verify-gguf: fail closed by default when no strong layer ran (magic-only verdict) — deferred until the next MAJOR release
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
调研方向
Read the deferral row at docs/roadmap/risks-and-decisions.md:259 and the exit-code rule in error-handling.md (rule 3) to see why the default stays at exit 0 for now. Do not start until #891 lands C-3 and C-4 and a MAJOR release is scheduled. Done means verify-gguf exits non-zero when no strong layer ran, the opt-out is documented, and the CHANGELOG names F-W20260729-TRUST-04-default.
由索引模型根据 Issue 内容生成。
描述
Deferred work ·
F-W20260729-TRUST-04-default· severity Medium · cost ~1 h · recorded indocs/roadmap/risks-and-decisions.md:259
Summary
Decision C-3 adds an opt-in --require-strong-layer (Phase 16 S5) and keeps a magic-only verdict at exit 0 by default, because flipping the default breaks the jq -e '.valid' recipe the manual recommends, which error-handling.md rule 3 makes a MAJOR change. This issue is the flip itself: a run in which neither the GGUF metadata parse nor a SHA-256 sidecar check happened must not exit 0 by default.
Condition or budget
The next MAJOR release, after Phase 16 S5 (#891) lands C-3 and C-4 — the exporter then writes the sidecar, so the strong layer is present by default and the flip is nearly free.
State today (re-checked 2026-10-09)
No --require-strong-layer flag exists yet, the exporter writes no .gguf.sha256, and a magic-only check still returns valid=True; the package is at 0.11.1rc1.
Related
#83 (magic-only acceptance; lists the non-zero exit only as an alternative) and #84 (the exporter writing the sidecar).
Acceptance criteria
- With no strong layer,
verify-ggufexits non-zero by default; the opt-out is documented and the release notes name the MAJOR change. - The row is removed from the deferral table and the CHANGELOG names its ID (the table's removal contract).
Recorded in docs/roadmap/risks-and-decisions.md (read at f94595f) and re-checked against the code on 2026-10-09. The row leaves that table only when the fix is on main, the CHANGELOG names its ID and any promised guard exists.
- 主要语言
- Python
- 星标
- 9
- 派生
- 1
- 平均合并
- 9 小时 14 分钟
- 30 天内合并 PR
- 37
环境准备
- 提供 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
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。
HodeTech/ForgeLM 的其他 Issue
-
area: site bug good first issue severity: low wave: 4
难度 2/5 1-3 小时 新手友好度 75/100
HodeTech/ForgeLM#961 · 1 条评论 ·
维护者通常 1 天内回复
-
area: dev-tooling bug severity: low source: roadmap wave: 4
难度 2/5 1-3 小时 新手友好度 70/100
维护者通常 1 天内回复
-
Trend-tracking jq recipe prints evaluation_completed=false as 'n/a', because jq's // operator treats false like null (F-W1-DOCS-TRAIN-16)可能已有人在做 @yuefdev 于 2 天前认领。 未关闭documentation good first issue severity: low source: review-2026-09 wave: 4
难度 1/5 1-3 小时 新手友好度 80/100
维护者通常 1 天内回复
-
documentation severity: medium source: review-2026-09 wave: 3
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
-
documentation good first issue severity: medium source: review-2026-09 wave: 3
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
相似的 Issue
-
Use mhctools predictor_required=False for saved-result rescoring可能已有人在做 @iskandr 今天认领。 未关闭
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 79/100
-
难度 1/5 1 小时以内 新手友好度 85/100
521xueweihan/HelloGitHub#4051 · 1 条评论 ·
-
bug documentation
难度 2/5 1-3 小时 新手友好度 78/100
awslabs/visual-asset-management-system#514 ·
维护者通常 1 天内回复
-
backend bug spec:REQ-053
难度 2/5 1-3 小时 新手友好度 66/100
nolte/kamerplanter#2547 ·
维护者通常 1 天内回复