[fix][broker] Fix delayed message index data loss when trimming overlapping bucket snapshots

[void-ptr974]

Motivation

When bucket trimming selects an immutable bucket that overlaps the first active ledger, RangeMap returns a clipped range key. The delete path still deletes the complete bucket snapshot, so the delayed-delivery index cannot be recovered after a broker restart.

Modifications

• Select only buckets whose complete immutable range ends before the first active ledger.
• Add a regression test for a bucket that overlaps the first active ledger.

Verifying this change

• Make sure that the change passes the CI checks.

This change added tests and can be verified as follows:

• Ran the bucket trimming tests covering complete deletion, delete failure, no orphaned buckets, and overlapping buckets.
• Ran pulsar-broker checkstyleMain and checkstyleTest.

Does this pull request potentially affect one of the following parts:

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment