Add AppVeyor configuration

[Daniel15]

The build doesn't actually work properly yet (a bunch of tests fail: https://ci.appveyor.com/project/Daniel15/jest/build/1.0.5/job/wjnpa9thok6twp5y) but it's worth at least starting to work on this. Once all the broken tests are fixed on Windows builds, having a continuous integration setup like this will prevent it from breaking again in the future.

Closes #1192

[Daniel15]

This is supposed to be stacked on top of #1463, but I forgot that Github is not Phabricator and doesn't support basic features such as that. The only change in this pull request is supposed to be the appveyor.yml file.

[codecov-io]

Current coverage is 90.31% (diff: 100%)

No coverage report found for master at 46868fc.

Powered by Codecov. Last update 46868fc...8425a27

[cpojer]

are you interested in fixing up all the tests to work on windows?

[Daniel15]

Sure, I can take a look when I get a chance.

[Daniel15]

@cpojer - What do you think of this approach for now? I simply make Windows builds skip all the tests that fail on Windows. It's not ideal and I do plan on going through and fixing the tests incrementally, but at least it makes a Windows build actually possible and will still help catch some regressions.

It works! https://ci.appveyor.com/project/Daniel15/jest/build/7/job/apivm49w0c7nm4hc

[cpojer]

let's just not support node 4 on windows. I don't think we'll be able to do this.

[Daniel15]

Are there any issues with Node 4 that would cause it to break?

[cpojer]

No but I'm deciding right now that the only node version we support on Windows will be the latest.

[cpojer]

Can we instead just dump the skipOnWindows file in scripts/ and add an @providesModule header? I'd like it not to be part of the actual jest packages.

[cpojer]

In testSetupFile.js you could also do:

jest.mock('skipOnWindows', () => {
  return implementation.
}, {virtual: true});

[Daniel15]

I'm having some trouble getting Jest running on my work laptop, so I'll do those changes tonight on my home PC instead 👍

[cpojer]

Thank you @Daniel15!

One problem we often run into is that Jest has an issue with test paths and can't find any more tests. The exit code in this case is 0, so it says everything is fine. Can we add an additional check to appveyor that makes sure at least one test was run?

[cpojer]

We have an integration test for that, however that uses Jest too, so if Jest breaks like this, we won't actually know :D

[Daniel15]

Hmm, that's tricky... I feel like the exit code should actually be non-zero
if Jest doesn't find any tests. Can we change that behaviour or add a
--fail-on-no-tests flag that'd do it?

Sent from my phone.

On Sep 3, 2016 1:53 AM, "Christoph Pojer" [email protected] wrote:

Thank you @Daniel15 https://github.com/Daniel15!

One problem we often run into is that Jest has an issue with test paths
and can't find any more tests. The exit code in this case is 0, so it says
everything is fine. Can we add an additional check to appveyor that makes
sure at least one test was run?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#1465 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAFnHaOB2ebBFJeeuWoUCeakdkgfoHMUks5qmTV0gaJpZM4JoL_3
.

[cpojer]

I'm happy to accept a PR that adds this flag, yes. I'm worried about this having a bad impact on www where our test infra might try to run something that isn't actually a test; we don't want to fail and require changes to the test infra there.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Please note this issue tracker is not a help forum. We recommend using StackOverflow or our discord channel for questions.