Future jdk9u updates & 9-critical-request

[Andrew Hughes]

On 7 February 2018 at 16:33, dalibor topic <dalibor.topic at oracle.com> wrote:
> On 06.02.2018 21:51, Andrew Hughes wrote:
>>
>> So we can expect more broken releases in the future?
>
>
> Not from Oracle, as Oracle doesn't plan to release another OpenJDK 9 Update.
>
> What platforms the next set of maintainers of the jdk9u forest decide to
> test their future releases on and how that squares with other people's
> expectations is a question for them, assuming someone steps up to the task,
> of course.
>
> What took place in the past when Red Hat took over leadership of update
> releases of 6 and 7 in OpenJDK was an immediate pruning of platforms that
> releases were tested against: from Windows, Solaris, OS X and Linux to just
> Linux.
>
> If a potential new maintainer for the jdk9u forest decided to adopt a
> similar approach to re-focusing their platform coverage, I would not be
> surprised if future jdk9u releases wouldn't build or work on platforms the
> next set of maintainers didn't chose to build and test on, either.
>
> But even if they did test their future release on all possible platforms, a
> source code release could fail to build downstream using a different, newer
> (or older) GCC version, a different version of a native library, etc.
>
> In short, someone is likely always going to be able to claim that a given
> release is broken for them in one way or another. They can then use the push
> approval process to contribute the corresponding changes back.
>

Actually we do test on Windows; I've just received news of a build failure
on the 7u171 we're working on.

I believe there are problems with this process in JDK projects maintained
both by Oracle and by others. We need to address these issues rather
than attempting to assert that the current process works well enough.

>> I find they are roughly similar for searching. Maybe you need a better
>> mail client if not?
>
>
> Indeed, I don't doubt that local search in your e-mail client works as well
> as in mine ;)
>
> But I am not aware of a mail client that lets me easily share a URL to a
> search of approved push requests like JBS does:
> https://bugs.openjdk.java.net/issues/?jql=labels%20%3D%209-critical-approved%20ORDER%20BY%20updated%20DESC
> .
>
> As with so many other things, it's the ability to share results that makes
> all the difference. ;)
>

Sharing was irrelevant in the use case I was describing, but this is already
vastly off-topic.

>> Well, it shows a couple of comments from Severin that would be the
>> equivalent
>> of an initial e-mail to the list. My concern is the lack of the part
>> after that; there's
>> no indication of who approved it or why.
>
>
> You need to go to the history tab to see who modified an issue in what way
> and when they did so.
>

It's the why I'd want to know, not the what or when.

>> Being an approval, that may be a bad example. Is there a case where an
>> issue
>> was rejected with some explanation?
>
>
> I'm not aware of a rejected push approval for jdk9u so far. I do agree that
> rejection should come with an explanation, though.
>

There are two issues I see with handling this process on bug reports,
one of which is potentially resolvable, and the other which I believe is
a disadvantage over the previous system of using e-mails:

1. There is little transparency as to why decisions are made on the bug
reports.  As we agree, rejections should come with an explanation. I
would also hope we would see the same discussion we used to see on
the mailing lists when the backport is non-trivial. This is all perfectly
possible, but it remains to be seen whether it happens or not. My
experience so far is not great.

2. Once you locate discussion on specific bug reports rather than
a general mailing list, there is less chance of input from casual observers.
I've often checked on the progress of a discussion on the 8u list and ended
up reading reviews of other bugs I would not otherwise be aware of.

Now I know you'll say people can search for the relevant tag instead,
but that still requires some direct action. You're not going to see these
things by chance when looking for something else, which I feel is a loss.

>
> cheers,
> dalibor topic
>
> --
> <http://www.oracle.com> Dalibor Topic | Principal Product Manager
> Phone: +494089091214 <tel:+494089091214> | Mobile: +491737185961
> <tel:+491737185961>
>
> ORACLE Deutschland B.V. & Co. KG | Kühnehöfe 5 | 22761 Hamburg
>
> ORACLE Deutschland B.V. & Co. KG
> Hauptverwaltung: Riesstr. 25, D-80992 München
> Registergericht: Amtsgericht München, HRA 95603
>
> Komplementärin: ORACLE Deutschland Verwaltung B.V.
> Hertogswetering 163/167, 3543 AS Utrecht, Niederlande
> Handelsregister der Handelskammer Midden-Niederlande, Nr. 30143697
> Geschäftsführer: Alexander van der Ven, Jan Schultheiss, Val Maher
>
> <http://www.oracle.com/commitment> Oracle is committed to developing
> practices and products that help protect the environment



-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Web Site: http://fuseyism.com
Twitter: https://twitter.com/gnu_andrew_java
PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222