Fix bugs with setup and teardown

[DmitrySharabin]

• Maintain proper sequencing when hooks (beforeEach/afterEach, beforeAll/afterAll) are used
• Fall back to parallel execution (using Promise.all) when no hooks exist

Now, we might end up (I already was bit by this bug) calling afterAll before afterEach for some tests. In most cases, beforeEach for all tests is called before the first test runs. There are probably some more issues we haven't encountered yet. This PR should fix those bugs.

[netlify]

✅ Deploy Preview for h-test ready!

Name	Link
🔨 Latest commit	64de46f
🔍 Latest deploy log	https://app.netlify.com/sites/h-test/deploys/679a538451c87d00088d02f9
😎 Deploy Preview	https://deploy-preview-68--h-test.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[LeaVerou]

The way I see it, we can still do parallel execution with these, as long as:

• For each test, beforeEach is always executed before run and afterEach after run
• beforeAll and afterAll are executed before and after all tests (regardless of their order)

[DmitrySharabin]

The way I see it, we can still do parallel execution with these, as long as:

• For each test, beforeEach is always executed before run and afterEach after run
• beforeAll and afterAll are executed before and after all tests (regardless of their order)

You are absolutely right. Previously, I was confused seeing beforeEach of a test before run of the previous one. But now I understand it's OK if the order of hooks for individual tests is preserved. It should be fixed now.

[LeaVerou]

This looks a bit too complicated and I have some reservations about whether the complexity is necessary, but I may be missing some of the nuances of this part of the code.

• allFinished is added dynamically, which means it could not be present when awaited from another part of the code
• if runAll() resolves when all tests have ran, why do we need allFinished? Can't we just Promise.allSettled() its result?
• Remember that in theory, tests can be run in isolation, without running the whole group. At first glance it seems to me that this wouldn't run beforeAll/afterAll, but I didn't look too closely.

[DmitrySharabin]

This looks a bit too complicated and I have some reservations about whether the complexity is necessary, but I may be missing some of the nuances of this part of the code.

• allFinished is added dynamically, which means it could not be present when awaited from another part of the code
• if runAll() resolves when all tests have ran, why do we need allFinished? Can't we just Promise.allSettled() its result?
• Remember that in theory, tests can be run in isolation, without running the whole group. At first glance it seems to me that this wouldn't run beforeAll/afterAll, but I didn't look too closely.

What an excellent feedback (as always)! Thank you.

I tried to address it:

• moved beforeEach/afterEach to TestResult.run, so if a test is skipped, those hooks are not called
• simplified the code (got rid of redundant allFinished in favor of Promise.allSettled)
• beforeAll/afterAll should now run when a test is running in isolation

I'll be grateful if you could find some time to review this PR one more time.