[8u] RFR: 8226575: OperatingSystemMXBean should be made container aware
Severin Gehwolf
sgehwolf at redhat.com
Tue Aug 25 15:42:47 UTC 2020
On Tue, 2020-08-25 at 16:24 +0100, Andrew Hughes wrote:
> On 14:29 Mon 24 Aug , Severin Gehwolf wrote:
> > On Fri, 2020-08-21 at 19:37 +0100, Andrew Hughes wrote:
> > > > > There is no Windows file in the 11u version. Why is it needed in 8u?
> > > > > In both, OperatingSystemImpl.java seems to be under
> > > > > src/jdk.management/unix/classes or src/solaris/classes/sun/management,
> > > > > rather than share/classes.
> > > >
> > > > Good point. There is a Windows specific OperatingSystemImpl.java with
> > > > references to native methods. I missed that. Note to self: No need to
> > > > change the native method impls as they've not changed
> > > > in src/windows/classes/sun/management/OperatingSystemImpl.java either.
> > > > Reverted in the updated webrev.
> > >
> > > Thanks. It took me a while to spot it as well. At first, I thought Windows
> > > was picking it up from one common shared Java file like the others, but
> > > then I looked at the paths again and spotted the unix/solaris usage.
> > >
> > > > Latest webrev here:
> > > > http://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-8226575/jdk8/02/webrev/
> > > >
> > >
> > > Thanks. JDK side looks good now.
> >
> > Thanks again for the review!
> >
> > > I'll wait for your response on
> > > whether to do the other change first for the HotSpot side.
> >
> > Please, lets not delay this any longer for a nice-to-have test cleanup.
> >
>
> Have you checked if the HotSpot test fix applies cleanly?
Yes. It depends on a number of fixes before it. Then it applies cleanly
but doesn't compile.
Can I consider this patch reviewed?
> If it does, we may as well quickly approve it and get it in.
Unfortunately not the case. It'll need a review.
> If not, there's no advantage to doing it first anyway, as it would
> need a review with or without these additional changes.
+1
> My concern is turning a simple patch into one which needs review
> because it is completed out of order.
Fair enough.
> Incidentally, this RFR may well have been posted a month ago, but it
> wasn't actionable without the dependent patch first being reviewed and
> approved.
Why? We already knew at the time that we want this patch (parity
patch). The initial webrev didn't change (with the dependency in or
not). As such, it could have been reviewed in my opinion.
Thanks,
Severin
More information about the jdk8u-dev
mailing list