Test out esbuild for minification

[NullVoxPopuli]

This is to demonstrate impact of esbuild vs terser on a laughably small app (compared to the size of the apps ya'll are working on if you're interested in this research 🙃 )

Process:

• checkout relevant commit (HEAD for esbuild, HEAD~1 for terser)
• yarn
• build once, throw that build away

• build 5+ times, recording the times

  All builds are ran with:

  time ember build -e production

  sourcemaps are disabled because those aren't properly implemented over here yet; https://github.com/NullVoxPopuli/ember-cli-esbuild/pull/1

  overall results:

esbuild is 34% faster ( 1 - (19.28 / 29.28) )

with esbuild

`time` of each run

### run 1 ``` real 0m19.175s user 0m22.053s sys 0m3.283s ``` ### run 2 ``` real 0m19.240s user 0m22.077s sys 0m3.187s ``` ### run 3 ``` real 0m18.945s user 0m21.675s sys 0m3.100s ``` ### run 4 ``` real 0m19.698s user 0m22.404s sys 0m3.278s ``` ### run 5 ``` real 0m19.366s user 0m21.901s sys 0m3.538s ```

Stats

Average time: 19.28s Variance: 0.07671s Coefficient of Variation: 0.01436

with terser

`time` of each run

### run 1 ``` real 0m29.035s user 0m48.452s sys 0m3.816s ``` ### run 2 ``` real 0m28.085s user 0m47.133s sys 0m4.140s ``` ### run 3 ``` real 0m31.346s user 0m54.739s sys 0m4.714s ``` ### run 4 ``` real 0m28.883s user 0m47.325s sys 0m4.172s ``` ### run 5 ``` real 0m29.079s user 0m47.197s sys 0m3.981s ```

Stats

Average time: 29.28s Variance: 1.458s Coefficient of Variation: 0.04124

Hardware:

• Ubuntu VM, VirtualBox
• "6 CPUs", i7-8700 @ 3.2GHz
• 17000MB RAM

System was mostly idle during these tests.

[github-actions[bot]]

🚀 Deployed on https://615d9e6d72917d1e0e6539f1--emberclear.netlify.app