feat: upgrade Yuku to 0.18.0 and adapt integrations

[fireairforce]

Summary

Upgrade the vendored Yuku parser and npm yuku-parser from 0.10.1 to 0.18.0, keeping the native engine, ESLint plugin parser, and Playground AST inspector aligned.

• Adapt visitor construction and TypeScript reference metadata. Preserve no-shadow, no-unused-vars, and no-use-before-define behavior for signature parameters, import aliases, and type-only exports.
• Use raw JSX text for comment diagnostics and remove duplicate entity decoding from the ESLint adapter. Cover scope bindings, JSX values and spans, deep traversal, and UTF-16 ranges with regressions.
• Migrate the Playground AST worker to yuku-parser with asynchronous @yuku-core/wasm initialization, including dependency pins and deployment asset checks.

Validation

• ReleaseFast native build; 2,276 Zig tests and 77 rule snapshots pass.
• 245 npm tests and type checks pass against the rebuilt native binary.
• 34 raw WASM ABI tests, 11 WASM package tests, and 14 Playground tests pass, together with type checks, production build, asset verification, and the browser AST worker smoke test.
• Formatting and diff checks pass.
• Functional CI passes on macOS arm64 with Node 20/24 and WebAssembly with Node 24.

Performance validation

CodSpeed's performance gate passes. The hosted walltime report warns about unknown/different execution environments, so its reported improvements should not be attributed to this upgrade. An unchanged-main control run under the current toolchain also increased versus the historical baseline.

A same-host differential check using pinned ReleaseFast builds and Node 24 validates each 1,000-file report: 120 alternating severity pairs average 485.52 ms on main versus 480.22 ms here (head/base 95% interval 0.959–1.019), and 40 native flat-config pairs average 571.27 ms versus 566.96 ms (0.966–1.017). These do not reproduce the earlier walltime alerts on macOS arm64; Linux hosted-runner measurements remain sensitive to environment differences.

[codspeed]

Merging this PR will improve performance by 61.29%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 5 improved benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	npm-wrapper/config-discovery/cached-1000-inputs	12.4 ms	7.5 ms	+66.66%
⚡	npm-wrapper/native-flat-config/1000-files	299 ms	179.7 ms	+66.35%
⚡	npm-wrapper/config-discovery/uncached-1000-inputs	163.4 ms	98.9 ms	+65.26%
⚡	utoo-lint/fixtures/codspeed	187.7 ms	118.7 ms	+58.15%
⚡	npm-wrapper/umi-severity-config/1000-files	281.3 ms	186.7 ms	+50.66%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing zoomdong/update-yuku-0.18.0 (b3c9f28) with main (c03bddd)}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-09T15:12:34.770853Z	b3c9f28	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.