GH-38997: [Java] Modularize format and vector

[jduo]

Rationale for this change

This PR depends on #39011 .

Make arrow-vector and arrow-format JPMS modules for improved compatibility with newer JDKs
and tools that require modules.

What changes are included in this PR?

• add module-info.java files for vector and format.

Are these changes tested?

Yes, these modules are run as JPMS modules in JDK9+ unit tests.

Are there any user-facing changes?

Yes, some test classes have moved and users can now run these as JPMS modules.

• Closes: [Java] Build arrow-format and arrow-vector as JPMS modules #38997

[github-actions]

Thanks for opening a pull request!

If this is not a minor PR. Could you open an issue for this pull request on GitHub? https:/apache/arrow/issues/new/choose

Opening GitHub issues ahead of time contributes to the Openness of the Apache Arrow project.

Then could you also rename the pull request title in the following format?

GH-${GITHUB_ISSUE_ID}: [${COMPONENT}] ${SUMMARY}

or

MINOR: [${COMPONENT}] ${SUMMARY}

In the case of PARQUET issues on JIRA the title also supports:

PARQUET-${JIRA_ISSUE_ID}: [${COMPONENT}] ${SUMMARY}

See also:

• Other pull requests
• Contribution Guidelines - How to contribute patches

[github-actions]

⚠️ GitHub issue #38997 has been automatically assigned in GitHub to PR creator.

[lidavidm]

This is on my backlog - I'm hoping to give this a review this week

[jduo]

Thanks for the review @lidavidm . This shouldn't go in yet until we're ready to bite the bullet on modularizing memory-core + making the command-line change though (or we figure out how to avoid the command-line change). I'm changing this back to draft for the time being.

[jduo]

Note that this one didn't require doc changes.

[lidavidm]

@davisusanibar can you help take a look at the nullness annotation issues here?

[jduo]

@davisusanibar can you help take a look at the nullness annotation issues here?

I fixed a null annotation issue with CountingAllocationListener in the parent patch (#39011).

[lidavidm]

Ah whoops, I thought I was looking at that one.

[lidavidm]

I think this just needs to be rebased?

[jduo]

I think this just needs to be rebased?

Yes, this is rebased now @lidavidm

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 5 benchmarking runs that have been run so far on merge-commit 1a622ec.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 3 possible false positives for unstable benchmarks that are known to sometimes produce them.