Skip to content

chore(ffi): the exporter pins moq on its own - #736

Merged
moq-bot[bot] merged 7 commits into
mainfrom
quest/m1/ffi-refusals/own-pin
Nov 14, 2026
Merged

moq-bot[bot] merged 7 commits into
mainfrom
quest/m1/ffi-refusals/own-pin

Conversation

@kixelated

@kixelated kixelated commented Nov 13, 2026 •

Copy link
Copy Markdown
Collaborator

Problem

The exporter's moq_rev read rust2-ts's pin from ts/fixtures/model/Cargo.toml, so moving the exporter's moq pin also moved the translator's. The moq pin bump (quest/m1/ffi-refusals/moq-pin.md) then waited on rind building for wasm32, which only the translator needs.

Approach

  • ffi/justfile holds its own 40-character revision, ffcb7cd9748422daa30e935ec464108b5a8ebb71, which is today's shared one, so the exported code is unchanged.
  • ts/justfile keeps its pin in ts/fixtures/model/Cargo.toml. Its comment no longer says the exporter follows it, and the manifest's comment no longer names just ffi.
  • quest/m1/ffi-refusals/own-pin.md is finished, so it is deleted with its links. The ffi-refusals README decision line drops its link.
  • The maintainer split the bump on 2026-11-12 (see Decisions). moq-pin.md is now the exporter's bump only: its Plan bumps ffi/justfile's pin, and its rind Waiting item is gone. Its moq-tokio Waiting item and its Required quests stay.
  • rust2-ts's bump is a new quest, quest/m2/ts-model/moq-pin.md (see Follow-ups), which carries the rind 0.4 wait as written.

Which pin each benchmark reads:

  • just ffi bench and the other just ffi exports (moq, moq-cpp, moq-swift, census) read the exporter's pin, through moq-export → standin → moq-pin.
  • go/justfile's bench reads no moq pin: its codec fixture is vendored in ts/fixtures/codec.
  • bench/build has its own corpus revision in bench/build/corpus.toml, which this PR does not touch.

Impact

Public API: none. Wire: none. No moq revision moves.

Alternatives

  • One pin for both, rejected by the maintainer on 2026-11-12. It keeps the exporter waiting on rind.
  • A shared revision file that both justfiles read. Rejected: it couples the two pins again.

Validation

  • quest check and quest check --base origin/main --branch quest/m1/ffi-refusals/own-pin: 62 documents ok, on af87f11ac, after merging main (68ef19591).
  • quest gate nix develop --command just ci origin/main on af87f11ac: exit 0. ci-select ran lint, quest, ffi, and ts.
  • just --evaluate moq_rev in ffi/ and in ts/ both print ffcb7cd9748422daa30e935ec464108b5a8ebb71, the exporter's and the model fixture's current revs.
  • just ffi moq-pin checks out target/moq at ffcb7cd9, and HEAD is verified.
  • Not run: just ffi moq, just ffi census, just ffi bench. They are CPU-heavy exports, and the revision is unchanged, so their inputs are unchanged.

Choices

  • The revision is a literal in ffi/justfile, not read from another file. No other recipe reads it, so the literal is the simplest form.
  • The new follow-up is sized XS: one manifest revision and its lock, regenerated expected output, and the census count. It has no Required, since just ts moq needs only rind.
  • Two comments the split made stale are fixed here: ts/fixtures/model/Cargo.toml said just ffi checks the model's pin out, and ts/README.md said "too" after the just ts recipes.
  • quest/m1/ffi-export/README.md now says the exporter's pin bump waited on rind, since it no longer does.
  • The superseded 2026-11-07 "one bump for both" decision stays in moq-pin.md, marked superseded, as the record.
  • kixel-ai/rust2#726 landed on main before this PR's merge, re-linking the rind item to kixel-ai. Merging main conflicted on that item, which this PR moves out of moq-pin.md. The conflict is resolved by deleting the item here, and the follow-up quest carries it with kixel-ai links, as docs: point rind links at kixel-ai #726's change moves with it.
  • Pullfrog's review thread: the #390 link in the follow-up quest pointed at moq-dev/rust2. It now points at kixel-ai/rust2, as the exporter's quest does.

Decisions

  • 2026-11-12, maintainer: ✅ split the moq pin bump: the exporter's pin stays in moq-pin.md, and rust2-ts's pin moves to its own quest. Rejected: keep one bump for both pins, which keeps the exporter waiting on rind.

Follow-ups

  • m2/ts-model/moq-pin - bumps rust2-ts's moq pin in ts/fixtures/model/Cargo.toml once moq's main passes just ts moq, which waits on rind 0.4 (moq-dev/moq#8432). No open decisions. Kept by the maintainer on 2026-11-12, in the round that picked this PR to merge.

Models

Size: XS
Implementer: Claude, claude-haiku-5-5, unknown effort
Implementer: Claude Code, claude-haiku-5-5, unknown effort (iterate, XS)

🤖 Generated with Claude Code

Review

Rewritten by quest review after each round.

  • Head: af87f11

  • Base: origin/main at fa311d1

  • Reviewer: Codex, gpt-6.1-sol, medium effort

  • Session: 01a1d652-8ed5-7a32-8191-9cfdd5c99851, round 2

  • Score: 10/10, The only new branch change corrects a documentation link, and no actionable findings remain at the head after excluding changes inherited from main.

  • Findings: none

Codex stands in for the local second reviewer, as the maintainer set on 2026-11-13: the round ran quest review --harness codex alone, since the OpenCode Go account behind opencode-go/deepseek-v4.1-flash is out of credit. Pullfrog's online review replaces that second reviewer.

Summary by CodeRabbit

  • Documentation
    • Clarified that the exporter and model fixture use separate moq revisions and have independent update plans.
    • Updated guidance on which commands update the model fixture and its census data.
    • Revised planning notes to distinguish the exporter’s outstanding checks from the model fixture’s wasm32 build blocker.
  • Chores
    • Set the exporter’s moq revision independently of the model fixture’s revision.

Summary by CodeRabbit

  • Chores
    • The exporter now uses its own fixed moq revision, separate from the model fixture’s revision.
  • Documentation
    • Clarified which commands use the model fixture’s pinned revision and which use the exporter’s pin.
    • Updated planning notes to track the exporter and model pin updates separately, including their distinct build dependencies and release sequencing.

@coderabbitai

coderabbitai Bot commented Nov 13, 2026 •

Copy link
Copy Markdown

Explain Change →

No actionable comments were generated in the recent review.

Walkthrough

The exporter now uses a fixed moq revision independent of the model fixture. Quest documentation separates the exporter and model-fixture pin updates. The model-fixture pin update remains planned; this diff does not change its pin.

Changes: Separate moq pin tracks

Layer / File(s) Summary
Exporter pin and quest
ffi/justfile, quest/m1/ffi-export/README.md, quest/m1/ffi-refusals/*
The exporter now uses a fixed moq revision. Its quest records the split from the model-fixture pin, revises the stated blockers, and removes the own-pin document and prerequisite. The exporter README changes the rind release wording to past tense.
Model fixture pin quest
quest/m2/ts-model/*, ts/justfile, ts/fixtures/model/Cargo.toml, ts/README.md
A new quest documents the planned model-fixture pin update, its census updates, and rind wasm32 build dependency. Comments distinguish the model-fixture revision from the exporter’s separate pin. The model-fixture pin is unchanged.

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Other

Merge Risk: Minimal · up to af87f

This change decouples the exporter's moq pin from the model fixture's pin and updates the related quest documentation. Both pins currently point to the same revision, so there is no expected behavior change. Merge risk is minimal.

Architecture risk: Low · up to f2340

The change affects 3 systems.

Changed systems: quest, ts, ffi

Architecture concerns
No architecture-level concerns identified.

Review details

Systems and components

  • observed — quest (service) was modified; 6 changed files map to changed impact.
  • observed — ts (service) was modified; 3 changed files map to changed impact.
  • observed — ffi (service) was modified; 1 changed file maps to changed impact.

Before / after behavior

  • observed — Modified behavior in ffi/justfile: The moq_rev pin no longer reads from just ts; it is set to a fixed revision. The comments now identify it as independent of the model fixture’s pin.
  • observed — Modified behavior in quest/m1/ffi-refusals/own-pin.md: Removed the document describing the proposed separate moq pins for the exporter and rust2-ts, including the existing pin source and the proposed relocation of the wasm32 build wait.
  • observed — Modified behavior in quest/m2/ts-model/README.md: The Plan adds a decision record separating the model fixture’s moq pin bump into its own quest, assigning that quest responsibility for updating the epic’s census count, and rejecting a combined bump for both pins.
  • observed — Modified behavior in quest/m2/ts-model/moq-pin.md: Adds a goal stating that the model fixture’s moq pin and the revision read by ts/justfile move to moq’s main, separately from the exporter’s pin. It describes the expected census and model-output updates and states that the translator’s output changes only where moq’s API does.
Pre-merge checks: Passed 4

Check name Status Explanation
Description Check Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check Passed The title "chore(ffi): the exporter pins moq on its own" matches the main change. The PR sets a fixed moq revision for the exporter in ffi/justfile, separate from the rust2-ts model fixture pin. The t…
Linked Issues check Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check Passed Check skipped because no linked issues were found for this pull request.
Full details: Title check

Explanation

The title "chore(ffi): the exporter pins moq on its own" matches the main change. The PR sets a fixed moq revision for the exporter in ffi/justfile, separate from the rust2-ts model fixture pin. The title is specific and concise, and it uses the conventional commit prefix. The title does not mention the documentation and quest updates, which is acceptable because they are secondary.

Finishing Touches

Generate unit tests (beta)

  • Commit to this branch
  • Create a new PR
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Usage-based review receipt

  • Mode: Continue automatically
  • Reviewed files: 1
  • Waived: $0.25 (charged $0.00)
  • View usage details

Note

This review exceeded your plan’s limits and used usage-based reviews—free during trial, billed after paid activation unless disabled. Manage usage-based reviews.

Recent review info

Run configuration

  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 5387538f-24d4-42fc-b31e-b53b69643876
Commits

Reviewing files that changed from the base of the PR and between f234091 and af87f11.

Files selected for processing: 4

  • quest/m1/ffi-export/README.md
  • quest/m1/ffi-refusals/moq-pin.md
  • quest/m2/ts-model/moq-pin.md
  • ts/README.md
Files with no reviewable changes: 1

  • quest/m1/ffi-refusals/moq-pin.md
Files skipped from review as they are similar to previous changes: 2

  • quest/m1/ffi-export/README.md
  • ts/README.md

Limit details: You’ve used all 10 included reviews currently available.

The exporter's `moq_rev` read rust2-ts's pin, so its moq bump waited on rind
building for wasm32, which only the translator needs. It now holds the
revision itself, starting at the same one, so nothing else moves.

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
@kixelated kixelated changed the title The exporter pins moq on its own chore(ffi): the exporter pins moq on its own Nov 13, 2026
kixelated and others added 2 commits November 12, 2026 17:44
The maintainer chose the split on 2026-11-12: the exporter's pin bumps in
moq-pin.md, and rust2-ts's pin bumps in m2/ts-model/moq-pin.md, which carries
the rind 0.4 wait. The exporter's bump no longer waits on rind.

Also drops the stale 'just ffi' from the model manifest's pin comment.

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
The new ts-model quest ended with a stray closing tag. The exporter's bump
still said its pin waited on 'just ts moq', which is the translator's wait.

Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
@kixelated
kixelated marked this pull request as ready for review November 13, 2026 02:04
@kixelated
kixelated enabled auto-merge November 13, 2026 02:20
@kixelated

Copy link
Copy Markdown
Collaborator Author

@pullfrog review this PR. It was opened or marked ready during the 01:10Z to 04:25Z window when no Pullfrog runs fired, so it never got a review.

(Written by Claude Opus 5.5 with high effort)

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ℹ️ No critical issues. One link nit is inline.

Reviewed changes

I reviewed the split of the moq pin: the exporter now pins its own revision, and rust2-ts's bump moves to its own quest.

  • ffi/justfile pins moq itself: moq_rev is now the literal ffcb7cd9748422daa30e935ec464108b5a8ebb71, the same revision as the model fixture's rev, so the exported code is unchanged. moq-pin is the only recipe that reads it, and it already uses its own target/moq checkout.
  • Comments in ts/ are updated: ts/justfile, ts/fixtures/model/Cargo.toml, and ts/README.md no longer say just ffi follows the model fixture's pin. No other doc in ffi/, ts/, go/, or bench/ still says it does.
  • The exporter's bump quest is narrowed: quest/m1/ffi-refusals/moq-pin.md now bumps only ffi/justfile. Its rind Waiting item and its own-pin requirement are gone, and the 2026-11-07 decision is marked superseded.
  • A new quest covers rust2-ts's bump: quest/m2/ts-model/moq-pin.md takes the rind 0.4 wait as it was written, and the ts-model epic records the 2026-11-12 decision.
  • The finished quest is deleted: own-pin.md is removed, and nothing in the tree still links to it.

Pullfrog  | Fix all ➔ | Fix 👍s ➔ | View workflow run | Using claude-opus-5-5 | 𝕏

Comment thread quest/m2/ts-model/moq-pin.md Resolved
kixelated and others added 2 commits November 13, 2026 15:24
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

✅ No new issues found.

Reviewed changes

I reviewed what changed since the prior Pullfrog review: two documentation-only commits.

  • Corrected links in quest/m2/ts-model/moq-pin.md: the #390 link and the rind links now point at kixel-ai, which resolves the earlier thread.
  • Matched the rind links in quest/m1/ffi-refusals/moq-pin.md: its rind #543 and #511 links also use kixel-ai.
  • Dropped a link in quest/m1/ffi-export/README.md: the async-start.md link is gone, which came in with the merge of main.

Pullfrog  | View workflow run | Using claude-sonnet-5-5 | 𝕏

@moq-bot
moq-bot Bot added this pull request to the merge queue Nov 14, 2026
Merged via the queue into main with commit a7db266 Nov 14, 2026
6 checks passed
@moq-bot
moq-bot Bot deleted the quest/m1/ffi-refusals/own-pin branch November 14, 2026 01:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant