[Performance] Use dynamic bitmaps in AlignedTVList

[jt2594838]

Description

Use dynamically selected bitmap implementations

Use BitMap.createBitMapDynamically for bitmap allocations created by AlignedTVList and LazyBitMap. With the default primitive_array_size=64, this selects the long-backed implementation and avoids a separate byte-array allocation.

Update bitmap memory accounting to use ramBytesUsed(), and grow the time-deletion bitmap with extend.

Preserve bitmap behavior

The long-backed implementation returns a copy from getByteArray(). Result-status bitmap updates therefore use merge instead of mutating the returned array. The public byte-array helper keeps its previous exact-length result, including byte-aligned sizes.

The all-value-deleted calculation reads each bitmap byte array once to avoid repeated temporary-array allocation.

Tests and performance

Added unit coverage for:

• reduced bitmap retained size;
• failed-status merging with non-byte-aligned offsets across a 64-row block boundary;
• exact byte-array length for byte-aligned result bitmaps.

Added an opt-in local standalone performance IT. Workload: 256 aligned INT64 measurements, 64 rows per tablet, one null per measurement per tablet, 50 warmup batches, and 5 rounds of 100 batches.

Three independent JVM runs were compared using the median reported by each run:

Implementation	Cross-run median	Throughput
new BitMap(64)	123.145 ms/round	13.30M points/s
createBitMapDynamically(64)	110.376 ms/round	14.84M points/s

The dynamic bitmap reduced round time by about 10.4% and increased throughput by about 11.6%.

---

This PR has:

• been self-reviewed.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage.
• added integration tests.
• been tested in a test IoTDB cluster.

---

Key changed/added classes

• AlignedTVList
• LazyBitMap
• AlignedTVListTest
• IoTDBAlignedTVListBitMapPerformanceIT