Git Cheatsheet

Commonly Used Commands

Discard the last commit (completely remove the commit from working dir and history)

git reset --hard HEAD^1

Throw away all local commits

git reset --hard origin/<branch_name>

Download all submodule while cloning the repo

git clone --recurse-submodules [URL]

Download all submodules in the repo

git submodule update --init --recursive

Completely remove a file (not only work dir but also in history commits) from the repo and repo history. This is really important if accidentally pushed confidential files like API keys.

git filter-branch --tree-filter 'rm -f my_file' HEAD
git push -f # this changes the entire commit history

Commit and Worktree

Each commit is a file snapshot of the repo, indexed by the a unique hash (SHA, 40 char long checksum hash).

Git history is managed as tree graphs (worktree), each commit is a node of the tree. When creating a new commit with git commit -m "[message]", we create a new node, and its parent is pointing to the previous commit.

branch is a line of development. It can be materialized as a snapshot of the tree. The working tree is associated at most one branch (it might be detached, which is not associated with any named branch), git branch [branch-name] creates such a new branch and git merge [branch-name] merges two branches into one.

HEAD is the reference to the current node (which commit and branch).

branch

Reference Examples

git reset [commit] undoes all commits after [commit], preserving changes locally
git reset --hard [commit] discards all history and back to the specified commit
git revert [commit] create a new commit, which the file changes are reverting the given commit.
git branch [branch-name] create a new branch
git checkout [branch-name] switch to the specified branch
git merge [branch-name] combine the specified branch into the current branch.
git branch -d [branch-name] delete branch
git tag [tag-name] create a tag from the current commit, usually for tagging specific versions for release.

Basic Commands Reference

Repos and Sync Remote

git init init repository
git clone [URL] clone repository
git fetch download the history (without updating the files) from all remote tracking branches
git merge combine remote tracking branch into current local branch
git pull update current local branch with new commits from corresponding remote branch
git push updates all local branch commits to corresponding remote branch

Commit Changes

git add [file] add files to be tracked
git status show the working tree status
git commit -m "[message]" create a commit (a file snapshot)

Inspect History

git log list commit history for the current branch
git diff [snap1] [snap2] diff the two branches/commits
git show [commit] outputs metadata and changes of the commit

Configure repo

git config --global user.name "[name]" config the name shown in the git metadata, --global means setup for all repos on this machine.
git config --local user.email "[email address]" config the email, --local means setup for this repo on this machine only.

`.gitignore`

Files to be excluded from being tracked (cache, build artifacts, binaries, etc.). Each line is a file pattern to be excluded.

Stash Temporary Changes

git stash maintain a stack of file changes snapshot

git stash save currently staged changes (without creating a commit) and push the snapshot to the stack, and revert them in the local dir
git stash pop pop the stack and write the stashed changes to the local dir
git stash drop discard the top of the stack

Submodules

Submodule is to link another repo into the current repo, so that we can use them as a library without copying and syncing their changes.

git submodules is implemented as .gitmodule file, where each entry in

[submodule "lib1"]
    path = lib1
    url = https://example.com/demo/lib1

git submodule add [url] add the repo as a submodule to current dir, it can also add specific branch/commit/tag, see the help doc.
git submodule update [submodule] update the submodule according to how it is configured.