In order to use Git effectively, it is necessary to understand how it works internally.

What Is A Git Repository?

The main things to remember are:

• A Git repository is a directed acyclic graph (DAG) of commit objects.
• Branches are pointers to individual commit objects

Some more details are:

• A commit object contains pointers to all the files in the repository
• A new commit object can be created on the base of one ore more other commit objects
• The commit object on which another commit object is based is its parents
• Each commit object knows its immediate parents
• At any time, one commit object is checked out, that is, its files are physically present in the file system

Pointers

• Branches, remote branches, tags, are just pointers to individual commit objects
• HEAD is a special pointer that points to a specific branch (and the branch points to a commit object)
• These pointers serve to “nail down” certain commit objects, so that these commit objects can be easily located and checked out

Types of Git objects

There are three types of Git objects:

• blob: single piece of data (e.g. of a single file); all your files in a repository are presented internally as blob objects
• tree: contains one or more pointers to blobs or other trees
• commit: wrapper around a tree object, including information like author, date, commit message, and parent commit object(s)