Repository navigation
chore(quest): follow-ups from #87 through #99 - #102
Conversation
Follow-ups from #98 and #94. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0115AT2ue6x9cT5e5J9SZtCn
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: 50e2c59
No actionable bugs found in these two follow-up plans. The comparison quest correctly distinguishes inherited defaults from overrides and derives ordering from partial_cmp, including unordered results, rather than incorrectly combining equality and ordering. Explicitly named generic instances with checked C names and refusal of unconfigured instances are a suitably bounded alternative to automatic ABI expansion; preserve the existing cross-export name-collision checks when implementing this.
Verified the scope against #94/#98 and the exporter’s current configuration/ABI rules. Plan/source inspection only; no implementation, tests, or census runs were executed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0115AT2ue6x9cT5e5J9SZtCn
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: 3999100
Reviewed the two new quests since 50e2c59; no actionable bugs found in the plans. The earlier comparison/generic-instance plans are unchanged. Separating Cow<str> from UTF-8 slicing/pattern support follows #96’s actual refusal boundaries and keeps the work bounded.
For implementation, a plain JS string is sufficient only while ownership-variant observations remain explicitly refused; supporting matches!/is_borrowed requires the tagged alternative already identified. Keep empty-pattern progress/endpoint behavior and byte-boundary failures in the planned native differential coverage rather than delegating to JS slicing/splitting semantics.
Plan/source inspection against #96 only; no implementation, tests, or census runs executed.
Replaces ts-shared-read's narrow fix, as #99 suggested. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0115AT2ue6x9cT5e5J9SZtCn
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: 2aa91ca
Reviewed the replacement of the narrow held-read optimization with whole-crate immutable-type sharing since 3999100, excluding the merged main changes. No actionable bugs found in the plan; the other quests are unchanged.
The broader analysis can remove several redundant copies through one rule rather than special-casing the decoder. Its stated fail-closed boundary is essential: include nested field/element and discriminant mutations, and keep copies where mutation can escape the analyzed crate. The planned cross-module/generic cases plus #83/#99 aliasing regressions are appropriate; retained Rust value semantics matter more than reaching the pre-fix timing target.
Plan/source inspection only, including #99’s store-analysis rationale. No tests, census runs, or benchmarks executed.
From #87's and #97's bench findings. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0115AT2ue6x9cT5e5J9SZtCn
kixelated
left a comment
There was a problem hiding this comment.
Automated review by review (OpenAI)
Reviewed commit: a89d428
Reviewed the new A/A benchmark-control quest since 2aa91ca, excluding upstream #96/#99 changes. No actionable bugs found in the plan; earlier quests are unchanged. Matching module layout and isolating routes directly addresses the reported profile-sharing/layout confounders instead of treating their systematic difference as timing noise.
For implementation, a separate process is the clearer isolation baseline; a realm-based alternative should demonstrate that the engine does not still share the relevant compiled-code/type profiles. Preserve comparable warm-up and randomized/repeated route scheduling so isolation does not simply replace cross-route contamination with run-order drift.
Plan/source inspection against the benchmark harness and #97’s report only; no benchmark measurements or tests run, and the reported engine effects were not independently reproduced.
Follow-up quests the maintainer chose from the #87, #94, #96, #97, #98, and #99 reports:
quest/m2/ts-cmp-defaults.md: the defaultne/lt/le/gt/gea crate's own impls inherit, plus[u8]equality (about 60 moq-netnerefusals).quest/m2/ts-cow-str.md:Cow<str>(139 moq-net refusals, the last one inietf/namespace.rs).quest/m2/ts-str-slice.md:&s[a..b]and&strpatterns, sorfind's byte offsets are usable.quest/m2/ts-shared-read.md, rewritten: share aCopyaggregate when nothing in the crate writes into that type in place, which removes fix(ts): copy Copy aggregates by type #83'sself.formcopy, fix(ts): copy what a container stores #99'scopy$Form, and map-key copies, in place of the narrow held-back-read fix.quest/m2/ts-bench-aa.md: the bench's A/A measures only noise, in JSC and Firefox too.quest/m1/ffi-generic.md: export a generic type at the instances the config names (about 67 partly covered moq-ffi items).Consumer: moq, through the translator's census and the exporter's coverage of moq-ffi.
Public API: none.
(written by Opus 5.5)
🤖 Generated with Claude Code
https://claude.ai/code/session_0115AT2ue6x9cT5e5J9SZtCn