fix(runtime): preserve declared receiver length semantics

[proggeramlug]

Closes #7853.

What changed

• add a property-semantic js_value_length_property_f64 sibling for guarded source-level .length reads
• preserve undefined and non-numeric length values instead of collapsing them to zero
• throw the ordinary catchable property-access TypeError for nullish receivers
• keep array-internal js_value_length_f64 behavior unchanged
• delegate fallback lookup through the normal dynamic property path so functions, typed arrays, native handles, accessors, inherited properties, and proxies keep their existing dispatch

Coverage

The parity regression passes real arrays, numbers, strings, numeric/string-valued array-like objects, functions, typed arrays, null, and undefined through an explicitly declared string[] receiver. Its output is byte-identical to Node.

Scope and performance

The inline Array/String layout path is unchanged; only its existing guarded slow arm calls the new helper. A quiet-M1-mini A/B used two compilers against one fixed runtime, seven alternating runs per cell: the honest-array length + loop was 0.21s vs 0.21s and length * was 0.22s vs 0.22s.

I also tested broadening the declared-number proof policy for every arithmetic use of .length; that changed the hot path and regressed those microbenchmarks substantially, so it is deliberately not part of this property-read fix. #7854 can remove its declared_only_array_locals exclusion once this fallback is available.

Tests

• direct Node/Perry parity comparison for test_gap_7853_declared_array_length_runtime_value.ts
• RUST_TEST_THREADS=1 cargo test -p perry-runtime --lib length_handle_band_tests (3 passed)
• cargo test -p perry-codegen --lib (877 passed)
• cargo fmt --all -- --check
• bash scripts/check_file_size.sh

Summary by CodeRabbit

• Bug Fixes

  • Fixed .length property reads to follow standard JavaScript behavior.
  • Missing or non-numeric length properties now remain undefined or retain their original value instead of becoming 0.
  • Accessing .length on null or undefined now throws a catchable TypeError.
  • Preserved expected behavior for arrays, strings, functions, typed arrays, and array-like objects.

• Tests

  • Added regression coverage for declared array and string length access across supported value types.

[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: cc127ce7-f631-40b3-a979-2b92dbef6d19

📥 Commits

Reviewing files that changed from the base of the PR and between 06ab783 and 1c3b7fc.

📒 Files selected for processing (7)

• changelog.d/7862-declared-array-length-fallback.md
• crates/perry-codegen/src/expr/property_get.rs
• crates/perry-codegen/src/gc_call_effects.rs
• crates/perry-codegen/src/runtime_decls/arrays.rs
• crates/perry-runtime/src/value/dynamic_object.rs
• crates/perry-runtime/src/value/mod.rs
• test-files/test_gap_7853_declared_array_length_runtime_value.ts

---

📝 Walkthrough

Walkthrough

The change adds property-semantic .length handling for declared array, string, and named receivers. Runtime code now preserves missing and non-numeric values and throws catchable TypeError exceptions for nullish receivers. Code generation and regression tests use the new helper.

Changes

Length property semantics

Layer / File(s)	Summary
Runtime length property helper crates/perry-runtime/src/value/dynamic_object.rs, crates/perry-runtime/src/value/mod.rs, crates/perry-codegen/src/runtime_decls/arrays.rs	Adds and exports js_value_length_property_f64. The helper handles nullish values, strings, and generic property lookup without numeric coercion.
Codegen fallback integration crates/perry-codegen/src/expr/property_get.rs	Routes TypedArray, string, and guarded array/string .length fallbacks through the new helper.
Audit and regression coverage crates/perry-codegen/src/gc_call_effects.rs, test-files/test_gap_7853_declared_array_length_runtime_value.ts, changelog.d/7862-declared-array-length-fallback.md	Classifies the helper as a conservative getter path, tests runtime values and nullish receivers, and documents the corrected semantics.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

• PerryTS/perry#6817: Both changes modify runtime .length handling, but this change addresses property-semantic fallback behavior.
• PerryTS/perry#7430: Both changes update specialized property-read fallbacks for nullish receivers.
• PerryTS/perry#7831: Both changes address code-generation assumptions about runtime values.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely describes the runtime fix for declared receiver length semantics.
Description check	✅ Passed	The description covers the change, related issue, scope, performance impact, and test commands, although it does not use the exact template headings.
Linked Issues check	✅ Passed	The changes satisfy issue [#7853] by preserving property semantics, throwing for nullish receivers, and leaving internal array length behavior unchanged.
Out of Scope Changes check	✅ Passed	The changelog, runtime, codegen, audit updates, and regression tests are all directly related to the declared-receiver length fix.

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/7853-declared-array-length

---

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.}