On the other questions. I’m no Jenkins specialist but this job
Jakarta Concurrency - Jenkins (eclipse.org) you can use to release the api jar to staging. You can dry run it etc. to see what it does.
What do you need to run to package up the TCK is it created as part of the master build? Using Jenkins to push to staging is easier as all the credentials are available to the Jenkins instance.
Steve
From: Nathan Rauh <nathan.rauh@xxxxxxxxxx>
Sent: 19 January 2022 16:26
To: Maxim Nesen <maxim.nesen@xxxxxxxxxx>; Dmitry Kornilov <dmitry.kornilov@xxxxxxxxxx>; Ed Bratt <ed.bratt@xxxxxxxxxx>; Steve Millidge (Payara) <steve.millidge@xxxxxxxxxxx>
Cc: cu developer discussions <cu-dev@xxxxxxxxxxx>
Subject: Re: : RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
As far as I'm aware, there is just one remaining TCK pull expected (hopefully today) and after that we should immediately get a 3.0.0 final build of
both the api jar and tck jar onto staging so that these can be pulled in to the Open Liberty beta that will come out in February, which we can then use as the compatible implementation.
I looked over these Jenkins jobs, and I'm worried that they either won't pass or won't build the tck jar within the project, and will need more updates, which will cause a delay that will make us
miss the February beta which is our shot at a compatible implementation. That said, I can easily build all of these artifacts manually with
mvn install and run with these in Open Liberty where the whole TCK passes. Is there a manual way to put artifacts built in this manner into staging, and would that be valid? Or is anyone able to update/fix the Jenkins jobs to successfully build the
api and tck jars in an official form that can be put onto staging as 3.0.0 final?
For reference, here is the error I currently see on recent Jenkins builds of concurrency-api-master-build:
https://ci.eclipse.org/cu/job/concurrency-api-master-build/68/console
ERROR: Build step failed with exception
java.lang.IllegalStateException: ID java is already used by another action: io.jenkins.plugins.analysis.core.model.ResultAction for Java Compiler
at io.jenkins.plugins.analysis.core.steps.IssuesPublisher.ensureThatIdIsUnique(IssuesPublisher.java:146)
at io.jenkins.plugins.analysis.core.steps.IssuesPublisher.attachAction(IssuesPublisher.java:97)
at io.jenkins.plugins.analysis.core.steps.IssuesRecorder.publishResult(IssuesRecorder.java:807)
at io.jenkins.plugins.analysis.core.steps.IssuesRecorder.record(IssuesRecorder.java:734)
at io.jenkins.plugins.analysis.core.steps.IssuesRecorder.perform(IssuesRecorder.java:697)
at io.jenkins.plugins.analysis.core.steps.IssuesRecorder.perform(IssuesRecorder.java:673)
at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:20)
at hudson.model.AbstractBuild$AbstractBuildExecution.perform(AbstractBuild.java:806)
at hudson.model.AbstractBuild$AbstractBuildExecution.performAllBuildSteps(AbstractBuild.java:755)
at hudson.model.Build$BuildExecution.post2(Build.java:178)
at hudson.model.AbstractBuild$AbstractBuildExecution.post(AbstractBuild.java:699)
at hudson.model.Run.execute(Run.java:1913)
at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:43)
at hudson.model.ResourceController.execute(ResourceController.java:99)
at hudson.model.Executor.run(Executor.java:431)
Build step 'Record compiler warnings and static analysis results' marked build as failure
Finished: FAILURE
Do we also need a concurrency-tck-master-build to build and package the jakarta.enterprise.concurrent-tck-3.0.0.jar jar from the main concurrency-api repo, or will concurrency-api-master-build do
that for us once it is working again?
From: Nathan Rauh/Rochester/IBM
To: "Maxim Nesen" <maxim.nesen@xxxxxxxxxx>
Cc: "cu developer discussions" <cu-dev@xxxxxxxxxxx>, "Dmitry Kornilov"
<dmitry.kornilov@xxxxxxxxxx>, "Ed Bratt" <ed.bratt@xxxxxxxxxx>, "Steve Millidge (Payara)" <steve.millidge@xxxxxxxxxxx>
Date: 01/18/2022 09:29 AM
Subject: Re: [External] : RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
To clarify, the TCK artifact in jakartaee-tck should not be used for EE 10 because the TCK has been ported over to the Concurrency project and the new tests for EE 10 have only been added to the
TCK within the Concurrency project itself. The old copy in jakartaee-tck would likely be removed after we have this all working.
From: "Maxim Nesen" <maxim.nesen@xxxxxxxxxx>
To: "Dmitry Kornilov" <dmitry.kornilov@xxxxxxxxxx>, "Steve
Millidge (Payara)" <steve.millidge@xxxxxxxxxxx>, "Ed Bratt" <ed.bratt@xxxxxxxxxx>, "Nathan Rauh" <nathan.rauh@xxxxxxxxxx>
Cc: "cu developer discussions" <cu-dev@xxxxxxxxxxx>
Date: 01/18/2022 07:51 AM
Subject: Re: [External] : RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
Hello everyone, regarding the state of infrastructure at
https://ci.eclipse.org/cu/, there is a bunch of jobs for Concurrency API/RI for Jakarta EE8/EE9 available. It is possible
to build SNAPSHOTs for Jakarta EE8/EE9, publish artifacts to ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside your organization.
ZjQcmQRYFpfptBannerEnd
Hello everyone,
regarding the state of infrastructure at https://ci.eclipse.org/cu/, there is a bunch of jobs for Concurrency
API/RI for Jakarta EE8/EE9 available. It is possible to build SNAPSHOTs for Jakarta EE8/EE9, publish artifacts to staging, and promote them to central afterward.
Also, there is a set of jobs to run TCK tests for Jakarta EE8/EE9 compatibility.
Presumably, it is possible to adapt jobs that manage builds and publishing/promoting of artifacts to the staging/central without major changes.
However, TCK jobs shall be implemented from scratch. There is already a staged TCK bundle for CU at
https://download.eclipse.org/ee4j/jakartaee-tck/jakartaee10/staged/eftl/jakarta-concurrency-tck-3.0.0.zipand
milestone for GF 7 at https://repo1.maven.org/maven2/org/glassfish/main/distributions/glassfish/7.0.0-M1/glassfish-7.0.0-M1.zip.
Those bundles (or newer if they are published before TCK jobs are implemented) shall be used while implementing jobs for TCK tests.
In my opinion, it would be enough from 1 to 2 weeks to adapt and implement the whole set of jobs for concurrency-ri 3.0.0 including jobs for TCK tests, builds, and staging/central manipulations.
Regards,
Maxim
Od: Dmitry Kornilov <dmitry.kornilov@xxxxxxxxxx>
Odesláno: pondělí 17. ledna 2022 18:00
Komu: Steve Millidge (Payara) <steve.millidge@xxxxxxxxxxx>; Ed Bratt <ed.bratt@xxxxxxxxxx>; Nathan Rauh <nathan.rauh@xxxxxxxxxx>;
Maxim Nesen <maxim.nesen@xxxxxxxxxx>
Kopie: cu developer discussions <cu-dev@xxxxxxxxxxx>
Předmět: RE: [External] : RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
Adding Maxim.
Maxim, can you please analyse a state of concurrency-ri?
-- Dmitry
From:Steve Millidge (Payara) <steve.millidge@xxxxxxxxxxx>
Sent: 14 января 2022 г. 10:54
To: Ed Bratt <ed.bratt@xxxxxxxxxx>; Nathan Rauh <nathan.rauh@xxxxxxxxxx>; Steve Millidge (Payara) <steve.millidge@xxxxxxxxxxx>
Cc: cu developer discussions <cu-dev@xxxxxxxxxxx>; Dmitry Kornilov <dmitry.kornilov@xxxxxxxxxx>
Subject: RE: [External] : RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
I’m not aware that any work has been done on concurrency-ri to bring it up to speed with the new capabilities in Concurrency 3.0 so I don’t think that is an option for this release.
From:Ed Bratt <ed.bratt@xxxxxxxxxx>
Sent: 13 January 2022 23:43
To: Nathan Rauh <nathan.rauh@xxxxxxxxxx>; Steve Millidge (Payara) <steve.millidge@xxxxxxxxxxx>
Cc: cu developer discussions <cu-dev@xxxxxxxxxxx>; DMITRY.KORNILOV <dmitry.kornilov@xxxxxxxxxx>
Subject: Re: [External] : RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
Nathan,
There is some CI infrastructure in place for the CU project (https://ci.eclipse.org/cu/). It might be useful
for completing this specification version.
The CU project consists of some bits of compatible implementation (stuff in
concurrency-ri) which I believe constitute a CI. As I understand it, concurrency utilities is of the class of components
that requires some additional infrastructure to host and validate the TCK. (for example, Enterprise Security has a Compatible Implementation that only work when integrated with something more substantial, like an EE Platform -- the same, I believe is true
of CU). That said, I see that the previous compatibility requests were made for Eclipse GlassFish -- I think you can decide if you want to repeat that, or just request a CCR for the (unfortunately named concurrency-ri project).
Based on previous releases it might be possible to either perform a local build of Eclipse GlassFish that includes the CU compatible implementation jar and API -- or, work with the GlassFish team to integrate your proposed final
CI and API JAR into GlassFish -- using that build, run the current TCK and you have what you need (the archived CI would be the the concurrency utilities implementation, not the GlassFish version that it was integrated into).
It looks like Dmitry Kornilov ran these tests last time they were run. He might recall these details more clearly.
Thanks,
-- Ed
PS, after this API is finalized, we should consider splitting it into an implementation and Spec. project. The Spec. project could then be moved into jakartaee GHO.
On 1/13/2022 6:45 AM, Nathan Rauh wrote:
I just noticed that Ed was somehow left off of the cc list when I replied yesterday. Sorry about that. I have just added him back on.
As far as I'm aware, every 3.0 enhancement for which a pull was submitted has been fully added to the specification and TCK, with nothing in a partial state except for the pending work to add TCK coverage for the new resource definition annotations in EJB container
and possibly JSPs, which are being worked on. We should create a 3.1 milestone in github to collect up a list of enhancements/issues that didn't make 3.0 (such as the one you mentioned) that could go into Jakarta EE 11. If everyone is happy with this outcome,
the main help that will be needed is the timely creation and publish to maven of the 3.0 RC1 image next week after those final TCK update(s) go in. Starting/continuing on your implementation is probably the best way to help at this point so that there is
a contingency plan if this doesn't make the Open Liberty beta in time. The odds became a bit more favorable because I just noticed that Kyle has successfully finished addressing all of the review comments on the TCK port issue, and so I've merged it in. Thist
just leaves the pending TCK scenario(s) for resource definition annotations to get in next week.
From: "Steve Millidge (Payara)"
<steve.millidge@xxxxxxxxxxx>
To: "Nathan Rauh"
<nathan.rauh@xxxxxxxxxx>, "cu developer discussions"
<cu-dev@xxxxxxxxxxx>
Cc: "DMITRY.KORNILOV"
<dmitry.kornilov@xxxxxxxxxx>
Date: 01/13/2022 06:38 AM
Subject: [EXTERNAL] RE: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
Hi Nathan, Dmitry
I should remember how to push artifacts to maven central assuming the old Jenkins jobs haven’t got “bit rot”. I was hoping to get an api change for this issue
MaxConcurrency annotation · Issue #136 · eclipse-ee4j/concurrency-api · GitHubbut
tbh I’m not going to get the time.
I haven’t been following this as closely as I should due to many and varied things. Are we OK with spec doc or do I need to pull some of my team in to drive this forward more?
Our team is starting to look at the implementation of this on Payara 6 alpha 3 but not sure we can meet the Feb deadline with a Compatible Implementation. We likely have more flexibility in pushing
out alphas than perhaps OpenLiberty. Is Feb 28th really a drop dead deadline for have a CI ready if the API is finalised?
Steve
From:Nathan Rauh
<nathan.rauh@xxxxxxxxxx>
Sent: 12 January 2022 19:26
To: cu developer discussions <cu-dev@xxxxxxxxxxx>
Cc: DMITRY.KORNILOV
<dmitry.kornilov@xxxxxxxxxx>; Steve Millidge (Payara)
<steve.millidge@xxxxxxxxxxx>
Subject: Re: [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
I can help provide some of the estimated timeline assuming that we will need to use Open Liberty as the compatible implementation upon which to certify. Toward the end of last year I sent an email
out querying if there were any other implementations that could be used as the compatible implementation that would meet the Feb 28 deadline, and thus far haven't heard back from any others.
Here's how the schedule would line up:
The pull to port over the old TCK bucket to the Concurrency project and merge it with the new TCK tests that have already been written for Concurrency 3.0 is currently in review state. It appears that all of the review comments thus far have been very minor
and so I expect it to be merged relatively soon, probably within the next few days.
After this, one of the participants is working on a remaining new TCK test to include the new resource definition annotations on an EJB. That will need to be submitted, reviewed and merged, hopefully early next week.
After that, an official RC1 build will need to be performed, and the spec API jar and TCK jar published to maven. I have no idea how to do either of these things, but hopefully Steve knows or either of you who are helping to mentor will know and can help with
it.
If that can all be done by January 19, then we should be able to get it into the next beta release of Open Liberty.
Every day beyond that point over the next week will have an increasingly lesser chance of making the February beta. I don't expect Open Liberty will hold up or change its release schedule for just this spec.
When the February beta comes out around the beginning of February, then we would be able to collect official results of the TCK running with the 3.0 RC1 release on the Open Liberty beta as the compatible implementation. It will take several days further to
get those results published to a publically accessible location, but it would be well before the Feb 28 deadline.
The main risk here is getting the 3.0 RC1 image created and published to Maven in time.
If another implementation with a more flexible schedule or more lenient approach to certification comes along, that would certainly help add some time. But absent that, we do at least have a path to meeting Feb 28, albeit a narrow one.
From: "Ed Bratt" <ed.bratt@xxxxxxxxxx>
To: "cu developer discussions" <cu-dev@xxxxxxxxxxx>, "Steve Millidge
(Payara)" <steve.millidge@xxxxxxxxxxx>
Cc: "DMITRY.KORNILOV" <dmitry.kornilov@xxxxxxxxxx>
Date: 01/12/2022 11:55 AM
Subject: [EXTERNAL] [cu-dev] Checking in -- progress towards finishing Concurrency Utilities 3.0?
Sent by: "cu-dev" <cu-dev-bounces@xxxxxxxxxxx>
Hi,
Dmitry Kornilov and I have been assigned by the Specification Committee to mentor you through the process of finalizing the Release PR for the upcoming release review of Jakarta Concurrency 3.0 for inclusion in EE10.
As a part of wave 1 in the Jakarta EE 10 Release Plan,
the release review PR could be made at any time. You may create the PR even if you don't have all the information required. Just mark it as DRAFT and work on it together with your mentor (me).
Are you on track with finishing up this specification version and creating a PR for this specification in the
specificationsrepository? Are you able to estimate
when the team anticipates starting this final step?
Is there anything blocking/delaying the release?
Is there anything Dmitry or I can help you or the project team with?
Thank you,
-- Ed_______________________________________________
cu-dev mailing list
cu-dev@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/cu-dev