nordisk/znippy-zoomies

lbzip2 — crate card

Reverse-engineered (from source — verify, don't trust the name)

• One-line pitch: A pure-Rust parallel bzip2 decompressor that scans block boundaries and decodes Burrows-Wheeler blocks across many cores.
• Does: One reader thread fills a ChunkRevolver ring buffer (~100 MB slots) with raw compressed bytes; a cheap bit-scan locates bzip2 block magic (0x314159265359 at arbitrary bit offsets) and dispatches sub-ranges to a worker pool; each worker independently decodes its BWT blocks and output is reassembled in order and streamed as Read. Modules cover the full bzip2 pipeline: bitreader, block/block_scan, bwt, huffman, mtf, chunk, reader, stream, plus par/parallel (rayon-free scoped-thread work-stealing map). Exposes BLOCK_MAGIC/FINAL_MAGIC constants.
• Key deps: memchr only (runtime). Deliberately rayon-free — parallelism is its own par::par_map, worker count from physical cores or LBZIP2_THREADS. Dev: bzip2 (reference/round-trip).
• Shape: lib + bin lbunzip2 (471 LoC); ~2.9k LoC. Feature timing. Benches under benches/ are registered as [[test]] --ignored profilers (autobenches off).
• Observed role in the workspace: Bzip2 decode backend for osm-katana (handles bzip2-compressed OSM XML). Version codename "Sleipner".

Fill in (one line each — only the human knows these)

• Primary consumer(s):
• Why it exists (vs the obvious alternative):
• Hard constraint (airgap / no-C-deps / no-FFI / pure-Rust / etc.):
• Release cadence & independence (rides a parent? own repo someday?):
• Status (production / prototype / experimental / stubbed):
• Must-NOT-change invariant:
• Biggest known gap or TODO:
• Anything a newcomer would get wrong about it: