[kixelated]

Problem

#204 and #205 left follow-ups that no quest owns.

Approach

Quest files only. I checked each item against main and the open PRs first:

• quest/m3/ts-byte-edits.md (new, m3): Vec<u8> and [u8] lack the Vec and slice edits other element types have (retain, dedup, insert, remove, swap_remove, last, reverse, swap, sorts). feat(ts): replace a captured &mut's object, and default a crate type in place #204 reported this as "Vec::retain through a place", but the cause is that library.rs skips u8 elements and maps nothing for them. It's in m3 because it's a generic std gap with no moq census site, which the tree plans there.
• quest/m1/ffi-export/iterators.md (new, m1 ffi-export child): Hops::iter, Versions::iter, and Patterns::iter are in moq's exclude list, which is the API a wrapper still has to cover. slices.md was deleted when feat(ffi): export &[T] of any type, and SocketAddrV4 and SocketAddrV6 #205 merged, and instance-keys.md (in progress in feat(ffi): export IntoIterator arguments and serde_json::Value #202) covers iterator arguments, so this is new. The quest leaves open which impl Iterator returns can be collected exactly.
• Item 1, replacing the whole object behind a &mut (6 moq-net refusals), is not in this PR. It's recorded via feat(ts): translate raw pointers, NonNull, MaybeUninit, and unions as typed pointers #201, whose quest/m2/ts-model/library-place.md now covers these refused sites alongside the silent *r = v case, since they share one root.

Impact

None: quest files only.

Alternatives

Folding the ffi iterators into instance-keys.md. That quest is in progress in #202 and is about arguments.

Follow-ups

None.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-29T03:54:26.730692Z	e3dc1cd	Draft marked ready
🔒 Security Review	✅ Completed 2026-10-29T03:54:57.604185Z	e3dc1cd	Draft marked ready

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[kixelated]

Automated review by review (OpenAI)

Reviewed commit: e3dc1cd

No actionable bugs found in this planning-only PR. The byte-edit gaps match library.rs's u8 exclusions, and all three returned-iterator methods remain in ffi/moq/ffi.toml's exclusions. Separating returned iterators from the existing IntoIterator-argument quest avoids conflating ownership and evaluation semantics.

Direction: reuse ByteVec/ByteSlice counterparts and the existing returned-slice ABI rather than add configuration or a new ABI family. Preserving callback cleanup/equality and requiring native differential tests is appropriate. For opaque iterator returns, the explicit finite/side-effect-free eligibility decision and span-based refusal are important: the plan does not assume every impl Iterator can safely be eagerly collected.

Verification: static quest, mapping and exclusion-list review only. No implementation was changed or tested, and no builds, differential tests or census were run.

(Written by OpenAI)