test,console: known_issues test for issue 831

[Trott]

Pull Request check-list

Please make sure to review and check all of these items:

• Does make -j8 test (UNIX) or vcbuild test nosign (Windows) pass with
  this change (including linting)?
• Is the commit message formatted according to [CONTRIBUTING.md][0]?

Affected core subsystem(s)

console, test

Description of change

This adds a test for the issue described in
#831.

I've added some metadata fields that I would personally find useful in
these tests.

Refs: #831

[Trott]

This is probably fragile, dependent on net.js always using the presence of the _handle property to determine if the socket has been closed or not. It does that today but I imagine there's no guarantee it will continue to work that way tomorrow. But I'm not sure there's a better way to have the problem fire reliably on CI. Maybe there's some child_process sorcery I'm unaware of?

[Trott]

The _handle checking in net.js that I refer to above is at

node/lib/net.js

Lines 645 to 648 in a498492

	if (!this._handle) {
	this._destroy(new Error('This socket is closed'), cb);
	return false;
	}

[Fishrock123]

@Trott I don't think that issue link is correct?

[Trott]

@Fishrock123 Argh. You are correct. The correct issue link is #831. Will fix above too...

[cjihrig]

I believe #947 is the same issue, or closely related. I have this failing test, which I haven't opened a PR for:

'use strict';
// Refs: https://github.com/nodejs/node/issues/947
const common = require('./test/common');
const assert = require('assert');
const cp = require('child_process');

if (process.argv[2] === 'child') {
  process.on('message', common.mustCall((msg) => {
    assert.strictEqual(msg, 'go');
    console.log('logging should not cause a crash');
    process.disconnect();
  }));
} else {
  const child = cp.fork(__filename, ['child'], {silent: true});

  child.on('close', common.mustCall((exitCode, signal) => {
    assert.strictEqual(exitCode, 0);
    assert.strictEqual(signal, null);
  }));

  child.stdout.destroy();
  child.send('go');
}

[jasnell]

Test LGTM, will be good to get this fixed.

[Trott]

@cjihrig That sure seems like the same thing and a more robust test. And I see you've included it in #5653. So I'm going to close this.