perf(next-plugin): profile and reuse Turbopack prewarm output

[owjs3901]

Summary

• add opt-in structured Turbopack timings for graph setup, production prewarm, coordinator extraction, snapshots, serialization, and writes
• reuse byte-identical singleCss prewarm output so Turbopack loaders do not repeat WASM extraction
• keep polymorphic React inference exact while preventing redundant native-prop inference and duplicate intrinsic-prop conditionals
• stabilize the existing benchmark.js comparison in the existing Publish workflow: remove Next memory-debug overhead, clean every sample, alternate order, and report six-sample medians

Validation

• wasm-pack release build completed after Smart App Control was disabled; no security-policy bypass was used
• Next plugin tests: 111 passed, 0 failed with the real WASM package
• React tests: 37 passed, 0 failed; React lint/build and TypeScript 6/7 type tests passed
• compile-time checks cover anchor/button event inference, invalid native props, required custom-component props, and bidirectional compatibility with the previous intrinsic/custom prop types
• clean CI run (benchmark, publish, landing-next-e2e all passed): https://github.com/dev-five-git/devup-ui/actions/runs/33193236872
• Codecov report: 100.0% overall, including 407/407 covered lines in coordinator.ts

Existing benchmark.js CI result

Six alternating cold builds on the same Ubuntu runner, using normal next build:

Phase	Devup UI singleCss	Tailwind Turbo	Result
Turbopack compile	4.0 s	4.1 s	Devup UI 0.1 s faster
Next config	104 ms	13 ms	Devup UI 91 ms slower
TypeScript	351 ms	257 ms	Devup UI 94 ms slower
Full wall time	6.82 s	6.76 s	Devup UI 60 ms / 0.9% slower
Emitted CSS	327 bytes	6,197 bytes	Devup UI 94.7% smaller

Raw wall samples:

• Tailwind Turbo: 6.79, 6.76, 6.71, 6.77, 6.68, 6.77 s
• Devup UI singleCss: 6.81, 6.87, 6.77, 6.76, 6.90, 6.83 s

The type changes reduced the Devup fixture's CI type check from roughly 840 ms to 351 ms without weakening inference. A Rust/WASM -O3 experiment was rejected: it produced a 7.61 s Devup median versus 6.95 s Tailwind and increased package-build time and WASM size. The compact s/-Oz settings remain.

Structured timings put CSS/sheet/class/file serialization below 1 ms, so batching/debouncing it would add complexity without addressing the measured bottleneck.

Benchmark limits

This is substantially fairer than the previous one-shot fixed-order benchmark, but it is still not proof that one library is generally faster. The fixtures are not semantically identical: Tailwind uses untyped class strings and emits its framework CSS, while Devup validates rich responsive/polymorphic props and emits much less CSS. The latest full-build median is still 60 ms slower; the two preceding clean runs were also 160 ms and 110 ms slower. This PR therefore does not claim that Devup currently beats Tailwind end-to-end. Disabling type checking or weakening inference would reverse the goal and was deliberately not done.

The local full pre-commit suite still has an unrelated Bun transform-cache issue that resolves df/devup-ui.css from another worktree (5055 pass, 7 fail); it does not reproduce on clean CI. bun.lock and the add-bench PR are excluded from this PR's changes.

[github-actions]

Changepacks

@devup-ui/next-plugin@1.0.83 - packages/next-plugin/package.json

Maybe you forgot to write the following files to the latest version

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
bindings/devup-ui-wasm/src/lib.rs	100.00% <100.00%> (ø)
libs/extractor/src/lib.rs	100.00% <100.00%> (ø)
packages/next-plugin/src/coordinator.ts	100.00% <100.00%> (ø)
packages/next-plugin/src/plugin.ts	100.00% <100.00%> (ø)
packages/next-plugin/src/profile.ts	100.00% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.