fix: close vite servers on all resolved projects

[surc54]

Description

I've been running into an issue where Vitest would hang after the testing is completed with the following output:

close timed out after 10000ms
Tests closed successfully but something prevents the main process from exiting
You can try to identify the cause by enabling "hanging-process" reporter. See https://vitest.dev/config/#reporters

Luckily, this repo has a reproducible setup at test/cli/fixtures/project -- in this directory, running npx vitest run --project space_1 causes the error above (you may need to try multiple times, but I've been getting it consistently).

The only other report I've found is by this commenter #2008 (comment) (the issue itself seems to be about a different time out issue).

I've tracked this down to a Vite web-socket server not being cleaned up at the end of the Vitest session.

This PR edits the Vitest#close() method to close the Vite dev servers on ALL resolved projects instead of just the filtered projects.

I tried to write a new test under test/cli but it seems that the Vite servers for the fixture projects do not get launched due to a port conflict with the Vitest instance running the actual test. I could not find a way to use a different port.

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

• 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.
• (N/A) 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

• (N/A) 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 fastidious-cascaron-4ded94 canceled.

Name	Link
🔨 Latest commit	4c840e5
🔍 Latest deploy log	https://app.netlify.com/sites/fastidious-cascaron-4ded94/deploys/661ef62a44c9a900088f4583

[AriPerkkio]

Interesting finding! How did you narrow the root cause to the websocket connections? Looks like Vitest's hanging-process reporter is unable to show those connections - I wonder if we could improve it to include them as well.

$ npx vitest run --project space_1 --reporter hanging-process --reporter basic

 RUN  v1.5.0 /Users/x/vitest/test/cli/fixtures/project

 ✓ |space_1| base.test.ts  (1 test) 1ms

 Test Files  1 passed (1)
      Tests  1 passed (1)
   Start at  08:15:25
   Duration  147ms (transform 12ms, setup 0ms, collect 6ms, tests 1ms, environment 0ms, prepare 45ms)

There are 12 handle(s) keeping the process running

# Tinypool
node:internal/async_hooks:202                                                                                    
node:internal/async_hooks:504                                                                                    
file:///Users/x/vitest/node_modules/.pnpm/[email protected]/node_modules/tinypool/dist/esm/index.js:37 
file:///Users/x/vitest/node_modules/.pnpm/[email protected]/node_modules/tinypool/dist/esm/index.js:58 
file:///Users/x/vitest/node_modules/.pnpm/[email protected]/node_modules/tinypool/dist/esm/index.js:945
file:///Users/x/vitest/packages/vitest/dist/vendor/cac.WCfoZ7tE.js:8776                             
file:///Users/x/vitest/packages/vitest/dist/vendor/cac.WCfoZ7tE.js:9425                             
file:///Users/x/vitest/packages/vitest/dist/vendor/cac.WCfoZ7tE.js:9450                             

# FILEHANDLE
node:internal/async_hooks:202

... FILEHANDLE x9 here

# FILEHANDLE
node:internal/async_hooks:202
close timed out after 10000ms
Tests closed successfully but something prevents the main process from exiting
You can try to identify the cause by enabling "hanging-process" reporter. See https://vitest.dev/config/#reporters

It's only reporting FILEHANDLE's for this case. The # Tinypool \n node:internal/async_hooks:202 part is coming from here https://github.com/tinylibs/tinypool/blob/f86e82927371b54ba26577d818b86393ccf902c2/src/EventEmitterAsyncResource.ts#L24 - we should probably add emitDestroy for those in Tinypool's pool.destroy() so that they wouldn't show up in the report.

[AriPerkkio]

Same logic should be used here:

vitest/packages/vitest/src/node/core.ts

Lines 908 to 917 in df6a432

	const runningServers = [this.server, ...this.projects.map(p => p.server)].filter(Boolean).length
	if (runningServers === 1)
	console.warn('Tests closed successfully but something prevents Vite server from exiting')
	else if (runningServers > 1)
	console.warn(`Tests closed successfully but something prevents ${runningServers} Vite servers from exiting`)
	else
	console.warn('Tests closed successfully but something prevents the main process from exiting')
	console.warn('You can try to identify the cause by enabling "hanging-process" reporter. See https://vitest.dev/config/#reporters')

Then the reproduction case would print following:

close timed out after 10000ms
Tests closed successfully but something prevents 2 Vite servers from exiting

[surc54]

@AriPerkkio Thanks for the review! I've updated that location to use this.resolvedProjects.

As for how I tracked it down, I used a debugger to run process._getActiveHandles() during that timeout period, which gave back something like this:

[ WriteStream, ReadStream, WriteStream, Server ]

The server caught my interest, and upon checking its address, I noticed that it was on port 24678. Searching the vite repository led me to the dev server code.