[WFCORE-7310] ear-exclusions-cascaded-to-subdeployments doesn't work correctly

[mskacelik]

https://issues.redhat.com/browse/WFCORE-7310

The fix is in that I used subDeploymentMap, which holds info for all the "global" sub-deployments, instead of result.getSubDeploymentSpecifications(), which only contains the explicitly set sub-deployments.

Tested with the reproducer in the issue:

And also tested with integration tests within WF:
wildfly/wildfly@main...mskacelik:wildfly:WFCORE-7310-TEST (WIP, Proof of Concept)
The test alone might spark another discussion if there is even a need for the original explicit subsystem test.

Test PR: wildfly/wildfly#19095

[wildfly-ci]

Hello, mskacelik. I'm waiting for one of the admins to verify this patch with /ok-to-test in a comment.

[yersan]

/ok-to-test

[yersan]

Hello @mskacelik , thanks for the PR. We need to have the testcase in place in order to fully verify this. I suggest to go ahead by creating the testcase PR (without the wildfly core version bump) and let it to fail as expected. Then, we can use our CI to run an Integration Job that verifies both of your topic branches, the one for core and the one for wildfly, and that one should pass.

[mskacelik]

Will do a bit of investigation on why the other tests failed.

[mskacelik]

@yersan hopefully fixed now, and created a PR for the tests as you wanted wildfly/wildfly#19095
The error for the test failures was my fault, due to my not understanding a trivial for loop and its implications. Some tests seem to still fail, but they seem to be failing even before the given commit.

[yersan]

Windows - JDK 17 (Pull Request) - merge failed test can be ignored since they are not related to this change. They are being addressed at https://issues.redhat.com/browse/WFCORE-7303

I have to take a closer look at tghis, after your changes I was expecting a possible failure on the EarJbossStructureCascadeExclusionsTestCase, but it passed.

[mskacelik]

I have to take a closer look at tghis, after your changes I was expecting a possible failure on the EarJbossStructureCascadeExclusionsTestCase, but it passed.

It actually should have passed, the PR only changes that the cascade is happening on all sub-deployments (implicit), so this includes the explicit (as described in the .xml file).

[yersan]

Linux Integration Job:
https://ci.wildfly.org/buildConfiguration/WF_WildFlyCoreIntegrationExperiments/512668

Windows Integration Job:
https://ci.wildfly.org/buildConfiguration/WF_WildFlyCoreIntegrationExperimentsWindows/512670

[mskacelik]

Linux Integration Job: https://ci.wildfly.org/buildConfiguration/WF_WildFlyCoreIntegrationExperiments/512668

Windows Integration Job: https://ci.wildfly.org/buildConfiguration/WF_WildFlyCoreIntegrationExperimentsWindows/512670

Linux TC jobs were successful. But the Windows ones, some tests regarding permissions seem to fail, but I am not really sure why that is.

[yersan]

Thanks @mskacelik

[yersan]

Linux TC jobs were successful. But the Windows ones, some tests regarding permissions seem to fail, but I am not really sure why that is.

@mskacelik
They are unrelated. Our CI is a bit sensible and we suffer from flaky tests, the trick there is to vompare whether tehy passed on other OS or inspecting the CI test history and verify that the faliled log trace has been produced on other PRs, meaning the current one is unrelated.