aarch32 project process

[Alex Kashchenko]

Hi Andrey,

On 06/02/2016 11:57 AM, Andrey Petushkov wrote:
> Dear Ed,
>
> Thank you for your letter! IMHO this all makes perfect sense and hope it
> helps us moving forward quicker.
> With this respect may I ask you to promote Sergey Nazarkin (snazarki) into
> committer role for the project. Although there are other people from Azul
> being listed as commiters, they are really busy with other stuff, so it
> would be nice to have at least one person who's really involved in the
> project. At the same time I believe Sergey has enough contributions to
> formally match the criteria.
>
> Best Regards,
> Andrey
>
> PS. Alex, so we're counting on you, as you seem to be the only other
> non-Azul person who's active on this project :)

I am reading Ed's email as a "call to arms" addressed to Red Hat. I am 
obviously non-qualified to comment on its essence, so I will comment 
only on myself to clarify the current state of things to non-RH 
participants.

While I am employed at RH as a Senior Software Engineer, aarch32 is only 
a part of my duties (and generally less than 1/2 of time) and in aarch32 
I am effectively doing a QA job (maintaining a "build farm" of various 
arm dev boards). I know how to build and debug jdk and can do some arm32 
assembler, but my hotspot experience is virtually non-existent (limited 
to aarch32 only). So if you (and Ed) are looking for the non-Azul 
committer (who effectively will be a reviewer for most of the changes) 
you most probably should look for someone else with a proper hotspot 
experience. And I will continue with testing and "fix-the-build"-type 
patches.

>
> On Wed, Jun 1, 2016 at 7:17 PM Edward Nevill <edward.nevill at gmail.com>
> wrote:
>
>> Hi all,
>>
>> I would like to discuss the aarch32 project process. IE. the process for
>> submitting patches for aarch32 and having those patches approved and
>> pushed.
>>
>> It would seem that I have become a defacto 'gatekeeper' and this was not
>> my intention. There are currently 6 committers on the project, however I
>> seem to be the only committer pushing patches.
>>
>> Unfortunately, I have been quite busy at work at the moment and as this
>> work is not sponsored by my employer, I have to try to find 'spare' time
>> to do it. I expect that I will have more time available for this work
>> after June 14.
>>
>> In order that my tardiness does not delay anyone I would like to propose
>> the following.
>>
>> - Since we have no formal Reviewers on the project any patch may be
>> approved by any other committer. I think it would be preferable that the
>> other committer does not work for the same company.
>>
>> - Any committer can then push a patch once it has been approved as
>> above. It would be good practice to allow 24 hours before pushing a
>> patch after it has been proposed in order to allow sufficient time to
>> review the patches.
>>
>> - It greatly eases the process of reviewing/pushing patches if the
>> patches are submitted using the 'webrev' tool (see
>> http://openjdk.java.net/guide/webrevHelp.html) and if the patch is
>> committed before running webrev. In this case the patch becomes a
>> changeset, in which case the reviewer/committer can simply do "hg import
>> <patch>" followed by "hg push". However, I understand that it is
>> sometimes not possible for contributors to create changesets if they do
>> not have OpenJDK IDs.
>>
>> - At the moment we are creating OpenJDK bugids for every change. This
>> has some value in that it is possible to look at a change, refer back to
>> the original bug report to see what the actual problem was. Also, if the
>> bug ID is mentioned in the header of the email to the list it makes it
>> easy to trace any discussion regarding the change. For this reason I
>> would like to keep the policy of having OpenJDK bugids, but I do
>> understand that it is an extra administrative overhead and am open to
>> persuasion.
>>
>> At the moment the list of committers are
>>
>> http://openjdk.java.net/census#adinn
>> http://openjdk.java.net/census#aph
>> http://openjdk.java.net/census#bae
>> http://openjdk.java.net/census#jjoyce
>> http://openjdk.java.net/census#vkarnauk
>> http://openjdk.java.net/census#enevill
>>
>> If there is a need for additional committers I am quite happy to propose
>> additional committers. Ideally any proposed committers should already
>> have OpenJDK ids and should have contributed some patches to OpenJDK (I
>> think the magic number is 8, but there is some flexibility with this,
>> especially for non trunk projects such as the aarch32 port).
>>
>> In the meantime I will try to get to reviewing the outstanding patches
>> ASAP.
>>
>> All the best,
>> Ed.
>>
>>
>>


-- 
-Alex