Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [jakartaee-platform-dev] TCK tests in the same repo as API and Spec

Hi Jan, I think you've raised some good questions -- I've picked out 2 basic questions from your note which I will try and answer.

1) How/where do we test multi-spec interactions?

Lets say we have an interaction between CDI + JSONB that needs to be tested. I propose we pick one of the specs of the involved technologies to house the test. This way we don't have any TCK tests in "no mans land". If there are optional integrations (e.g. do X IF the runtime is CDI-capable) we can control this in a number of ways such as: A) do a Class.forName() of some CDI API to see if CDI is available in the test or B) control test inclusion with sysprops/env vars like `mvn test -Dcdi=false` to disable CDI-capable tests.

2) What standard test framework do we settle on? We don't want every spec doing their own thing because it will make running all TCK tests very difficult.

Fully agree. I propose we do the following:
 - Create a new entry point in the jarkartaee-tck repo which will be a maven-based project that can test the new "externalized" TCK tests. Then we can incrementally migrate TCK tests on a per-spec level to the new externalized form.
 - Externalized TCK tests MUST use JUnit + Arquillian, even if the technology is standalone (like JSONB). This way, application servers can still certify that the "standalone" technologies still work when deployed in applications on the servers. We can create a "no-op" Arquillian container that standalone implementations can use (e.g. Yasson for JSONB) for when components want to verify TCK tests in standalone mode. Although MicroProfile does not have a deployment spec, its TCK tests still use Arquillian to create JEE deployments such as .war or .ear files.
 - Create a new umbrella entry point so we can run all TCK tests in one shot:  the old ones based on the JavaTest harness, and the new/externalized ones based on JUnit+Arquillian

- Andy

On Mon, Feb 3, 2020 at 11:53 AM Jan Supol <jan.supol@xxxxxxxxxx> wrote:

> +1 Kevin! I think the TCK in Jakarta raises the bar higher for people
> to run TCK freely. As a contributor, I would like to be able to add
> TCKs whenever I propose API/Spec changes. Otherwise, we will run out
> of sync in no time. Therefore, I support what Andy proposed to split
> TCK and rehome them to be with where it should belong. I speak
> directly from my own experience of working on MicroProfile. It is
> great to add tests together with spec or api changes to demonstrate
> the full piece.

This is exactly what I miss on Microprofile, the tests that are
mandatory if and only if two of the specs should work together (i.e. not
optional tests). Those were known as the CTS only tests, i.e. were not
part of the standalone TCKs (For instance CDI & JAX-RS, CDI & WebSocket,
BeanValidation & JAX-RS). Standalone  TCKs are what the Microprofile
has, it does not have the full platform test. If the tests should still
be part of the platform TCK, and the tests are in the Spec repository,
how to make sure the tests are run for full platform and not for the
standalone TCK?

Using the approach Andy used for JSON-B works fine. What the approach
does not have, and what Microprofile does not have, is the deployment.
Jakarta EE TCK has the concept of deployment (not necessarily JSR-88),
and the deployment behaviour is described in various Specs, such as
Servlet of WebSockets. How the concept of the deployment would work with
Andy's approach is not clear. Should the deployment be tested in the
future? Should it be left on the Spec to define the deployment
requirements/API or should it more common to every Spec? Currently, the
TCK framework does that and vendors implement the porting package that
takes care of the deployment. Every TCK uses the same mechanism. But
with having a different test framework in each TCK, the deployment needs
to be thought through.

As for releasing TCKs separately:
> 1. When releasing a new version of a spec, we release API/Spec/TCK and
> they have either major or minor version changes. They share the same
> version. With the same version, we know which version of TCK we should
> execute for the version of Spec you choose.
> 2. If we found any bugs in TCK, we can do a patch release for the
> project, which produces a micro/patch version release of API/Spec/TCK
> so that runtimes can execute the new TCK while end users are not
> affected.
> The above fits in well with the general understanding of
> major/minor/micro release.

It is great that you think about standalone TCK. However, that is the
easy part. The hard part is to think about the full platform. The
approach Andy described is to JUnitize the tests. The other approach
will use some different framework. It is fairly fine to have separate
TCKs like this as long as the Full platform TCK is no longer a
requirement. Before the split, it needs to be known whether the full
platform TCK is what is required, or whether the set of standalone TCKs,
such as in Microprofile, is enough. If the full platform TCK is needed
for verifying the compatible implementation, the separation needs to be
done with the idea of putting all parts together. Otherwise, each Spec
will take their part, convert it to use some fancy test framework and
say: "Hey, its not my business to make the full platform TCK possible.
Someone else should take care".

Even the Microprofile has a unified Arquillian based test environment.
Using JUnit in one TCK, Arquillian in other, and JavaTest in the next
brings a challenge for the team that ought to run the TCKs. Are we cool
with having it different in every TCK?

> On Mon, Feb 3, 2020 at 1:40 PM Kevin Sutter <sutter@xxxxxxxxxx
> <mailto:sutter@xxxxxxxxxx>> wrote:
>     Jan,
>     You bring up some excellent points.  Having a well thought out
>     plan before jumping in with both feet would be nice.  But, on the
>     other hand, we don't want to stifle the excited effort that Andy
>     is demonstrating. It's a tough line.  Maybe we need to allow one
>     or two teams to experiment with the processes and see what it
>     takes.  And, then we can take those experiences and develop a plan
>     that works across all the projects.
For the experiment, grabbing JSON-B, or JSON-P is easy. Those are
basically standalone only technologies. A Servlet based technology seems
more appropriate for experimenting.
>     ---------------------------------------------------
>     Kevin Sutter
>     STSM, MicroProfile and Jakarta EE architect @ IBM
>     e-mail: sutter@xxxxxxxxxx <mailto:sutter@xxxxxxxxxx>     Twitter:
>      @kwsutter
>     phone: tl-553-3620 (office), 507-253-3620 (office)
>     LinkedIn:
>     From: Jan Supol <jan.supol@xxxxxxxxxx <mailto:jan.supol@xxxxxxxxxx>>
>     To: jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     Date: 01/31/2020 15:02
>     Subject: [EXTERNAL] Re: [jakartaee-platform-dev] TCK tests in the
>     same repo as API and Spec
>     Sent by: jakartaee-platform-dev-bounces@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev-bounces@xxxxxxxxxxx>
>     ------------------------------------------------------------------------
>     The discussion on the various email lists is quite extensive, and I
>     admit I do not follow every thread that is discussed. But I am
>     waiting
>     for the plan from any of the committee we have since the
>     transition to
>     Eclipse Foundation. And I am not aware about any clear statement
>     about
>     what the final goal is. For start:
>     - Is the splitting really what is recommended?
>     - Should the Platform Jakarta EE TCK still be available in the future
>     after the split?
>     - Should the interoperability between Specs tests (the tests that
>     require multiple Specifications to cooperate as it is described in
>     one
>     or the other specification, the tests that used to be part of the
>     full
>     CTS only) still be part of the combined (?) Jakarta EE TCK?
>     - Is there a recommended framework to be used in the future with the
>     TCKs (such as the existing JavaTest + perhaps Arquillian similarly to
>     Microprofile + ?)?
>     Those should be goals presented by the Jakarta representatives before
>     the TCK repo is started to be ripped apart. Not on email, but
>     clearly on
>     the main Jakarta informational page (whatever that is now).
>     -- Jan
>     On 31.01.2020 20:34, Lance Andersen wrote:
>     > I also agree with Bill.  There has to be thought  given to CTS
>     > structure first IMHO so that there is consistency in the test
>     > structure to make it easier to pull in the standalone tests  and
>     >  where the platform specific tests are maintained for the
>     standalone
>     > technologies.
>     >
>     > This will take some time to flush this out.  During the Java EE
>     days,
>     > we did find it easier to work out of one master workspace, but
>     part of
>     > that was due to the fact there was one team primarily
>     responsible  for
>     > the test development for all of the Java EE technologies.
>     >
>     >> On Jan 31, 2020, at 2:01 PM, dmitry.kornilov@xxxxxxxxxx
>     <mailto:dmitry.kornilov@xxxxxxxxxx>
>     >> <mailto:dmitry.kornilov@xxxxxxxxxx> wrote:
>     >>
>     >> Bill, you are right. On the other hand the process of splitting
>     has
>     >> to be started somehow. It will never start with an assumption that
>     >> some developer with TCK knowledge will take an initiative and
>     start
>     >> working on it. There are not too many developers like this and
>     no one
>     >> wants to take a risk and responsibility.
>     >> On the other hand I don’t think that it’s a right time to do it
>     now
>     >> as part of Jakarta EE 9 release. I already proposed thatwe will
>     not
>     >> move JSONB TCK tests from CTS yet. We will do it after Jakarta
>     EE 9
>     >> is released. We will keep them in sync until that time. Users
>     may use
>     >> TCK tests in JSONB repo as more convenient way of checking
>     >> compatibility. But it cannot be used officially for compliance
>     >> testing for now.
>     >> -- Dmitry
>     >> *From:*jakartaee-platform-dev-bounces@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev-bounces@xxxxxxxxxxx>
>     >>
>     <mailto:jakartaee-platform-dev-bounces@xxxxxxxxxxx><jakartaee-platform-dev-bounces@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev-bounces@xxxxxxxxxxx>
>     >> <mailto:jakartaee-platform-dev-bounces@xxxxxxxxxxx>>*On Behalf
>     >> Of*Bill Shannon
>     >> *Sent:*Friday, January 31, 2020 7:06 PM
>     >> *To:*jakartaee-platform developer discussions
>     >> <jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     >> <mailto:jakartaee-platform-dev@xxxxxxxxxxx>>; Andy Guibert
>     >> <andy.guibert@xxxxxxxxx <mailto:andy.guibert@xxxxxxxxx>
>     <mailto:andy.guibert@xxxxxxxxx>>
>     >> *Subject:*Re: [jakartaee-platform-dev] TCK tests in the same
>     repo as
>     >> API and Spec
>     >>
>     >> This has been discussed numerous times over the last 3 years.
>     No one
>     >> disagrees with the goal of splitting up the TCK repo.  Whether the
>     >> individual spec TCKs are in the same repo as the API classes or
>     in a
>     >> different repo is just a detail.  The challenge is in splitting up
>     >> the existing TCK repo such that the resulting TCK for each spec is
>     >> functionally identical to the existing standalone TCKs for the
>     specs,
>     >> and that the platform TCK is somehow created by combining all the
>     >> individual TCKs to produce a new platform TCK that is functionally
>     >> identical to the existing platform TCK.
>     >>
>     >> Needless to say, this is not a small job.
>     >>
>     >> No one with deep knowledge of the existing TCKs has come
>     forward with
>     >> a detailed plan for how to achieve the above. Without that, we're
>     >> all just wishing and hoping.
>     >>
>     >> And I strongly encourage you to*not*just start hacking on the
>     TCK to
>     >> create what you want for your spec.  We need to solve the larger
>     >> problem and a bunch of uncoordinated hacks to individual TCKs will
>     >> not get us there.
>     >>
>     >> Thanks.
>     >>
>     >> P.S.  You should be able to find some of the previous history
>     of this
>     >> discussion in the jakartaee-tck-dev mailing list.
>     >>
>     >> Andy Guibert wrote on 1/30/20 10:45 AM:
>     >>> Hi all,
>     >>> Currently all of the Jakarta EE TCK tests are housed in one
>     big repo
>     >>> and they use a custom test framework from the Java EE days:
>     >>>
>     >>>  It is more convenient to have the TCK tests in the same repo
>     as the
>     >>> API and spec docs because as the technologies change over time
>     all 3
>     >>> parts (tck/api/spec) can be updated in the same PR. Additionally,
>     >>> implementations can then consume the TCK tests as Maven artifacts
>     >>> and run/verify that they pass the TCK. This is what
>     MicroProfile has
>     >>> done and it works very well.
>     >>> As an example, I've started this off with the JSON-B TCK test
>     here:
>     >>> jsonb-api pr:
>     >>> yasson (impl) pr:
>     >>> Just wanted to share this with the wider dev community and
>     encourage
>     >>> other specs to follow suit as time permits.
>     >>> Cheers, Andy
>     >>> -- IBM
>     >>>
>     >>>
>     >>> _______________________________________________
>     >>> jakartaee-platform-dev mailing list
>     >>> jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>      <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     >>> To change your delivery options, retrieve your password, or
>     unsubscribe from this list, visit
>     >>>
>     >> _______________________________________________
>     >> jakartaee-platform-dev mailing list
>     >> jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     >> <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     >> To change your delivery options, retrieve your password, or
>     >> unsubscribe from this list, visit
>     >>
>     >
>     > <>
>     >
>     <><>
>     > <>Lance
>     > Andersen| Principal Member of Technical Staff | +1.781.442.2037
>     > Oracle Java Engineering
>     > 1 Network Drive
>     > Burlington, MA 01803
>     > Lance.Andersen@xxxxxxxxxx <mailto:Lance.Andersen@xxxxxxxxxx>
>     <mailto:Lance.Andersen@xxxxxxxxxx>
>     >
>     >
>     >
>     >
>     > _______________________________________________
>     > jakartaee-platform-dev mailing list
>     > jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     > To change your delivery options, retrieve your password, or
>     unsubscribe from this list, visit
>     >
>     _______________________________________________
>     jakartaee-platform-dev mailing list
>     jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     To change your delivery options, retrieve your password, or
>     unsubscribe from this list, visit
>     _______________________________________________
>     jakartaee-platform-dev mailing list
>     jakartaee-platform-dev@xxxxxxxxxxx
>     <mailto:jakartaee-platform-dev@xxxxxxxxxxx>
>     To change your delivery options, retrieve your password, or
>     unsubscribe from this list, visit
> --
> Thanks
> Emily
> _______________________________________________
> jakartaee-platform-dev mailing list
> jakartaee-platform-dev@xxxxxxxxxxx
> To change your delivery options, retrieve your password, or unsubscribe from this list, visit
jakartaee-platform-dev mailing list
To change your delivery options, retrieve your password, or unsubscribe from this list, visit

Back to the top