Proposal: Out-of-order weekly release? Just another attempt

[Oleg Nenashev]

Hi,

According to my previous experience, I assume that nobody in this list cares about the Weekly releases. And you're doing right, LTS is the only way to get enough stability. But I want to try again and to propose out-of-order release (2.30).

Why?

• Many issues opened to JENKINS-39414, blocker for any instances having Ruby Runtime
  
• LTS baseline selection is coming
• Starting from 2.26 weekly releases are screwed up by regressions
• If we do not resolve the fallout, we may have to select 2.25 as the new LTS baseline
• In such case major improvements like JNLP4 won't get into the release.
  
• According to Vivek, it's also going to impact the anticipated BlueOcean public release

What to fix?

• JENKINS-39414 - has not been completely fixed in 2.29
• Stapler needs to be fixed again, PR #2622
• JENKINS-39465 - Critical bug, which blocks the JNLP4 adoption
• PR #2621 has two +1s
• Not important: Several other minor bugfixes, which could be integrated by tomorrow
  

When?

• I propose to do it tomorrow (Nov 8)

What do you think about it?

Best regards,
Oleg

[Daniel Beck]

> On 07.11.2016, at 19:11, Oleg Nenashev <o.v.ne...@gmail.com> wrote:
>
> What do you think about it?

There's also JENKINS-39555. 2.30 without a fix for that seems pointless.

Screwing up a weekly due to a big change once in a while happens, but we should work towards making 2.30 not suck, LTS baseline or not.

Given the scope and that the past weekly didn't fix it, an out of order release would be useful, assuming we can get all of this fixed before Thursday or so.

[Arnaud Héritier]

+1 for an additional release

(+2 to use LTS in production :-P)

--
You received this message because you are subscribed to the Google Groups "Jenkins Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-dev+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-dev/1da7822a-5270-4cf6-8b5e-f322dfb37cb0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

-----

Arnaud Héritier

http://aheritier.net

Mail/GTalk: aheritier AT gmail DOT com

Twitter/Skype : aheritier

[Oleg Nenashev]

Yes, JENKINS-39555 is a must-have. Hopefully your ongoing triangulation will find a hotfix for it.

In general, we must include ATH in our weekly release Pipeline. It should have caught all issues we see.


понедельник, 7 ноября 2016 г., 19:17:10 UTC+1 пользователь Arnaud Héritier написал:

To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-de...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-dev/1da7822a-5270-4cf6-8b5e-f322dfb37cb0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Samuel Van Oort]

I'm looking into JENKINS-39555 now -- for some reason the usual JIRA notification wasn't delivered.  

[Samuel Van Oort]

+1 - needs an out of order release.  This causes 2 major breakages.   

A small aspect of JENKINS-39404 is responsible for JENKINS-39555, which badly breaks the whole pipeline plugin suite (and goodness knows what else). 

I have a fix for that which works, but am ensuring it doesn't have side effects.

On Monday, November 7, 2016 at 1:11:19 PM UTC-5, Oleg Nenashev wrote:

[Jesse Glick]

On Mon, Nov 7, 2016 at 1:28 PM, Oleg Nenashev <o.v.ne...@gmail.com> wrote:
> we must include ATH in our weekly release Pipeline. It should
> have caught all issues we see.

Confirmed that ATH would have caught JENKINS-39555, since the symptoms
are so severe. In general though ATH only covers a tiny fraction of
plugin functionality; functional tests would catch far more
regressions (including this one), which is why I consider
`plugin-compatibility-tester` important.

[Oleg Nenashev]

So we have all 3 patches ready for review (internal reviewbybees process have been addressed):

• https://github.com/jenkinsci/jenkins/pull/2624
  
• https://github.com/jenkinsci/jenkins/pull/2622
  
• https://github.com/jenkinsci/jenkins/pull/2621
  

Would be great if somebody else reviews them. If we get enough +1s, we may be probably able to release the fix even today.

Thanks in advance,
Oleg


понедельник, 7 ноября 2016 г., 21:00:24 UTC+1 пользователь Samuel Van Oort написал:

[Oleg Nenashev]

Hi,

So we have at least 4 +1s in all PRs + a kind of "the approach looks good to me " from orrc in IRC. I think it's enough to proceed taking the current community ratings into account (2 positive vs. 9 negative).

Kohsuke is available to spin the release, and I think it's better to do it sooner than later.

BR, Oleg

понедельник, 7 ноября 2016 г., 21:57:28 UTC+1 пользователь Oleg Nenashev написал:

[Baptiste Mathus]

+1. I've already seen many users asking for help about that on IRC and on ml. Releasing earlier totally makes sense.

Thanks a lot for the hard work Oleg. The community owes you much.

--

You received this message because you are subscribed to the Google Groups "Jenkins Developers" group.

To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-dev+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-dev/90974857-f373-46d2-9e33-6a4d851217e1%40googlegroups.com.

[Oleg Nenashev]

Hi,

All the fixes are integrated. Changelogs for 2.30 are posted as well (here).
We are ready to roll out.

Thanks a lot for the hard work Oleg. The community owes you much.

Ah, thanks. People can always say thanks by contributing stuff to the Jenkins project ^_^

BR, Oleg

вторник, 8 ноября 2016 г., 0:15:46 UTC+1 пользователь Baptiste Mathus написал:

To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-de...@googlegroups.com.

[Oleg Nenashev]

Hi,

The 2.30 release has been published. Thanks a lot to all contributors!

The new LTS baseline selection is coming soon. My proposal would be to be more conservative for the next two weeks and to avoid merging potentially breaking changes (e.g. this guy), which need more soaking in weekly releases. In the Jenkins core we have the "on-hold" label, which can be used for indicating such PRs.

BR, Oleg


вторник, 8 ноября 2016 г., 0:33:02 UTC+1 пользователь Oleg Nenashev написал: