Git

Config

git config --global user.name "username"

git config --global user.email "[email protected]"

git config --global color.ui auto (Colorization of command line output)

git config --global core.excludesfile ~/.gitignore_global (Create global ignore file at ~/.gitignore_global)

Setup

git clone <url> (Download existing repository)

git init <reponame> (Initialize new repository)

git status (Check status of repository)

git remote add origin <url> (Specify remote repository)

Workflow

Describing the above picture: "git pull fetches remote changes into the local clone, and merges them into the current working files. git checkout replaces the current working files with files from a branch. git checkout --track creates a local branch from a remote branch, links them, and replaces the current working files with files from that branch. git fetch downloads changes from a remote repository into the local clone git reset makes the current branch point to some specific revision or branch. git reset --hard makes the current branch point to some specific revision or branch, and replaces the current working files with the files from that branch. git merge merges files from a given branch into the current branch. git push uploads changes from local branches to the respective remote repositories. git add puts current working files into the stage (Aka index or cache) git commit commits staged changes to a local branch git commit -a commits all modified files to a local branch (shorthand for git add and git commit)" - Daniel Kinzler

git commit -m "<message>" (Otherwise you type the message into prompt)

git commit --amend -v (Amend a commit)

git rm

git rm <filename> (Removes file)

git rm --cached <filename> (Removes file from index, but not from filesystem)

git rm -r --cached (Remove all files from index, recursivly)

.gitignore

.gitignore will not automatically untrack already commited files. To untrack already commited files you must remove them from the index, but not from filesystem! All subsequent commits will ignore the files.

Copy
# Ignore Mac system files
.DS_store
# Ignore node_modules folder
node_modules
# Ignore all text files
*.txt
# Ignore files related to API keys
.env
# - relative path understood by git
/log/*.log
# - relative path understood by the linux shell
./log/*.log

.gitignore can also be used by other aplications (such as tar), but the syntax required is different, as noted above.

git show

git show

git show --oneline

git show --oneline <commit>

git show --oneline <commit>~1 (Parent of <commit>)

git show --oneline <commit>~2 (Grandparent of <commit>)

Since branches have two commits:

git show --oneline <commit>^1(1st parent of <commit>)

git show --oneline <commit>^2(2nd parent of <commit>)

Combined:

git show --oneline <commit>~3^2(2nd parent of the great-grandparent of <commit>)

git reset and git revert

git reset (Cancels commits and leaves files uncommitted)

git reset --soft <commit> (Code appears uncommitted as it would from <commit>)

git reset --hard HEAD (Back to the most recently-committed state)

git reset --hard <commit> (Back to <commit>)

git revert (Creates a revert commit, otherwise same as reset)

git branch

git branch (See branches)

git branch <newbranch> (Create a new branch)

git branch -d <branch> (Delete <branch>)

git branch -D <branch> (Force delete <branch>)

git branch -D $(git branch --merged | grep -v \* | xargs) (Force delete all --merged branches)

git push origin <local_branch>:<remote_branch>

git push origin :<branch_name> (Delete a branch while pushing to a remote server)

git checkout

git checkout -b <branch> (Change working directory to that of <branch>)

git checkout <commit> (Change working directory state to that of <commit>)

git stash

git stash (Save code without making a commit)

git stash list (List stashes)

git stash apply (Recall stashed state)

git stash apply stash@{1} (Recall second before last stash; count starting with 0 here)

git drop <stash> (Remove <stash>)

git stash clear (Remove all stashes)

git checkout . (Resets all uncommitted code)

git stash <branch> (Copy stash to <branch>)

git merge

git merge <branchname>

git merge --squash (Combine commits)

git diff --merge (Diff between the state of the source file in the staging area, and the versions from the parent branches)

git diff and patch

git diff file1.txt file2.txt

• “-” for each line that is in the first file but not in the second
• “+” for each line that is in the second file but not the first

git diff file1.txt file2.txt > patch.txt (Create patch file to reconcile differences between file1.txt and file2.txt)

git diff <first-branch> <second-branch> (Differences between two branches)

git log

git log

git log --pretty=oneline

git log --graph --oneline

git log --follow <file> (Version history including renames)

git log --merge (Designed to be used during a merge conflict)

Similar to git log

git shortlog (Uses summary lines to create changelog-like output)

git format-patch --stdout (Converts a series of commits to a series of emails)

git bisect

Search between good and bad commits to find a problem.

git bisect start (Start git bisect) git bisect bad <commit> (Tell git bisect <commit> is the bad one) git bisect good <commit> (Tell git bisect <commit> is the good one) git bisect reset (Reset to latest state)

git rebase

An advanced topic, but just to give the gist...

Do not (lightly) rebase changes in already shared repository as it will change the commit-ids and make combining those changes difficult.

git rebase -i HEAD~# (Change the commit messages, the order or the number of commits)

git rebase -i --onto <newbase> --root [<branch>] (If running Git locally and have not pushed your branch to a remote repo)

git rebase -i [--onto <newbase> | --keep-base] <upstream> [<branch>] (To push upstream)

git rebase -i --abort (Abort the rebase!)