feat(rust): #259 slice 3 — own-bridge, OwnIR facts → Layer 2 lowering - #301
Merged
Conversation
…lay, mechanism pins
New crate own-bridge (deps: own-ir, own-lowered — both data leaves; the DAG
map widened deliberately). Production API is a pure transformation:
OwnIr -> Result<LoweredDocument, BridgeError>; no filesystem/CLI/analysis.
Red now, green next:
* tests/replay.rs — the acceptance contract: every rust_replay manifest
case lowered FROM FACTS (golden is expected output only, never an input)
must be byte-identical to the Python golden, twice (determinism); plus
independent LOWERED_VERSION, ledger/tree equality, and the exact
{tolerant_unknown_kind} Python-only set.
* tests/mechanisms.rs — metamorphic/negative pins the plain replay cannot
prove: record order drives the global mint counters; kill-on-rebind
removes the old mapping; a sig-carrying overload call channels $consume
WITHOUT unmapping; a sig-less merged-may call discharges at the kill
site and unmaps; each hoisting negative gate (nested depth, while body,
early return) blocks the hoist while the positive control hoists; an
unknown flow op fails loud with Python's exact rejection text.
The skeleton lowers everything to an empty document, so all 10 tests fail.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MdYFKUaygHz1T9H1qJ7BqK
…m facts The lowering port, restricted to the frozen semantics the shared fixtures exercise. own_bridge::lower(&OwnIr) re-serializes the typed document to a JSON value once and walks it with the same key-by-key logic as Python's to_module — every membership/default/truthiness decision stays textually comparable to the reference: * routing R1–R6 (unresolved skip, self-subscribe skip, static capture with released mitigation, returned_fresh provenance, DI-typed capture with region derivation, token kinds with released Release); * global sub_/cap_ minting and the shared parg_/loc_ counter; capture params at line 0 under their source region; lifetimes emitted iff any capture; DI registrations via the services graph; * MOS: skeleton building (explicit effects, definite/partial release, single-unconditional-forward rule, early-return guard, fresh/forward return inference with the Tier A/Tier B boundary and the BCL factory table), overload merge at (key, index) granularity + per-sig overload keys, SCC-condensation fixpoint solve, cycle-safe return chase; * flow lowering: localmap, kill-on-rebind, alias_join (old binding dies first), unmapped-reference silence, hoisting with all four gates, the $consume/$borrow channel (overloads and may/unknown contracts), the top-level kill-site discharge + unmap vs the in-branch whole-body untrack, fresh-result minting, and the fail-loud vocabulary error with Python's exact message text (kill sites keyed by node identity in the value tree — the id(n) analog). One loud deliberate divergence: a present-but-unknown resource kind is a BridgeError, not Python's silent Subscription fallback — the tolerant-door contract is #294 and this crate refuses to pre-empt it in either direction (tolerant_unknown_kind stays Python-only, asserted in the replay). All 26 rust_replay cases lower byte-identically from facts (golden used only as expected output), twice (determinism); all 9 mechanism pins green; clippy 0 warnings; full workspace + python suites green. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MdYFKUaygHz1T9H1qJ7BqK
…alidation The two follow-ups recorded on PR #300's review threads (deferred to the next ledger change, which this slice is): * own-lowered's minimum-coverage floor rises 25 -> 26 to cover the handles_null_metadata fixture (exact ledger integrity stays with the manifest == facts == goldens equality; the floor is the shrink guard); * manifest 'rules' must be a non-empty array of non-empty strings — enforced in the Python harness (verified red on violation) and in the own-lowered typed replay. Kept separate from the lowering commits so the own-bridge diff stays readable. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MdYFKUaygHz1T9H1qJ7BqK
📝 WalkthroughWalkthroughThe PR adds the Rust ChangesOwnIR bridge and Layer 2 parity
Estimated code review effort: 4 (Complex) | ~60 minutes Sequence Diagram(s)sequenceDiagram
participant OwnIR
participant own_bridge
participant MOS
participant own_lowered
participant GoldenFixtures
OwnIR->>own_bridge: parse and lower facts
own_bridge->>MOS: solve ownership summaries
MOS-->>own_bridge: transfer and return contracts
own_bridge->>own_lowered: emit LoweredDocument
own_lowered->>GoldenFixtures: canonicalize output
GoldenFixtures-->>own_lowered: compare byte-exact golden
Possibly related issues
Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 4✅ Passed checks (4 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
The normative documents described the pre-slice-3 world. Recorded now: * spec/Bridge.md §6 Layer 2: #299 (Python emitter + fixture foundation), #300 (typed own-lowered surface/emitter, presence-aware metadata, per-document version gate), #301 (own-bridge constructs Layer 2 from facts and reproduces all 26 shared goldens byte-exact; golden is expected output only); tolerant_unknown_kind stays Python-only under #294 (the Rust bridge fails loud on an unknown kind instead of adopting the tolerant fallback); Layer 1, Layer 3, analysis wiring and #259 as a whole remain open. * spec/BridgeBehaviorMatrix.md: the Layer 2 summary paragraph updated to the same completed state. * tests/test_lowered_fixtures.py docstring: future tense ("will replay", "until the Rust emitter exists") moved to the present. * BR-L7 wording nit folded in (recorded on #297): condition (4) guards the post-merge reference/discharge, not only a release — the use-only hoist fixture is in the gate too. Docs-only; no Rust code, facts, manifest or golden changes. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MdYFKUaygHz1T9H1qJ7BqK
Owner
Author
|
Ревизия по review (REQUEST CHANGES, docs-only) — применена, head
Локально: python suite + ruff зелёные, полный Rust workspace 115/0. Жду зелёный CI на Generated by Claude Code |
Owner
Author
|
Независимый review — финальный verdict на head
Merge авторизован reviewer'ом. Последовательность: ready → дождаться CodeRabbit post-draft review → перепроверка (exact head / mergeable clean / CI green / нет новых blocking threads) → merge commit без изменения head. Generated by Claude Code |
PhysShell
marked this pull request as ready for review
July 19, 2026 02:15
There was a problem hiding this comment.
🧹 Nitpick comments (1)
rust/crates/own-bridge/tests/replay.rs (1)
68-74: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winMissing
rulesnon-emptiness check (present in the sibling own-lowered validator).
own-lowered/tests/replay.rsandtests/test_lowered_fixtures.pyboth now assert each case'srulesis a non-empty array of non-empty strings, but this file — which explicitly claims to independently enforce manifest-contract properties — doesn't.Vec<String>doesn't guarantee non-emptiness at the type level, so an empty/blankrulesentry would silently pass here.♻️ Proposed fix
for case in &manifest.cases { assert!( listed.insert(case.name.clone()), "duplicate manifest case name: {}", case.name ); + assert!( + !case.rules.is_empty() && case.rules.iter().all(|r| !r.is_empty()), + "{}: 'rules' must be a non-empty array of non-empty strings", + case.name + ); }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@rust/crates/own-bridge/tests/replay.rs` around lines 68 - 74, Extend the manifest validation loop in the replay test to assert that every case’s rules collection is non-empty and that each rule is a non-empty string after trimming. Use the case’s existing rules field alongside the duplicate-name check, preserving the validator’s independent enforcement of manifest-contract properties.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@rust/crates/own-bridge/tests/replay.rs`:
- Around line 68-74: Extend the manifest validation loop in the replay test to
assert that every case’s rules collection is non-empty and that each rule is a
non-empty string after trimming. Use the case’s existing rules field alongside
the duplicate-name check, preserving the validator’s independent enforcement of
manifest-contract properties.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 08637902-cf51-411d-9471-fed54548ce9e
⛔ Files ignored due to path filters (1)
rust/Cargo.lockis excluded by!**/*.lock
📒 Files selected for processing (12)
rust/Cargo.tomlrust/crates/own-bridge/Cargo.tomlrust/crates/own-bridge/src/lib.rsrust/crates/own-bridge/src/lower.rsrust/crates/own-bridge/src/mos.rsrust/crates/own-bridge/tests/mechanisms.rsrust/crates/own-bridge/tests/replay.rsrust/crates/own-diagnostics/tests/dag.rsrust/crates/own-lowered/tests/replay.rsspec/Bridge.mdspec/BridgeBehaviorMatrix.mdtests/test_lowered_fixtures.py
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 878cfdf0ea
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".
13 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Что и зачем
Третий slice #259 в авторизованной границе: Rust lowering facts → Layer 2 — новый pure transformation crate
own-bridge(OwnIr -> Result<LoweredDocument, BridgeError>; без ФС, CLI, диагностики и analysis side effects — fixture I/O живёт только в integration tests). Портownlang/ownir.py::to_module, ограниченный семантикой, которую тренируют 26rust_replay: truefixtures:released-митигирование),returned_freshprovenance, DI-типизированный capture с регионами изservices, token-путь всех kind'ов creleased→Release;sub_/cap_и общий счётчикparg_/loc_; capture-параметры с line 0 под регионом источника; lifetimes ровно при any-capture;alias_join(старый binding умирает первым), тишина unmapped-ссылок, hoisting со всеми четырьмя negative gates, канал$consume/$borrow(overload'ы и may/unknown-контракты), top-level kill-site discharge+unmap против in-branch whole-body untrack, fresh-result minting, fail-loud vocabulary с точным текстом Python-ошибки (kill-sites по identity узла в value-дереве — аналогid(n)).Вход — типизированный
own_ir::OwnIr(acceptance-цепочкаfacts → own-ir parse → own-bridge → own-lowered emitter); внутри walk словарный, поверх однократной re-сериализации (round-trip — закреплённое свойство own-ir), чтобы каждое membership/default/truthiness-решение читалось против референса строка-в-строку.Одна громкая сознательная дивергенция: present-but-unknown resource kind —
BridgeError, а не тихий Python-fallback вSubscription: tolerant door — открытое решение #294, и crate отказывается его предвосхищать в любую сторону (tolerant_unknown_kindостаётся единственнымrust_replay: false, что утверждается тестом).DAG расширен сознательно:
own-bridge → {own-ir, own-lowered}— обе стрелки в data-leaves, обратные запрещены;own-loweredостаётся leaf.Отдельным маленьким коммитом — два ledger follow-up с review #300: floor
>= 26и валидацияrulesкак непустого массива непустых строк (Python harness, проверено красным + Rust replay). BR-L7 wording nit не взят — spec/Bridge.md в этом slice не трогается.История: red
ff58433(skeleton + падающие replay/mechanism тесты) → green6531d70(порт до 26/26) → choreb6465d0(ledger).Тип изменения
Как проверено
own-bridge/tests/replay.rs— acceptance contract: для каждого manifest-кейсаrust_replay: trueцепочкаfacts.json → OwnIr::from_json → lower → to_canonical_json== golden byte-for-byte, 26/26, дважды (детерминизм); golden — только expected output, никогда не вход; LOWERED_VERSION == 1, exact ledger/tree equality, skip-set == ровно{tolerant_unknown_kind}— всё утверждается самим Rust-тестомown-bridge/tests/mechanisms.rs— метаморфические/негативные пины: порядок records предсказуемо меняет глобальный mint; kill-on-rebind убирает старый mapping; precise sig-канал$consumeНЕ unmap'ит (use после канала жив), merged-may — unmap'ит (release молчит); каждый hoisting negative gate блокирует hoist при зелёном positive control; unknown flow op — fail-loud с точным Python-текстомff58433)cargo fmt --check,cargo clippy --workspace --all-targets— 0 warnings; полный workspace — 115 passed / 0 failedpython tests/run_tests.py,ruff— зелёные;ownir.py/to_moduleне изменялись; fixtures/goldens не изменялись (кроме нуля — ledger diff только в harness-валидации)rules→ FAIL harness)Связанные issue
Refs #259 (slice 3 — НЕ закрывает issue), #258, #294 (не решается — unknown-kind теперь громкий отказ в Rust, поведение не зафиксировано), #296 (rules-валидация закрыта здесь как ledger follow-up), #250.
Чеклист
Draft — не мержить, не self-accept. Гейт: независимый review — подтверждение, что (1) порт семантически точен против
to_module(routing/minting/MOS/flow), (2) golden нигде не используется как вход построения, (3) MOS не расширен и контракт не изменён, (4) дивергенция по unknown-kind не решает #294, (5) DAG-запись обоснована и own-lowered остался leaf, (6) ledger-коммит не маскирует lowering diff.🤖 Generated with Claude Code
https://claude.ai/code/session_01MdYFKUaygHz1T9H1qJ7BqK
Generated by Claude Code
Summary by CodeRabbit
New Features
Tests
Documentation