On Thu, Apr 23, 2020 at 2:18 AM James Perkins <jperkins@redhat.com> wrote:
>
> Okay all here's what I've got https://github.com/resteasy/Resteasy/compare/master...jamezp:github-actions [1]. This passed all jobs [2]. This does skip the MP TCK tests on Windows for now. We can address that later if desired.
Awesome!
>
> One question I have is how to deal with [3]. I need to file a JIRA still, but it seems we need a way to gracefully shutdown. With what I've currently done in the commit it uses a new thread instead of being bound to the HttpServer thread. We could add a constructor property and use the `HttpServer.stop(int)` with that value. Any opinions on this?
Yes we can deal that in separate JIRA issue IMHO.
Excellent. I submit a PR
https://github.com/resteasy/Resteasy/pull/2367 which does contain a fix. It seems we should likely allow for the delay to be set somehow, but we already don't allow this in any of the other HttpJaxrsServer's.
>
> Once the above is decided and I file a JIRA I can file a PR against master and any other branch we want to use GitHub Actions for CI on.
>
> [1]: https://github.com/resteasy/Resteasy/compare/master...jamezp:github-actions
> [2]: https://github.com/jamezp/Resteasy/actions/runs/84947652
> [3]: https://github.com/resteasy/Resteasy/commit/8c77f5b1943fc2c6196f3a249c15d8aac7c5a30e
>
> On Wed, Apr 22, 2020 at 7:38 AM James Perkins <jperkins@redhat.com> wrote:
>>
>> I've had luck with Linux on both Java 8 and Java 11. Windows has been a bit more of an issue, but I may be close at least with Java 8 on Windows. I'm currently down to two failures that I'm testing a fix for now. One of the fixes does need some discussion.
>>
>> It might make sense to disable the MicroProfile TCK for Windows as really we should only need to test that on one environment. It only seems to be a bit of an issue on Windows with Java 11. That is something I can look at though and then we could re-enable if it gets fixed upstream. The errors seem to be timing issues so it could just be racy tests.
>>
>> On Tue, Apr 21, 2020 at 6:32 PM Wei Nan Li <weli@redhat.com> wrote:
>>>
>>> My suggestion is that we can submit a initial working setup on single
>>> environment(jdk8, for example), to make it start working, and then
>>> gradually evolving our setup until we can get rid of TravisCI.
>>>
>>> @Alessio Soldano @Ronald Sigal wdyt?
>>>
>>> On Tue, Apr 21, 2020 at 11:34 PM James Perkins <jperkins@redhat.com> wrote:
>>> >
>>> > Hello All,
>>> > I've submitted a draft for an example of how we could use GitHub Actions to run CI. The main benefit over TravisCI is we can run Windows jobs as well. I'm curious what others think of this approach.
>>> >
>>> > One thing to note is I've seen some issues, especially on Java 8 on Windows, with the resteasy-jdk-http tests. We may need to ignore the test in that environment which I can make a change to do if we notice it being an issue.
>>> >
>>> > There's also another commit I'd like to get in regardless of using GitHub Actions which allows various tests to pass on other environments such as Windows.
>>> >
>>> > [1]: https://github.com/resteasy/Resteasy/pull/2362
>>> >
>>> > --
>>> > James R. Perkins
>>> > JBoss by Red Hat
>>> > _______________________________________________
>>> > resteasy-dev mailing list
>>> > resteasy-dev@lists.jboss.org
>>> > https://lists.jboss.org/mailman/listinfo/resteasy-dev
>>>
>>>
>>>
>>> --
>>> Cheers,
>>> Wei Nan | JBoss
>>>
>>
>>
>> --
>> James R. Perkins
>> JBoss by Red Hat
>
>
>
> --
> James R. Perkins
> JBoss by Red Hat
--
Cheers,
Wei Nan | JBoss