bit

A simplified version control system written in Rust. Built from scratch to understand how Git works under the hood — from object hashing and storage to repository initialization.

Why?

Most of us use git every day without questioning what happens under the hood. This project is a learning exercise to answer questions like:

• How is a file stored as a blob object?
• What does content-addressable storage really mean?
• What actually happens when you run git init?

Features

• bit init — initializes a repository (.bit directory with objects, refs, HEAD, and index)
• bit object create <path> — creates a blob object from a file (SHA-1 hash + zlib compression, same scheme as Git)
• bit object cat-object <hash> — reads and decompresses an object by its hash
• CLI stubs for clone, commit, and push (work in progress)

Building

cargo build

Usage

# Initialize a repository
bit init

# Create an object from a file
bit object create path/to/file.txt

# Read an object back by its hash
bit object cat-object <hash>

How it works

Object storage

Like Git, bit stores content-addressable objects:

1. A header is built as blob <size>\0<content>
2. The header + content are hashed with SHA-1
3. The object is compressed with zlib and stored at .bit/objects/<first 2 chars of hash>/<remaining hash>

Repository discovery

bit walks up the filesystem from the current directory until it finds a .bit directory, mirroring how Git finds the repo root.

Project structure

src/
├── main.rs            # CLI entry point (clap)
├── internals/
│   ├── objects.rs     # Object creation and reading (blob/tree)
│   └── database.rs    # Object database (in progress)
├── phases/
│   ├── init.rs        # Repo initialization
│   └── index.rs       # Index/staging (in progress)
├── hashing/           # Hashing logic (in progress)
├── traits/            # Shared traits
└── utils/
    └── path.rs        # Filesystem utilities

Roadmap

• Tree objects for directories
• Commit objects
• Staging area / index
• clone support
• commit support
• push support

Technologies

• Rust (edition 2024)
• clap — CLI argument parsing
• sha1 — hashing
• flate2 — zlib compression
• const-hex — hex encoding
• strum — enum serialization