test_runner: ensure proper teardown when tests run without isolation

[pmarchini]

This should address #57234.

Adding the teardown in case of isolation: none addresses this specific issue, and I see no potential negative effects in having it.

@cjihrig, was there a reason why we avoided it in the first place?

[nodejs-github-bot]

Review requested:

• @nodejs/test_runner

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.23%. Comparing base (9df0ff7) to head (dd7af59).
⚠️ Report is 1207 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #57394      +/-   ##
==========================================
- Coverage   90.23%   90.23%   -0.01%     
==========================================
  Files         630      630              
  Lines      185200   185214      +14     
  Branches    36233    36243      +10     
==========================================
+ Hits       167108   167120      +12     
- Misses      11056    11057       +1     
- Partials     7036     7037       +1     

Files with missing lines	Coverage Δ
lib/internal/test_runner/runner.js	89.52% <100.00%> (+0.01%)	⬆️

... and 22 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[pmarchini]

These changes have broken the following test:

node/test/parallel/test-runner-extraneous-async-activity.js

Lines 52 to 68 in 96457b4

	{
	const child = spawnSync(process.execPath, [
	'--test',
	'--test-isolation=none',
	fixtures.path('test-runner', 'async-error-in-test-hook.mjs'),
	]);
	const stdout = child.stdout.toString();
	assert.match(stdout, /Error: Test hook "before" at .+async-error-in-test-hook\.mjs:3:1 generated asynchronous activity after the test ended/m);
	assert.match(stdout, /Error: Test hook "beforeEach" at .+async-error-in-test-hook\.mjs:9:1 generated asynchronous activity after the test ended/m);
	assert.match(stdout, /Error: Test hook "after" at .+async-error-in-test-hook\.mjs:15:1 generated asynchronous activity after the test ended/m);
	assert.match(stdout, /Error: Test hook "afterEach" at .+async-error-in-test-hook\.mjs:21:1 generated asynchronous activity after the test ended/m);
	assert.match(stdout, /pass 1$/m);
	assert.match(stdout, /fail 0$/m);
	assert.match(stdout, /cancelled 0$/m);
	assert.strictEqual(child.status, 1);
	assert.strictEqual(child.signal, null);
	}

I'm taking a look!

[mcollina]

lgtm

[inoway46]

cc @pmarchini

I opened #62056 to fix #60020, but I did not realize this PR was already addressing the same underlying problem, so I closed mine.

I added a regression test in #62056 based on the cluster worker reproduction from #60020, so if it's helpful, please feel free to reuse it.