test: #695 — behavioral parity fixtures for runtime core FFI surface

[TheHypnoo]

Summary

Converts ~555 inventory entries from test_ffi_surface_runtime_core.ts into deterministic behavioral TypeScript parity tests, the follow-up tracked by #695 from #694.

Adds six focused fixtures, each with an @covers manifest so test-coverage/regen_ts_surface_inventory.py moves the listed FFI names out of the inventory:

• test_compat_arrays_iterators.ts — array creation/mutation, higher-order iteration, immutable variants (toSorted/toReversed/toSpliced/with), flat/flatMap, Array.of/from/isArray
• test_compat_objects_symbols.ts — Object.keys/values/entries, descriptors, freeze/seal/preventExtensions, Symbol.for/keyFor, class instanceof and getters, Object.groupBy, spread/rest, Map and Set with insertion-order iteration
• test_compat_strings_regex_json.ts — char-code helpers, slice/substring, case/trim/pad/repeat/split/replace, btoa/atob, regex exec/test/flags, named groups, matchAll, JSON.parse/stringify with replacer/reviver
• test_compat_url_date_math.ts — URL accessors and mutation, URLSearchParams read paths, fixed-epoch Date getters/setters in UTC, Math constants and trig/log/exp helpers, Number/BigInt parsing and arithmetic
• test_compat_async_errors.ts — Promise.{resolve,reject,all,allSettled,race,any,withResolvers}, async/await with chained then/catch/finally, error subclasses with cause, AggregateError, try/catch/finally, structuredClone, async generators
• test_compat_buffers_typed.ts — Buffer.{alloc,allocUnsafe,from,concat,compare,equals,includes}, numeric read/write helpers across all widths/endians, BigInt reads/writes, Uint*/Int* typed arrays, typeof/truthiness, closure-arity exercises

Inventory + coverage

• test_ffi_surface_runtime_core.ts: 915 → 360 unique FFI names.
• test-coverage/COVERAGE.md: TypeScript FFI surface coverage stays at 100% (1790/1790).

Known gaps deliberately not exercised

Behaviors that currently diverge from Node and stay in the inventory / gap suite for follow-up:

• Multi-arg Array.prototype.concat
• Array.prototype.lastIndexOf
• Array.prototype.entries().next().value tuple
• Custom Symbol.iterator in for…of
• DataView.prototype.setInt32 / setFloat32
• String.prototype.replace(regex, fn)
• URLSearchParams ordering after delete
• BigInt large left-shift (1n << 60n)
• Object.create(proto) prototype chain lookup (null-proto path is exercised instead)

Test plan

• ./test-coverage/regen_ts_surface_inventory.py
• ./test-coverage/audit.sh --markdown — 100% TS FFI surface coverage
• ./run_parity_tests.sh --filter test_compat_ — 7/7 PASS
• ./run_parity_tests.sh --filter test_ffi_surface — 5/5 PASS (no regression)

Notes

Per CLAUDE.md and contributor guidance, this PR does not bump the workspace version or edit CHANGELOG.md / CLAUDE.md — the maintainer handles that at merge time.

Checklist

• I have NOT bumped the workspace version or edited CLAUDE.md / CHANGELOG.md
• My commit follows the loose conventional prefix style used in the log
• I've read CONTRIBUTING.md and agree to the Code of Conduct