git-lfs-filter

Clean and smudge filters and the filter-process protocol for Git LFS.

Git invokes content filters whenever a file moves between the working tree and a git blob: a clean filter runs on the way in (git add) and a smudge filter runs on the way out (git checkout). LFS hooks into both ends. Clean hashes the working-tree bytes, hands them to the local LFS store, and emits a small pointer file (which is what git ends up storing); smudge takes the pointer back from git, looks up the real bytes (fetching from the server if they're not local), and writes the content into the working tree.

This crate implements both filters plus the long-running filter-process protocol, which modern git uses by default: one subprocess handles many files in a single session over a pkt-line-framed connection. The three entry points are clean, smudge, and filter_process, invoked as the bodies of git lfs clean, git lfs smudge, and git lfs filter-process respectively.

Pointer extensions chain external programs between the raw bytes and the stored object. On clean, content passes through each registered extension in priority order, with each stage's input OID recorded in the resulting pointer; smudge undoes the chain in reverse to reconstruct the original bytes. Used for case-inverters, content-defined chunking, encryption shims, or similar transforms; configured via lfs.extension.<name>.* in git config.