repro: vmForks bug

[sheremet-va]

Description

To run the reproduction: pnpm run node-error-repro. This should run tests and create files with all the output from those tests.

What happens:

• Vitest spawns several workers to run tests inside of them. One worker will run several tests one after another when it's free.
• At one point the worker (child_process in the reproduction, because worker_thread crashes the whole process) crashes while trying to import something. I've added console logs in our module runner implementation - it only logs the module if it's not cached already. Every time it crashes it imports the same file it looks like (because the end of the debug is always the same in all files - node_modules/loupe/lib/helpers.js). I don't think that file is the problem, but one of its parents maybe? Usually, it happens at least on the second test run after we create a new executor:
  

  vitest/packages/vitest/src/runtime/workers/vm.ts

  Line 77 in 0499a31

  const executor = await startVitestExecutor({

  
  Vitest never manually clears the cache, we rely on gc, but we also reuse fileMap cache:
  

  vitest/packages/vitest/src/runtime/workers/vm.ts

  Line 15 in 0499a31

  const fileMap = new FileMap()

• The problem is most likely with how we implement VM executor - it was based on Jest's with some modifications, but I am not smart enough to figure out why. The error started happening in Node 22.1, but the changelog doesn't really mention related changes

[netlify]

✅ Deploy Preview for vitest-dev ready!

Name	Link
🔨 Latest commit	cf668f0
🔍 Latest deploy log	https://app.netlify.com/sites/vitest-dev/deploys/673f7f877ab6be0008c6ebe4
😎 Deploy Preview	https://deploy-preview-6452--vitest-dev.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.

[AriPerkkio]

The test/cli/test/vm-threads.test.ts is also failing on CI and locally now. The failure does not happen with Node 22.0, but happens with all versions above 22.1.

ari ~/x/vitest/test/cli (main) $ pnpm run test vm-threads --run

> @vitest/test-cli@ test /Users/x/x/vitest/test/cli
> vitest "vm-threads" "--run"


 RUN  v2.1.1 /Users/x/x/vitest/test/cli

 ❯ test/vm-threads.test.ts (1)
 ❯ test/vm-threads.test.ts (1)
   ⠏ importing files in restricted fs works correctly

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Unhandled Errors ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

Vitest caught 1 unhandled error during the test run.
This might cause false positive tests. Resolve unhandled errors to make sure your tests are not affected.

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Unhandled Error ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
Error: Worker exited unexpectedly
 ❯ ChildProcess.<anonymous> ../../node_modules/.pnpm/[email protected]/node_modules/tinypool/dist/index.js:136:34
 ❯ ChildProcess.emit node:events:531:35
 ❯ ChildProcess._handle.onexit node:internal/child_process:293:12

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

 Test Files   (1)
      Tests   (1)
     Errors  1 error
   Start at  09:34:59
   Duration  683ms (transform 43ms, setup 0ms, collect 161ms, tests 0ms, environment 0ms, prepare 46ms)

 ELIFECYCLE  Test failed. See above for more details.

[AriPerkkio]

Oh wow, does Node 22.11 fix this? Can't crash locally with that. 22.10 keeps crashing.

Edit: Nope, it's much more stable but crashes after couple of minutes of re-running test/core with vmThreads.