v2.5.2

Highlights

• Archive extraction is significantly faster and more reliable, with batch indexing and async build jobs replacing the previous per-file synchronous approach
• Task history UI is streamlined with a cleaner table and removed legacy component

Added / Changed

• Added FastUncompressedBytes and FastFileEntryCount to the archive extractor interface, avoiding expensive full decompression passes for compressed tarballs when an exact count or size is not required
• Added AddRawBatch to the DroneDB adapter, indexing multiple files in a single native database transaction instead of opening a connection per file
• Rewrote archive extraction to extract all files first, then index them in batches of 250, and enqueue per-file build jobs asynchronously rather than building derivatives inline
• Added incremental disk space guards during extraction with automatic rollback if the server runs out of free space
• [Hub] Improved task history table with a cleaner UI and removed the legacy BuildHistory component

Fixed

• Archive extraction no longer blocks until all derivatives are built, reducing task execution time for large archives
• Compressed tarball extraction no longer triggers a full decompression pass just to count entries or estimate size

Full Changelog: v2.5.1...v2.5.2