Let's take your first few steps in your Git (with GitHub) journey.
0. Take a peek at the full picture(?). Optionally, if you are the sort who prefers to have some sense of the full picture before you get into the nitty-gritty details, watch the video in the panel below:
1. The first step is to install SourceTree, which is Git + a GUI for Git. If you prefer to use Git via the command line (i.e., without a GUI), you can install Git instead.
2. Next, initialize a repository. Let us assume you want to version control content in a specific directory. In that case, you need to initialize a Git repository in that directory. Here are the steps:
Create a directory for the repo (e.g., a directory named things).
Windows: Click File → Clone/New…. Click on Create button.
Mac: New... → Create New Repository.
Enter the location of the directory (Windows version shown below) and click Create.
Go to the things folder and observe how a hidden folder .git has been created.
Windows: you might have to configure Windows Explorer to show hidden files.
Open a Git Bash Terminal.
If you installed SourceTree, you can click the Terminal button to open a GitBash terminal.
Navigate to the things directory.
Use the command git init which should initialize the repo.
$ git init
Initialized empty Git repository in c:/repos/things/.git/
You can use the command ls -a to view all files, which should show the .git directory that was created by the previous command.
$ ls -a
. .. .git
You can also use the git status command to check the status of the newly-created repo. It should respond with something like the following:
git status
# On branch master
#
# Initial commit
#
nothing to commit (create/copy files and use "git add" to track)
After initializing a repository, Git can help you with revision controlling files inside the working directory. However, it is not automatic. It is up to you to tell Git which of your changes (aka revisions) should be committed to its memory for later use. Saving changes into Git's memory in that way is often called committing and a change saved to the revision history is called a commit.
Working directory: the root directory revision-controlled by Git (e.g., the directory in which the repo was initialized).
Commit (noun): a change (aka a revision) saved in the Git revision history.
(verb): the act of creating a commit i.e., saving a change in the working directory into the Git revision history.
Here are the steps you can follow to learn how to work with Git commits:
1. Do some changes to the content inside the working directory e.g., create a file named fruits.txt in the things directory and add some dummy text to it.
2. Observe how the file is detected by Git.
The file is shown as ‘unstaged’.
You can use the git status command to check the status of the working directory.
git status
# On branch master
#
# Initial commit
#
# Untracked files:
# (use "git add <file>..." to include in what will be committed)
#
# a.txt
nothing added to commit but untracked files present (use "git add" to track)
3. Stage the changes to commit: Although Git has detected the file in the working directory, it will not do anything with the file unless you tell it to. Suppose you want to commit the current changes to the file. First, you should stage the file.
Stage (verb): Instructing Git to prepare a file for committing.
Select the fruits.txt and click on the Stage Selected button.
fruits.txt should appear in the Staged files panel now.
You can use the stage or the add command (they are synonyms, add is the more popular choice) to stage files.
git add fruits.txt
git status
# On branch master
#
# Initial commit
#
# Changes to be committed:
# (use "git rm --cached <file>..." to unstage)
#
# new file: fruits.txt
#
4. Commit the staged version of fruits.txt.
Click the Commit button, enter a commit message e.g. add fruits.txt into the text box, and click Commit.
Use the commit command to commit. The -m switch is used to specify the commit message.
git commit -m "Add fruits.txt"
You can use the log command to see the commit history.
git log
commit 8fd30a6910efb28bb258cd01be93e481caeab846
Author: … < … @... >
Date: Wed Jul 5 16:06:28 2017 +0800
Add fruits.txt
Note the existence of something called the master branch. Git allows you to have multiple branches (i.e. it is a way to evolve the content in parallel) and Git auto-creates a branch named master on which the commits go on by default.
5. Do a few more commits.
Make some changes to
fruits.txt(e.g. add some text and delete some text). Stage the changes, and commit the changes using the same steps you followed before. You should end up with something like this.
Next, add two more files
colors.txtandshapes.txtto the same working directory. Add a third commit to record the current state of the working directory.
6. See the revision graph: Note how commits form a path-like structure aka the revision tree/graph. In the revision graph, each commit is shown as linked to its 'parent' commit (i.e., the commit before it).
To see the revision graph, click on the History item on the menu on the right edge of SourceTree.
The gitk command opens a rudimentary graphical view of the revision graph.
Each Git commit is uniquely identified by a hash e.g., d670460b4b4aece5915caf5c68d12f560a9fe3e4. As you can imagine, using such an identifier is not very convenient for our day-to-day use. As a solution, Git allows adding a more human-readable tag to a commit e.g., v1.0-beta.
Here's how you can tag a commit in a local repo (e.g. in the samplerepo-things repo):
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.
The added tag will appear in the revision graph view.
To add a tag to the current commit as v1.0,
git tag –a v1.0
To view tags
git tag
To learn how to add a tag to a past commit, go to the ‘Git Basics – Tagging’ page of the git-scm book and refer the ‘Tagging Later’ section.
Remember to push tags to the repo. A normal push does not include tags.
# push a specific tag
git push origin v1.0b
# push all tags
git push origin --tags
After adding a tag to a commit, you can use the tag to refer to that commit, as an alternative to using the hash.
Git can show you what changed in each commit.
To see which files changed in a commit, click on the commit. To see what changed in a specific file in that commit, click on the file name.
git show < part-of-commit-hash >
Example:
git show 251b4cf
commit 5bc0e30635a754908dbdd3d2d833756cc4b52ef3
Author: … < … >
Date: Sat Jul 8 16:50:27 2017 +0800
fruits.txt: replace banana with berries
diff --git a/fruits.txt b/fruits.txt
index 15b57f7..17f4528 100644
--- a/fruits.txt
+++ b/fruits.txt
@@ -1,3 +1,3 @@
apples
-bananas
+berries
cherries
Git can also show you the difference between two points in the history of the repo.
Select the two points you want to compare using Ctrl+Click. The differences between the two selected versions will show up in the bottom half of SourceTree, as shown in the screenshot 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.
The diff command can be used to view the differences between two points of the history.
git diff: shows the changes (uncommitted) since the last commit.git diff 0023cdd..fcd6199: shows the changes between the points indicated by commit hashes.
Note that when using a commit hash in a Git command, you can use only the first few characters (e.g., first 7-10 chars) as that's usually enough for Git to locate the commit.git diff v1.0..HEAD: shows changes that happened from the commit tagged asv1.0to the most recent commit.
Git can load a specific version of the history to the working directory. Note that if you have uncommitted changes in the working directory, you need to stash them first to prevent them from being overwritten.
Double-click the commit you want to load to the working directory, or right-click on that commit and choose Checkout....
Click OK to the warning about ‘detached HEAD’ (similar to below).
The specified version is now loaded to the working folder, as indicated by the HEAD label. HEAD is a reference to the currently checked out commit.
If you checkout a commit that comes before the commit in which you added the .gitignore file, Git will now show ignored files as ‘unstaged modifications’ because at that stage Git hasn’t been told to ignore those files.
To go back to the latest commit, double-click it.
Use the checkout <commit-identifier> command to change the working directory to the state it was in at a specific past commit.
git checkout v1.0: loads the state as at commit taggedv1.0git checkout 0023cdd: loads the state as at commit with the hash0023cddgit checkout HEAD~2: loads the state that is 2 commits behind the most recent commit
For now, you can ignore the warning about ‘detached HEAD’.
Here's a scenario you can try in order to learn how to pull commits from another repo to yours.
1. Clone a repo (e.g., the repo used in [Git & GitHub → Clone]) to be used for this activity.
2. Delete the last few commits to simulate cloning the repo a few commits ago.
Right-click the target commit (i.e. the commit that is 2 commits behind the tip) and choose Reset current branch to this commit.
Choose the Hard - … option and click OK.
This is what you will see.
Note the following (cross refer the screenshot above):
Arrow marked as a: The local repo is now at this commit, marked by the master label.
Arrow marked as b: The origin/master label shows what is the latest commit in the master branch in the remote repo.
Use the reset command to delete commits at the tip of the revision history.
git reset --hard HEAD~2
Now, your local repo state is exactly how it would be if you had cloned the repo 2 commits ago, as if somebody has added two more commits to the remote repo since you cloned it.
3. Pull from the other repo: To get those missing commits to your local repo (i.e. to sync your local repo with upstream repo) you can do a pull.
Click the Pull button in the main menu, choose origin and master in the next dialog, and click OK.
Now you should see something like this where master and origin/master are both pointing the same commit.
git pull origin
You can also do a fetch instead of a pull in which case the new commits will be downloaded to your repo but the working directory will remain at the current commit. To move the current state to the latest commit that was downloaded, you need to do a merge. A pull is a shortcut that does both those steps in one go.
Working with multiple remotes
When you clone a repo, Git automatically adds a remote repo named origin to your repo configuration. As you know, you can pull commits from that repo. As you know, a Git repo can work with remote repos other than the one it was cloned from.
To communicate with another remote repo, you can first add it as a remote of your repo. Here is an example scenario you can follow to learn how to pull from another repo:
Open the local repo in SourceTree. Suggested: Use your local clone of the
samplerepo-thingsrepo.Choose
Repository→Repository Settingsmenu option.Add a new remote to the repo with the following values.
Remote name: the name you want to assign to the remote repo e.g.,upstream1URL/path: the URL of your repo (ending in.git) that. Suggested:https://github.com/se-edu/samplerepo-things-2.git(samplerepo-things-2is another repo that has a shared history withsamplerepo-things)Username: your GitHub username
Now, you can pull from the added repo as you did before but choose the remote name of the repo you want to pull from (instead of
origin):
If the
Remote branch to pulldropdown is empty, click theRefreshbutton on its right.If the pull from the
samplerepo-things-2was successful, you should have received one more commit into your local repo.
Navigate to the folder containing the local repo.
Set the new remote repo as a remote of the local repo.
command:git remote add {remote_name} {remote_repo_url}
e.g.,git remote add upstream1 https://github.com/johndoe/foobar.gitNow you can pull from the new remote.
e.g.,git pull upstream1 master
Given below is a scenario you can try in order to learn how to fork a repo:.
0. Create a GitHub account if you don't have one yet.
1. Go to the GitHub repo you want to fork e.g., samplerepo-things
2. Click on the 
button on the top-right corner. In the next step, choose to fork to your own account or to another GitHub organization that you are an admin of.
GitHub does not allow you to fork the same repo more than once to the same destination. If you want to re-fork, you need to delete the previous fork.
Given below is a scenario you can try in order to learn how to push commits to a remote repo hosted on GitHub:
1. Fork an existing GitHub repo (e.g., samplerepo-things) to your GitHub account.
2. Clone the fork (not the original) to your computer.
3. Commit some changes in your local repo.
4. Push the new commits to your fork on GitHub
Click the Push button on the main menu, ensure the settings are as follows in the next dialog, and click the Push button on the dialog.
Tags are not included in a normal push. Remember to tick Push all tags when pushing to the remote repo if you want them to be pushed to the repo.
Use the command git push origin master. Enter your Github username and password when prompted.
Tags are not included in a normal push. To push a tag, use this command: git push origin <tag_name> e.g. git push origin v1.0
You can push to repos other than the one you cloned from, as long as the target repo and your repo have a shared history.
- Add the GitHub repo URL as a remote, if you haven't done so already.
- Push to the target repo.
Push your repo to the new remote the usual way, but select the name of target remote instead of origin and remember to select the Track checkbox.
Push to the new remote the usual way e.g., git push upstream1 master (assuming you gave the name upstream1 to the remote).
You can even push an entire local repository to GitHub, to form an entirely new remote repository. For example, you created a local repo and worked with it for a while but now you want to upload it onto GitHub (as a backup or to share it with others). The steps are given below.
1. Create an empty remote repo on GitHub.
Login to your GitHub account and choose to create a new Repo.
In the next screen, provide a name for your repo but keep the
Initialize this repo ...tick box unchecked.
Note the URL of the repo. It will be of the form
https://github.com/{your_user_name}/{repo_name}.git.
e.g.,https://github.com/johndoe/foobar.git(note the.gitat the end)
2. Add the GitHub repo URL as a remote of the local repo. You can give it the name origin (or any other name).
3. Push the repo to the remote.
Push each branch to the new remote the usual way but use the -u flag to inform Git that you wish to i.e., remember which branch in the remote repo corresponds to which branch in the local repotrack the branch.
e.g., git push -u origin master
Git supports branching, which allows you to do multiple parallel changes to the content of a repository.
A Git branch is simply a named label pointing to a commit. The HEAD label indicates which branch you are on. Git creates a branch named master by default. When you add a commit, it goes into the branch you are currently on, and the branch label (together with the HEAD label) moves to the new commit.
Given below is an illustration of how branch labels move as branches evolve.
- There is only one branch (i.e.,
master) and there is only one commit on it. - A new commit has been added. The
masterand theHEADlabels have moved to the new commit. - A new branch
fix1has been added. The repo has switched to the new branch too (hence, theHEADlabel is attached to thefix1branch). - A new commit (
c) has been added. The current branch labelfix1moves to the new commit, together with theHEADlabel. - The repo has switched back to the
masterbranch.
- A new commit (
d) has been added. Themasterlabel has moved to that commit. - The repo has switched back to the
fix1branch and added a new commit (e) to it. - The repo has switched to the
masterbranch and thefix1branch has been merged into themasterbranch, creating a merge commitf. The repo is currently on themasterbranch.
Follow the steps below to learn how to work with branches. You can use any repo you have on your computer (e.g. a clone of the samplerepo-things) for this.
0. Observe that you are normally in the branch called master.
git status
on branch master
1. Start a branch named feature1 and switch to the new branch.
Click on the Branch button on the main menu. In the next dialog, enter the branch name and click Create Branch.
Note how the feature1 is indicated as the current branch.
You can use the branch command to create a new branch and the checkout command to switch to a specific branch.
git branch feature1
git checkout feature1
One-step shortcut to create a branch and switch to it at the same time:
git checkout –b feature1
2. Create some commits in the new branch. Just commit as per normal. Commits you add while on a certain branch will become part of that branch.
Note how the master label and the HEAD label moves to the new commit (The HEAD label of the local repo is represented as in SourceTree).
3. Switch to the master branch. Note how the changes you did in the feature1 branch are no longer in the working directory.
Double-click the master branch.
git checkout master
4. Add a commit to the master branch. Let’s imagine it’s a bug fix.
To keep things simple for the time being, this commit should not involve the same content that you changed in the feature1 branch. To be on the safe side, this commit can change an entirely different file.
5. Switch back to the feature1 branch (similar to step 3).
6. Merge the master branch to the feature1 branch, giving an end-result like the following. Also note how Git has created a merge commit.
Right-click on the master branch and choose merge master into the current branch. Click OK in the next dialog.
git merge master
The objective of that merge was to sync the feature1 branch with the master branch. Observe how the changes you did in the master branch (i.e. the imaginary bug fix) is now available even when you are in the feature1 branch.
Instead of merging master to feature1, an alternative is to rebase the feature1 branch. However, rebasing is an advanced feature that requires modifying past commits. If you modify past commits that have been pushed to a remote repository, you'll have to force-push the modified commit to the remote repo in order to update the commits in it.
7. Add another commit to the feature1 branch.
8. Switch to the master branch and add one more commit.
9. Merge feature1 to the master branch, giving and end-result like this:
Right-click on the feature1 branch and choose Merge....
git merge feature1
10. Create a new branch called add-countries, switch to it, and add some commits to it (similar to steps 1-2 above). You should have something like this now:
Avoid this common rookie mistake!
Always remember to switch back to the master branch before creating a new branch. If not, your new branch will be created on top of the current branch.
11. Go back to the master branch and merge the add-countries branch onto the master branch (similar to steps 8-9 above). While you might expect to see something like the following,
... you are likely to see something like this instead:
That is because Git does a fast forward merge if possible. Seeing that the master branch has not changed since you started the add-countries branch, Git has decided it is simpler to just put the commits of the add-countries branch in front of the master branch, without going into the trouble of creating an extra merge commit.
It is possible to force Git to create a merge commit even if fast forwarding is possible.
Tick the box shown below when you merge a branch:
Use the --no-ff switch (short for no fast forward):
git merge --no-ff add-countries
Pushing a branch to a remote repo
Here's how to push a branch to a remote repo:
Here's how to push a branch named add-intro to your own fork of a repo named samplerepo-pr-practice:
Normally: git push {remote repository} {branch}. Examples:
git push origin masterpushes themasterbranch to the repo namedorigin(i.e., the repo you cloned from)git push upstream-repo add-intropushes theadd-introbranch to the repo namedupstream-repo
If pushing a branch you created locally to the remote for the first time, add the -u flag to get the local branch to track the new upstream branch:
e.g., git push -u origin add-intro
See git-scm.com/docs/git-push for details of the push command.
Merge conflicts happen when you try to combine two incompatible versions (e.g., merging a branch to another but each branch changed the same part of the code in a different way).
Here are the steps to simulate a merge conflict and use it to learn how to resolve merge conflicts.
0. Create an empty repo or clone an existing repo, to be used for this activity.
1. Start a branch named fix1 in the repo. Create a commit that adds a line with some text to one of the files.
2. Switch back to master branch. Create a commit with a conflicting change i.e. it adds a line with some different text in the exact location the previous line was added.
3. Try to merge the fix1 branch onto the master branch. Git will pause mid-way during the merge and report a merge conflict. If you open the conflicted file, you will see something like this:
COLORS
------
blue
<<<<<< HEAD
black
=======
green
>>>>>> fix1
red
white
4. Observe how the conflicted part is marked between a line starting with <<<<<< and a line starting with >>>>>>, separated by another line starting with =======.
Highlighted below is the conflicting part that is coming from the master branch:
blue
<<<<<< HEAD
black
=======
green
>>>>>> fix1
red
This is the conflicting part that is coming from the fix1 branch:
blue
<<<<<< HEAD
black
=======
green
>>>>>> fix1
red
5. Resolve the conflict by editing the file. Let us assume you want to keep both lines in the merged version. You can modify the file to be like this:
COLORS
------
blue
black
green
red
white
6. Stage the changes, and commit.
The PR review stage is a dialog between the PR author and members of the repo that received the PR, in order to refine and eventually merge the PR.
Given below are some steps you can follow when reviewing a PR.
1. Locate the PR:
- Go to the GitHub page of the repo.
- Click on the Pull requests tab.
- Click on the PR you want to review.
2. Read the PR description. It might contain information relevant to reviewing the PR.
3. Click on the Files changed tab to see the diff view.
4. Add review comments:
- Hover over the line you want to comment on and click on the icon that appears on the left margin. That should create a text box for you to enter your comment.
- To give a comment related to multiple lines, click-and-drag the icon. The result will look like this:
- To give a comment related to multiple lines, click-and-drag the icon. The result will look like this:
- Enter your comment.
- This page @SE-EDU/guides has some best practices PR reviewers can follow.
- To suggest an in-line code change, click on this icon:
The comment will look like this to the viewers:
- After typing in the comment, click on the Start a review button (not the Add single comment button. This way, your comment is saved but not visible to others yet. It will be visible to others only when you have finished the entire review.
- Repeat the above steps to add more comments.
5. Submit the review:
- When there are no more comments to add, click on the Review changes button (on the top right of the diff page).
- Type in an overall comment about the PR, if any. e.g.,
Overall, I found your code easy to read for the most part except a few places where the nesting was too deep. I noted a few minor coding standard violations too. Some of the classes are getting quite long. Consider splitting into smaller classes if that makes sense.LGTMis often used in such overall comments, to indicateLooks good to merge.
nitis another such term, used to indicate minor flaws e.g.,LGTM, almost. Just a few nits to fix.. - Choose
Approve,Comment, orRequest changesoption as appropriate and click on the Submit review button.
Let's look at the steps involved in merging a PR, assuming the PR has been reviewed, refined, and approved for merging already.
Preparation: If you would like to try merging a PR yourself, you can create a dummy PR in the following manner.
- Fork any repo (e.g., samplerepo-pr-practice).
- Clone in to your computer.
- Create a new branch e.g., (
feature1) and add some commits to it. - Push the new branch to the fork.
- Create a PR from that branch to the
masterbranch in your fork. Yes, it is possible to create a PR within the same repo.
1. Locate the PR to be merged in your repo's GitHub page.
2. Click on the Conversation tab and scroll to the bottom. You'll see a panel containing the PR status summary.
3. If the PR is not merge-able in the current state, the Merge pull request will not be green. Here are the possible reasons and remedies:
- Problem: The PR code is out-of-date, indicated by the message This branch is out-of-date with the base branch. That means the repo's
masterbranch has been updated since the PR code was last updated.- If the PR author has allowed you to update the PR and you have sufficient permissions, GitHub will allow you to update the PR simply by clicking the Update branch on the right side of the 'out-of-date' error message. If that option is not available, post a message in the PR requesting the PR author to update the PR.
- Problem: There are merge conflicts, indicated by the message This branch has conflicts that must be resolved. That means the repo's
masterbranch has been updated since the PR code was last updated, in a way that the PR code conflicts with the currentmasterbranch. Those conflicts must be resolved before the PR can be merged.- If the conflicts are simple, GitHub might allow you to resolve them using the Web interface.
- If that option is not available, post a message in the PR requesting the PR author to update the PR.
3. Merge the PR by clicking on the Merge pull request button, followed by the Confirm merge button. You should see a Pull request successfully merged and closed message after the PR is merged.
- You can choose between three merging options by clicking on the down-arrow in the Merge pull request button. If you are new to Git and GitHub, the
Create merge commitoptions are recommended.
Next, sync your local repos (and forks). Merging a PR simply merges the code in the upstream remote repository in which it was merged. The PR author (and other members of the repo) needs to pull the merged code from the upstream repo to their local repos and push the new code to their respective forks to sync the fork with the upstream repo.