Skip to content

What is the architecture of Git?

What is a two-tree architecture?

Other version control systems use a two-tree architecture. The two trees refer to the repository and the working directory:

flowchart LR
    A[repository]
    B[working directory]

You "check out" files from the repository and edit them within your working directory:

stateDiagram-v2
    s1 : repository
    s2 : working directory
    s1 --> s2: checkout
    s2 --> s1: commit

After those changes are made, you commit the changes back to the repository.

What is a three-tree architecture?

Git uses a three-tree architecture. It has:

  • A repository that's tracked by Git.
  • A working directory that contains changes that may not be tracked by Git yet.
  • A staging index in between the repository and working directory that contains changes we plan to commit to the repository.

flowchart LR
    A[repository]
    B[staging index]
    C[working directory]
In Git, you have a copy of the repository in your working directory where you can edit files. When you're ready for Git to start tracking your changes, you add files to the staging index. Then, you commit your changes to the repository along with a message.

Sources
  • Skoglund, K. (2019). Git Essential Training: The Basics [Video]. LinkedIn Learning. https://www.linkedin.com/learning/git-essential-training-the-basics
  • Pro Git