Learning how to use GitHub

Don't know any git or GitHub? Then this is for you! It's a very simple hands-on tutorial.

Contents

Making changes
Ignoring files
Handling conflicts
Working with collaborators
Further

Learning objectives

Store version-controlled projects online
Work with collaborators on the same project

Not covered

Command-line git tools
Working with branches

Required

GitHub GUI (install for Windows or Mac)
GitHub account (get an account here)
Good text-editor (e.g. Atom, but notepad or similar will work)
Internet connection

Instructions

A. Making changes

Basic git usage is to have a repository online, make changes to a local copy on your machine, and then push those changes to the online version. This allows you to have a backup of all your code, have a history of all the steps you've taken for developing your code (which allows you to revert to previous versions if anything goes wrong) and work with collaborators.

Repo -- or repository is a folder that is under version-control. It can be held online and across multiple computers, each copy is kept up-to-date by syncing. When under version-control, nothing is ever lost. All versions and deleted files are kept in the .git/ folder and, in theory, everything is recoverable. A repo can contain both folders and files.

1. Fork this repo

Every repo on GitHub can be forked. Look for the fork button in the top right corner of this page to copy this repo to your own account.

Fork -- copy someone else's repo to your own account. Once its in your account, its yours to do what you want with (but make sure you read the licensing of a repo before forking, usually stored in a LICENSE.txt)

2. Clone this repo to your local machine

In your GitHub GUI, select the + sign in the top-left corner. And clone the recently forked repo. This will create a complete copy of the repository on your machine.

Clone -- copy a repo, to which you have access, to your own computer.

3. Make a change

With the repo on your machine, you can now make changes to it. Open the script.md file and adapt the poem to your liking or make any other change you like to the file e.g. replace sheep with goat:

## The goat

*The goat is alone.  
A single purple dot,  
in fields purple.*

4. Commit and sync this repo

Now you've made a change to script.md, you'll see the GUI tell you what changes you have made. Commit your change with a message e.g. "I made the poem better". Then click the sync button in the top-right corner and check your copy online: https://github.com/[YOUR ACCOUNT NAME]/Teaching-github. Also check out the history of script.md online or on your GUI.

Commit -- A change with a message (e.g. corrected infinite loop in myscript.sh) that can then be pushed to GitHub. All changes made must be committed in order to sync them.

B. Ignoring files

Git will version control everything placed in the repo. This is not always helpful as you do not necessarily want all files and folders to be synced and not everything needs to be under version-control (e.g. automatically generated files). To control what is under version-control use the .gitignore file.

.gitignore -- a text file that contains a list of files (or file patterns) that should be ignored.

1. Open the .gitignore and ignore .csv files

In your text editor, open the .gitignore file and add *.csv under 'Ignored files'. This tells git to ignore any file that ends with .csv. It is useful to ignore data files (like csv files) for multiple reasons: GitHub has a 1GB repo limit; it is pointless version controlling files that are automatically generated, this will make the .git/ folder very large over time and slow-down simple git processes; large git repos take longer to perform simple tasks like cloning and syncing.

# Ignored files
.DS_Store
*.csv

Now you've edited a file, commit it with a message in your GUI e.g. 'Now ignoring csv files'

.DS_Store -- in the .gitignore you will already find a file name. This is a Mac generated file, it is not part of the coding project and should be ignored. The file helps Macs perform spotlight searches.

2. Add a .csv file to the repo

Create a new file in the repo called 'test.csv'. Does the git detect it?

. prefix -- in UNIX systems this makes a file invisible. If you cannot find your .gitignore open a terminal and use ls -a to see all files in a folder. You can also use a simple terminal text editor like nano to modify the file (nano .gitignore).

C. Handling conflicts

When multiple people work on a repo at once, conflicts can occur if more than one person has made a change to a file. These are termed merge conflicts and must be resolved before a repo can be synced.

1. Creating a conflict

You now have two copies of your repo, one on your machine and another on GitHub. Go to the GitHub repo and modify the script.md, use the pencil icon in the top-right. Edit the 'I expect this line to be a conflict' line to whatever you like. Commit your change directly on the GitHub version.

Now repeat what you did for the copy on your machine. Change the line you changed online to something different. Commit the change and try to sync the repo. This will create a 'merge-conflict' because your local copy has a committed change that conflicts with a committed change on the remote repo.

To avoid conflicts it's best to commit and sync often; small and frequent is the aim of the game.

2. Conflict resolution

Git has recognised the conflict and has added to the file in conflict, something like this:

<<<<<<< HEAD
    [This is the text that is in your local copy]
=======
    [This is the test in the remote copy]
>>>>>>> master

It shows the lines in conflict and all you need to do is fix it manually. Either by merging the two lines or choosing one over the other.

Once you've corrected the file, commit the change. Git suggests a message for you, something like: 'Merge remote-tracking branch origin/master'.

D. Working with collaborators

Now you have a repo, you know how to commit changes to it and resolve conflicts, you just need to have some collaborators with whom to work.

Go to the online repo
Click on settings in the panel on the right
Click on collaborators
Add collaborators using their GitHub account names

Any collaborator can now sync ('push access') their changes to the repo online.

Further

I've only introduced the GitHub GUI, with git command-line tools you're able to do a lot more. That said, 95% of what you use git and GitHub for is covered here (forking, cloning, committing and syncing). To learn more see the following links:

GitHub is constantly growing and inventing new things, here are some useful things I know about:

Pay for private repos that you and collaborators can see
If you're a student you can apply for free privates repos through their education package
Highlight issues with your code for others to see (and hopefully fix!)
You can create wikis to explain your code
Create organisations that share repos which members of the group can push to
Copy and paste snippets of code between people with gist
Host websites
GitHub supports emoji (:sheep::cow::horse:)