fix(runtime): #5834 — WeakMap/WeakSet constructor + instance-identity test262 fixes

[proggeramlug]

Summary

Fixes the built-ins/WeakMap (11) and built-ins/WeakSet (7) test262 failures from the #5834 worklist — a coherent single-root subcluster: WeakMap/WeakSet constructor semantics and instance identity.

• Constructor iterable handling: js_weakmap_init_iterable/js_weakset_init_iterable previously drained the whole iterable eagerly via a native helper and never actually read/called the set/add adder through a real property Get. Rewrote both to mirror js_map_from_iterable/js_set_from_iterable: fetch the adder only when iterable is non-null/undefined (matching spec step order, so a poisoned WeakMap.prototype.set accessor getter must not fire for new WeakMap()/new WeakMap(null)), honor an accessor descriptor on set/add via a new collection_iter::builtin_prototype_adder helper, and drive the iterable with lazy per-item stepping (iterator_next_value) so an abrupt Get/adder-call closes the iterator (IteratorClose) before rethrowing.
• Arity-gated dispatch bug: try_weak_method_dispatch's "add"/"set"/etc. arms were gated on args.len(), so an under-arity call like s.add() fell through to a no-op instead of running WeakSet.prototype.add's CanBeHeldWeakly check (which must throw TypeError for undefined). Now always dispatches, padding missing positions with undefined.
• Instance identity: Object.getPrototypeOf(new WeakMap()) and (new WeakMap()).constructor had no arm for WeakMap/WeakSet instances (a reserved, non-declared-class class_id), so both fell through to the receiver itself / undefined. Added arms in prototype.rs's collection_prototype closure and get_field_by_name_tail.rs's constructor special case.
• instanceof: x instanceof WeakMap/WeakSet had no runtime probe for real instances — this was a documented, deliberate gap in perry-codegen/src/expr/instance_misc1.rs ("no runtime probe for real instances yet"). Added a probe in js_instanceof (compile-time-literal path) plus the global_builtin_constructor_class_id counterpart for the dynamic x instanceof ctorVar form.

Test plan

• scripts/test262_subset.py --dir built-ins/WeakMap built-ins/WeakSet: 18/18 originally-failing cases now pass; 100% parity (158/158 judged), 0 regressions.
• cargo fmt --all -- --check clean.
• bash scripts/check_file_size.sh clean.
• cargo test --release -p perry-runtime -p perry-codegen: 1107 passed / 1 failed. The 1 failure (object::tests::builtin_prototype_methods_reject_dynamic_new, asserting Date.prototype.toJSON) is unrelated to this change (no Date/global_this code touched) and matches a known pre-existing test-isolation flake in this suite (shared global builtin-population state races under parallel test threads).

No version bump / CHANGELOG update, per the #5834 worklist's code-only-PR instructions.

Refs #5834.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved WeakMap/WeakSet handling for constructor, prototype, and instanceof checks, including correct resolution for internally-reserved class-id cases.
  • Corrected getPrototypeOf behavior so getPrototypeOf(new WeakMap())/getPrototypeOf(new WeakSet()) no longer falls back to returning the receiver itself.
  • Fixed WeakMap/WeakSet method dispatch and iterable initialization for missing arguments, lazy consumption, and proper iterator cleanup on abrupt failures.

[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: d8589464-7e5b-451b-80a9-632a955f23f7

📥 Commits

Reviewing files that changed from the base of the PR and between de67ad5 and 5de1979.

📒 Files selected for processing (1)

• crates/perry-runtime/src/weakref.rs

🚧 Files skipped from review as they are similar to previous changes (1)

• crates/perry-runtime/src/weakref.rs

---

📝 Walkthrough

Walkthrough

Runtime changes add reserved class-id handling for WeakMap/WeakSet across instanceof checks, constructor property resolution, prototype lookup, method dispatch, and iterable initialization. A new builtin_prototype_adder helper resolves accessor properties, and codegen comments are clarified.

Changes

WeakMap/WeakSet reserved-id handling

Layer / File(s)	Summary
Codegen comment clarification crates/perry-codegen/src/expr/instance_misc1.rs	Comments updated to state WeakMap/WeakSet use runtime probing for real instances, while DataView still relies on class-chain matching.
instanceof and constructor class-id mapping crates/perry-runtime/src/object/instanceof.rs, crates/perry-runtime/src/object/field_get_set/get_field_by_name_tail.rs, crates/perry-runtime/src/object/object_ops/prototype.rs	Adds reserved class-id handling for WeakMap/WeakSet in constructor-id lookup, instanceof probing, constructor property reads, and prototype resolution.
Builtin prototype adder and weak method dispatch crates/perry-runtime/src/collection_iter.rs, crates/perry-runtime/src/weakref.rs	Adds builtin_prototype_adder and updates weak method dispatch to pad missing arguments, constrain dispatch by receiver class id, and fall through for unsupported method names.
WeakMap and WeakSet iterable initialization crates/perry-runtime/src/weakref.rs	Rewrites WeakMap and WeakSet iterable initialization to use lazy constructor_iter consumption with iterator closing on abrupt completion.
Estimated code review effort: 4 (Complex)	~50 minutes

Possibly related PRs

• PerryTS/perry#5468: Also changes crates/perry-runtime/src/weakref.rs around WeakMap/WeakSet reserved ids and weak-collection behavior.
• PerryTS/perry#5667: Also modifies crates/perry-runtime/src/object/instanceof.rs and related instanceof resolution behavior.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly matches the PR's main theme: WeakMap/WeakSet constructor and instance-identity fixes for runtime behavior and test262.
Description check	✅ Passed	The summary and test plan are detailed and relevant, but the template's separate Changes/Checklist sections aren't filled out explicitly.
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 worktree-5834-test262-weakmap-weakset

---

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.}

[coderabbitai]

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

crates/perry-runtime/src/weakref.rs (1)

802-834: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

Gate weak-method dispatch by receiver class crates/perry-runtime/src/weakref.rs:802-833 — try_weak_method_dispatch still routes only on method_name, so a WeakMap receiver can hit "add"/a WeakSet receiver can hit "set" etc. and call the wrong helper instead of throwing the incompatible-receiver TypeError. Add a class/method pairing check before dispatching.

🤖 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 `@crates/perry-runtime/src/weakref.rs` around lines 802 - 834,
`try_weak_method_dispatch` is dispatching solely by `method_name`, which can let
a WeakMap receiver reach WeakSet helpers (and vice versa) instead of rejecting
incompatible calls. Update the dispatch logic in `try_weak_method_dispatch` to
verify the receiver’s `class_id` matches the requested weak method before
calling `js_weakmap_set`, `js_weakset_add`, `js_weakmap_get`, `js_weakmap_has`,
or `js_weakmap_delete`, and return `None` for mismatched class/method pairs so
the normal incompatible-receiver `TypeError` path is preserved.

🤖 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.

Outside diff comments:
In `@crates/perry-runtime/src/weakref.rs`:
- Around line 802-834: `try_weak_method_dispatch` is dispatching solely by
`method_name`, which can let a WeakMap receiver reach WeakSet helpers (and vice
versa) instead of rejecting incompatible calls. Update the dispatch logic in
`try_weak_method_dispatch` to verify the receiver’s `class_id` matches the
requested weak method before calling `js_weakmap_set`, `js_weakset_add`,
`js_weakmap_get`, `js_weakmap_has`, or `js_weakmap_delete`, and return `None`
for mismatched class/method pairs so the normal incompatible-receiver
`TypeError` path is preserved.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: c1fb4d93-1f66-453a-9acb-51a6078616f5

📥 Commits

Reviewing files that changed from the base of the PR and between 9b6993b and de67ad5.

📒 Files selected for processing (6)

• crates/perry-codegen/src/expr/instance_misc1.rs
• crates/perry-runtime/src/collection_iter.rs
• crates/perry-runtime/src/object/field_get_set/get_field_by_name_tail.rs
• crates/perry-runtime/src/object/instanceof.rs
• crates/perry-runtime/src/object/object_ops/prototype.rs
• crates/perry-runtime/src/weakref.rs