Switiching from '!!' to '===' for 'process.env.D365_UI_TEST_HEADLESS'

[julianYaman]

Hello,
I thought to move my suggestion to the "Discussions" section since it is not really an issue and IMO does not need a fix with a PR.

Today, I was working with environment variables to launch the browser in the beforeAll statement and had problems with using the approach from the README

const response = await xrmTest.launch("chromium", {
    headless: !!process.env.D365_UI_TEST_HEADLESS,
    args: [
      "--disable-setuid-sandbox",
      "--disable-infobars",
      "--start-fullscreen",
      "--window-position=0,0",
      "--window-size=1920,1080",
      '--user-agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.198 Safari/537.36"',
    ],
});

I use dotenv and a local .env file to set the value of D365_UI_TEST_HEADLESS by writing D365_UI_TEST_HEADLESS=false.

My workaround was to use this approach:

const response = await xrmTest.launch("chromium", {
    headless: process.env.D365_UI_TEST_HEADLESS === "true",
    args: [
      "--disable-setuid-sandbox",
      "--disable-infobars",
      "--start-fullscreen",
      "--window-position=0,0",
      "--window-size=1920,1080",
      '--user-agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.198 Safari/537.36"',
    ],
});

I decided to use this approach after reading this on StackOverflow.

Would it be better to switch from !! to ===?
It would definitely mean that developers are only allowed to set 'true' explicitly, otherwise they would get false as the value for D365_UI_TEST_HEADLESS.

What do you think about this?

[imartinflores]

hey @julianYaman ! I am not very experienced in JS, but if that's suggested in StackOverflow and works better, then it should be ok to update it!

[julianYaman]

Also thought this. Seems that using !! may not be recommended, because of the explanation in the StackOverflow post:

Any string which isn't the empty string will evaluate to true by using them.

Hope this is ok for @DigitalFlow to change it too.

[DigitalFlow]

Hi @julianYaman and @imartinflores,

you're totally right, the === approach is better.
There is an open task for migrating to dotenv as well, as the current parsing is not optimal either.

We'll implement these two changes in one commit 👍

Kind regards,
Florian

[julianYaman]

Great to hear that! 😄