What is git anyway? Short answer: version control. Long answer: Think of it as having unlimited undos on your code. Or better yet, if you’re collaborating with another developer, it allows you to both work on the same project without overwriting each other’s code. Awesome!
This has saved my butt on more than one occasion. I’ve vowed to always use source control, regardless of how big or small the project is.
So, now that you’re interested in git, where do you begin? To get started, you’ll need to install git on your local machine (don’t worry, I’ll explain). Eventually, you’ll want to use an online service to store your code, like GitHub, BitBucket, or Beanstalk (more on that below, too).
Installing git Locally
I use a Mac, so unfortunately, I can only speak to that. However, if you’re using another operating system try looking here.
The easiest way to to install Git on your Mac (if you’re on Mavericks or older) is to install the XCode Command Line Tools. This is as simple as trying to run git from the Terminal.
Remember, when you’re working in the Command Line, the dollar sign ($) is the prompt and signifies a new line. No need to copy it.
If you don’t have git installed already, it will prompt you to do so.
If you want a more up to date version, or would rather run the binary installer, you can grab that here.
The first thing you’ll want to do is configure your git settings.
$ git config --global user.name "John Doe" $ git config --global user.email "email@example.com" $ git config --global color.ui auto
Git is typically run from the Terminal. However, as a designer, I tend to be a little leary of the Terminal. Granted I’ve gotten more comfortable, but I still prefer a GUI (Graphical User Interface) when I can get one.
I think the easiest thing to do is learn the vocabulary. Then, it doesn’t matter whether it’s the Terminal or GUI, the same concepts apply no matter what.
Learning the vocabulary
Here’s a cheat sheet you can print out and keep handy.
Let’s walk through this as you would for a real project.
You’ll want to create a folder on your computer for your project. In my user folder, I have a folder called Code. Then, a subfolder called GIT. I keep all my project folders there.
If you’re trying to do everything from the command line, these commands create a new folder and then navigates inside:
$ mkdir new_project $ cd new_project
Let’s initiatize our git repository. (All your git projects are called repositories.)
$ git init
So what does initializing really do? It creates an empty repository in a hidden folder, .git.
If you ever wanted to remove the repository, all you’d need to do is delete that .git subdirectory:
$ git rm -rf .git
Notice, I said empty. It did not add the current content within your folder. You have to tell it to do that yourself.
All the content within your project folder is considered the working copy. Once you get to a place where you want to save those files, you’ll stage your changes and then commit them to your local repository.
The advantage here is not all of the files you’ve modified since your last commit have to be staged and committed.
Let’s get an overview of everything that has changed since we last committed:
Git will put your changes into three main categories:
- Changes not staged for commit
- Changes to be committed
- Untracked files
You can add all these changes by using:
git add .
If you don’t want to add the changes you’ve made, but cherry pick files, you can list them out:
git add assets/dist/css/main.css assets/dist/js/production.js
If we removed a file, we’d need to confirm that too:
git rm something.html
Now, let’s commit our updates:
git commit -m "Initial commit"
The first commit, I usually title as “Initial commit” but as you continue to make changes and commit your code, you’ll want these messages to be meaningful. This will make it easier for you (or your teammates) to understand the changes you made later down the road.
Your message should include information like, “What was the motivation for the change? How does this differ from the previous version?”
If you need a commit message longer than 50 characters, you can leave out the “-m” parameter and Git will open an editor application for you to write a longer message.
Here are a few guidelines from Tower’s eBook, Learn Version Control with Git:
- Related Changes A commit should only contain changes from a single topic. Don’t mix up contents from different topics. This will make it harder to understand what happened.
- Completed Work Never commit something that is half-done. If you need to save your current work temporarily in something like a clipboard, you can use Git’s “Stash” feature. But don’t eternalize it in a commit.
- Tested Work You shouldn’t commit code that you think is working. Test it well — and before you commit it to the repository.
- Short and Descriptive Messages A good commit also needs a good message.
If you want to review a history of all the commits that have been made to a project:
You’ll see a list in chronological order, beginning with the newest item.
If there are more items than what can be displayed on one page, you’ll see a : at the bottom of the screen. You can go to the next page by hitting the <SPACE> key and quit with “q”.
You’ll notice from the log, every commit has:
- Commit Hash This is that crazy long string of letters and numbers (i.e. aa093b890c78e9d0869a3f267b2530cf2cbeb83f)
- Author Name and Email Remember when we set our name and email address within the git configurations above? This is where it gets used.
- Commit Message
Now that we’ve made a bunch of commits, let’s publish (or push) them online. First we need to tell Git where our remote repository is.
There’s several services you can use, three of the most popular ones are:
Regardless of the service you use, when you create a repository, they will give you an SSH address and HTTPS address.
You can use either:
$ git remote add origin firstname.lastname@example.org:ahhacreative/git_blog_post.git
$ git remote add origin https://email@example.com/ahhacreative/git_blog_post.git
In this case, we named our remote repository “origin.” This is the default within Git. However, I could just as easily name it something else. I usually name mine based on where the remote repository is. This makes it easy within Tower (my GUI of choice) to tell where it’s being saved:
From the command line:
$ git remote add BITBUCKET firstname.lastname@example.org:ahhacreative/git_blog_post.git
You can see a list of all your remotes:
git remote -v
You’ll notice there are two URLs listed (fetch and push). One is for read access (fetch) and the other is for write (push). Usually, the two URLs are the same, but you could use different URLs for security and performance issues.
Now, we need to push our code up to our remote:
$ git push origin master
Hopefully, some of these things are starting to look familiar. Origin is the name of the remote repository and master is the name of our branch. (We haven’t talked about branching yet. It’s OK, just trust me for now, but it’s coming.)
Let’s pause for a moment. I just want to take a moment to point out: you’re more than halfway there! It wasn’t that hard, was it?! You know everything you need to know to save, commit, and publish your files online! There’s still more topics to cover, but you already know the basics. That’s worth celebrating!
OK. So far, we’ve just published our changes online. But, what if we’re collaborating with another developer? They’re publishing their changes too. How do we pull down there code?
First you may want to see what you’re pulling:
$ git fetch origin $ git log origin/master
If you want to integrate these changes into your local working copy:
$ git pull
If you have multiple remotes and branches, you may need to specify:
$ git pull origin master
OK, so let’s (finally) talk about branching.
When you’re coding, there are usually several different contexts or “topics” you’re working with:
- bug fixes
In the real world, these all happen simultaneously. You’re trying to fix a bug while a teammate is working on the new about page design.
If branching didn’t exist, how would you integrate some of these changes, but not all of them? You need to post the code for the bug fix, but your teammate isn’t ready to launch the new about page. You find out the new “members only” code you’ve been working on is going in a different direction, but you’ve already integrated that code in with everything else. How do you separate it out?
You can create a branch to handle each of these topics. It stays separate from all the other contexts. When you make changes, they only apply to the current active branch.
Your current branch can be referred to as the active branch, checked out branch, or HEAD branch. They all mean the same thing.
Whenever you run
git status, it will tell you what branch you’re currently on.
You may have noticed from my screenshots, that my Terminal shows me what branch I’m in and changes color based on whether there are uncommitted files or not.
To create a new branch:
$ git branch feature/about-page
I like to name my branches my contexts. So, you’ll notice, I prepended my branch name (about-page) with “feature.” This is nice because within Tower, it treats these contexts as folders:
We’ve created a branch, but it’s still not our current branch. To change branches, you’ll need to checkout:
$ git checkout -b feature/about-page
To view a list of all our branches:
$ git branch -v
-v (verbose) flag, provides more information, than just
Now, that you know how to create and change branches. What happens when you’re ready to commit that code to your main (master) branch?
First, you’ll want to navigate to your master branch. Then, merge in your new branch:
$ git checkout master $ git merge feature/about-page
Once a branch is merged and deployed, you no longer need it. To delete a local branch:
$ git branch -d feature/about-page
Remember when we talked about not committing half done work? How do you get around that?
git-tower.com explains it well:
Think of the Stash as a clipboard on steroids: it takes all the changes in your working copy and saves them for you on a new clipboard. You’re left with a clean working copy, i.e. you have no more local changes.
To stash your local changes:
$ git stash
To see an overview of all your current stashes:
$ git stash list
The most recent Stashes will always be listed first.
When you want to apply a Stash, you can run:
$ git stash pop
This will pull the latest Stash and clear it from your clipboard. Or
$ git stash apply <StashName>
This pulls the specified Stash, but unlike
pop, it remains saved. To delete it, you’ll need to run
git stash drop <StashName>
These things are all great, but what if you’re not starting from scratch? What if you’re joining a project that already exists?
Then, you’ll want to use
$ git clone https://email@example.com/ahhacreative/git_blog_post.git
This will automatically set the remote to the “origin.” Anytime you
pull, it will grab any updates made to the repository.
Sometimes, you’ll want to grab the code from a repository, but start with a clean history. To do so, simply add the flag
$ git clone --bare https://firstname.lastname@example.org/ahhacreative/git_blog_post.git
There will be times when you don’t want to put all your code within your repository. For example, if you’re using bower, you probably don’t want the bower_components folder. Or, if you’re running grunt or gulp, you don’t need node_modules. A teammate only needs to run
bower init or
npm install to get those files. Committing them would only bloat your repository.
You can create a file called
.gitignore and list all the files or folders to exclude. For example, here’s my .gitignore file for a WordPress project:
There are several services that you can use to host your remote repositories.
GitHub is probably the most popular. If you make your repositories public, then you can create an unlimited number of repositories for free. You start paying when you create private repositories.
GitHub also has an excellent issue tracker built in among several other great resources.
Even if you don’t plan on paying for a GitHub account, I would highly recommend signing up for an account anyway. GitHub has practically become a standard for developers. Meaning, if you’re applying for a programming job, your potential employer might ask to see examples on GitHub.
BitBucket does everything that GitHub does, but is missing the community that surrounds GitHub.
The advantage to using BitBucket is that it allows you to have an unlimited number of private repositories and 5 collaborators for free.
BeanStalk is another great service. Like GitHub, you’re paying based on the number of repositories you use.
It’s missing a lot of the features that the other two services provide, but the biggest benefit it has is built in deployment. Meaning, you can push changes to your repository and it will automatically send those updates to your server via FTP (or however you specify). That’s almost enough to make me switch.
In the meantime, I’m using dploy.io instead. It’s made by the same company (WildBit) that makes Beanstalk. Dploy.io allows you to deploy commits from BitBucket or GitHub. Pricing is based on the number of repositories you’ve connected.
OK, now that you know the correct terminology, I want to introduce a few GUIs:
I use Tower. IMHO, it’s the best. Just from glancing at the screenshot, you can see the buttons along the top for most of the vocabulary terms covered in this post. If you want a little more information, check out this screencast (from the Tower app team):
GitHub has their own GUI. Unlike Tower, it’s free. I’ve downloaded it, but I rarely use it since I have Tower.
At the end of the day, it’s whatever you’re the most comfortable with. They all accomplish the same thing. My thing is don’t let the tools get in the way of making cool stuff.
Granted, we’ve covered the basics, it should be enough to get you up and going. However, if you want to dig a little deeper, here are a few additional resources to check out: