Bug: MessageChannel in Scheduler prevents Jest test from exiting

[victor-homyakov]

React version: any
Scheduler version: any up to current (0.23.0)

Steps To Reproduce

1. Create Jest unit test.
2. Select JSDom as test environment so that runtime will not have setImmediate function.
3. JSDom still does not implement MessageChannel MessageChannel is not defined jsdom/jsdom#2448. If MessageChannel is required to test some important functionality, one can add an implementation from Node.js as recommended in comment MessageChannel is not defined jsdom/jsdom#2448 (comment)

   window.MessageChannel = require('node:worker_threads').MessageChannel;

4. Add Scheduler or React as a dependency and require it in the test or one of the files under test.
5. Run test

Link to code example: https://github.com/victor-homyakov/scheduler-jest-jsdom-example

The current behavior

Test is endless. Jest won't stop. Console shows a message

Jest did not exit one second after the test run has completed.

'This usually means that there are asynchronous operations that weren't stopped in your tests. Consider running Jest with `--detectOpenHandles` to troubleshoot this issue.

Running Jest with --detectOpenHandles outputs:

Jest has detected the following 1 open handle potentially keeping Jest from exiting:

  ●  MESSAGEPORT

    > 1 | require('scheduler');
        |                     ^

      at node_modules/scheduler/cjs/scheduler.development.js:569:17

The expected behavior

Code at

react/packages/scheduler/src/forks/Scheduler.js

Line 621 in 5426af3

channel.port1.onmessage = performWorkUntilDeadline;

should unref the handle:

  channel.port1.onmessage = performWorkUntilDeadline;
  // Allow the thread to exit if this is the only active handle in the event system
  if (channel.port1.unref) {
    channel.port1.unref();
  }

[s100]

I'm seeing the same issue during some experiments with Mocha testing. The suggested fix works for me. I will produce a minimal repro if that's considered desirable.

@victor-homyakov Are you aware of any kind of workaround for this?

[victor-homyakov]

@victor-homyakov Are you aware of any kind of workaround for this?

The workaround is to use patched MessageChannel in JSDom, something like

const MessageChannelOriginal = require('node:worker_threads').MessageChannel;
class MessageChannel {
    constructor() {
        const channel = new MessageChannelOriginal();
        // Allow the thread to exit if this is the only active handle in the event system
        if (channel.port1.unref) {
            channel.port1.unref();
        }
        this.channel = channel;
    }
    get port1() {
        return this.channel.port1;
    }
    // ...
}
window.MessageChannel = MessageChannel;

[s100]

Thank you very much for the workaround! In practice I found that simply unrefing the port immediately on creation did not do the trick, because as soon as scheduler calls channel.port1.onmessage = performWorkUntilDeadline the port becomes reffed again. I had to do something more complex:

const MessageChannelOriginal = global.MessageChannel
global.MessageChannel = class {
  constructor () {
    const channel = new MessageChannelOriginal()
    this.port1 = new Proxy(channel.port1, {
      set (port1, prop, value) {
        const result = Reflect.set(port1, prop, value)
        if (prop === 'onmessage') {
          port1.unref()
        }
        return result
      }
    })
    this.port2 = channel.port2
  }
}

Alternatively, as I happen to be using Mocha, I found that the --exit flag also worked.

[github-actions]

This issue has been automatically marked as stale. If this issue is still affecting you, please leave any comment (for example, "bump"), and we'll keep it open. We are sorry that we haven't been able to prioritize it yet. If you have any new additional information, please include it with your comment!

[s100]

Bump, this is still an issue for us.

[github-actions]

This issue has been automatically marked as stale. If this issue is still affecting you, please leave any comment (for example, "bump"), and we'll keep it open. We are sorry that we haven't been able to prioritize it yet. If you have any new additional information, please include it with your comment!

[victor-homyakov]

Bump

[victor-homyakov]

Bump

[github-actions]

This issue has been automatically marked as stale. If this issue is still affecting you, please leave any comment (for example, "bump"), and we'll keep it open. We are sorry that we haven't been able to prioritize it yet. If you have any new additional information, please include it with your comment!

[s100]

Bump

[petrloukota]

Bump