repsel: one safety gate and one write walker for both numeric proofs, plus the #7770 super-chain test and bench numbers

[proggeramlug]

Follow-up to #7774 (merged as 2e5bf44). These two commits were pushed to the PR branch but were not in the rebase that got merged, so they are absent from main; the branch was deleted on merge, so this restores them. No behavior change — a consolidation, a coverage gap, and two efficiency items, all from the code review of #7774.

1. One safety gate, not two (the load-bearing one)

prove_group_numeric_fields had grown its own copy of the 'cand loop's entire Pass-3 obligation set: ThisFlowAnalysis construction, ctor_chain_safe, prototype_is_stable, field/method name-ambiguity, per-method method_safe. That verdict is what licenses a bare unchecked load double, so the two copies silently drifting is a miscompile, not a missed optimization — and nothing diffed them against each other.

Extracted as chain_this_flow_verdict, now the single implementation both callers use. Tightening an obligation is a one-place change again.

2. The super-chain resolution path had no test at group scope

prove_numeric_fields resolves a parent constructor's parameters through super(...) argument lists under the caller's already-resolved environment. #7774 made that path reachable through the group MEET (multiple provenance news) for the first time, and every fixture in the new test file had extends: None — so a wrong index or an unresolved caller env there would have granted an unsound claim with zero red tests.

super_chain_params_resolve_under_the_group_meet covers both directions: all-numeric push sites prove the derived and the inherited field; one string at the super-feeding position drops exactly z group-wide while x survives.

3. One write walker, not three

collect_numeric_by_construction_locals had a hand-rolled full Stmt-tree write collector nearly identical to not_bigint_locals::collect_writes — the "two traversals that must agree, drifting by one Stmt variant" bug class that ptr_shape_elements.rs's own doc warns this file cannot afford. Generalized collect_writes to record a no-init Let as None (fine for the non-BigInt fixpoint, fatal for the numeric one) and deleted the duplicate.

4. Two efficiency items

• The group proof ran its full this-flow walk before discovering the class has no raw-f64-declared field anywhere on the chain (proven arrays of string/object-only records paid for a moot proof). Now bails first.
• group_members() rebuilt its map twice per collector run — once for the numeric proof, once for the pre-existing group-integrity pass. Computed once, passed to both.

5. The perf numbers for #7770

The merged fragment stops before the bench results. Restores them: on the pinned quiet mini (interleaved arms, best-of-15, two independent runs, outputs byte-verified against Node before timing) the issue's read loop goes 101/102 ms → 91/90 ms, while batch.ts, suite/04_array_read and suite/09_method_calls are unchanged. Includes the trap worth recording — benchmarking this needs the array and its read loop in ONE function, because an array crossing a function boundary is the #7766 shape #7770 does not address (both arms keep all 4 js_number_coerce sites, and the "no regression" is vacuous).

Validation

Rebased onto current main (after #7771 and #7765, which touch neighbouring element-shape code): cargo test -p perry-codegen --lib 850/850 green, cargo fmt --check and the 2000-line file-size gate clean.

Summary by CodeRabbit

• Performance

  • Improved numeric-field optimization, with an estimated 10–12% target-loop improvement in applicable cases.
  • Reduced unnecessary analysis while preserving byte-identical output where expected.

• Bug Fixes

  • Correctly handles uninitialized local bindings and values passed through superclass constructors.
  • Prevents invalid numeric promotion when incompatible values, such as strings, are introduced.

• Tests & Documentation

  • Added coverage for inherited constructors, numeric fields, validation behavior, and benchmark results.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 748f02a7-0bf7-4064-8bbd-3f88d02ce469

📥 Commits

Reviewing files that changed from the base of the PR and between 762d10f and 5852fa2.

📒 Files selected for processing (6)

• changelog.d/7774-element-group-numeric-proof.md
• changelog.d/7788-shared-this-flow-gate.md
• crates/perry-codegen/src/collectors/not_bigint_locals.rs
• crates/perry-codegen/src/collectors/ptr_shape.rs
• crates/perry-codegen/src/collectors/ptr_shape_group_numeric_tests.rs
• crates/perry-codegen/src/collectors/ptr_shape_numeric.rs

---

📝 Walkthrough

Walkthrough

The PR centralizes this-flow and write analysis, reuses element-group membership, improves group-wide numeric proofing, and adds super(...) inheritance coverage.

Changes

Element-group numeric proof

Layer / File(s)	Summary
Shared group and flow analysis crates/perry-codegen/src/collectors/ptr_shape.rs	Element-group membership is cached and reused. Constructor, prototype, method-resolution, and this-flow checks use chain_this_flow_verdict.
Group numeric proof and write tracking crates/perry-codegen/src/collectors/ptr_shape_numeric.rs, crates/perry-codegen/src/collectors/not_bigint_locals.rs	Group proofing uses explicit membership and skips groups without raw-f64 candidates. Numeric-local analysis reuses collect_writes, which now records uninitialized let bindings as None.
Inheritance validation and benchmark record crates/perry-codegen/src/collectors/ptr_shape_group_numeric_tests.rs, changelog.d/*.md	Tests cover numeric fields flowing through super(...). Changelogs record validation results, benchmark data, and unchanged cases.

Estimated code review effort: 4 (Complex) | ~45 minutes

Possibly related PRs

• PerryTS/perry#6860: Introduces the local BigInt analysis refined by this PR.
• PerryTS/perry#7669: Modifies related element-shape numeric proofing in perry-codegen.
• PerryTS/perry#7774: Provides the group-wide numeric proof infrastructure extended by this PR.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the shared safety gate, write walker, super-chain test, and benchmark updates.
Description check	✅ Passed	The description clearly covers the changes, related issue, validation commands, benchmark results, and no-behavior-change intent.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
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.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch repsel/7770-followup-shared-gate

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}