Given below is a general introduction to revision control, adapted from bryan-mercurial-guide:

A revision is a state of a piece of information at a specific time that is a result of some changes to it e.g., if you modify the code and save the file, you have a new revision (or a new version) of that file. Some seem to use this term interchangeably with version while others seem to distinguish the two -- here, let us treat them as the same, for simplicity.
Revision Control Software (RCS) are the software tools that automate the process of Revision Control i.e., managing revisions of software artifacts. RCS are also known as Version Control Software (VCS), and by a few other names.

Git is the most widely used RCS today. Other RCS tools include Mercurial, Subversion (SVN), Perforce, CVS (Concurrent Versions System), Bazaar, TFS (Team Foundation Server), and Clearcase.

Github is a web-based project hosting platform for projects using Git for revision control. Other similar services include GitLab, BitBucket, and SourceForge.

Installing Git

Git is a free and open source software used for revision control. To use Git, you need to install Git on your computer.

PREPARATION: Install Git

Windows

Download the Git installer from the official Git website.
Run the installer and make sure to select the option to install Git Bash when prompted.

Screenshots given below provide some guidance on the dialogs you might encounter when installing Git. In other cases, go with the default option.

---

When running Git commands, we recommend Windows users to use the Git Bash terminal that comes with Git. To open Git Bash terminal, hit the key and type git-bash.

It may be possible that the installation didn't add a shortcut to the Start Menu. You can navigate to the directory where git-bash.exe is (most likely C:\Program Files\Git\git-bash.exe), double click git-bash.exe to open Git Bash.
You can also right-click it and choose Pin to Start or Pin to taskbar.

SIDEBAR: Git Bash Terminal

Git Bash is a terminal application that lets you use Git from the command line on Windows. Since Git was originally developed for Unix-like systems (like Linux and macOS), Windows does not come with a native shell that supports all the commands and utilities commonly used with Git.

Git Bash provides a Unix-like command-line environment on Windows. It includes:

• A Bash shell (Bash stands for Bourne Again SHell), which is a widely used command-line interpreter on Linux and macOS.
• Common Unix tools and commands (like ls, cat, ssh, etc.) that are useful when working with Git and scripting.

When copy-pasting text onto a Git Bash terminal, you will not be able to use the familiar Ctrl+V key combo to paste. Instead, right-click on the terminal and use the Paste menu option.

---

MacOS

Install homebrew if you don't already have it, and then, run brew install git

---

Linux

Use your Linux distribution's package manager to install Git. Examples:

• Debian/Ubuntu, run sudo apt-get update and then sudo apt-get install git.

• Fedora: run sudo dnf update and then sudo dnf install git.

---

Verify Git is installed, by running the following command in a terminal.

git --version

⤷

git version 2._._

The output should display the version number.

---

Configuring user.name and user.email

Git needs to know who you are to record changes properly. When you save a snapshot of your work in Git, it records your name and email as the author of that change. This ensures everyone working on the project can see who made which changes. Accordingly, you should set the config settings user.name and user.email before you start Git for revision control.

git config --global user.name "Your Name"
git config --global user.email "your_email@example.com"

git config --global user.name
git config --global user.email

Interacting with Git: CLI vs GUI

Git is fundamentally a command-line tool. You primarily interact with it through its by typing commands. This gives you full control over its features and helps you understand what’s really happening under the hood.

clients for Git also exist, such as Sourcetree, GitKraken, and the built-in Git support in editors like Intellij IDEA and VS Code. These tools provide a more visual way to perform some Git operations.

If you're new to Git, it's best to learn the CLI first. The CLI is universal, always available (even on servers), and helps you build a solid understanding of Git’s concepts. You can use GUI clients as a supplement — for example, to visualise complex history structures.

Mastering the CLI gives you confidence and flexibility, while GUI tools can serve as helpful companions.

Normally, we use Git to manage a revision history of a specific folder, which gives us the ability to revision-control any file in that folder and its subfolders.

To put a folder under the control of Git, we initialise a repository (short name: repo) in that folder. This way, we can initialise repos in different folders, to revision-control different clusters of files independently of each other e.g., files belonging to different projects.

You can follow the hands-on practical below to learn how to initialise a repo in a folder.

cd my-projects
mkdir things

cd things

git status

fatal: not a git repository (or any of the parent directories): .git

git init

Initialized empty Git repository in things/.git/

Initialising a repo results in two things:

• a)
  First, Git now recognises this folder as a Git repository, which means it can now help you track the version history of files inside this folder.

git status

On branch master

No commits yet

nothing to commit (create/copy files and use "git add" to track)

• b)
  Second, Git created a hidden subfolder named .git inside the things folder. This folder will be used by Git to store metadata about this repository.

A Git-controlled folder is divided into two main parts:

1. The repository – stored in the hidden .git subfolder, which contains all the metadata and history.
2. The working directory – everything else in that folder, where you create and edit files.

Git considers new files that you add to the working directory as 'untracked' i.e., Git is aware of them, but they are not yet under Git's control. The same applies to files that existed in the working folder at the time you initialised the repo.

A Git repo has an internal space called the staging area which it uses to build the next snapshot. Another name for the staging area is the index.

We can stage an untracked file to tell Git that we want its current version to be included in the next snapshot. Once you stage an untracked file, it becomes 'tracked' (i.e., under Git's control).

In the example below, you can see how staging files change the status of the repo as you go from (a) to (c).

HANDS-ON: Adding untracked files

1 First, add a file (e.g., fruits.txt) to the things folder.

Here is an easy way to do that with a single terminal command.

echo -e "apples\nbananas\ncherries\n" > fruits.txt

⤷

things/fruits.txt

apples
bananas
cherries

When using the echo command to write to text files from git-bash, you might see a warning LF will be replaced by CRLF the next time Git touches it when Git interacts with such a file. This warning is caused by the way line endings are handled differently by Git and Windows. You can simply ignore it, or suppress it in future by running the following command:

git config --global core.safecrlf false

2 Stage the new file.

CLI

2.1 Check the status of the folder using the git status command.

git status

⤷

On branch master

No commits yet

Untracked files:
  (use "git add <file>..." to include in what will be committed)

  fruits.txt
nothing added to commit but untracked files present (use "git add" to track)

2.2 Use the add command to stage the file.

git add fruits.txt

You can replace the add with stage (e.g., git stage fruits.txt) and the result is the same (they are synonyms).

2.3 Check the status again. You can see the file is no longer 'untracked'.

git status

⤷

On branch master

No commits yet

Changes to be committed:
  (use "git rm --cached <file>..." to unstage)

      new file:   fruits.txt

As before, don't worry if you don't understand the content of the output (we'll unpack it in a later lesson). The point to note is that the file is no longer listed as 'untracked'.

---

Sourcetree

2.1 Note how the file is shown as ‘unstaged’. The question mark icon indicates the file is untracked.

If the newly-added file does not show up in Sourcetree UI, refresh the UI (: F5
| ⌥+R)

2.2 Stage the file:

Select the fruits.txt and click on the Stage Selected button.

Staging can be done using tick boxes or the ... menu in front of the file.

2.3 Note how the file is staged now i.e., fruits.txt appears in the Staged files panel now.

If Sourcetree shows a \ No newline at the end of the file message below the staged lines (i.e., below the cherries line in the above screenshot), that is because you did not hit enter after entering the last line of the file (hence, Git is not sure if that line is complete). To rectify, move the cursor to the end of the last line in that file and hit enter (like you are adding a blank line below it). This new change will now appear as an 'unstaged' change. Stage it as well.

---

done!

If you modify a staged file, it goes into the 'modified' state i.e., the file contains modifications that are not present in the copy that is waiting (in the staging area) to be included in the next snapshot. If you wish to include these new changes in the next snapshot, you need to stage the file again, which will overwrite the copy of the file that was previously in the staging area.
The example below shows how the status of a file changes when it is modified after it was staged.

Alice

Alice

Alice

Alice
Bob

Alice
Bob

Alice
Bob

HANDS-ON: Re-staging 'modified' files

1 First, add another line to fruits.txt, to make it 'modified'.

Here is a way to do that with a single terminal command.

echo "dragon fruits" >> fruits.txt

⤷

things/fruits.txt

apples
bananas
cherries
dragon fruits

2 Now, verify that Git sees that file as 'modified'.

CLI

Use the git status command to check the status of the working directory.

$ git status

⤷

On branch master

No commits yet

Changes to be committed:
(use "git rm --cached <file>..." to unstage)
new file:   fruits.txt

Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git restore <file>..." to discard changes in working directory)
modified:   fruits.txt

Note how fruits.txt now appears twice, once as new file: ... (representing the version of the file we staged earlier, which had only three lines) and once as modified: ... (representing the latest version of the file which now has a fourth line).

---

Sourcetree

Note how fruits.txt appears in the Staged files panel as well as 'Unstaged files'.

---

3 Stage the file again, the same way you added/staged it earlier.

4 Verify that Git no longer sees it as 'modified', similar to step 2.

done!

Git does not track empty folders. You can test this by adding an empty subfolder inside the things folder (e.g., things/more-things) and checking if it shows up as 'untracked' (it will not). If you add a file to that folder (e.g., things/more-things/food.txt) and then staged that file (e.g., git add more-things/food.txt), the folder will now be included in the next snapshot.

Saving a snapshot is called committing and a saved snapshot is called a commit.

A Git commit is a full snapshot of your working directory based on the files you have staged, more precisely, a record of the exact state of all files in the staging area (index) at that moment -- even the files that have not changed since the last commit. This is in contrast to other revision control software that only store the in a commit. Consequently, a Git commit has all the information it needs to recreate the snapshot of the working directory at the time the commit was created.
A commit also includes metadata such as the author, date, and an optional commit message describing the change.

HANDS-ON: Creating your first commit

Assuming you have previously staged changes to the fruits.txt, go ahead and create a commit.

CLI

1 First, let us do a sanity check using the git status command.

git status

⤷

On branch master

No commits yet

Changes to be committed:
(use "git rm --cached <file>..." to unstage)
  new file:   fruits.txt

2 Now, create a commit using the commit command. The -m switch is used to specify the commit message.

git commit -m "Add fruits.txt"

⤷

[master (root-commit) d5f91de] Add fruits.txt
 1 file changed, 5 insertions(+)
 create mode 100644 fruits.txt

3 Verify the staging area is empty using the git status command again.

git status

⤷

On branch master
nothing to commit, working tree clean

Note how the output says nothing to commit which means the staging area is now empty.

---

Sourcetree

Click the Commit button, enter a commit message (e.g. add fruits.txt) into the text box, and click Commit.

---

done!

Git commits form a timeline, as each corresponds to a point in time when you asked Git to take a snapshot of your working directory. Each commit links to at least one previous commit, forming a structure that we can traverse.
A timeline of commits is called a branch. By default, Git names the initial branch master -- though many now use main instead. You'll learn more about branches in future lessons. For now, just be aware that the commits you create in a new repo will be on a branch called master (or main) by default.

gitGraph
    %%{init: { 'theme': 'default', 'gitGraph': {'mainBranchName': 'master (or main)'}} }%%
    commit id: "Add fruits.txt"
    commit id: "Update fruits.txt"
    commit id: "Add colours.txt"
    commit id: "..."

Git can show you the list of commits in the Git history.

HANDS-ON: Viewing the list of commits

1 View the list of commits, which should show just the one commit you created just now.

CLI

You can use the git log command to see the commit history.

git log

⤷

commit d5f91de... (HEAD -> master)
Author: ... <...@...>
Date:   ...

Add fruits.txt

Use the Q key to exit the output screen of the git log command.

Note how the output has some details about the commit you just created. You can ignore most of it for now, but notice it also shows the commit message you provided.

---

Sourcetree

Expand the BRANCHES menu and click on the master to view the history graph, which contains only one node at the moment, representing the commit you just added. For now, ignore the label master attached to the commit.

---

2 Create a few more commits (i.e., a few rounds of add/edit files -> stage -> commit), and observe how the list of commits grows.

CLI

Here is an example list of bash commands to add two commits while observing the list of commits

$ echo "figs" >> fruits.txt  # add another line to fruits.txt
$ git add fruits.txt  # stage the updated file
$ git commit -m "Insert figs into fruits.txt"  # commit the changes
$ git log  # check commits list

$ echo "a file for colours" >> colours.txt  # add a colours.txt file
$ echo "a file for shapes" >> shapes.txt  # add a shapes.txt file
$ git add colours.txt shapes.txt  # stage both files in one go
$ git commit -m "Add colours.txt, shapes.txt"  # commit the changes
$ git log  # check commits list

The output of the final git log should be something like this:

commit 18300... (HEAD -> master)
Author: ... <...@...>
Date:   ...

    Add colours.txt, shapes.txt

commit 2beda...
Author: ... <...@...>
Date:   ...

    Insert figs into fruits.txt

commit d5f91...
Author: ... <...@...>
Date:   ...

    Add colours.txt, shapes.txt

---

Sourcetree

To see the list of commits, click on the History item (listed under the WORKSPACE section) on the menu on the right edge of Sourcetree.

After adding two more commits, the list of commits should look something like this:

---

done!

A repo you have on your computer is called a local repo. A remote repo is a repo hosted on a remote computer and allows remote access. Some use cases for remote repositories:

• a)
  as a backup of your local repo
• b)
  as an intermediary repo to work on the same files from multiple computers
• c)
  for sharing the revision history of a codebase among team members of a multi-person project

It is possible to set up a Git remote repo on your own server, but an easier option is to use a remote repo hosting service such as GitHub.

GitHub is a web-based service that hosts Git repositories and adds collaboration features on top of Git. Two other similar platforms are GitLab and Bitbucket. While Git manages version control locally, such platforms make it easier for individuals and teams to work together by providing shared access to repositories, issue tracking, pull requests, and permission controls. They are widely used in both open-source and commercial software development. Here we'll be using GitHub.

On GitHub, a Git repo can be put in one of two spaces:

• A GitHub user account represents an individual user. It is created when you sign up for GitHub and includes a username, profile page, and personal settings. With a user account, you can create your own repositories, contribute to others’ projects, and manage collaboration settings for any repositories you own.
• A GitHub organisation (org for short) is a shared account used by a group such as a team, company, or open-source project. Organisations can own repositories and manage access to them through teams, roles, and permissions. Organisations are especially useful when managing repositories with shared ownership or when working at scale.

Every GitHub user must have a user account, even if they primarily work within an organisation.

Before you can interact with GitHub from your local Git client, you need to set up authentication. In the past, you could simply enter your GitHub username and password, but GitHub no longer accepts passwords for Git operations. Instead, you’ll use a more secure method — such as a Personal Access Token (PAT) or SSH keys — to prove your identity.

Git supports two main protocols for communicating with GitHub: HTTPS and SSH.

• With HTTPS, you connect over the web and authenticate using your GitHub username and a Personal Access Token.
• With SSH, you connect using a cryptographic key pair you generate on your machine. Once you add your public key to your GitHub account, GitHub recognises your machine and lets you authenticate without typing anything further.

GitHub associates a commit to a user based on the email address in the commit metadata. When you push a commit, GitHub checks if the email matches a verified email on a GitHub account. If it does, the commit is shown as authored by that user. If the email doesn’t match any account, the commit is still accepted but won’t be linked to any profile.

GitHub provides a no-reply email (e.g., 12345678+username@users.noreply.github.com) that you can use as your Git user.email to hide your real email while still associating commits with your GitHub account.

GitHub offers its own clients to make working with GitHub more convenient.

• The GitHub Desktop app provides a GUI for performing GitHub operations from your desktop, without needing to visit the GitHub web UI.
• The GitHub CLI (gh) brings GitHub-specific commands to your terminal, letting you perform operations on GitHub from your commandline.

You can create a remote repository based on an existing local repository, to serve as a remote copy of your local repo. For example, suppose you created a local repo and worked with it for a while, but now you want to upload it onto GitHub. The first step is to create an empty repository on GitHub.

A Git remote is a reference to a repository hosted elsewhere, usually on a server like GitHub, GitLab, or Bitbucket. It allows your local Git repo to communicate with another remote copy — for example, to upload locally-created commits that are missing in the remote copy.

By adding a remote, you are informing the local repo details of a remote repo it can communicate with, for example, where the repo exists and what name to use to refer to the remote.

The URL you use to connect to a remote repo depends on the protocol — HTTPS or SSH:

• HTTPS URLs use the standard web protocol and start with https://github.com/ (for GitHub users). e.g.,

  https://github.com/username/repo-name.git
  

• SSH URLs use the secure shell protocol and start with git@github.com:. e.g.,

  git@github.com:username/repo-name.git
  

A Git repo can have multiple remotes. You simply need to specify different names for each remote (e.g., upstream, central, production, other-backup ...).

HANDS-ON: Add a remote to a repo

Add the empty remote repo you created on GitHub as a remote of a local repo you have.

CLI

1 In a terminal, navigate to the folder containing the local repo things your created earlier.

2 List the current list of remotes using the git remote -v command, for a sanity check. No output is expected if there are no remotes yet.

3 Add a new remote repo using the git remote add <remote-name> <remote-url> command.
i.e., if using HTTPS, git remote add origin https://github.com/{YOUR-GITHUB-USERNAME}/things.git

4 List the remotes again to verify the new remote was added.

git remote -v

⤷

origin  https://github.com/johndoe/things.git (fetch)
origin  https://github.com/johndoe/things.git (push)

The same remote will be listed twice, to show that you can do two operations (fetch and push) using this remote. You can ignore that for now. The important thing is the remote you added is being listed.

---

Sourcetree

1 Open the local repo in Sourcetree.

2 Choose Repository → Repository Settings menu option.

3 Add a new remote to the repo with the following values.

• Remote name: the name you want to assign to the remote repo i.e., origin
• URL/path: the URL of your remote repo
  i.e., https://github.com/{YOUR-GITHUB-USERNAME}/things.git
• Username: your GitHub username
  
  

4 Verify the remote was added by going to Repository → Repository Settings again.

---

5 Add another remote, to verify that a repo can have multiple remotes. You can use any name (e.g., backup and any URL for this).

done!

You can push content of one repository to another. Pushing can transfer Git history (e.g., past commits) as well as files in the working directory. Note that pushing to a remote repo requires you to have write-access to it.

When pushing to a remote repo, you typically need to specify the following information:

• The name of the remote (e.g., origin).
• The name of your current local branch (e.g., master).

If this is the first time you are pushing this branch to the remote repo, you can also ask Git to track this remote/branch pairing (e.g., remember that this local master branch is tracking the master branch in the repo origin i.e., local master branch is tracking upstream origin/master branch), so in future you can push the same remote/branch without needing to specify them again.

HANDS-ON: Pushing a local repo to an empty remote repo

Here, we assume you already have a local repo that is connected to an empty remote repo, from previous hands-on practicals:

CLI

# format: git push -u <remote-repo-name> <branch-name>
git push -u origin master

Explanation:

• push: the Git sub-command that pushes the current local repo content to a remote repo
• origin: name of the remote
• master: branch to push
• -u (or --set-upstream): the flag that tells Git to track that this local master is tracking origin/master branch

---

Sourcetree

Click the Push button on the buttons ribbon at the top.

In the next dialog, ensure the settings are as follows, ensure the Track option is selected, and click the Push button on the dialog.

---

done!

The push command can be used repeatedly to send further updates to another repo e.g., to update the remote with commits you created since you pushed the first time.

HANDS-ON: Pushing to send further updates to a repo

Add a few more commits to your local repo, and push those commits to the remote repo, as follows:

1 Commit some changes in your local repo.

CLI

Use the git commit command to create commits, as you did before.

Optionally, you can run the git status command, which should confirm that your local branch is 'ahead' by one commit (i.e., the local branch has one new commit that is not in the corresponding branch in the remote repo).

git status

⤷

On branch master
Your branch is ahead of 'origin/master' by 1 commit.
  (use "git push" to publish your local commits)

nothing to commit, working tree clean

---

Sourcetree

Create commits as you did before.

Before pushing the new commit, Sourcetree will indicate that your local branch is 'ahead' by one commit (i.e., the local branch has one new commit that is not in the corresponding branch in the remote repo).

---

2 Push the new commits to your fork on GitHub.

CLI

To push the newer commit(s) to the remote, any of the following commands should work:

• git push origin master
• git push origin
  (due to tracking you set up earlier, Git will assume you are pushing themaster branch)
• git push
  (due to tracking, Git will assume you are pushing to the remote origin and to the branch master i.e., origin/master)

---

Sourcetree

To push, click the Push button on the top buttons ribbon, ensure the settings are as follows in the next dialog, and click the Push button on the dialog.

---

done!

Note that you can push between two repos only if those repos have a shared history among them (i.e., one should have been created by copying the other).

You can specify which files Git should ignore from revision control. While you can always omit files from revision control simply by not staging them, having an 'ignore-list' is more convenient, especially if there are files inside the working folder that are not suitable for revision control (e.g., temporary log files) or files you want to prevent from accidentally including in a commit (files containing confidential information).

A repo-specific ignore-list of files can be specified in a .gitignore file, stored in the root of the repo folder.

The .gitignore file itself can be either revision controlled or ignored.

• To version control it (the more common choice – which allows you to track how the .gitignore file changes over time), simply commit it as you would commit any other file.
• To ignore it, simply add its name to the .gitignore file itself.

The .gitignore file supports file patterns e.g., adding temp/*.tmp to the .gitignore file prevents Git from tracking any .tmp files in the temp directory.

# Ignore all log files
*.log

# Ignore node_modules folder
node_modules/

# Don’t ignore main.log
!main.log

HANDS-ON: Adding a file to the ignore-list

1 Add a file into your repo's working folder that you presumably do not want to revision-control e.g., a file named temp.txt. Observe how Git has detected the new file.
Add a few other files with .tmp extension.

2 Configure Git to ignore those files:

CLI

Create a file named .gitignore in the working directory root and add the following line in it.

.gitignore

temp.txt

Observe how temp.txt is no longer detected as 'untracked' by running the git status command (but now it will detect the .gitignore file as 'untracked'.

Update the .gitignore file as follows:

.gitignore

temp.txt
*.tmp

Observe how .tmp files are no longer detected as 'untracked' by running the git status command.

---

Sourcetree

The file should be currently listed under Unstaged files. Right-click it and choose Ignore.... Choose Ignore exact filename(s) and click OK.
Also take note of other options available e.g., Ignore all files with this extension etc. They may be useful in future.

Note how the temp.text is no longer listed under Unstaged files. Observe that a file named .gitignore has been created in the working directory root and has the following line in it. This new file is now listed under Unstaged files.

.gitignore

temp.txt

Right-click on any of the .tmp files you added, and choose Ignore... as you did previously. This time, choose the option Ignore files with this extension.

Note how .temp files are no longer shown as unstaged files, and the .gitignore file has been updated as given below:

.gitignore

temp.txt
*.tmp

---

3 Optionally, stage and commit the .gitignore file.

done!

Files recommended to be omitted from version control

• Binary files generated when building your project e.g., *.class, *.jar, *.exe (reasons: 1. no need to version control these files as they can be generated again from the source code 2. Revision control systems are optimized for tracking text-based files, not binary files.
• Temporary files e.g., log files generated while testing the product
• Local files i.e., files specific to your own computer e.g., local settings of your IDE
• Sensitive content i.e., files containing sensitive/personal information e.g., credential files, personal identification data (especially, if there is a possibility of those files getting leaked via the revision control system).

A fork is a copy of a remote repository created on the same hosting service such as GitHub, GitLab, or Bitbucket. On GitHub, you can fork a repository from another user or organisation into your own space (i.e., your user account or an organisation you have sufficient access to). Forking is particularly useful if you want to experiment with a repo but don’t have write permissions to the original -- you can fork it and work on your own remote copy without affecting the original repository.

You can clone a repository to create a full copy of it on your computer. This copy includes the entire revision history, branches, and files of the original, so it behaves just like the original repository. For example, you can clone a repository from a hosting service like GitHub to your computer, giving you a complete local version to work with.

Cloning a repo automatically creates a remote named origin which points to the repo you cloned from.

The repo you cloned from is often referred to as the upstream repo.

https://github.com/se-edu/samplerepo-things  # GitHub project URL
https://github.com/se-edu/samplerepo-things.git # the repo URL

git clone https://github.com/se-edu/samplerepo-things.git  # if using HTTPS
git clone git@github.com:se-edu/samplerepo-things.git  # if using SSH

git clone https://github.com/foo/bar.git my-bar-copy  # also specifies a dir to use

There are two steps to bringing down changes from a remote repository into a local repository: fetch and merge.

• Fetch is the act of downloading the latest changes from the remote repository, but without applying them to your current branch yet. It updates metadata in your repo so that repo knows what has changed in the remote repo, but your own local branch remain untouched.
• Merge is what you do after fetching, to actually incorporate the fetched changes into your local branch. It combines your local branch with the changes from the corresponding branch from the remote repo.

HANDS-ON: Fetch and merge from a remote

1 Clone the repo se-edu/samplerepo-finances. It has 3 commits. Your clone now has a remote origin pointing to the remote repo you cloned from.

2 Change the remote origin to point to samplerepo-finances-2. This remote repo is a copy of the one you cloned, but it has two extra commits.

CLI

git remote set-url origin https://github.com/se-edu/samplerepo-finances-2.git

---

Sourcetree

Go to Repository → Repository settings ... to update remotes.

---

3 Verify the local repo is unaware of the extra commits in the remote.

CLI

git status

⤷

On branch master
Your branch is up to date with 'origin/master'.

nothing to commit, working tree clean

---

Sourcetree

The revision graph should look like the below:

If it looks like the below, it is possible that Sourcetree is auto-fetching data from the repo periodically.

---

3 Fetch from the new remote.

CLI

Use the git fetch <remote> command to fetch changes from a remote. If the <remote> is not specified, the default remote origin will be used.

git fetch origin

⤷

remote: Enumerating objects: 8, done.
... # more output ...
   afbe966..cc6a151  master     -> origin/master
 * [new tag]         beta       -> beta

---

Sourcetree

Click on the Fetch button on the top menu:

---

4 Verify the fetch worked i.e., the local repo is now aware of the two missing commits. Also observe how the local branch ref of the master branch, the staging area, and the working directory remain unchanged after the fetch.

CLI

Use the git status command to confirm the repo now knows that it is behind the remote repo.

git status

⤷

On branch master
Your branch is behind 'origin/master' by 2 commits, and can be fast-forwarded.
  (use "git pull" to update your local branch)

nothing to commit, working tree clean

---

Sourcetree

Now, the revision graph should look something like the below. Note how the origin/master ref is now two commits ahead of the master ref.

---

5 Merge the fetched changes.

CLI

Use the git merge <remote-tracking-branch> command to merge the fetched changes. Check the status and the revision graph to verify the branch tip has now moved by two more commits.

git merge origin/master

git status
git log --oneline --decorate

---

Sourcetree

To merge the fetched changes, right-click on the latest commit on origin/remote branch and choose Merge.

In the next dialog, choose as follows:

The final result should be something like the below (same as the repo state before we started this hands-on practical):

---

Note that merging the fetched changes can get complicated if there are multiple branches or the commits in the local repo conflict with commits in the remote repo. We will address them when we learn more about Git branches, in a later lesson.

done!

Pull is a shortcut that combines fetch and merge — it fetches the latest changes from the remote and immediately merges them into your current branch. In practice, Git users typically use the pull instead of the fetch-then-merge.

HANDS-ON: Pull from a remote

1 Similar to the previous hands-on practical, clone the repo se-edu/samplerepo-finances (to a new location).
Change the remote origin to point to samplerepo-finances-2.

2 Pull the newer commits from the remote, instead of a fetch-then-merge.

CLI

Use the git pull <remote> <branch> command to pull changes.

git pull origin master

The following works too. If the <remote> and <branch> are not specified, Git will pull to the current branch from the remote branch it is tracking.

git pull

---

Sourcetree

Click on the Pull button on the top menu:

In the next dialog, choose as follows:

---

3 Verify the outcome is same as the fetch + merge steps you did in the previous hands-on practical.

done!

You can pull from any number of remote repos, provided the repos involved have a shared history. This can be useful when the upstream repo you forked from has some new commits that you wish to bring over to your copies of the repo (i.e., your fork and your local repo).

git pull upstream master

git push origin master

git remote add other-upstream https://github.com/se-edu/samplerepo-finances-2.git
git pull other-upstream master
git push origin master

The Git data model consists of two types of entities: objects and refs (short for references). In this lesson, you will encounter examples of both.

A Git revision graph is visualisation of a repo's revision history, contains examples of both objects and refs. First, let us learn to work with simpler revision graphs consisting of one branch, such as the one given below.

Nodes in the revision graph represent commits.

• A commit is one of four main types of Git objects (blobs, trees, and annotated tags are the other three, to be covered later).
• A commit identified by its SHA value. A SHA (Secure Hash Algorithm) value is a unique identifier generated by Git to represent each commit. It is produced by using SHA-1 (i.e., one of the algorithms in the SHA family of cryptographic hash functions) on the entire content of the commit. It's a 40-character hexadecimal string (e.g., f761ea63738a67258628e9e54095b88ea67d95e2) that acts like a fingerprint, ensuring that every commit can be referenced unambiguously.
• A commit is a full snapshot of the working directory, constructed based on the previous commit, and the changes staged. The previous commit a commit is based on is called the parent commit (some commits can have multiple parent commits -- we’ll cover that later).

Edges in the revision graph represent links between a commit and its parent commit(s) In some revision graph visualisations, you might see arrows (instead of lines) showing how each commit points to its parent commit.

Git uses refs to name and keep track of various points in a repository’s history. These refs are essentially 'named-pointers' that can serve as bookmarks to reach a certain point in the revision graph using the ref name.

In the revision graph above, there are two refs master and  ←HEAD.

• master is a branch ref. A branch points to the latest commit on a branch (in this visualisation, the commit shown alongside the ref is the one it points to i.e., C3). When you create a new commit, the ref of the branch moves to the new commit.
• ←HEAD is a special ref. Normally, it points to the current branch (in this example, it is pointing to the master branch), and moves together with the branch ref.

In the revision graph above you see a third type of ref ( origin/master). This is a remote-tracking branch ref that represents the state of a corresponding branch in a remote repository (if you previously set up the branch to 'track' a remote branch). In this example, the master branch in the remote origin is also at the commit C3 (which means you have not created new commits after you pushed to the remote).

If you now create a new commit C4, the state of the revision graph will be as follows:

Explanation: When you create C4, the current branch master moves to C4, and HEAD moves along with it. However, the master branch in the remote origin remains at C3 (because you have not pushed C4 yet). The origin/master ref will move to C4 after you push your repo to the remote again.

HANDS-ON: View the revision graph

Use Git features to examine the revision graph of a simple repo. For this, use a repo with just a few commits and only one branch for this hands-on practical.

CLI

1 First, use a simple git log to view the list of commits.

git log

⤷

commit f761ea63738a... (HEAD -> master, origin/master)
Author: ... <...@...>
Date:   Sat ...

    Add colours.txt, shapes.txt

commit 2bedace69990...
Author: ... <...@...>
Date:   Sat ...

    Add figs to fruits.txt

commit d5f91de5f0b5...
Author: ... <...@...>
Date:   Fri ...

    Add fruits.txt

For comparison, given below the visual representation of the same revision graph. As you can see, the log output shows the refs slightly differently, but it is not hard to see what they mean.

C3 master←HEAD origin/masterAdd colours.txt, shapes.txt

|

C2Add figs to fruits.txt

|

C1Add fruits.txt

SIDEBAR: Working with the 'less' pager

Some Git commands — such as git log— may show their output through a pager. A pager is a program that lets you view long text one screen at a time, so you don’t miss anything that scrolls off the top. For example, the output of git log command will temporarily hide the current content of the terminal, and enter the pager view that shows output one screen at a time. When you exit the pager, the git log output will disappear from view, and the previous content of the pager will reappear.

command 1
output 1

git log

→

commit f761ea63738a...
Author: ... <...@...>
Date:   Sat ...

    Add colours.txt

By default, Git uses a pager called less. Given below are some useful commands you can use inside the less pager.

Command	Description
q	Quit less and return to the terminal
↓ or j	Move down one line
↑ or k	Move up one line
Space	Move down one screen
b	Move up one screen
G	Go to the end of the content
g	Go to the beginning of the content
/pattern	Search forward for pattern (e.g., /fix)
n	Repeat the last search (forward)
N	Repeat the last search (backward)
h	Show help screen with all less commands

If you’d rather see the output directly, without using a pager, you can add the --no-pager flag to the command e.g.,

git --no-pager log

It is possible to ask Git to not use less at all, use a different pager, or fine-tune how less is used. For example, you can reduce Git's use of the pager (recommended), using the following command:

git config --global core.pager "less -FRX"

Explanation:

• -F : Quit if the output fits on one screen (don’t show pager unnecessarily)
• -R : Show raw control characters (for coloured Git output)
• -X : Keep content visible after quitting the pager (so output stays on the terminal)

2 Use the --oneline flag to get a more concise view. Note how the commit SHA has been truncated to first seven characters (first seven characters of a commit SHA is enough for Git to identify a commit).

git log --oneline

⤷

f761ea6 (HEAD -> master, origin/master) Add colours.txt, shapes.txt
2bedace Add figs to fruits.txt
d5f91de Add fruits.txt

3 The --graph flag makes the result closer to a graphical revision graph. Note the * that indicates a node in a revision graph.

git log --oneline --graph

⤷

* f761ea6 (HEAD -> master, origin/master) Add colours.txt, shapes.txt
* 2bedace Add figs to fruits.txt
* d5f91de Add fruits.txt

The --graph option is more useful when examining a more complicated revision graph consisting of multiple parallel branches.

---

Sourcetree

Click the History to see the revision graph.

• In some versions of Sourcetree, the HEAD ref may not be shown -- it is implied that the HEAD ref is pointing to the same commit the currently active branch ref is pointing.
• If the remote-tracking branch ref (e.g., origin/master) is not showing up, you may need to enable the Show Remote Branches option.

---

Observe how the revision graph changes as you add a commit, and push that commit to the remote repo.

For example, we can update the fruits.txt in the things repo as follows, and commit it with the message Update fruits list.

fruits.txt

apples
bananas
cherries
dragon fruits
elderberries
figs

→
[update file as...]

fruits.txt

apples, apricots
bananas
blueberries
cherries
dragon fruits
figs

CLI

After creating the new commit, the output of git log --oneline --decorate should be of the form:

e60deae (HEAD -> master) Update fruits list
f761ea6 (origin/master) Add colours.txt, shapes.txt
2bedace Add figs to fruits.txt
d5f91de Add fruits.txt

After pushing the new commit to the remote, the remote-tracking branch ref should move to the new commit:

e60deae (HEAD -> master, origin/master) Update fruits list
f761ea6 Add colours.txt, shapes.txt
2bedace Add figs to fruits.txt
d5f91de Add fruits.txt

---

Sourcetree

After creating the new commit, the branch ref (master) will move to the new commit, but the remote-tracking branch ref (origin/master) will remain at the previous commit. Sourcetree will also indicate that your local branch is 1 commit ahead (i.e., has one more commit than) the remote-tracking branch.

After pushing the new commit to the remote, the remote-tracking branch ref should move to the new commit:

---

done!

When you examine a commit, normally what you see is the 'changes since the previous commit'. This should not be interpreted as Git commits contain only the changes. As you recall, a Git commit contains a full snapshot of the working directory. However, tools used to examine commits show only the changes, as that is the more informative part.

Git shows changes included in a commit by dynamically calculating the difference between the snapshots stored in the target commit and the parent commit. This is because Git commits stores snapshots of the working directory, not changes themselves.

To address a specific commit, you can use its SHA (e.g., e60deaeb2964bf2ebc907b7416efc890c9d4914b). In fact, just the first few characters of the SHA is enough to uniquely address a commit (e.g., e60deae), provided the partial SHA is long enough uniquely identify the commit (i.e., only one commit has that partial SHA).
Naturally, a commit can be addressed using any ref pointing to it too (e.g., HEAD, master).
Another related technique is to use the <ref>~<n> notation (e.g., HEAD~1) to address the commit that is n commits prior to the commit pointed by <ref> i.e., "start with the commit pointed by <ref> and go back n commits".
A further shortcut of this notation is to use HEAD~, HEAD~~, HEAD~~~, ... to mean HEAD~1, HEAD~2, HEAD~3 etc.

Git uses the diff format to show file changes in a commit. The diff format was originally developed for Unix, later extended with headers and metadata to show changes between file versions and commits. Here is an example diff showing the changes to a file.

diff --git a/fruits.txt b/fruits.txt
index 7d0a594..f84d1c9 100644
--- a/fruits.txt
+++ b/fruits.txt
@@ -1,6 +1,6 @@
-apples
+apples, apricots
 bananas
 cherries
 dragon fruits
-elderberries
 figs
@@ -20,2 +20,3 @@
 oranges
+pears
 raisins
diff --git a/colours.txt b/colours.txt
new file mode 100644
index 0000000..55c8449
--- /dev/null
+++ b/colours.txt
@@ -0,0 +1 @@
+a file for colours

A Git diff can consist of multiple file diffs, one for each changed file. Each file diff can contain one or more hunk i.e., a localised group of changes within the file — including lines added, removed, or left unchanged (included for context).

diff --git a/fruits.txt b/fruits.txt
index 7d0a594..f84d1c9 100644
--- a/fruits.txt
+++ b/fruits.txt

@@ -1,6 +1,6 @@
-apples
+apples, apricots
 bananas
 cherries
 dragon fruits
-elderberries
 figs

@@ -20,2 +20,3 @@
 oranges
+pears
 raisins

diff --git a/colours.txt b/colours.txt
new file mode 100644
index 0000000..55c8449
--- /dev/null
+++ b/colours.txt

@@ -0,0 +1 @@
+a file for colours

HANDS-ON: View specific commits

View contents of specific commits in a repo (e.g., the things repo):

CLI

1 Locate the commits to view, using the revision graph.

git log --oneline --decorate

⤷

 e60deae (HEAD -> master, origin/master) Update fruits list
 f761ea6 Add colours.txt, shapes.txt
 2bedace Add figs to fruits.txt
 d5f91de Add fruits.txt

2 Use the git show command to view specific commits.

git show  # shows the latest commit

⤷

commit e60deaeb2964bf2ebc907b7416efc890c9d4914b (HEAD -> master, origin/master)
Author: damithc <...@...>
Date:   Sat Jun ...

    Update fruits list

diff --git a/fruits.txt b/fruits.txt
index 7d0a594..6d502c3 100644
--- a/fruits.txt
+++ b/fruits.txt
@@ -1,6 +1,6 @@
-apples
+apples, apricots
 bananas
+blueberries
 cherries
 dragon fruits
-elderberries
 figs

To view the parent commit of the latest commit, you can use any of these commands:

git show HEAD~1
git show master~1
git show e60deae  # first few characters of the SHA
git show e60deae.....  # run git log to find the full SHA and specify the full SHA

To view the one two commits prior to the latest commit, you can use git show HEAD~2 etc.

---

Sourcetree

Click on the commit. The remaining panels (indicated in the image below) will be populated with the details of the commit.

---

done!

git config --global alias.lod "log --oneline --graph --decorate"

alias gs='git status'
alias glod='git log --oneline --graph --decorate'

Git lets you tag commits with names, making them easy to reference later. This is useful when you want to mark specific commits -- such as releases or key milestones (e.g., v1.0 or v2.1). Using tags to refer to commits is much more convenient than using SHA hashes. In the diagram below, v1.0 and interim are tags.

A tag stays fixed to the commit. Unlike branch refs or HEAD, tags do not move automatically as new commits are made. As you see below, after adding a new commit, tags stay in the previous commits while master←HEAD have moved to the new commit.

Git supports two kinds of tags:

1. A lightweight tag is just a ref that points directly to a commit, like a branch that doesn’t move.
2. An annotated tag is a full Git object that stores a reference to a commit along with metadata such as the tagger’s name, date, and a message.

Annotated tags are generally preferred for versioning and public releases, while lightweight tags are often used for less formal purposes, such as marking a commit for your own reference.

HANDS-ON: Adding tags

0 Preparation: fork and clone the samplerepo-preferences. Use the cloned repo on your computer for the following steps.

CLI

1 Add a lightweight tag to the current commit as v1.0:

git tag v1.0

2 Verify the tag was added. To view tags:

git tag

⤷

v1.0

To view tags in the context of the revision graph:

git log --oneline --decorate

⤷

507bb74 (HEAD -> master, tag: v1.0, origin/master, origin/HEAD) Add donuts
de97f08 Add cake
5e6733a Add bananas
3398df7 Add food.txt

3 Use the tag to refer to the commit e.g., git show v1.0 should show the changes in the tagged commit.

4 Add an annotated tag to an earlier commit. The example below adds a tag v0.9 to the commit HEAD~2 with the message First beta release. The -a switch tells Git this is an annotated tag.

git tag -a v0.9  HEAD~2 -m "First beta release"

5 Check the new annotated tag. While both types of tags appear similarly in the revision graph, the show command on an annotated tag will show the details of the tag and the details of the commit it points to.

git show v0.9

⤷

tag v0.9
Tagger: ... <...@...>
Date:   Sun Jun ...

First beta release

commit ....999087124af... (tag: v0.9)
Author: ... <...@...>
Date:   Sat Jun ...

    Add figs to fruits.txt

diff --git a/fruits.txt b/fruits.txt
index a8a0a01..7d0a594 100644
# rest of the diff goes here

---

Sourcetree

Right-click on the commit (in the graphical revision graph) you want to tag and choose Tag….

Specify the tag name e.g., v1.0 and click Add Tag.

Configure tag properties in the next dialog and press Add. For example, you can choose whether to make it a lightweight tag or an annotated tag (default).

Tags will appear as labels in the revision graph, as seen below. To see the details of an annotated tag, you need to use the menu indicated in the screenshot.

---

done!

If you need to change what a tag points to, you must delete the old one and create a new tag with the same name. This is because tags are designed to be fixed references to a specific commit, and there is no built-in mechanism to 'move' a tag.

HANDS-ON: Deleting/moving tags

Move the v1.0 tag to the commit HEAD~1, by deleting it first and creating it again at the destination commit.

CLI

Delete the previous v1.0 tag by using the -d switch. Add it again to the other commit, as before.

git tag -d v1.0
git tag v1.0 HEAD~1

---

Sourcetree

The same dialog used to add a tag can be used to delete and even move a tag. Note that the 'moving' here translates to deleting and re-adding behind the scene.

---

done!

Pushing commits to a remote does not push tags automatically. You need to push tags specifically.

HANDS-ON: Pushing tags to a remote

Push tags you created earlier to the remote.

You can go to your remote on GitHub link https://github.com/{USER}/{REPO}/tags (e.g., https://github.com/johndoe/samplerepo-prefrences/tags) to verify the tag is present there.

Note how GitHub assumes these tags are meant as releases, and automatically provides zip and tar.gz archives of the repo (as at that tag).

CLI

1 Push a specific tag in the local repo to the remote (e.g., v1.0) using the git push <origin> <tag-name> command.

git push origin v1.0

In addition to verifying the tag's presence via GitHub, you can also use the following command to list the tags presently in the remote.

git ls-remote --tags origin

2 Delete a tag in the remote, using the git push --delete <remote> <tag-name> command.

git push --delete origin v1.0

3 Push all tags to the remote repo, using the git push <remote> --tags command.

git push origin --tags

---

Sourcetree

To push a specific tag, use the following menu:

To push all tags, you can tick the Push all tags option when pushing commits:

---

done!

Git's diff feature can show you what changed between two points in the revision history. Given below are some use cases.

Usage 1: Examining changes in the working directory
Example use case: To verify the next commit will include exactly what you intend it to include.

HANDS-ON: Examining staged and unstaged changes

Preparation For this, you can use the things repo you created earlier. If you don't have it, you can clone a copy of a similar repo given here.

1 Do some changes to the working directory. Stage some (but not all) changes. For example, you can run the following commands.

echo -e "blue\nred\ngreen" >> colours.txt
git add .  # a shortcut to stage all changes
echo "no shapes added yet" >> shapes.txt

2 Examine the staged and unstaged changes.

CLI

The git diff command shows unstaged changes in the working directory (tracked files only). The output of the diff command, is a diff view (introduced in this lesson).

git diff

⤷

diff --git a/shapes.txt b/shapes.txt
index 5c2644b..949c676 100644
--- a/shapes.txt
+++ b/shapes.txt
@@ -1 +1,2 @@
a file for shapes
+no shapes added yet!

The git diff --staged command shows the staged changes (same as git diff --cached).

git diff --staged

---

Sourcetree

Select the two commits: Click on one commit, and Ctrl-Click (or Cmd-Click) on the second commit. The changes between the two selected commits will appear in the other panels, as shown below:

---

done!

Usage 2: Comparing two commits at different points of the revision graph
Example use case: Suppose you’re trying to improve the performance of a piece of software by experimenting with different code tweaks. You commit after each change (as you should). After several commits, you now want to review the overall effect of all those changes on the code.

HANDS-ON: Comparing two commits

Compare two commits in a repo (e.g., the things repo).

CLI

You can use the git diff <commit1> <commit2> command for this.

• You may use any valid way to refer to commits (e.g., SHA, tag, HEAD~n etc.).
• You may also use the .. notation to specify the commit range too e.g., 0023cdd..fcd6199, HEAD~2..HEAD

git diff v0.9 HEAD

⤷

diff --git a/colours.txt b/colours.txt
new file mode 100644
index 0000000..55c8449
--- /dev/null
+++ b/colours.txt
@@ -0,0 +1 @@
+a file for colours
# rest of the diff ...

Swap the commit order in the command and see what happens.

git diff HEAD v0.9

⤷

diff --git a/colours.txt b/colours.txt
deleted file mode 100644
index 55c8449..0000000
--- a/colours.txt
+++ /dev/null
@@ -1 +0,0 @@
-a file for colours
# rest of the diff ...

As you can see, the diff is directional i.e., dif <commit1> <commit2> shows what changes you need to do to go from the <commit1> to <commit2>. If you swap <commit1> and <commit2>, the output will change accordingly e.g., lines previously shown as 'added' will now be shown as 'deleted'.

---

Sourcetree

Select the two commits: Click on one commit, and Ctrl-Click (or Cmd-Click) on the second commit. The changes between the two selected commits will appear in the other panels, as shown below:

The same method can be used to compare the current state of the working directory (which might have uncommitted changes) to a point in the history.

---

done!

Usage 3: Examining changes to a specific file
Example use case: Similar to other use cases but when you are interested in a specific file only.

HANDS-ON: Examining changes to a specific file

Examine the changes done to a file between two different points in the version history (including the working directory).

CLI

Add the -- path/to/file to a previous diff command to narrow the output to a specific file. Some examples:

git diff -- fruits.txt               # unstaged changes to fruits.txt
git diff --staged -- src/main.java   # staged changes to src/main.java
git diff HEAD~2..HEAD -- fruits.txt  # changes to fruits.txt between commits

---

Sourcetree

Sourcetree UI shows changes to one file at a time by default; just click on the file to view changes to that file. To view changes to multiple files, Ctrl-Click (or Cmd-Click) on multiple files to select them.

---

done!

Suppose you added a new feature to a software product, and while testing it, you noticed that another feature added two commits ago doesn’t handle a certain edge case correctly. Now you’re wondering: did the new feature break the old one, or was it already broken? Can you go back to the moment you committed the old feature and test it in isolation, and come back to the present after you found the answer? With Git, you can.

To view the working directory at a specific point in history, you can check out a commit created at that point.

When you check out a commit, Git:

1. Updates your working directory to match the snapshot in that commit, overwriting current files as needed.
2. Moves the HEAD ref to that commit, marking it as the current state you’re viewing.

Checking out a specific commit puts you in a "detached HEAD" state: i.e., the HEAD no longer points to a branch, but directly to a commit (see the above diagram for an example). This isn't a problem by itself, but any commits you make in this state can be lost, unless certain follow-up actions are taken. It is perfectly fine to be in a detached state if you are only examining the state of the working directory at that commit.

To get out of a "detached HEAD" state, you can simply check out a branch, which "re-attaches" HEAD to the branch you checked out.

HANDS-ON: Checking out some commits

Checkout a few commits in a local repo (e.g., the things repo), while examining the working directory to verify that it matches the state when you created the corresponding commit:

CLI

1 Examine the revision tree, to get your bearing first.

git log --oneline --decorate

Reminder: You can use aliases to reduce typing Git commands.

⤷

e60deae (HEAD -> master, origin/master) Update fruits list
f761ea6 (tag: v1.0) Add colours.txt, shapes.txt
2bedace (tag: v0.9) Add figs to fruits.txt
d5f91de Add fruits.txt

2 Use the checkout <commit-identifier> command to check out a commit other than the one currently pointed by HEAD. You can use any of the following methods:

• git checkout v1.0: checks out the commit tagged v1.0
• git checkout 0023cdd: checks out the commit with the hash 0023cdd
• git checkout HEAD~2: checks out the commit 2 commits behind the most recent commit.

git checkout HEAD~2

⤷

Note: switching to 'HEAD~2'.

You are in 'detached HEAD' state.
# rest of the warning about the detached head ...

HEAD is now at 2bedace Add figs to fruits.txt

3 Verify HEAD and the working directory have updated as expected.

• HEAD should now be pointing at the target commit
• The working directory should match the state it was in at that commit (e.g., files added after that commit -- such as shapes.txt should not be in the folder).

git log --one-line --decorate

⤷

2bedace (HEAD, tag: v0.9) Add figs to fruits.txt
d5f91de Add fruits.txt

HEAD is indeed pointing at the target commit.

But note how the output does not show commits you added after the checked-out commit.

The --all switch tells git log to show commits from all refs, not just those reachable from the current HEAD. This includes commits from other branches, tags, and remotes.

git log --one-line --decorate --all

⤷

e60deae (origin/master, master) Update fruits list
f761ea6 (tag: v1.0) Add colours.txt, shapes.txt
2bedace (HEAD, tag: v0.9) Add figs to fruits.txt
d5f91de Add fruits.txt

4 Go back to the latest commit by checking out the master branch again.

git checkout master

---

Sourcetree

In the revision graph, double-click the commit you want to check out, or right-click on that commit and choose Checkout....

Click OK to the warning about ‘detached HEAD’ (similar to below).

The specified commit is now loaded onto the working folder, as indicated by the HEAD label.

To go back to the latest commit on the master branch, double-click the master branch.

---

If you check out a commit that comes before the commit in which you added a certain file (e.g., temp.txt) to the .gitignore file, and if the .gitignore file is version controlled as well, Git will now show it under ‘unstaged modifications’ because at Git hasn’t been told to ignore that file yet.

done!

If there are uncommitted changes in the working directory, Git proceeds with a checkout only if it can preserve those changes.

• Example 1: There is a new file in the working directory that is not committed yet.
  → Git will proceed with the checkout and will keep the uncommitted file as well.
• Example 2: There is an uncommitted change to a file that conflicts with the version of that file in the commit you wish to check out.
  → Git will abort the checkout, and the repo will remain in the current commit.

The Git stash feature temporarily set aside uncommitted changes you’ve made (in your working directory and staging area), without committing them. This is useful when you’re in the middle of some work, but need to switch to another state (e.g., checkout a previous commit), and your current changes are not yet ready to be committed or discarded. You can later reapply the stashed changes when you’re ready to resume that work.

Suppose you realise your last few commits have gone in the wrong direction, and you want to go back to an earlier commit and continue from there — as if the “bad” commits never happened. Git’s reset feature can help you do that.

Git reset moves the tip of the current branch to a specific commit, optionally adjusting your staged and unstaged changes to match. This effectively rewrites the branch's history by discarding any commits that came after that point.

Resetting is different from the checkout feature:

• Reset: Lets you start over from a past state. It rewrites history by moving the branch ref to a new location.
• Checkout: Lets you explore a past state without rewriting history. It just moves the HEAD ref.

There are three types of resets: soft, mixed, hard. All three moves the branch pointer to a new commit but they vary based on what happens to the staging area and the working directory.

• soft reset: Moves the cumulative changes from the discarded commits in to the staging area, waiting to be committed again. Any staged and unstaged changes that existed before the reset will remain untouched.
• mixed reset: Cumulative changes from the discarded commits, and any existing staged changes, are moved into the working directory.
• hard reset: All staged and unstaged changes are discarded. Both the working directory and the staging area are aligned with the target commit (as if no changes were done after that commit).

HANDS-ON: Resetting to past commits

1 First, set the stage as follows (e.g., in the things repo):
i) Add four commits that are supposedly 'bad' commits.
ii) Do a 'bad' change to one file and stage it.
iii) Do a 'bad' change to another file, but don't stage it.

B4 master←HEADAdd incorrect.txt

|

B3Incorrectly update fruits.txt

|

B2Incorrectly update shapes.txt

|

B1Incorrectly update colours.txt

|

C4Update fruits list

|

The following commands can be used to add commits B1-B4:

echo "bad colour" >> colours.txt
git commit -am "Incorrectly update colours.txt"

echo "bad shape" >> shapes.txt
git commit -am "Incorrectly update shapes.txt"

echo "bad fruit" >> fruits.txt
git commit -am "Incorrectly update fruits.txt"

echo "bad line" >> incorrect.txt
git add incorrect.txt
git commit -m "Add incorrect.txt"

echo "another bad colour" >> colours.txt
git add colours.txt

echo "another bad shape" >> shapes.txt

Now we have some 'bad' commits and some 'bad' changes in both the staging area and the working directory. Let's use the reset feature to get rid of all of them, but do it in three steps so that you can learn all three types of resets.

2 Do a soft reset to B2 (i.e., discard last two commits). Verify,

• the master branch is now pointing at B2, and,
• the changes that were in the discarded commits (i.e., B3and B4) are now in the staging area.

CLI

Use the git reset --soft <commit> command to do a soft reset.

git reset --soft HEAD~2

You can run the following commands to verify the current status of the repo is as expected.

git status                    # check overall status
git log --oneline --decorate  # check the branch tip
git diff                      # check unstaged changes
git diff --staged             # check staged changes

---

Sourcetree

Right-click on the commit that you want to reset to, and choose Reset <branch-name> to this commit option.

In the next dialog, choose Soft - keep all local changes.

---

3 Do a mixed reset to commit B1. Verify,

• the master branch is now pointing at B1.
• the staging area is empty.
• the accumulated changes from all three discarded commits (including those from the previous soft reset) are now appearing as unstaged changes in the working directory.
  Note how incorrect.txt appears as an 'untracked' file -- this is because unstaging a change of type 'add file' results in an untracked file.

CLI

Use the git --mixed reset <commit> command to do a mixed reset. The --mixed flag is the default, and can be omitted.

git reset HEAD~1

Verify the repo status, as before.

---

Sourcetree

Similar to the previous reset, but choose the Mixed - keep working copy but reset index option in the reset dialog.

---

4 Do a hard reset to commit C4. Verify,

• the master branch is now pointing at C4 i.e., all 'bad' commits are gone.
• the staging area is empty.
• there are no unstaged changes (except for the untracked files incorrect.txt -- Git leaves untracked files alone, as untracked files are not meant to be under Git's control).

CLI

Use the git --hard reset <commit> command.

git reset --hard HEAD~1

Verify the repo status, as before.

---

Sourcetree

Similar to the previous reset, but choose the Hard - discard all working copy changes option.

---

done!

Rewriting history can cause your local repo to diverge from its remote counterpart. For example, if you discard earlier commits and create new ones in their place, and you’ve already pushed the original commits to a remote repository, your local branch history will no longer match the corresponding remote branch. Git refers to this as a diverged history.

To protect the integrity of the remote, Git will reject attempts to push a diverged branch using a normal push. If you want to overwrite the remote history with your local version, you must perform a force push.

git reset --hard HEAD~1
echo "water" >> drinks.txt
git add .
git commit -m "Add drinks.txt"

git log --oneline --graph --all

* fc1d04e (HEAD -> master) Add drinks.txt
| * e60deae (upstream/master, origin/master) Update fruits list
|/
* f761ea6 (tag: v1.0) Add colours.txt, shapes.txt
* 2bedace (tag: v0.9) Add figs to fruits.txt
* d5f91de Add fruits.txt

git push origin master

To https://github.com/.../things.git
 ! [rejected]        master -> master (non-fast-forward)
error: failed to push some refs to 'https://github.com/.../things.git'
hint: Updates were rejected because the tip of your current branch is behind
hint: its remote counterpart. If you want to integrate the remote changes,
hint: ...

git push -f origin master

git push -force-with-lease origin master

When a past commit introduced a bug or an unwanted change, but you do not want to modify that commit — because rewriting history can cause problems if others have already based work on it — you can instead revert that commit.

Reverting creates a new commit that cancels out the changes of the earlier one i.e., Git computes the opposite of the changes introduced by that commit — essentially a reverse diff — and applies it as a new commit on top of the current branch. This way, the problematic changes are reversed while preserving the full history, including the "bad" commit and the "fix".

mkdir pioneers
cd pioneers
git init

echo "hacked the matrix" >> neo.txt
git add .
git commit -m "Add Neo"

echo "father of theoretical computing" >> alan-turing.txt
git add .
git commit -m "Add Turing"

echo "created COBOL, compiler pioneer" >> grace-hopper.txt
git add .
git commit -m "Add Hopper"

git revert HEAD~2

Crafting a commit involves two aspects:

1. What changes to include in it: deciding what changes belong together in a single commit — this is about commit granularity, ensuring each commit represents a meaningful, self-contained change.
2. How to include those changes: carefully staging just those changes — this is about using Git’s tools to precisely control what ends up in the commit.

Git can let you choose not just which files, but which specific changes within those files, to include in a commit. Most Git tools — including the command line and many GUIs — let you interactively select which "hunks" or even individual lines of a file to stage. This allows you to separate unrelated changes and avoid committing unnecessary edits. If you make multiple changes in the same file, you can selectively stage only the parts that belong to the current logical change.

This level of control is particularly useful when:

• You noticed and fixed a small, unrelated issue while working on something else.
• You experimented with multiple approaches in the same file and now want to commit only the final, clean solution.
• You want your commit history to clearly separate concerns, even when the edits touch the same files.

HANDS-ON: Stage changes selectively

You can use any repo for this.

1 Do several changes to some tracked files. Change multiple files. Also change multiple locations in the same file.

2 Stage some changes in some files while keeping other changes in the same files unstaged.

CLI

As you know, you can use git add <filename> to stage changes to an entire file.

To select which hunks to stage, you can use the git add -p command instead (-p stands for 'by patch'):

git add -p

This command will take you to an interactive mode in which you can go through each hunk and decide if you want to stage it. The video below has contains a demonstration of how this feature works:

---

Sourcetree

To stage a hunk, you can click the Stage button above the hunk in question:

To stage specific lines, select the lines first before clicking the `Stage` button above the hunk in question:

Unstaging can be done similarly:

---

Most git operations can be done faster through the CLI than equivalent Git GUI clients, once you are familiar enough with the CLI commands.

However, selective staging is one exception where a good GUI can do better than the CLI, if you need to do many fine-grained staging operations (e.g., frequently staging only parts of hunks).

done!

Every commit you make in Git also includes a commit message that explains the change. While one-line messages are fine for small or obvious changes, as your revision history grows, good commit messages become an important source of information — for example, to understand the rationale behind a specific change made in the past.

A commit message is meant to explain the intent behind the changes, not just what was changed. The code (or diff) already shows what changed. Well-written commit messages make collaboration, code reviews, debugging, and future maintenance easier by helping you and others quickly understand the project’s history without digging into the code of every commit.

A complete commit message can include a short summary line (the subject) followed by a more detailed body if needed. The subject line should be a concise description of the change, while the body can elaborate on the context, rationale, side effects, or other details if the change is more complex.

Here is an example commit message:

Find command: make matching case-insensitive

Find command is case-sensitive.

A case-insensitive find is more user-friendly because users cannot be
expected to remember the exact case of the keywords.

Let's,
* update the search algorithm to use case-insensitive matching
* add a script to migrate stress tests to the new format

Following a style guide makes your commit messages more consistent and fit-for-purpose. Many teams adopt established guidelines. These style guides typically contain common conventions that Git users follow when writing commit messages. For example:

• Keep the subject line (the first line) under 50–72 characters.
• Write the subject in the imperative mood (e.g., Fix typo in README rather than Fixed typo or Fixes typo).
• Leave a blank line between the subject and the body, if you include a body.
• Wrap the body at around 72 characters per line for readability.

Git has a powerful tool called interactive rebasing which lets you review and reorganise your recent commits. With it, you can reword commit messages, change their order, delete commits, combine several commits into one (squash), or split a commit into smaller pieces. This feature is useful for tidying up a commit history that has become messy — for example, when some commits are out of order, poorly described, or include changes that would be clearer if split up or combined.

mkdir samplerepo-sitcom
cd samplerepo-sitcom
git init

echo "Aspiring actress" >> Penny.txt
git add .
git commit -m "C1: Add Penny.txt"

echo "Scientist" >> Sheldon.txt
git add .
git commit -m "C3: Add Sheldon.txt"

echo "Comic book store owner" >> Stuart.txt
git add .
git commit -m "C2: Add Stuart.txt"

echo "Engineer" >> Stuart.txt
git commit -am "X: Incorrectly update Stuart.txt"

echo "Engineer" >> Howard.txt
git add .
git commit -m "C4: Adddd Howard.txt"