Parallelize InMemoryIndex construction with bucket put loop.

[dmkozh]

Description

Build the LiveBucketIndex on an async worker thread while the put loop in mergeInMemory runs on the main thread. Both only read mergedEntries as const, so they need no synchronization.

Checklist

• Reviewed the contributing document
• Rebased on top of master (no merge commits)
• Ran clang-format v8.0.0 (via make format or the Visual Studio extension)
• Compiles
• Ran all tests
• If change impacts performance, include supporting evidence per the performance document

[Copilot]

Pull request overview

This PR parallelizes level-0 LiveBucket::mergeInMemory work by constructing the LiveBucketIndex on a worker thread while the main thread writes merged entries to the bucket file, then passes the pre-built index into BucketOutputIterator::getBucket() to avoid rebuilding it.

Changes:

• Spawn LiveBucketIndex construction via std::async during LiveBucket::mergeInMemory while the merge output is written to disk.
• Extend BucketOutputIterator::getBucket() to accept an optional pre-built index (preBuiltIndex) and prefer it when provided.
• Wire the pre-built index from mergeInMemory into getBucket().

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
src/bucket/LiveBucket.cpp	Starts async index construction during in-memory merge and forwards the built index into bucket adoption.
src/bucket/BucketOutputIterator.h	Adds an optional preBuiltIndex parameter to getBucket() with a default nullptr.
src/bucket/BucketOutputIterator.cpp	Uses preBuiltIndex when provided to avoid recomputing an index from in-memory state or disk.