Skip to main content

​
A real Git

At its core, Git is a key-value store. When a file is added to a commit, its contents are hashed using the SHA-1 algorithm. That hash becomes the key used to store and retrieve the file’s data. Git commands fall into two categories:

Porcelain commands

High-level, user-friendly commands you use every day.Examples: git add, git status, git commit, git stash

Plumbing commands

Low-level commands that give direct access to Git internals.Examples: git hash-object, git cat-file, git ls-files, git rev-parse, git ls-remote

​
Plumbing commands

​
hash-object

git hash-object hashes the contents of a file and returns its SHA-1 hash. 
git hash-object <filename>

# Example
git hash-object hello
The returned hash (e.g. 492f47831b16c8217339fcb1449345b424c72fcb) is used as the storage key. The first two characters form the folder name inside .git/objects/, and the rest is the file name within that folder.

​
cat-file

git cat-file reads and displays the contents of a stored Git object by its hash. 
git cat-file -p <object-hash>

# Example
git cat-file -p 492f47831b16c8217339fcb1449345b424c72fcb
The -p flag pretty-prints the contents.

​
Git object types

The .git/objects/ folder contains three types of objects:
A commit object stores a snapshot pointer, author information, a timestamp, and a reference to the previous commit. Each commit is identified by a unique hash.
A tree object represents a directory. It maps file names to blob (or other tree) hashes, mirroring the structure of your project’s file system.
A blob object stores the raw contents of a file. For example, when you hash a file called hello, its contents are stored as a blob.
Understanding Git’s object model helps you reason about what happens under the hood when you run everyday commands like git add and git commit.

Build docs developers (and LLMs) love

Get started for freeTalk to us