Fix sizing of Maps

[kilink]

Fix several instances where Maps were sized incorrectly when the number of elements to insert is known; specifying the capacity without taking into account the load factor meant it would always result in a rehash / resize operation.

To this end, add a simple Maps utility class that handles the capacity calculation, much like the static HashMap.newHashMap helper that was added in JDK19.

Checklist

Before we can review and merge your changes, please go through the checklist below. If you're still working on some items, feel free to submit your pull request as a draft—our CI will help guide you through the remaining steps.

✅ Required checks

• License: I confirm that my changes are submitted under the Apache License, Version 2.0.

• Commit signatures: All commits are signed and verifiable. (See GitHub Docs on Commit Signature Verification).

• Code formatting: The code is formatted according to the project’s style guide.

• Build & Test: I verified that the project builds and all unit tests pass.

  How to build the project

  Run: ./mvnw verify

  See the build instructions for details.

🧪 Tests (select one)

• I have added or updated tests to cover my changes.
• No additional tests are needed for this change.

📝 Changelog (select one)

• I added a changelog entry in src/changelog/.2.x.x. (See Changelog Entry File Guide).
• This is a trivial change and does not require a changelog entry.

[vy]

Thanks so much for the patch. Would you mind sharing some context, please? Did you observe a particular performance issue?

…

Message ID: ***@***.***>

[kilink]

Did you observe a particular performance issue?

I don't recall how I stumbled upon it or if it showed up in any performance profiles (it definitely has for other libraries). It's just a common thing I've noticed across various projects (e.g., spring and grpc-java) since the HashMap constructor is a foot gun and it's not easy to do correctly until Java 19. Since the intent here was to presize the map, and we know how many elements to insert, it just seemed to make sense to size it so we always avoid the resize / rehash operations.