site: show bounded field-evidence gaps and collection path - #367
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c7760396d9
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| stable_retest = plan.get("stableRetest") | ||
| if not isinstance(stable_retest, dict) or stable_retest.get("state") != "not-publicly-documented" or not isinstance(stable_retest.get("requirements"), list) or len(stable_retest["requirements"]) != 3 or any(not isinstance(item, str) or len(item.strip()) < 20 for item in stable_retest["requirements"]): | ||
| errors.append("device-evidence.json: incomplete current-stable retest intake requirements") | ||
| if any(isinstance(p, dict) and p.get("lastRetest") is not None for p in profiles) and isinstance(stable_retest, dict) and stable_retest.get("state") == "not-publicly-documented": | ||
| errors.append("device-evidence.json: retest declaration changed; refresh coverage plan and page claims") |
There was a problem hiding this comment.
When a profile records any valid lastRetest, this check rejects the still-correct not-publicly-documented state even if that retest used an older release; meanwhile the preceding validation rejects every alternative state. For example, a retest on 1.6.20 while stable is 1.6.40 cannot be represented without failing CI. Compare the retest version with the current stable version and support the documented state transition rather than treating every retest as current-stable evidence.
AGENTS.md reference: AGENTS.md:L178-L178
Useful? React with 👍 / 👎.
| attributes: | ||
| label: Evidence type | ||
| description: A pull request or implementation history alone is not proof of a new physical-relay test. | ||
| options: [Sanitized field capture or test log, Sanitized ARSAS diagnostic and reproducible steps, Engineering or implementation history only, Planned test - no result yet] |
There was a problem hiding this comment.
When a reporter selects Planned test - no result yet, no actual test date or sanitized diagnostic exists, but both fields remain required and the date instructions explicitly demand an actual test date. The advertised path therefore forces users either to enter fabricated evidence or abandon the form; planned tests need a separate form/optional fields, or this option should be removed from the evidence intake.
AGENTS.md reference: AGENTS.md:L58-L65
Useful? React with 👍 / 👎.
1 participant
R6.3 — Evidence Coverage & Gap Visibility
Builds on merged R6.2 (#366), using rollback checkpoint
checkpoint/site-r6.2-evidence-freshness-2026-09-22at9092b600fa2bdc53ca6cd7b901efe8b560be4afb.Not declaredversusNot tested.Scope: 5 website/validator/issue-form files. No ARSAS runtime changes and no new vendor/conformance claim.