JDK-8191948, 8u211 and push confusion

[David Holmes]

Hi Aleksey,

On 22/02/2019 4:03 am, Aleksey Shipilev wrote:
> Hi,
> 
> I am so confused how do 8u tags correspond with actual fixes delivered.
> 
> Sifting through Christoph list:
>    https://bugs.openjdk.java.net/issues/?filter=36394
> 
> Look at this bug:
>    https://bugs.openjdk.java.net/browse/JDK-8191948
> 
> It has a single backport with fix version "8u211":
>    https://bugs.openjdk.java.net/browse/JDK-8214637

It actually has multiple backports but most of them are not public.

> ...which is apparently Oracle-internal 8u branch? The backport issue indeed has no hgupdater comment
> that goes with pushes to repositories, which seems to be usual for these internal pushes. 

The hgupdater comment is non-public hence you can't see it.

> Yet, here is the push:
>    http://hg.openjdk.java.net/jdk8u/jdk8u/jdk/rev/b88f126a9690
> 
> ...which apparently got into jdk8u202-b01?!
> 
> http://hg.openjdk.java.net/jdk8u/jdk8u/jdk/file/jdk8u202-b01/src/share/classes/com/sun/tools/jdi/ReferenceTypeImpl.java#l1114
> 
> How so? Is it the hgupdater bug? Was 8191948 backport pushed accidentally to the 8u202?

For some reason the 8u202 backport issue is non-public, despite the 
changesets going to the OpenJDK 8u forests. I'm querying that to see if 
it can be changed.

At the same time the 8u211 Oracle backport is a public issue for a 
non-public push.

Both were created by hgupdater but I have no idea why it chose the 
settings it did. There may be a glitch there that needs to be looked into.

HTH,
David

> -Aleksey
> 
>