Also a big thank you to all that contributed to refactoring and
moving TCK tests to other Specifications. There was a lot of
challenging work all around (e.g. the refactoring of Faces tests
in the Platform TCK and Faces TCK). Alwin Joseph spent a lot of
time on refactoring Platform tests on the other SPECs as did
others. Alwin has always been a big contributor to the Platform
TCK and I wanted to call him out specifically which I didn't do
below.
Big thank you to Platform TCK Contributors
A big thank you for all of those that created pull requests for
EE 10 that were merged. Below is a sorted list of those that
contributed those pull requests:
Alwin Joseph
Andrea Boriero
Arjan Tijms
Aymen Naghmouchi
Bauke Scholtz
Brian Decker
Christian Beikov
Darran Lofthouse
David Kral
David Matějček
Ed Bratt
Gurunandan Rao
Jean-Louis Monteiro
Joakim Erdfelt
Jorge Bescos Gascon
Kalin Chan
Kyle Aure
Lukas Jungmann
Mark Thomas
Ondro Mihályi
Piotr Żygieło
Scott Kurz
Scott Marlow
Scott Stark
Suhrid Karthik
Tomáš Kraus
Volodymyr Siedlecki
Also a big thank you to Gurunandan Rao for running most of the
GlassFish test runs (from what I recall), others pitched in as
well but I definitely saw Guru's name as the author of most of
the TCK
test results.
Also a big thank you to Scott Stark for contributing the Core
Profile TCK!
Also a big thank you to all of the other contributors that
created issues, showed up to Platform TCK calls and participated
on this email list. All of you made time for contributing to
the TCK effort which helped!
Also a big thank you all of the EE 10 implementations
regardless of their current state. We should all create issues
+ pull requests for correcting TCK User Guides + improving tests
as we run into issues.
Preparing for next EE release
We have a lot to talk about regarding modernizing the Platform
TCK tests. One suggestion is that we keep the `master` branch
referencing EE 10 as long as possible so we can validate the
work that we do to refactor TCKs by running the Platform TCK
against an EE 10 implementation. Once we start adding in new
SPEC APIs, we will lose the ability to validate that the
Platform TCK code is correct until we have implementations for
the next EE release to test against (I think for EE 10 we were
mostly blocked on such testing from early fall of 2021 to May of
2022).
Perhaps we can have the standalone TCK build script use new
SPEC APIs (to support the ballot process) while having the Full
Platform/Web Profile TCK build script continue to use the EE 10
SPEC APIs.
We also want to fit in time for adding tests for new EE
features but hopefully those will only be done on the refactored
TCKs.
I think we learned a lot during the EE 10 release cycle that we
can carry forward into the next EE development cycle. I have
heard critical feedback that we didn't have enough early
participation on the TCKs split out to separate SPEC
repositories, which I agree with. One suggestion for
improvement is to keep refactored TCK tests in the Platform TCK
until all of the relevant tests are also refactored (or
removed). There are other options as well but making this
change will ensure we don't require those currently running
Platform TCK tests needing to switch their TCK configuration
settings to match a completely different Standalone TCK
configuration file (this was a big pain point).
EE 10 TCK Challenges
https://github.com/eclipse-ee4j/jakartaee-tck/tree/10.0.x
is created for dealing with any EE 10 Platform TCK challenges.
We need to update the (EFTL) jobs on https://ci.eclipse.org/jakartaee-tck/job/10/
to use the new branch (we might want to create the https://ci.eclipse.org/jakartaee-tck/job/11
jobs against the `master` branch as well). I didn't see a need
for the EPL jobs since we don't (yet) push Platform TCKs to
maven central.
Scott