Implement Concurrency 3.0

[jungm]

Implements the missing gap to pass the Jakarta Concurrency 3.0 TCK including:

• Asynchronous Interceptor
• Defining Resources via annotations/web.xml
• Rework ApplicationThreadContextProvider as well as implement TxThreadContextProvider
• ManagedConnection invocation handler can now enroll already created connections in a Transaction
• ManagedThreadFactory is now wired up to ContextService

Important: This change disables SXC for parsin web.xml, I wasn't able to get the sxc-maven-plugin to run on java 17/jaxb 4 in a meaningful time, probably needs some more work

Thanks to @rzo1 a full build for this branch is already running here

TCK results

tck/concurrency-standalone:

[INFO] Tests run: 147, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 176.575 s - in TestSuite
[INFO] 
[INFO] Results:
[INFO] 
[INFO] Tests run: 147, Failures: 0, Errors: 0, Skipped: 0

tck/concurrency-sigtest:

[INFO] Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 4.950 s -- in ee.jakarta.tck.concurrent.spec.signature.SignatureTests
[INFO] 
[INFO] Results:
[INFO] 
[INFO] Tests run: 1, Failures: 0, Errors: 0, Skipped: 0

[rzo1]

Note: We are currently looking into SXC to avoid to having to disable it here. Current status is, that we created a Jakarta-migrated version of SXC and @jungm started to integrated this version on a separate branch. If this works out, we can update it here.

[jungm]

IMO we should not regenerate the SXC accessors in this PR, would make the diff absolutely huge and hard to review.
Think it's best to first merge both the SXC rework and this PR, then regenerate accessors on main branch

[jungm]

Also there is currently one test failure left (org.apache.openejb.threads.ManagedScheduledExecutorServiceTest.triggerRunnableSchedule)

[rzo1]

IMO we should not regenerate the SXC accessors in this PR, would make the diff absolutely huge and hard to review. Think it's best to first merge both the SXC rework and this PR, then regenerate accessors on main branch

Yup. This is done in a separate PR but my comments aim to avoid merging this PR until we have re-generated the accessors because disabled something tends to be forgotten in the code base :)

From my POV I would prefer:

1. Integrate SXC work
2. Regenerate stuff
3. Rebase this PR
4. Regenerate concurrency related stuff
5. Be Happy

Or are there any objections in doing so?

I am also open to

1. Integrate SXC
2. Integrate Concurrency
3. Re-generate accessors + re-enable SXC (but I can remember a discussion on the dev list, that disabling SXC isn't a particulary good idea :) )

[jungm]

+1 for your preferred suggestion, just don't want this PR to become like a 10k+ diff that's impossible to review

[jungm]

I merged main into this branch, regenerated the SXC accessors and enabled SXC again for parsing web.xml. Also the test I mentioned in a comment above should be fixed now.

TCK is still passing on my machine and full build is green: https://ci-builds.apache.org/job/Tomee/view/tomee-10.x/job/pull-request-manual/130/