fix: add throat when running spec files

[sheremet-va]

Fixes VITETS_POOL_ID logic.

I noticed no performance differences with this.

I think this might also help with segfalts and Node errors? I have a feeling that running them all might cause it, but I don't have anything to back up that claim.

[netlify]

❌ Deploy Preview for vitest-dev failed.

Built without sensitive environment variables

Name	Link
🔨 Latest commit	629bf65
🔍 Latest deploy log	https://app.netlify.com/sites/vitest-dev/deploys/62a82842bfcd4e000afc8366

[antfu]

https://www.npmtrends.com/p-limit-vs-p-throttle-vs-promise-throttle-vs-throat

Maybe p-limit?

[sheremet-va]

https://www.npmtrends.com/p-limit-vs-p-throttle-vs-promise-throttle-vs-throat

Maybe p-limit?

Sure. Replaced it. Also added comment to PR, if you didn't notice.

[sheremet-va]

Closed in favor of #1482

[antfu]

I guess tinypool already has the queueing mechanism, probably this might be redundant. About the Segfault, I guess it's about some deep bug of Node, where even if this makes it less likely to appear (I doubt tho), it doesn't really fix the root cause. 🤔

[sheremet-va]

I guess tinypool already has the queueing mechanism, probably this might be redundant. About the Segfault, I guess it's about some deep bug of Node, where even if this makes it less likely to appear (I doubt tho), it doesn't really fix the root cause. 🤔

It has, but we can't hook up to it.

I guess it's about some deep bug of Node, where even if this makes it less likely to appear

Doesn't hurt to try. This bug is not reproducible with Jest, so we might have something wrong

[sheremet-va]

It has, but we can't hook up to it.

Maybe we can add this functionality to tinypool?

Doesn't hurt to try.

react-enzym failed once in the new PR, but fixed in the second one. I fear this change might introduce flaky tests now :P