issue
-
check:passed streams:add
难度 2/5 1-3 小时 新手友好度 65/100
维护者通常 1 天内回复
-
type: enhancement
难度 2/5 1-3 小时 新手友好度 62/100
franciscoBSalgueiro/en-croissant#1061 ·
维护者通常 3 天内回复
-
S: triage
难度 2/5 1-3 小时 新手友好度 72/100
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 65/100
维护者通常 1 天内回复
-
comp/desktop desktop/shell os/windows P3 tool/file type/bug
难度 2/5 1-3 小时 新手友好度 76/100
NousResearch/hermes-agent#145928 ·
维护者通常 1 天内回复