Need help testing new Selector implementation on Windows

[Simone Bordet]

Hi,

On Thu, May 20, 2021 at 9:47 AM Alan Bateman <Alan.Bateman at oracle.com> wrote:
>
> On 19/05/2021 22:28, Simone Bordet wrote:
> > We have setup a Jetty 10 + OpenJDK 17+22 on our Jenkins:
> > https://jenkins.webtide.net/job/nightlies/job/jetty-10.0.x-windows-nightly-jdk17/
> >
> > There are a few tests failing mostly due to filesystem differences
> > (deleting files that are in use, etc.) so it's not the cleanest of the
> > builds, but we'll try to spend some time on it to improve it.
> > The NIO code is tested quite thoroughly though.
> > I'm assuming the new implementation is the default and there is
> > nothing to configure to enable it, right?
> >
> Yes, it's the default so no system property or other switch.
>
> It's good to see that this Jenkins jobs is setup, thanks for that. The
> call for additional helping test is mostly to make sure that swapping
> out the underlying implementation doesn't change behavior or introduce a
> regression. It would be a bonus to get more performance testing. We
> expect the new implementation to scale significantly much better than
> the old implementation as the number of registered socket channels
> increases.

We are doing some performance testing, but I can't promise we will do
them on Windows :)

In case we do, I'll follow up.

-- 
Simone Bordet
---
Finally, no matter how good the architecture and design are,
to deliver bug-free software with optimal performance and reliability,
the implementation technique must be flawless.   Victoria Livschitz