nordisk/znippy-zoomies

linflate — crate card

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

• One-line pitch: A fast pure-Rust DEFLATE (inflate) decompressor with SIMD match-copy, branchless refill, and a segment-aware API — the shared codec core for the workspace's zip/jar/gzip crates.
• Does: Full-buffer, zero-copy inflate following libdeflate's architecture with zlib-ng's branchless bit refill and SIMD match copy, plus a thread-local table pool. Primary API inflate_into(compressed, &mut output) -> Result<written, InflateError> (caller allocates uncompressed_size + OVERWRITE_HEADROOM for SIMD overwrite slack) and a segment-aware inflate_segment. InflateError enumerates the failure modes. Modules: bitreader, tables, fixed, copy (SIMD match copy), fastloop, deflate_scan.
• Key deps: none at runtime (pure std). Dev: miniz_oxide, criterion.
• Shape: lib only; ~2.5k LoC (lib 648, tables 494, copy 367, fastloop 356, deflate_scan 307, bitreader 301). Benches inflate and deflate_scan.
• Observed role in the workspace: The DEFLATE engine shared by lgz, ljar, and lzip-parallel (each re-exports or depends on it). ljar's doc note: linflate "grew out of" ljar's embedded inflate, was extracted to be shared, and fixed the refill_slow bits==56 corruption bug the old copies carried.

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: