feat: P-022 step 3 — port own-cfg (AST → CFG lowering) with exact CFG-JSON parity#203
Merged
Merged
Conversation
…-JSON parity
Third strangler-fig step of the P-022 Rust core migration. `own-cfg` lowers the
`own-syntax` AST to a control-flow graph and projects the canonical CFG-JSON
seam, mirroring `ownlang/{cfg,cfg_json,buffers}.py` line-for-line. Python stays
authoritative; this port introduces no behaviour change.
Crate (rust/crates/own-cfg, depends on own-syntax only, per the P-022 DAG — no
analysis/diagnostics edge):
* ir.rs — Kind/Symbol/Instr/Block/Cfg/Signature; symbol identity is a
newtype arena index (replacing Python's id(sym)), blocks are
threaded by BlockId so lowering never aliases a &mut.
* buffers.rs — buffers.py resolve/validate_policies/BufferInfo, value-for-value.
* builder.rs — the _Builder scope resolver + AST→CFG lowering + collect_*.
* json.rs — cfg_json.py projection plus a serializer that reproduces
json.dumps(indent=2, sort_keys=True, ensure_ascii=True)
byte-for-byte (the property the seam rests on).
Parity gate (mirrors the step-2 syntax fixtures):
* tests/test_cfg_fixtures.py (Python-authoritative) sweeps corpus/, examples/,
tests/fixtures/ plus curated cases, and freezes each input's canonical CFG
JSON + (line, code) diagnostics into tests/fixtures/cfg_parity.json; the
Python suite fails if it is stale.
* rust/crates/own-cfg/tests/parity.rs replays every case and asserts the
byte-identical CFG JSON, the identical diagnostics, and — for rejected
sources — the identical parser error text. 66 cases (60 corpus + 6 curated)
green; the Rust output also matches `python -m ownlang cfg --format json`
end-to-end.
Diagnostics carry code + line only at this step: the CFG-JSON seam does not
carry diagnostic message text (a verdict-layer contract pinned later by the
SARIF oracle), so the human message is deferred to the analysis/diagnostics
step, keeping this step's surface fully covered by its own oracle.
Closes #197
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Cg6DVXahkyY68Ruu7f76rG
📝 WalkthroughWalkthroughThis PR adds a new Changesown-cfg crate: CFG lowering with Python parity
Estimated code review effort: 4 (Complex) | ~75 minutes Sequence Diagram(s)sequenceDiagram
participant Caller
participant build_module
participant Builder
participant resolve
participant canonical_json
Caller->>build_module: build_module(module)
build_module->>Builder: build_cfg(fn_decl, resources, signatures, policies, kinds)
Builder->>resolve: resolve buffer intent options
resolve-->>Builder: BufferInfo, diagnostics
Builder-->>build_module: Cfg, diagnostics
build_module-->>Caller: Vec<Cfg>, Vec<Diag>
Caller->>canonical_json: canonical_json(cfgs)
canonical_json-->>Caller: byte-canonical JSON string
🚥 Pre-merge checks | ✅ 4✅ Passed checks (4 passed)
✨ Finishing Touches🧪 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 |
![chatgpt-codex-connector[bot]](https://cdn.github.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: 9334faa4ad
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".
| // Python dict iterates in insertion order; a HashMap does not, so sort to | ||
| // keep the diagnostic stream stable.) | ||
| let mut names: Vec<&String> = policies.keys().collect(); | ||
| names.sort(); |
There was a problem hiding this comment.
When a caller validates multiple bad policy blocks whose names are not already sorted (for example policy Z before policy A), this sort emits the later A diagnostic before the earlier Z one. Python's authoritative validate_policies iterates policy declarations in insertion/source order, so the Rust port no longer matches the (line, code) diagnostic stream and users/fixtures see diagnostics out of source order; keep declaration order alongside the lookup map instead of sorting names.
Useful? React with 👍 / 👎.
…n order Python's `validate_policies` iterates `policies.values()` in dict insertion (source-declaration) order; the Rust port sorted policy names alphabetically for HashMap determinism, which reorders the (line, code) diagnostic stream when policy blocks aren't already alphabetized. Sort by declaration line instead — that IS the source order — so the port matches Python. Add a unit test with deliberately non-alphabetical policy names. Not observable on the CFG-JSON seam (validate_policies is a `check`-pipeline step, ported here for later use), but a divergence from Python all the same. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Cg6DVXahkyY68Ruu7f76rG
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 шаг strangler-fig миграции P-022: новый крейт
rust/crates/own-cfgпонижает AST изown-syntaxв CFG и проектирует канонический CFG-JSON seam, зеркаляownlang/{cfg,cfg_json,buffers}.pyстрока-в-строку. Питон остаётся эталоном — поведение не меняется; расхождение считается багом Rust-стороны. Крейт зависит только отown-syntax(без рёбер на analysis/diagnostics, как требует DAG P-022).Состав:
ir.rs—Kind/Symbol/Instr/Block/Cfg/Signature; идентичность символа выражена newtype-индексом арены (вместоid(sym)), блоки протянуты черезBlockId, чтобы понижение никогда не алиасило&mut.buffers.rs—resolve/validate_policies/BufferInfoизbuffers.py, значение-в-значение.builder.rs— резолвер областей_Builder+ понижение AST→CFG +collect_*.json.rs— проекцияcfg_json.pyплюс сериализатор, воспроизводящийjson.dumps(indent=2, sort_keys=True, ensure_ascii=True)байт-в-байт (на этом и держится байтовая идентичность seam).Гейт паритета (по образцу syntax-фикстур шага 2):
tests/test_cfg_fixtures.py(питон — эталон) прогоняетcorpus/,examples/,tests/fixtures/+ курируемые кейсы и замораживает канонический CFG-JSON и диагностики(line, code)вtests/fixtures/cfg_parity.json;own-cfg/tests/parity.rsреплеит каждый кейс и требует байт-идентичный CFG-JSON, те же диагностики и — для отвергнутых источников — тот же текст ошибки парсера. 66 кейсов (60 корпус + 6 курируемых) зелёные; вывод Rust также сходится сpython -m ownlang cfg --format jsonend-to-end.Диагностики на этом шаге несут только
code + line: CFG-JSON seam не содержит текста сообщений (это контракт verdict-слоя, закрепляемый позже SARIF-оракулом), поэтому человекочитаемый текст отложен до шага analysis/diagnostics — при этом поверхность данного шага полностью покрыта собственным оракулом. PR не трогает схему OwnIR, так что отложенный OwnIR-rename сюда не подмешивается.Тип изменения
Как проверено
python tests/run_tests.py(включая новыеtest_cfg_fixtures.py, 66 кейсов в синхроне)ruff check .иmypy— чистоcd rust && cargo fmt --check && cargo clippy --all-targets && cargo test— 0 warnings, все тесты зелёныеpython scripts/oracle_exact.py --selftest— PASSСвязанные issue
Closes #197
Чеклист
feat:)🤖 Generated with Claude Code
https://claude.ai/code/session_01Cg6DVXahkyY68Ruu7f76rG
Generated by Claude Code
Summary by CodeRabbit
New Features
Tests