{"schema_version":1,"metadata":{"branch":"release","release":"8.0","architecture":"amd64","generated_at":"2026-10-21T02:35:53.703506+00:00","source_url":"https://cdn.openbsd.org/pub/OpenBSD/8.0/packages/amd64/sqlports-7.55.tgz","source_sha256":"393d2bd6f5b23c8ba1809ec63f3e8310d7787d329b7350089609a078eb01a387","package_count":12098,"source_kind":"sqlports"},"package":{"name":"mkz-0.1.2","path":"archivers/mkz","url":"/packages/release/archivers/mkz/","comment":"reversible columnar pre-pass + zstd archiver","homepage":"https://github.com/gitspoked/mkz","maintainer":"Matt Klein <mk@ntele.net>","description":"mkz is a tar-style archiver whose backend is a reversible, schema-free\nauto-columnar transform followed by zstd. It is not a new compressor: it is\na structural pre-pass that makes zstd measurably smaller on structured,\nline-oriented data (logs, CSV, JSONL, telemetry) and is provably never worse\nanywhere else: each block keeps whichever of raw-zstd or transformed-zstd\nis smaller.\n\nBoth creating and extracting stream in constant memory (roughly one block,\nindependent of archive or member size), and every archive carries a SHA-256\nof the reconstructed original, verified on extract, so corruption is always\ndetected and reported (streaming extraction is not yet atomic, so already-\nwritten files may remain on failure).\n\nThis is the pure-C build: it depends only on libzstd (SHA-256 and base-95 are\nvendored) and is byte-compatible with the reference implementation.\n","description_source":null,"description_unresolved":[],"package_architecture":"amd64","stem":"mkz","readme":null,"dependencies":[{"path":"archivers/zstd","type":"library","package_spec":"","url":"/packages/release/archivers/zstd/"}],"reverse_dependencies":{"count":0,"url":null},"categories":["archivers"],"flavors":[],"only_for_architectures":[],"not_for_architectures":[]}}