Throughput of in-flight message renewal degrades while some handle data locks have expired

### Before Creating the Enhancement Request

- [x] I have confirmed that this should be classified as an enhancement rather than a bug/feature.

### Summary

In ReceiptHandleGroup.HandleData.lock(long timeoutMs), the current implementation always blocks for the full timeoutMs (default 3 seconds) on semaphore.tryAcquire(timeoutMs, ...) before checking whether the lock has expired.

This means even when the lock holder has been stuck for well over the expiration threshold (9 seconds), the calling thread still wastes up to 3 seconds waiting on the semaphore before it discovers it can force-acquire via the expiration path.

This unnecessary blocking impacts proxy throughput in scenarios like message renewal and could lead to significant duplicate consumptions due to untimely renewal.

### Motivation

Optimize in-flight message renewal throughput in lock expired scenario.

### Describe the Solution You'd Like

Add a non-blocking tryAcquire() as the fast path, and check the expiration condition before falling through to the blocking tryAcquire(timeout, unit)

### Describe Alternatives You've Considered

/

### Additional Context

_No response_

Start at ReceiptHandleGroup.HandleData.lock(long timeoutMs) and inspect how the semaphore and expiration condition are currently checked. Verify the lock-renewal path, then run the relevant existing tests for ReceiptHandleGroup or lock expiration; done means expired locks no longer wait through the full timeout before the expiration path can acquire them.

Written by the indexing model from the issue text.