Inspecting Commits

git show displays a single commit: its details and the changes it made. To use it — and almost every other Git command — you need ways to name commits: hashes, branch names, HEAD, and relative references like HEAD~2.

git show

With no argument, git show describes the commit you're on:

git show

Output

commit 8d99c4fbe68c42d6ab76dba6ef77e497a6980698 (HEAD -> main)
Author: Sam Rivera <sam@example.com>
Date:   Mon Mar 2 09:42:00 2026 +0000

    Add cheese to omelette

diff --git a/omelette.txt b/omelette.txt
index b4927c2..671eb8e 100644
--- a/omelette.txt
+++ b/omelette.txt
@@ -1 +1 @@
-Omelette: eggs, butter
+Omelette: eggs, butter, cheese

Give it any reference to see a different commit. --stat lists files instead of the full diff, and --oneline shortens the header:

git log --oneline

Output

8d99c4f (HEAD -> main) Add cheese to omelette
cfe25ef Add sugar to pancakes
90c0089 Add omelette recipe
c4f5e47 Add pancake recipe

git show --stat --oneline HEAD~1

Output

cfe25ef Add sugar to pancakes
 pancakes.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Naming commits

Reference	Means
`a1b2c3d`	A commit hash — any unique prefix of at least 4 characters
`main`, `feature/login`	The latest commit on that branch
`HEAD`	The commit you're on now
`HEAD~1` or `HEAD~`	Its parent — one commit back
`HEAD~3`	Three commits back, following first parents
`HEAD^2`	The second parent of a merge commit
`v1.0`	A tag (lesson 39)
`main@{yesterday}`, `HEAD@{2}`	Where a branch or `HEAD` pointed in the past (lesson 20)

~ walks back along the line of history; ^ picks between parents, which only matters for merge commits. For ordinary commits HEAD^ and HEAD~ mean the same thing. These work everywhere: main~2, a1b2c3d~1 and so on.

git rev-parse turns any reference into its full hash — useful to check what a reference means:

git rev-parse --short HEAD~2

Output

90c0089

On Windows: in Command Prompt, ^ is a special character — write HEAD~1 instead of HEAD^, or quote it as "HEAD^". Git Bash and PowerShell don't have this problem.

A file at a past commit

reference:path shows a file exactly as it was in that commit, without changing anything in your working tree:

git show HEAD~3:pancakes.txt

Output

Pancakes: flour, milk, eggs

To list which files existed in a commit, use git ls-tree:

git ls-tree --name-only HEAD~3

Output

pancakes.txt

Ranges

Many commands accept a range of commits. A..B means "commits reachable from B but not from A" — in a straight line of history, the commits after A up to and including B:

git log --oneline HEAD~3..HEAD

Output

8d99c4f (HEAD -> main) Add cheese to omelette
cfe25ef Add sugar to pancakes
90c0089 Add omelette recipe

With branches this answers useful questions. main..feature lists the commits on feature that aren't on main yet — what a merge would bring in. And origin/main..main lists commits you haven't pushed (lesson 35).

git log --oneline main..brunch

Output

7ab5594 (brunch) Add feta to shakshuka
1b53e7f Add shakshuka recipe

Describing a commit

A few more commands report on commits without changing anything:

Command	Shows
`git show --name-status HEAD`	Each file with A (added), M (modified) or D (deleted)
`git branch --contains a1b2c3d`	Which branches include a commit
`git merge-base main brunch`	The last commit two branches have in common
`git cat-file -p HEAD`	The raw commit object (lesson 42)

git branch --contains HEAD~1

Output

  brunch
* main

Try it yourself

In a repository with a few commits, print the file you changed most recently as it was two commits ago, without checking anything out. Then list the commits between HEAD~3 and HEAD with their changed files.

Previousgit diff Next.gitignore