Add CodSpeed performance benchmarks

[codspeed]

Summary

This PR sets up continuous performance testing for htmlnano using CodSpeed with vitest benchmarks.

Changes

• Benchmark suite (bench/htmlnano.bench.ts): 9 benchmarks covering the core minification pipeline:
  • Safe preset: small, medium, and large HTML documents
  • Max preset: small, medium, and large HTML documents
  • Individual modules: collapseWhitespace, removeComments, and minifyJs in isolation
• Vitest configuration (vitest.config.mts): Configured with the @codspeed/vitest-plugin for CodSpeed integration. Includes a small Vite transform plugin to handle the existing module.exports compatibility shim in src/index.ts when running under ESM.
• CI workflow (.github/workflows/codspeed.yml): GitHub Actions workflow that runs benchmarks on pushes to master, pull requests, and manual dispatches. Uses OIDC authentication and simulation mode.
• README badge: Added CodSpeed performance badge alongside existing badges.

Dependencies added

• vitest (dev)
• @codspeed/vitest-plugin (dev)

Next steps

• Once merged, CodSpeed will start tracking performance on every PR and push to master.
• Performance regressions will be reported directly on pull requests.
• Visit the CodSpeed dashboard to view results and configure thresholds.

[netlify]

✅ Deploy Preview for htmlnano canceled.

Name	Link
🔨 Latest commit	31d78a2
🔍 Latest deploy log	https://app.netlify.com/projects/htmlnano/deploys/69b3f237332fd50008b0d682

[codspeed]

Congrats! CodSpeed is installed 🎉

🆕 9 new benchmarks were detected.

You will start to see performance impacts in the reports once the benchmarks are run from your default branch.

Detected benchmarks

• collapseWhitespace only: 7.1 ms
• medium HTML: 19.9 ms
• minifyJs only: 13.9 ms
• large HTML: 61.3 ms
• large HTML: 59.2 ms
• small HTML: 2.4 ms
• medium HTML: 21.7 ms
• removeComments only: 7.2 ms
• small HTML: 2.7 ms

---

[maltsev]

This was a PoC. It works, but I think it isn't really needed at this point. We might get back to it later, though.