perf(codegen): inline trusted boxed capture access; reject legacy numerics in strict eval

[proggeramlug]

Lands #8702 and #8698, plus a stale-test fix.

#8702 — inline trusted boxed capture access

Loads compiler-validated raw box capture pointers once at entry to private exact-arrow clones and reads/writes the box cell directly, keeping public and indirect closure bodies on the checked runtime accessors.

I verified the safety argument rather than taking it on trust, since caching a raw heap pointer across a collection is precisely how this codebase has been bitten before. The claim holds: box cells are malloc-side, not GC-heap — closure/box_captures.rs opens with "Lifetime bridge between GC closures and malloc-side async box cells" — so they genuinely do not move under evacuation. The crate already carries closure_box_captures_owner_moved to rekey when the closure relocates, which matches the PR's "retains each non-moving box cell for the invocation even if the closure relocates". TDZ keeps its TAG_TDZ test plus the trusted getter on the cold path, and direct stores keep the child-shading barrier.

#8698 — reject legacy numerics in strict eval

Re-lexes constant direct-eval bodies under inherited strict mode and surfaces the deferred legacy decimal/octal diagnostics as SyntaxError, leaving modern octal literals and numeric-looking text in comments/strings valid. The PR carried no changelog fragment and no skip-changelog label, so I added one rather than block on it.

Stale test fixed — red on main since #8653

typed_feedback_guards_direct_class_method_specialization asserted call double @js_class_field_add. #8653 deliberately stopped emitting that for this shape when it restored the guarded field-init fast path and reverted the #8648 3.11× shapes regression — so the test has been red since, and invisible to per-PR CI because codegen integration suites only run when the diff names them. #8702's author flagged it as pre-existing; I confirmed that independently on clean main before touching it.

The assertion now pins the actual emitted shape rather than the old helper:

assert!(ir.contains("js_typed_feedback_class_field_set_guard"));
assert!(ir.contains("class_field_set.fast"));
assert!(ir.contains("class_field_set.fallback"));
assert!(ir.contains("call void @js_class_field_set_fallback"));
assert!(!ir.contains("call double @js_class_field_add"));

js_class_field_set_fallback on the slow arm is what preserves DefineField semantics (an inherited setter must not run), so the negative assertion is safe.

Validation

• All 30 lint-job checkers pass
• perry-runtime --lib (RUST_TEST_THREADS=1): 2655 passed, 0 failed
• perry-codegen --lib: 1214 passed, 0 failed
• perry-codegen --tests (all 28 integration suites): 0 failures
• perry-hir --lib: 331 passed, 0 failed

An intermediate run showed 2 GC failures in gc::tests::handle_bound_method_name::*. They are not from this stack: I reproduced them on clean main and isolated the cause to forcing CARGO_INCREMENTAL=0 onto the perry-dev profile, which declares incremental = true — same commit gives 2655/0 with incremental on and 2 failures with it off. CI's actual invocation (CARGO_INCREMENTAL=0, default test profile) gives 2655/0, so CI is unaffected. Worth knowing those two tests compare literal addresses and are therefore build-config sensitive.

No version bump.

Summary by CodeRabbit

• Bug Fixes

  • Strict-mode direct eval now correctly raises SyntaxError for legacy decimal and octal numeric literals.
  • Numeric-looking text inside comments and strings, along with valid modern numeric forms, continues to work correctly.
  • Class-field initialization now follows the correct guarded path, improving behavior across supported execution scenarios.

• Performance

  • Improved execution efficiency for certain closures that access captured boxed values, while preserving safety checks and expected runtime behavior.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 882142f8-73e2-4f79-8a1d-c5694669e2e2

📥 Commits

Reviewing files that changed from the base of the PR and between 8224d87 and 74b3587.

📒 Files selected for processing (12)

• changelog.d/8698-strict-eval-legacy-numerics.md
• changelog.d/8702-inline-trusted-box-capture-access.md
• crates/perry-codegen/src/codegen/closure.rs
• crates/perry-codegen/src/codegen/entry.rs
• crates/perry-codegen/src/codegen/function.rs
• crates/perry-codegen/src/codegen/method.rs
• crates/perry-codegen/src/codegen/trusted_box_callback_tests.rs
• crates/perry-codegen/src/expr/literals_vars.rs
• crates/perry-codegen/src/expr/mod.rs
• crates/perry-codegen/src/target_layout.rs
• crates/perry-codegen/tests/typed_feedback.rs
• crates/perry-hir/src/lower/const_fold_fn.rs

---

📝 Walkthrough

Walkthrough

The PR adds trusted boxed-capture pointer access across closure code generation, adds strict direct-eval validation for legacy numeric literals, updates related changelog entries, and changes a typed-feedback test to expect guarded class-field initialization.

Changes

Trusted boxed-capture access

Layer / File(s)	Summary
Capture pointer contract and target layout crates/perry-codegen/src/expr/mod.rs, crates/perry-codegen/src/target_layout.rs	FnCtx stores trusted capture pointers. Closure header sizing now supports LP64 and ILP32 targets with coverage for supported target triples.
Trusted closure pointer loading crates/perry-codegen/src/codegen/closure.rs, crates/perry-codegen/src/codegen/entry.rs, crates/perry-codegen/src/codegen/function.rs, crates/perry-codegen/src/codegen/method.rs	Trusted closures load boxed-capture pointers from closure storage and pass them into FnCtx. Other context constructors initialize an empty map.
Direct capture operations and validation crates/perry-codegen/src/expr/literals_vars.rs, crates/perry-codegen/src/codegen/trusted_box_callback_tests.rs, changelog.d/8702-inline-trusted-box-capture-access.md	Trusted capture reads, writes, and updates use cached pointers, retain TDZ fallback handling, and emit box-parent write barriers. Tests verify the generated pointer access and helper usage.

Strict direct-eval numeric validation

Layer / File(s)	Summary
Strict eval diagnostic and tests crates/perry-hir/src/lower/const_fold_fn.rs, changelog.d/8698-strict-eval-legacy-numerics.md	Constant strict direct-eval folding reparses source with a strict directive, detects legacy numeric diagnostics, and synthesizes a runtime SyntaxError. Tests cover rejected legacy literals and accepted modern, string, and comment forms.

Typed-feedback expectation

Layer / File(s)	Summary
Guarded class-field feedback assertion crates/perry-codegen/tests/typed_feedback.rs	The test now expects guarded class-field initialization with fast and fallback paths, and rejects the unconditional js_class_field_add call.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Suggested reviewers: thehypnoo, jdalton

Sequence Diagram(s)

sequenceDiagram
  participant TrustedClosure as Trusted closure codegen
  participant FnCtx
  participant CaptureBody as Capture body lowering
  TrustedClosure->>TrustedClosure: Load boxed-capture pointers from closure storage
  TrustedClosure->>FnCtx: Pass trusted_box_capture_ptrs
  CaptureBody->>FnCtx: Read cached capture pointer
  CaptureBody->>CaptureBody: Load or store capture cell

Loading

sequenceDiagram
  participant ConstFoldDirectEval
  participant StrictEvalDiagnostic
  participant StrictParser
  ConstFoldDirectEval->>StrictEvalDiagnostic: Check constant eval source
  StrictEvalDiagnostic->>StrictParser: Reparse source with use strict
  StrictParser-->>StrictEvalDiagnostic: Return legacy numeric diagnostic
  StrictEvalDiagnostic-->>ConstFoldDirectEval: Report diagnostic result
  ConstFoldDirectEval->>ConstFoldDirectEval: Synthesize runtime SyntaxError

Loading

✨ 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 merge/b14

---

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.}