And the original benchmarks of bun vs deno had bun using a native-code http server against a deno using a js-code http server, afaik.
And the bun-specific-fork-of-react vs deno + normal react, was suspect too.
And the bun-specific-fork-of-react vs deno + normal react, was suspect too.
When you import “react-dom/server” in browsers or Deno, it loads “react-dom/server.browser.js”
When you import “react-dom/server” in Bun, it will load “react-dom/server.bun.js”.
Its not so much of a useful thing to just focus so much on what you use to benchmark.
Reminds of:
https://www.theregister.com/2021/01/28/nvidia_tpc_benchmark/
https://www.zdnet.com/article/futuremark-reverses-nvidia-did...
Like, you wouldn't be impressed if it was react bun vs preact deno, right?