fix(typechecker): reject when tsc fails to start

[mmkal]

Description

Give a better error message when trying to use typecheck without typescript installed.

Previously, it was working ok with pnpm vitest run but just hanging on pnpm vitest (the TS task never completed). This was because we weren't listening for the 'error' event, and set throwOnError: false when spawning the process.

The fix was to have a Promise.race between the process spawn event and the error event, with the error event rejecting with a helpful message.

Please don't delete this checklist! Before submitting the PR, please make sure you do the following:

• _ there isn't an issue for this exactly, but I talked about it with @sheremet-va _ It's really useful if your PR references an issue where it is discussed ahead of time. If the feature is substantial or introduces breaking changes without a discussion, PR might be closed.
• Ideally, include a test that fails without this PR but passes with it.
• Please, don't make changes to pnpm-lock.yaml unless you introduce a new test example.

Tests

• Run the tests with pnpm test:ci.

Documentation

• If you introduce new functionality, document it. You can run documentation with pnpm run docs command.

Changesets

• Changes in changelog are generated from PR name. Please, make sure that it explains your changes in an understandable manner. Please, prefix changeset messages with feat:, fix:, perf:, docs:, or chore:.

[netlify]

✅ Deploy Preview for vitest-dev ready!

Built without sensitive environment variables

Name	Link
🔨 Latest commit	00e70a8
🔍 Latest deploy log	https://app.netlify.com/sites/vitest-dev/deploys/670834fbb86bc4000852a81f
😎 Deploy Preview	https://deploy-preview-6561--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.

[hi-ogawa]

Thanks for the PR! Indeed this hangs on main currently:

$ pnpm -C examples/typecheck test -- --typecheck.checker tsc-bad
...
 DEV  v2.1.1 /home/hiroshi/code/others/vitest/examples/typecheck

 ·  TS  test/type.test-d.ts (1)
 ✓ test/normal.test.ts (1)

I'm testing your PR locally and while it shows a better error, I'm seeing a weird hanging process node (vitest 1) which eventually OOM. On my linux PC, this is how my terminal looks like:

$ pnpm -C examples/typecheck test -- --typecheck.checker tsc-bad
...
 DEV  v2.1.1 /home/hiroshi/code/others/vitest/examples/typecheck


⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ Unhandled Rejection ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
Error: Spawning typechecker failed - is typescript installed?
 ❯ ChildProcess.<anonymous> ../../packages/vitest/dist/chunks/RandomSequencer.C4LDd-jj.js:879:16
 ❯ Object.onceWrapper node:events:634:26
 ❯ ChildProcess.emit node:events:531:35
 ❯ ChildProcess._handle.onexit node:internal/child_process:292:12
 ❯ onErrorNT node:internal/child_process:484:16
 ❯ process.processTicksAndRejections node:internal/process/task_queues:82:21

Caused by: Error: spawn tsc-bad ENOENT
 ❯ ChildProcess._handle.onexit node:internal/child_process:286:19
 ❯ onErrorNT node:internal/child_process:484:16
 ❯ process.processTicksAndRejections node:internal/process/task_queues:82:21

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
Serialized Error: { errno: -2, code: 'ENOENT', syscall: 'spawn tsc-bad', path: 'tsc-bad', spawnargs: [ '--noEmit', '--pretty', 'false', '-p', '/home/hiroshi/code/others/vitest/examples/typecheck/tsconfig.vitest-temp.json', '--watch' ] }

$ ... wait for a while after terminal is back ...

<--- Last few GCs --->

[37137:0x438b1840]    41388 ms: Mark-Compact 4043.2 (4139.0) -> 4030.9 (4141.8) MB, 560.84 / 1.08 ms  (average mu = 0.253, current mu = 0.049) allocation failure; scavenge might not succeed
[37137:0x438b1840]    42441 ms: Mark-Compact 4046.8 (4141.8) -> 4034.4 (4145.5) MB, 1021.84 / 0.60 ms  (average mu = 0.120, current mu = 0.029) allocation failure; scavenge might not succeed


<--- JS stacktrace --->

FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory
----- Native stack trace -----

 1: 0xb86ecf node::OOMErrorHandler(char const*, v8::OOMDetails const&) [node (vitest 1)]
 2: 0xef74d0 v8::Utils::ReportOOMFailure(v8::internal::Isolate*, char const*, v8::OOMDetails const&) [node (vitest 1)]
 3: 0xef77b7 v8::internal::V8::FatalProcessOutOfMemory(v8::internal::Isolate*, char const*, v8::OOMDetails const&) [node (vitest 1)]
 4: 0x1109355  [node (vitest 1)]
 5: 0x11211d8 v8::internal::Heap::CollectGarbage(v8::internal::AllocationSpace, v8::internal::GarbageCollectionReason, v8::GCCallbackFlags) [node (vitest 1)]
 6: 0x10f72f1 v8::internal::HeapAllocator::AllocateRawWithLightRetrySlowPath(int, v8::internal::AllocationType, v8::internal::AllocationOrigin, v8::internal::AllocationAlignment) [node (vitest 1)]
 7: 0x10f8485 v8::internal::HeapAllocator::AllocateRawWithRetryOrFailSlowPath(int, v8::internal::AllocationType, v8::internal::AllocationOrigin, v8::internal::AllocationAlignment) [node (vitest 1)]
 8: 0x10d5ad6 v8::internal::Factory::NewFillerObject(int, v8::internal::AllocationAlignment, v8::internal::AllocationType, v8::internal::AllocationOrigin) [node (vitest 1)]
 9: 0x1531906 v8::internal::Runtime_AllocateInYoungGeneration(int, unsigned long*, v8::internal::Isolate*) [node (vitest 1)]
10: 0x76d50d4d9ef6 

Do you see something similar?

Also I noticed tsconfig.vitest-temp.json is not cleaned up currently.

[mmkal]

Thanks for the tip, have removed the Promise.race.

Re the OOM - yes I was able to reproduce that just now. But, I was also able to reproduce it on main, so not sure it's related - but it's possible that this change makes it more frequent or something?

Re tsconfig.vitest-temp.json - reproing that too with your command. It looks like we only call typechecker.clear() when the startup promise doesn't reject. Will look into whether there's a clean way to always call it.

[mmkal]

Resolved the temp file issue by explicitly calling await this.clear() before rejecting. I don't love that as a solution but what I originally tried, which was calling .clear() from packages/vitest/src/node/pools/typecheck.ts - didn't seem to work. Maybe that's not what manages the typechecker?

[hi-ogawa]

Re the OOM - yes I was able to reproduce that just now. But, I was also able to reproduce it on main, so not sure it's related - but it's possible that this change makes it more frequent or something?

Thanks for confirming. I don't fully understand why, but it looks like this startTypechecker is not awaited here and this might be making a situation worse:

vitest/packages/vitest/src/node/pools/typecheck.ts

Line 145 in 7dd2ecf

startTypechecker(project, files)

If I do promises.push(startTypechecker(project, files)) here, it's looking better though not sure if it's breaking something else. Can you try it and see how CI goes?

[hi-ogawa]

Can you add test in test/typescript? Not the realistic one, but verifying something like { typecheck: { checker: "bad-checker" }} and stderr should be more than enough.

[hi-ogawa]

It looks like this.clear is not needed anymore since pool gets closed properly now.