Preparation of update releases

[Volker Simonis]

On Fri, Oct 19, 2018 at 6:35 PM Phil Race <philip.race at oracle.com> wrote:
>
> I've seen this too .. some CPU fixes that I did in 2017 and were in 11 GA
> were somehow just synced again into 11u creating duplicate changesets
> for the fix and consequently additional backport records in JBS.
>

Yes, that's exactly what I've observed as well.
Can somebody provide an explanation for this?

> -phil.
>
> On 10/19/2018 09:04 AM, Volker Simonis wrote:
> > I think one of the problems is that 11.0.1 brings in a lot of changes
> > which are already in 11, but with a different hash. E.g.
> >
> > simonis at simonis:/OpenJDK/jdk-jdk11u$  hg log -k 8191907
> > changeset:   51235:9c1534b2e82c
> > user:        mdoerr
> > date:        Tue Nov 28 01:08:26 2017 +0300
> > summary:     8191907: PPC64 and s390 parts of JDK-8174962: Better
> > interface invocations
> > Reviewed-by: goetz
> >
> > changeset:   48649:d9fcb7ba8133
> > user:        mdoerr
> > date:        Tue Nov 28 01:08:26 2017 +0300
> > summary:     8191907: PPC64 and s390 parts of JDK-8174962: Better
> > interface invocations
> > Reviewed-by: goetz
> >
> > "48649:d9fcb7ba8133" is the change which was already in jdk 11 and for
> > some reason it was brought in again with 11.0.1 (as
> > "51235:9c1534b2e82c"). That's strange because 8191907 was pushed in
> > Nov. 2017 which should have been long before jdk 11.0.1 was cloned
> > from the jdk11 repo and the jdk11 repo should have already contained
> > 8191907 at that time. So why has this change been push one more time
> > to 11.0.1 ?
>