Skip to content

Lesson

Commits

How a commit object is built, how it links to its parent, and how the branch and HEAD move.

Beginner 6 min read

Learning objectives

  • Describe the parts of a commit object
  • Explain parent links and the commit graph
  • Understand how committing moves the branch and HEAD

What a commit stores

A commit object records a tree (the full snapshot of your files), one or more parent pointers, the author and committer, a timestamp, and your message. The commit's hash is derived from all of this, so any change produces a different hash.

The parent link

Each commit points back to the commit that came before it. Following these links backward walks your history. This is what makes the history a graph rather than a flat list.

Moving forward

After git commit, the current branch pointer moves to the new commit, and because HEAD points at that branch, HEAD moves too. Your staging area is cleared.

Try it yourself

interactive visualization

Empty folder

Run git init to create a repository.

Working directory

Clean — no uncommitted changes.

git add ↓

Staging area

Nothing staged. Run git add.

git commit ↓

Repository

No commits yet.

1/7Empty repositoryOpen in visualizer

Make two commits and watch the second link to the first as its parent.

Commands used

terminal
git init
git add a.ts
git commit -m "c1"
git add b.ts
git commit -m "c2"
git log

Common mistakes

  • • Writing vague messages like 'updates' or 'fix stuff'.
  • • Bundling many unrelated changes into one commit.

Best practices

  • • Write imperative messages: 'add', 'fix', 'refactor'.
  • • Keep commits atomic — one logical change each.

Mini challenge

Make three commits, then use git log --oneline to read the chain of parents.

Open the visualizer to try it

Summary

  • A commit = snapshot + parent + metadata + message.
  • Commits form a graph via parent links.
  • Committing moves the branch and HEAD forward.
Next lessonBranches