Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: incorrect timers-promisified test case #37425

Closed
wants to merge 10 commits into from

Conversation

ttzztztz
Copy link
Contributor

@ttzztztz ttzztztz commented Feb 18, 2021

This test case is incorrect, and it caused too many unit-test failures

https://ci.nodejs.org/job/node-test-commit-linuxone/nodes=rhel7-s390x/25697/testReport/
https://ci.nodejs.org/job/node-test-commit-linuxone/nodes=rhel7-s390x/25708/testReport/
...

You can see it in this page: https://ci.nodejs.org/job/node-test-commit-linuxone/nodes=rhel7-s390x/
This is probably caused by a wrong time-sequence execution order. I'm trying to work on this.

Analysis

  setPromiseTimeout(1).then(() => pre = true);  // L1
  const iterable = setInterval(() => {}, 2); // L2
  //...
  iterator.next().then(...); // L3
  setPromiseTimeout(3).then(() => post = true); // L4

Assume we execute this code at time 0
Time=0, FinishExecute=L1, callback L1 is queued to execute at Time=1
Time=1, FinishExecute=L2, callback L2 is queued to execute at Time=3, 5, 7, ...
Time=2, FinishExecute=L3
Time=11, FinishExecute=L4, callback L3 is queued to execute at Time=14

In this case, this unit test will throw an error 'second interval ran too early', but it is OK.

My solution

First, the timeout is 10, 20, 30 to tolerate the time for code-execution
Second, use the Promise.all to wait for these three promises

I think it would work.

Fixes: #37395

@nodejs-github-bot nodejs-github-bot added the test Issues and PRs related to the tests. label Feb 18, 2021
@Trott
Copy link
Member

Trott commented Feb 18, 2021

@benjamingr @Linkgoron @nodejs/timers

@Trott
Copy link
Member

Trott commented Feb 18, 2021

Refs: #37395

@Linkgoron
Copy link
Member

In retrospect, I'm not 100% sure that the test is really necessary, as there are already some timing tests.

However, while the test fix looks fine, I'm not sure what the policy is about test performance though, is ~400ms for a test OK?

@ttzztztz
Copy link
Contributor Author

ttzztztz commented Feb 18, 2021

In retrospect, I'm not 100% sure that the test is really necessary, as there are already some timing tests.

However, while the test fix looks fine, I'm not sure what the policy is about test performance though, is ~400ms for a test OK?

After looking at other test cases, I found this

const TIMEOUT = common.platformTimeout(100)

The definition of the function platformTimeout is

function platformTimeout(ms) {
  const multipliers = typeof ms === 'bigint' ?
    { two: 2n, four: 4n, seven: 7n } : { two: 2, four: 4, seven: 7 };

  if (process.features.debug)
    ms = multipliers.two * ms;

  if (isAIX)
    return multipliers.two * ms; // Default localhost speed is slower on AIX

  if (process.arch !== 'arm')
    return ms;

  const armv = process.config.variables.arm_version;

  if (armv === '6')
    return multipliers.seven * ms;  // ARMv6

  if (armv === '7')
    return multipliers.two * ms;  // ARMv7

  return ms; // ARMv8+
}

400ms is a long time, I admit. So maybe I can adjust it to common.platformTimeout(...) so that the timeout can be adjusted in different platforms.

Copy link
Contributor

@aduh95 aduh95 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit

test/parallel/test-timers-promisified.js Outdated Show resolved Hide resolved
test/parallel/test-timers-promisified.js Outdated Show resolved Hide resolved
test/parallel/test-timers-promisified.js Outdated Show resolved Hide resolved
ttzztztz and others added 3 commits February 18, 2021 17:17
Co-authored-by: Antoine du Hamel <duhamelantoine1995@gmail.com>
Co-authored-by: Antoine du Hamel <duhamelantoine1995@gmail.com>
Co-authored-by: Antoine du Hamel <duhamelantoine1995@gmail.com>
@aduh95
Copy link
Contributor

aduh95 commented Feb 18, 2021

LinuxOne CI is passing: https://ci.nodejs.org/job/node-test-commit-linuxone/nodes=rhel7-s390x/25734/

400ms is a long time, I admit. So maybe I can adjust it to common.platformTimeout(...) so that the timeout can be adjusted in different platforms.

That's sounds like a solid plan! Can you make this change?

@ttzztztz
Copy link
Contributor Author

LinuxOne CI is passing: https://ci.nodejs.org/job/node-test-commit-linuxone/nodes=rhel7-s390x/25734/

400ms is a long time, I admit. So maybe I can adjust it to common.platformTimeout(...) so that the timeout can be adjusted in different platforms.

That's sounds like a solid plan! Can you make this change?

Implemented 51921f73890790322f58c1b263265c39d10c9c01, waiting for the test result.

@ttzztztz
Copy link
Contributor Author

@aduh95 can you run the test-linuxone test for the new changes once again?

@nodejs-github-bot
Copy link
Collaborator

@aduh95 aduh95 added author ready PRs that have at least one approval, no pending requests for changes, and a CI started. fast-track PRs that do not need to wait for 48 hours to land. labels Feb 19, 2021
@aduh95
Copy link
Contributor

aduh95 commented Feb 19, 2021

Fast-track?


setPromiseTimeout(3).then(() => post = true);
const time_unit = common.platformTimeout(50);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Optional nit: Since the problem we've been seeing is with the test being flaky on a fast machine, I wonder if common.platformTimeout() isn't needed here. It's usually for giving slow machines more time, but the problem we're seeing right now is the other way around--our LinuxONE host is too efficient.

@Trott
Copy link
Member

Trott commented Feb 20, 2021

@nodejs/timers

@nodejs-github-bot
Copy link
Collaborator

targos pushed a commit to targos/node that referenced this pull request Feb 25, 2021
PR-URL: nodejs#37425
Fixes: nodejs#37395
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>
Reviewed-By: Richard Lau <rlau@redhat.com>
@targos
Copy link
Member

targos commented Feb 25, 2021

Landed in 63794b4

@targos targos closed this Feb 25, 2021
targos pushed a commit that referenced this pull request Feb 28, 2021
PR-URL: #37425
Fixes: #37395
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>
Reviewed-By: Richard Lau <rlau@redhat.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
author ready PRs that have at least one approval, no pending requests for changes, and a CI started. fast-track PRs that do not need to wait for 48 hours to land. test Issues and PRs related to the tests.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

investigate flaky test-timers-promisified
7 participants