`test-buffer-tostring-rangeerror` is flaky (timeout)

[aduh95]

Test

test-buffer-tostring-rangeerror

Platform

Linux x64

Console output

not ok 1244 parallel/test-buffer-tostring-rangeerror
  ---
  duration_ms: 120073.76400
  severity: fail
  exitcode: -15
  stack: |-
    timeout
    (node:2934155) [DEP0005] DeprecationWarning: Buffer() is deprecated due to security and usability issues. Please use the Buffer.alloc(), Buffer.allocUnsafe(), or Buffer.from() methods instead.
    (Use `node --trace-deprecation ...` to show where the warning was created)
  ...

Build links

• https://ci.nodejs.org/job/node-test-commit-linux-containered/43366/nodes=ubuntu2204_sharedlibs_withoutintl_x64/

Additional information

No response

[aduh95]

I'm able to reproduce locally: running the test 100'000 times, I got 6 timeouts. If I add a console.log("pass") at the end of the file, it gets printed even when the test times out. Adding process.exit(); does not help.

[fahrradflucht]

I also investigated this a bit earlier. I was able to reproduce it quite consistently with just a 1000 runs and reduced it to this:

'use strict';
require('../common');

const assert = require('assert');
const { constants: { MAX_STRING_LENGTH } } = require('buffer');

const len = MAX_STRING_LENGTH + 1;
const message = {
  code: 'ERR_STRING_TOO_LONG',
  name: 'Error',
};
assert.throws(() => Buffer(len).toString('utf8'), message);
assert.throws(() => Buffer(len).toString('utf8'), message);

Removing either the second buffer.toString call, the ../common import, adding a console.log('pass') at the end or using a debug build will make it not reproduce anymore. 🙃 🤔

[aduh95]

Could we try to add --no-warnings CLI flag and see if that would have any effect? EDIT: nope, I tried with both --no-warnings and --no-deprecation, and I was still getting timeouts

[fahrradflucht]

I got curious again by #52959 (comment) and indeed this doesn't reproduce on 07f481c but does on 91661ec so it looks like something in #52293 introduced this flakiness as well / aka probably it's the same root cause.

[lpinca]

@fahrradflucht can you try with the --jitless flag?

[fahrradflucht]

@fahrradflucht can you try with the --jitless flag?

It indeed does not reproduce with --jitless on 91661ec.

[fahrradflucht]

I got curious again by #52959 (comment) and indeed this doesn't reproduce on 07f481c but does on 91661ec so it looks like something in #52293 introduced this flakiness as well / aka probably it's the same root cause.

Actually, I'm sorry, I think I have to take this back. I now reproduced a bunch of times on 07f481c as well. It might happen less prior to the v8 update, but it can be reproduced as well.

[jakecastelli]

what environment did you guys reproduce it locally? @aduh95 @fahrradflucht