Frequently asked Git and GitHub interview questions on commits, the staging area, branches, merge and rebase, reset and revert, conflicts, pull requests, workflows, the reflog and Git internals, with hints and answers.
22 questions · Updated
One is a tool, the other a website.
Git is a version control tool that runs on your computer and works offline. GitHub is a website that hosts Git repositories, and adds pull requests, issues and automation. GitLab and Bitbucket are alternatives to GitHub.
A draft of the next commit.
The staging area (or index) holds what will go into the next commit. git add copies changes there, and git commit saves them. It lets you split your work into small, focused commits, even when you have changed several things at once.
How is the ID calculated?
A commit is a snapshot of the project with an author, date, message and parent commit. Its ID is a hash of all of that, so any change to the content or history gives a different ID.
Which one touches your branch?
git fetch downloads new commits and updates origin/main, but it does not change your branch or files. git pull fetches, then merges or rebases those commits into your current branch.
How big is it on disk?
A branch is a lightweight pointer to a commit, which moves forward with each new commit. Because it is only a small reference, creating one is instant. Branches let you work on features without affecting main.
Markers, then add, then commit.
Open the conflicted files, choose the right content, and delete the <<<<<<<, ======= and >>>>>>> markers. Run git add on each file, then git commit to finish the merge. git merge --abort cancels it.
Ignore rules only apply to some files.
.gitignore only applies to untracked files. Stop tracking the file with git rm --cached <file> and commit that. The file stays on disk, but it remains in old commits.
One keeps history, the other rewrites it.
merge joins two branches, adding a merge commit if they diverged, and leaves existing commits unchanged. rebase replays your commits on top of another branch, giving a straight history, but it creates new commit IDs. Do not rebase commits that others have used.
Which one is safe on a shared branch?
git reset moves the branch back, removing commits from it, which rewrites history. git revert adds a new commit that undoes an earlier one, so it is safe on shared branches.
Where do the changes end up?
All three move the branch. --soft leaves the changes staged. --mixed, the default, leaves them in your files but unstaged. --hard discards them completely.
An urgent interruption.
When you need a clean working tree but your work is not ready to commit, for example to switch branches for an urgent fix. git stash saves the changes, and git stash pop restores them. Add -u to include untracked files.
One commit, not a whole branch.
It copies a single commit to the current branch as a new commit. It is commonly used to bring a bug fix to a release branch without merging everything else.
Two common ways, plus a GitHub button.
Use git rebase -i HEAD~n and mark commits as squash or fixup. Or use git reset --soft HEAD~n and commit again. On GitHub, Squash and merge does it when merging a pull request.
What if a teammate pushed?
--force overwrites the remote branch no matter what. --force-with-lease refuses if the remote has commits you have not seen, so it does not silently delete a teammate's work.
From branch to delete.
Pull the latest main, create a branch, commit and push. Open a pull request, let the automated checks run and get a review. Push any fixes to the same branch, merge, delete the branch, and pull main.
How often do they release?
GitHub Flow uses short branches off an always-deployable main. Git Flow adds develop, release and hotfix branches for scheduled releases. Trunk-based development merges into main at least daily, with feature flags hiding unfinished work.
Git records where HEAD has been.
Run git reflog to find the commit, then git reset --hard HEAD@{n} or git branch rescue <id>. The reflog is local and its entries expire, but commits normally stay recoverable for weeks.
Binary search.
Use git bisect. Mark a bad commit and a good commit, and Git checks out commits in between, halving the range each time. git bisect run <test command> automates it.
What comes first, the key or the history?
Revoke the key first, because it must be treated as leaked. Then remove it from history with git filter-repo if needed, force-push, and have the team re-clone. Add secret scanning, .gitignore rules and hooks to prevent a repeat.
There are four types.
Blobs hold file contents, trees hold folder listings, commits point to a tree and parents, and annotated tags point to other objects. Each is identified by a hash of its content, so identical files are stored once.
What does a commit point to?
Snapshots. A commit points to a full tree, and unchanged files reuse existing blobs. Differences are calculated on demand. Packfiles use delta compression, but only to save storage.
Which three commits?
Git compares both branch tips with their common ancestor. A part changed on one side only takes that change. A part changed differently on both sides is a conflict.