This article was updated Dec 09, 2023

1. History of Git

Git is an open source software for distributed version control. It was created in 2005 by Linus Torvalds (creator of Linux) as a replacement for BitKeeper to manage Linux kernel source code.

2. Git application

Git is usually used for coordinating work among programmers collaboratively developing source code during software development. Its goals include speed, data integrity, and support for distributed, non-linear workflows (thousands of parallel branches running on different systems).

GitHub, GitLab and Bitbucket are a cloud based hosting services based on Git software.

3. Git installation (Windows)

  • Download latest version of Git Bash from official Git sites: https://git-scm.com
  • Follow the steps and install the package
  • This is how I personally proceed the installation procedure:
    • TrueType: yes
    • Default editor: Notepad++ (up to user preference)
    • PATH environment: Git from command line and from 3rd-party software
    • Install MSYS2 (ridk install): yes
    • Use OpenSSL library: yes
    • Windows-style: no
    • Unix-style line ending: yes
    • MinTTY terminal (Linux style terminal): yes

4. Git configuration

4.1 Three levels of Git configuration:

  • System level configuration
    • Config storage: Program Files\Git\etc\gitconfig*
    • See actual settings: git config --system --list
    • Change settings: git config --system <options>
  • User level configuration
    • Config storage: *$HOME\.gitconfig*
    • See actual settings: git config --global --list
    • Change settings: git config --global <options>
  • Project level configuration
    • Config storage: project_folder\.git\config
    • See actual settings: git config --local --list
    • Change settings: git config --local <options>

4.2 Essential configuration of Git:

To make Git work properly, it is necessary to set Username and Email in user level configuration:

  • Open Git Bash
  • Set your Username with command:
    git config --global user.name "Your Name"
  • Set your Email with command:
    git config --global user.email "you@example.com"

4.3 Optional configuration of Git:

  • You can set your default text editor by command:
    • Vim: git config --global core.editor "vim"
    • Notepad: git config core.editor notepad
    • Notepad++: git config --global core.editor "'C:/Program Files/Notepad++/notepad++.exe' -multiInst -notabbar -nosession -noPlugin"
    • Visual Studio Code: git config --global core.editor "code --wait"
    • Atom: git config --global core.editor "atom --wait"
    • Sublime Text: git config --global core.editor "'C:/Program Files (x86)/sublime text 3/subl.exe' -w"
    • Other text editors
  • You can allow colors in Git with command:
    git config --global color.ui true
  • You can configure Git to handle line endings (CRLF vs. LF) automatically so you can collaborate effectively with people who use different operating systems with command:
    git config --global core.autocrlf true (in Windows ecosystem)
    git config --global core.autocrlf input (in Linux and macOS ecosystems)

4.4 Optional configuration of Bash

Bellow you can see my personal Bash settings.

  • Open Git Bash and open “.bashrc” config by command: 
      notepad.exe ~/.bashrc
  • You can paste your setting in “.bashrc”, e.g. something like this: 
      # Use Notepad++ instead of classic Notepad
  alias notepad="/c/Program\ Files/Notepad++/notepad++.exe"
	
  # Enable color support for ls, dir, grep...
  alias ls='ls --color=auto'
  alias dir='dir --color=auto'
  alias vdir='vdir --color=auto'
  alias grep='grep --color=auto'
  alias fgrep='fgrep --color=auto'
  alias egrep='egrep --color=auto'
	
  # Allowing colors in shell
  case "$TERM" in
      xterm-color) color_prompt=yes;;
  esac

  # Ask before copy, move, remove
  alias cp="cp -i"
  alias mv="mv -i"
  alias rm="rm -i"
	
  # Prevents accidentally clobbering files.
  alias mkdir='mkdir -p'

  # Shell history: remember last 1000 commands
  export HISTSIZE=1000

  # Shell history: ignore duplicates and commands starting with space
  export HISTCONTROL=$HISTCONTROL:ignoreboth

  # Shell history: make Bash append to history rather than overwrite it:
  shopt -s histappend

  # Shell history: after each command, append to the history file and reread it
  PROMPT_COMMAND="${PROMPT_COMMAND:+$PROMPT_COMMAND$'\n'}history -a; history -c; history -r"

5. Getting started

5.1 Initialize a local repository:

To get started, create a new project and start to track changes in a project folder.

  • Create project folder with command:
    mkdir First_Project
  • Go to project folder with command:
    cd First_Project
  • Initialize a repository inside project folder with command:
    git init

“git init” creates a “.git” folder inside project directory where all changes are recorded.

5.2 Commit the change

When you finish your changes in your project folder and you are ready to record new version, commit the changes in local repository.

  • First, make your changes in the project folder. You can use your favourite text editor.
  • Add all changes in project folder to the staging index with command:
    git add .
    or
    git add --all

  • Commit the changes to repository with command:
    git commit -m "Initial commit"

    The message parameter “-m” is the way to describe the change. It should be a short single-line summary with less than 50 characters

5.3 View the commit log

Show all commits done inside the project:
git log

Show last 5 commits done inside the project:
git log -n 5

Show all commits done from 1.1.2021:
git log --since=2021-01-01

Show all commits done until 1.1.2021:
git log --until=2021-01-01

Show all commits from David (don’t need to write full name):
git log --author="David"

Show all commits containing “Init” (works with Regex):
git log --grep="Init"

6. Git concepts and architecture

6.1 The 3-trees

Git uses a 3-trees architecture:

  1. Working directory - contains changes that may not be tracked by Git yet
  2. Staging index - contains changes that we are about to commit into repository
  3. Repository - committed files which are tracked by Git

3-trees

6.2 Hash values

  • Git uses SHA-1 hash algorithm to keep data integrity (40-character hexadecimal string - contains 0-9 and a-f)
  • Git generates a checksum for each change set
  • Checksum algorithms convert data into a simple number
  • Same data always equals same checksum

6.3 HEAD

  • Pointer to tip current branch in repository
  • Last state of repository, what was last checked out

7. Make changes to files

7.1 Add new files

  • First, add new file(s) in your repository
  • To see new/untracked files in your repository, use command:
    git status
  • To add specific untracked files to staging index, use command:
    git add <file>
  • To add all files (incl. untracked) files to staging index, use command:
    git add .

7.2 Edit files

When you edit a file in repository, you have to treat it as a new file with following steps:

git add <file>
git commit -m "<message>"

7.3 View changes with diff

To see changes to file(s) use the command:
git diff

This command by default compares working directory against staging index.

7.4 View only staged changes

As soon as you use git add <file>, you won’t see the changes of the file with git diff

To see changes in staging area, use the command:
git diff --staged
or
git diff --cached

These two commands compares repository against staging index

7.5 Mark word changes

By default git diff mark all the lines where change happened. To see changed words, use command:
git diff --color-words

7.6 Delete files

To delete file from Git repository, use the command:

git rm <file>
git commit -m "<message>"

7.7 Move and rename files

To move or rename the file from Git repository, use the command:

git mv <file> <file_moved_or_renamed>
git commit -m "<message>"

7.8 Stage and commit shortcut

There is a shortcut for git add and git commit. To stage and commit all changes to tracked files use command:
git commit -a

To provide a comment about change use:
git commit -am "<message>"

It does not include untracked files

7.9 View a commit

To view what was changed in the past, you can use a command:
git show <hash>

You don’t need to write all 40-character hash. Usually first 6 characters of SHA-1 hash is enough.

To view exact word changes, use this command:
git show <hash> --color-words

To see the last changes, you can use HEAD instead of hash value:
git show HEAD
or
git show HEAD --color-words

7.10 Compare commits

To compare 2 different commits and see the changes, use the command:
git diff <hash_1>..<hash_2>

To view exact word changes, use this command:
git diff <hash_1>..<hash_2> --color-words

7.11 Multiline commit messages

To add a multiline commit message follow these steps:

  • Write a command git commit -a
  • Your default text editor will automatically open a file named COMMIT_EDITMSG
  • Write a multiline commit message to the top part of COMMIT_EDITMSG, save it and close it
  • You can check the the the message with command git log -n 1

To see only first lines of commit messages, use command:
git log --oneline

7.12 Make atomic commits

  • It is a standard approach to commit changes per partes by topic
  • Make small commits that affect a single aspect
  • It is a best practice and it improves collaboration
  • It’s easier to understand, to work with and to find bugs

8. Undo changes

8.1 Undo working directory changes

If you want undo changes in working directory and go back to version from repository, use command:
git checkout -- <file>

8.2 Unstage files

If you want to undo changes in staging index and go back to version from repository, use command:
git reset HEAD <file>

8.3 Change last commit

If you do a commit and forget to add some information, you can additionally add (amend) a change to the last commit by command:

git add <amended_file>
git commit --amend -m "<message>"

This works only for last commit, cannot change older commits in this way

8.4 Retrieve old versions

  • Git only allows editing (amending) the most recent commit
  • Edits which undo older changes are new commits because of data integrity
  • To retrieve the old version of specific file from current branch to staging directory, use the command git checkout <hash> -- <file>
  • With command git diff --staged you can see it is going to revert the change which was done in the past.
  • If you want to revert the commit of specific file, use command git commit -m <message>
  • If you don’t want to revert the commit, use the command git reset HEAD <file> to clean the staging directory and then use git checkout -- <file> to clean the working directory

8.5 Revert a commit

  • It means to do the opposite of what specific commit did
  • To revert a specific commit, use command git revert <hash>
  • Your default text editor will automatically open a file named COMMIT_EDITMSG
  • Write a multiline commit message to the top part of COMMIT_EDITMSG, save it and close it
  • You can check the log of last change with command git log -n 1
  • You can check last changes with git show HEAD
  • If you use atomic commits, it is very easy revert specific change

8.6 Remove untracked files

To remove files which are not tracked by Git yet or files in working directory, you can use one of these commands:

  • To see what would be removed use command git clean -n
  • To removed files step by step use command git clean -i
  • To remove all untracked files use command git clean -f (f like “force”)

9. Ignore files

9.1 Use .gitignore files

To ignore changes of specific files (e.g. temporary files, logs, databases, big files, compressed files, videos, compiled source code, etc.), you can write these files in .gitignore file. Changes made in these files will be ignored by Git. If the .gitignore file does not exist yet, create it in project directory and commit it to see the changes in .gitignore file.

Examples of .gitignore file:

  • logs/*.txt will ignore all txt file in folder log
  • *.php together with !index.php will ignore all php files except index.php file
  • logs/ will ignore all files in folder logs
  • If the line starts with # it is a comment and it won’t do anything, e.g. # Comment
  • Blank lines are skipped

To find .gitignore examples for specific projects, go to page: https://github.com/github/gitignore

E.g. for Jekyll projects it is recommended to ignore these files and folders:

_site/
.sass-cache/
.jekyll-cache/
.jekyll-metadata

9.2 Globally ignore files

  • Ignore files in all repositories
  • Settings are not tracked in repository
  • User specific instead of repository specific

Use e.g. this command to define where to find global settings:
git config --global core.excludefile ~/.gitignore_global

9.3 Ignore tracked files

If a file is already tracked follow these steps to ignore it:

  • You have to make it untracked first with command git rm --cached <file>
  • Then you put it in .gitignore file
  • Finally you commit that the file will be removed from tracking with command git commit -m "<message>

9.4 Track empty directory

By default Git does not track empty directory. To let him track it you have to put some file in it. The convention is to create a blank file .gitkeep in this folder directory and commit it.

10. Branches

Branches create another line of development in the project. By default the main branch in Git is set to master, but it can by changed in Git configuration. E.g. cloud services as Github, Gitlab and DevOps use main as a default branch.

10.1 Local branches

Local branches are branches on your machine.

  • To see local branches in repository: git branch or git branch --list
  • To create new local branch (based on main branch): git branch <local-branch-name>
  • To switch to another local branch: git checkout <another-branch>
  • To create and switch to new local branch: git checkout -b <local-branch-name>
  • To sync another branch with changes on main branch: git checkout <another-branch> and then git merge main
  • To rename a branch: git branch -m <branch-before> <branch-after>
  • To remove a local branch safely: git branch -d <local-branch-name> (it will delete the branch only if it has already been pushed and merged with the remote branch)
  • To force delete the specified local branch, even if it has unmerged changes: git branch -D <local-branch-name>
  • To set default branch in git configuration: git config --global init.defaultBranch <default-branch-name> (it influences only new repositories)

10.2 Remote branches

Remote branches are branches on remote machine/server (e.g. GitHub, GitLab, BitBucket, DevOps, etc.)

  • To add remote repo to a local repo config: git remote add origin https://<server>/<user>/<repo>.git
  • To verify new remote repo: git remote -v
  • To push local repo branch to a remote repo for the first time: git push -u origin <local-branch-name> (-u argument will create a link between local and remote branch)
  • To see all branches (local and remote): git branch -a
  • To remove a remote branch: git push origin --delete <remote-branch-name>
  • When you (or someone else) remove a remote branch and you want to synchronize this change to your local repository: git fetch --prune

11. Git life cycle

git-life-cycle

12. Further resources