Backports to 1.509.2 done, please post RC

[Vojtech Juranek]

Hi,
backports to 1.509.2 are done and passed all automated tests [1].
As discussed on past governance meeting [2], only soaked fixes (contained in main release) were backported:

JENKINS-17402 - Failed to instantiate class hudson.plugins.copyartifact.CopyArtifact
JENKINS-17721 - StringIndexOutOfBoundsException in PackageResult.findCorrespondingResult
JENKINS-17724 - /about no longer shows third-party licenses
JENKINS-17728 - NPE from MatrixConfiguration.newBuild

Kohsuke, could you please port RC?
Thanks
Vojta

[1] https://ci.jenkins-ci.org/view/Jenkins%20core/job/jenkins_lts_branch/90/
[2] http://meetings.jenkins-ci.org/jenkins/2013/jenkins.2013-05-15-18.00.html

[domi]

I'm sad to hear that JENKINS-17508 is not included in the backports
Without this one we have to stick to 1.480.3 which is a real bummer :(
If the only reason to not have this one in is the timeframe, then I'm more into waiting another week with this release until we have this one.
/imod

[Jesse Glick]

On 05/18/2013 03:28 AM, domi wrote:
> Without this one we have to stick to 1.480.3 which is a real bummer

I can think of several bugs off the top of my head which will be prime candidates for 1.509.3 once fixed in trunk. So long as vjuranek is willing to produce new LTS
versions I see no harm in sticking to the “train” model.

JENKINS-17508 could probably be fixed for you by just updating the Maven plugin, by the way. And we should look into splitting this plugin out into its own repository so
it can have an independent lifecycle and not hold core releases hostage. As far as I know the only thing stopping that from happening is compatibility in the test
harness, which is an annoyance but probably not a blocker.

[domi]

Thats true, Jesse
and splitting off the maven plugin from the main repo was a topic quite a couple of times already - maybe we should have a go on this then.
But if we do this, would it not make sense to split of the MavenTask too? Maybe a MavenBuilder/task-plugin an then the maven-plugin which extends/uses the first one.
This way we could still have basic features used by both maven integrations and have releases independent from the core (e.g. things like the SettingsProvider).
regards Domi

> --
> 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-de...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

[Vojtech Juranek]

> I can think of several bugs off the top of my head which will be prime
> candidates for 1.509.3 once fixed in trunk. So long as vjuranek is willing
> to produce new LTS versions I see no harm in sticking to the “train” model.

yes, this is intended (to produce another minor LTS). IIRC we chose 1.509 as a
base line for LTS with awareness that it contains bugs and we will produce
minor versions more often to included as much fixes as possible. Therefore I
also wanted to do an exception and backport fixes regardless they are already
in mail release line or not, but Kohsuke had different opinion [1] (and as I
hadn't any strong opinion on this, I agreed to backport only soaked fixes)

[1] http://meetings.jenkins-ci.org/jenkins/2013/jenkins.2013-05-15-18.00.log.html

[Kohsuke Kawaguchi]

So I completely dropped the ball on this one, and my apologies.

I'm going to produce the bits tonight, but given that there are still some more issues like JENKINS-15652, what if we backport some more fixes, skip 1.509.2, and call that 1.509.3 (or call that the new 1.509.2)?

I'll produce 1.509.2 RC with the current bits but while I wait to hear from others, I'll do some backporting toward 1.509.3 RC.

2013/5/19 Vojtech Juranek <vjur...@redhat.com>

--
Kohsuke Kawaguchi

[Kohsuke Kawaguchi]

I backported two more bugs and pushed 'stable-1.509.3'.

I've also expanded backend-commit-history-parser to assist the
backporting process. See
https://wiki.jenkins-ci.org/display/JENKINS/Backporting+toolkit+for+LTS
for more details.

On 06/19/2013 09:55 AM, Kohsuke Kawaguchi wrote:
> So I completely dropped the ball on this one, and my apologies.
>
> I'm going to produce the bits tonight, but given that there are still
> some more issues like JENKINS-15652, what if we backport some more
> fixes, skip 1.509.2, and call that 1.509.3 (or call that the new 1.509.2)?
>
> I'll produce 1.509.2 RC with the current bits but while I wait to hear
> from others, I'll do some backporting toward 1.509.3 RC.
>
>

> 2013/5/19 Vojtech Juranek <vjur...@redhat.com <mailto:vjur...@redhat.com>>

>
>
> > I can think of several bugs off the top of my head which will be
> prime
> > candidates for 1.509.3 once fixed in trunk. So long as vjuranek
> is willing
> > to produce new LTS versions I see no harm in sticking to the

> �train� model.

>
> yes, this is intended (to produce another minor LTS). IIRC we chose
> 1.509 as a
> base line for LTS with awareness that it contains bugs and we will
> produce
> minor versions more often to included as much fixes as possible.
> Therefore I
> also wanted to do an exception and backport fixes regardless they
> are already
> in mail release line or not, but Kohsuke had different opinion [1]
> (and as I
> hadn't any strong opinion on this, I agreed to backport only soaked
> fixes)
>
> [1]
> http://meetings.jenkins-ci.org/jenkins/2013/jenkins.2013-05-15-18.00.log.html
>
>
>
>
> --
> Kohsuke Kawaguchi
>

> --
> 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-de...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
Kohsuke Kawaguchi | CloudBees, Inc. | http://cloudbees.com/
Try Jenkins Enterprise, our professional version of Jenkins

[Vojtech Juranek]

On Wednesday 19 June 2013 11:51:58 Kohsuke Kawaguchi wrote:
> I backported two more bugs and pushed 'stable-1.509.3'.

sounds good to me. I'll be completely offline for once week since Saturaday or
so, so there could be some delay with testing (but given we have already good
amount of selenium tests, I'll maybe finished it this by Friday)

> I've also expanded backend-commit-history-parser to assist the
> backporting process. See
> https://wiki.jenkins-ci.org/display/JENKINS/Backporting+toolkit+for+LTS
> for more details.

great, thanks for that!

[Kohsuke Kawaguchi]

All right, then I'll incorporate these changes into this evening 1.509.2 RC.

[Kohsuke Kawaguchi]

Posted

--
Kohsuke Kawaguchi

[Vojtech Juranek]

I didn't spot any issues [1] (*). I'll be completely offline for one week since
now, so from my PoV, I'm fine to release it

(*) I had some issue with updated Jira plugin and hadn't enough time to
investigate it, but seems to me like a configuration bug in my environment

[1] https://wiki.jenkins-ci.org/display/JENKINS/LTS+1.509.x+RC+Testing

On Thursday 20 June 2013 20:51:04 Kohsuke Kawaguchi wrote:
> Posted
>
> On Wednesday, June 19, 2013, Kohsuke Kawaguchi wrote:
> > All right, then I'll incorporate these changes into this evening 1.509.2
> > RC.
> >
> > On 06/19/2013 12:49 PM, Vojtech Juranek wrote:
> >> On Wednesday 19 June 2013 11:51:58 Kohsuke Kawaguchi wrote:
> >>> I backported two more bugs and pushed 'stable-1.509.3'.
> >>
> >> sounds good to me. I'll be completely offline for once week since
> >> Saturaday or
> >> so, so there could be some delay with testing (but given we have already
> >> good
> >> amount of selenium tests, I'll maybe finished it this by Friday)
> >>
> >> I've also expanded backend-commit-history-parser to assist the
> >>
> >>> backporting process. See

> >>> https://wiki.jenkins-ci.org/**display/JENKINS/Backporting+**
> >>> toolkit+for+LTS<https://wiki.jenkins-ci.org/display/JENKINS/Backporting+
> >>> toolkit+for+LTS> for more details.

[Jerome Lacoste]

On Saturday, June 22, 2013 5:26:49 AM UTC-5, vjuranek wrote:

I didn't spot any issues [1] (*). I'll be completely offline for one week since
now, so from my PoV, I'm fine to release it

After many users of only using latest, I have now started using LTS versions :)

Is it expected that http://jenkins-ci.org/changelog-stable doesn't reflect the contents of the branch ? The upcoming changes is empty.

Thanks!

Jerome

[James Nord (jnord)]

Issues I have found:

Weather Icons.

Icons on folders are randomly missing (pointing to empty.png) going into the folder (and its child shows enabled jobs have status).

e.g

 

 

On a folder if you click on a weather link you get taken to jobname/lastbuild – but folders don’t get built and this results in a 404.

 

After a period of inactivity (overnight) when navigating to a page with lots of jobs I see a high page load time (minutes) (also have the latest dashboard plugin installed).

So could be a fault in m2release, but would only expect the delay if I was viewing the dashboard.

 

Otherwise seems to be holding up much better than 1.509.1

 

/James

 

 

--

[Jesse Glick]

On 06/28/2013 08:56 AM, James Nord (jnord) wrote:
> Issues I have found:

Are these regressions from 1.509.1, or broken also in 1.509.1? And are they filed in JIRA? Issues specific to folders would be better sent to CloudBees support (though
the bug might ultimately be determined to be in Jenkins core).

[James Nord (jnord)]

> From: jenkin...@googlegroups.com [mailto:jenkinsci-
> d...@googlegroups.com] On Behalf Of Jesse Glick
> Sent: 08 July 2013 23:36
> To: jenkin...@googlegroups.com
> Subject: Re: Backports to 1.509.2 done, please post RC

Hi Jesse,

As you are aware I had some major issues with 1.509.1 so the issues mentioned may have already been present in the prior LTS but I may not have noticed as I was pre-occupied with the "OMG it doesn't work and is in production" issues that were not apparent on my test system :-(

I will raise a CB support ticket about the folder + weather.

> > "After a period of inactivity (overnight) when navigating to a page with lots of jobs I see a high
> > page load time (minutes) (also have the latest dashboard plugin installed)."

So the above was very likely in 1.509.1 but not prior LTS and there are numerous reports / JIRAs about it - however I thought they had been traced back to dashboard plugin (and some others) and had been fixed - so I did not expect to see this anymore.
I haven't yet raised a JIRA for this as it may well be down to my m2Release plugin - where the "lastRelease" permalink is not using the new symlink support in 1.509 to avoid having to walk the build history. However this functionality is only exposed on the "job" page - or a dashboard - so I did not expect to see any wait until one of those pages was actually displayed (can anyone confirm if my expectations are correct in this regard - if they are I will raise a JIRA and point CB support at it :-) )

/James

[Jesse Glick]

On 07/09/2013 05:45 AM, James Nord (jnord) wrote:
> I thought they had been traced back to dashboard plugin (and some others) and had been fixed

There is an open bug in the dashboard plugin related to lazy loading; I have a PR filed for it but it has not been reviewed.

> this functionality is only exposed on the "job" page - or a dashboard - so I did not expect to see any wait until one of those pages was actually displayed

That would be my expectation too. Thread dumps can help pinpoint problems like these.

[Jesse Glick]

On 06/19/2013 02:51 PM, Kohsuke Kawaguchi wrote:
> I backported two more bugs

Still there are tons of lts-candidate bugs which are (a) important, (b) fixed in the trunk for a while. JENKINS-16023 [1] comes to mind, which was in 1.518 eight days (>
1 week) since you wrote this message. There are others like JENKINS-17125 [2] which were a bit too late. Is there a 1.509.3 plan?


[1] https://issues.jenkins-ci.org/browse/JENKINS-16023
[2] https://issues.jenkins-ci.org/browse/JENKINS-17125

[Jesse Glick]

On 06/20/2013 11:51 PM, Kohsuke Kawaguchi wrote:
> Posted

BTW the changelog [1] omits mention of JENKINS-17721 [2] which the git log says was backported.

I also took the liberty of replacing lts-candidate with 1.509.2-fixed on these seven bugs.


[1] http://jenkins-ci.org/changelog-stable
[2] https://issues.jenkins-ci.org/browse/JENKINS-17721