The latest updates on your projects. Learn more about Vercel for Git ↗︎

I'm not fan of the provided example because it's based on experimental api.

maybe do same things as this test (do with jest). But I mean based on hand written variable
https://github.com/nodejs/nodejs.org/blob/main/apps/site/util/__tests__/detectOS.test.mjs

Hmm, I think that example isn't quite as good. The one I have is unknowable.

I'm happy to just switch the assertion (and far less work 😅)

EDIT: ohhh! Wait, is the experimental API you're concerned about globSync? (Rather than assert.partialDeepStrictEqual)

EDIT: ohhh! Wait, is the experimental API you're concerned about globSync? (Rather than assert.partialDeepStrictEqual)

Yup and and using a variable is easier because the reader can copy past and run.

using a variable is easier because the reader can copy past and run

A variable for what?

A variable for what?

import { isBiscuit } from "factory";

const valid_tests_case = [
  "cookies",
  "speculoos"
];

test("isBiscuit", (t) => {
   for() {
   // loop for dynamic test 
   t.test(`${e} - should be true`)
   }
});

in this case it's use Biscuit because it's fun

But here it's test user agent

Hmm, I've given this some more thought, and I think the experimental API used is okay because it does not affect the main issue of the article. Thoughts on that?

How about both options? The user-agent one is simple, and the glob is advanced; I know I frequently see docs cover a simple example but not an advanced one, which leaves an advanced user on their own to figure out something complicated (which is very frustrating).

Lighthouse Results

Unit Test Coverage Report

Lines	Statements	Branches	Functions
	87.1% (736/845)	72.13% (233/323)	86.33% (139/161)

Unit Test Report