[jdk8u-dev] RFR: 8285497: Add system property for Java SE specification maintenance version [v2]

[Joe Darcy]

On Wed, 24 Aug 2022 18:23:33 GMT, Andrew John Hughes <andrew at openjdk.org> wrote:

> > > Change looks good and nearly the same as that in 8u42. However, one copyright header change was missed in `System.c`, presumably due to 8189761 already having bumped it to 2019. Can we fix this please?
> > > Also, please change the title to "Backport 31a63ba5f255e09349b3842984ac5bb3ad8e6c0b" so SKARA correctly identifies this as a backport.
> > 
> > 
> > I've done as you've requested.
> > > See https://wiki.openjdk.org/display/SKARA/Backports
> > 
> > 
> > Yes, I am familiar with the basic operation of Skara. I don't see anything on the referenced page that requires use of a Skara Backport PR when doing a backport. The impetus for adding the backport feature to Skara, a feature enabled several months after JDK mainline switched to Skara, was to reduce the overhead of the common case of backport a fix to an earlier release where few to no changes are needed.
> > Before sending out this PR, I looked for, but did not find, documentation for what procedures 8u wanted to use for a situation like this.
> 
> Thanks for the feedback. I can improve the documentation on the 8u wiki to make this  #clearer. If Skara doesn't recognise the issue as a backport, it doesn't seem to like the referenced bug. Prior to you updating the title, it was issuing a warning that the bug was not open (as we'd expect), so I'm not sure it would correctly resolve it with a backport issue when pushed. Using the `"Backport <x>"` nomenclature just seems to make things work well in all cases.

I also noticed the previous message about "bug not being open" before the PR was changed to a backport. I assume that is a opportunity-for-improvement for the bots to say "we noticed we'll need to create a new backport record if this gets pushed," which the bots should certainly be able to do.

My general advice on using JBS includes "don't create a backport record before it is needed." I didn't try to create an 8-pool record to see if it would making the warning go away since I assumed the right thing would happen on the push with respect to JBS.

-------------

PR: https://git.openjdk.org/jdk8u-dev/pull/100