jdk10/hs -> jdk10/jdk10 integrations?

[Volker Simonis]

On Fri, Jul 7, 2017 at 12:43 AM, Joseph D. Darcy <joe.darcy at oracle.com> wrote:
> Hi Aleksey,
>
> On 7/6/2017 10:08 AM, Aleksey Shipilev wrote:
>>
>> Hi Joe,
>>
>> On 07/06/2017 06:56 PM, joe darcy wrote:
>>>
>>> On 7/6/2017 9:23 AM, Aleksey Shipilev wrote:
>>>>
>>>> I am trying to understand the relationship between jdk10/hs and
>>>> jdk10/jdk10.
>>>> Naively, I would assume it has the same relationship as jdk9/dev and
>>>> jdk9/jdk9:
>>>> e.g. the development is going on in /dev or /hs, and then jdkX/jdkX
>>>> serves as
>>>> integration+release repository, and then regular integrations are
>>>> happening to
>>>> jdkX/jdkX.
>>>
>>> The relationship that is naive to assume is the correct one, but isn't
>>> fully
>>> implemented yet since, as you've noticed, there haven't been any hs ->
>>> master
>>> integrations in JDK 10 yet. (We're working on some new internal
>>> infrastructure
>>> to support integrations and it isn't fully in place.)
>>>
>>> Having integrations occurring is a precondition for the planned repo
>>> consolidation, which is targeted for September:
>>>
>>> http://mail.openjdk.java.net/pipermail/jdk10-dev/2017-June/000371.html
>>
>> So, that means jdk10/hs -> jdk10/jdk10 integration would start happening
>> *before* repo consolidation, right?
>
>
> Correct; both hs and client will need to integration to jdk10/jdk10 before
> the consolidation.
>

What about the other way round? It seems that jdk10/hs and
jdk10/client are still synced with jdk10/jdk10. Is that correct?

I'm asking because I want to push a change to the top-level test
directory. So if I push that to jdk10/jdk10 it will eventually arrive
in jdk10/hs even before the jdk10/hs -> jdk10/jdk10 integration would
start, right?

Regards,
Volker

>>
>> We can use jdk10/hs as the basis for Shenandoah, because we will have to
>> integrate through that anyway. Too bad jdk10/sandbox does not have
>> jdk10/hs
>> changes too, it seems:
>>    http://hg.openjdk.java.net/jdk10/hs/hotspot/rev/ae91ec8b554a
>>    http://hg.openjdk.java.net/jdk10/sandbox/hotspot/log?rev=ae91ec8b554a
>>
>> Which means whatever we are doing in sandboxes is not getting exposed to
>> jdk10
>> mainline work...
>>
>
> The sandbox is being synced with jdk10/jdk10.
>
> Cheers,
>
> -Joe
>