Git deep-dive: "git init"
In this article:
This article kicks off a multi-part series taking a closer look at the common parts of git we often take for granted.
And what better way to start the series than with the command that starts all git repositories:
git init starts a new repository! You want source control in Git? Just run the command
git init and BAM!– that directory is now a git repository.
Here's what it looks like in action:
Once this is done, you can do all the cool things you do in git, such as
merge, and more!...But those commands are for another article. We're here to focus a little more on the usage of
git init. Let's take a closer look...
git init actually doing when it initializes a repository? When you run the command, it creates a hidden directory called
.git/ contains all the behind-the-scenes data that makes git source control tick. All the branch information, past changes, current status, and more is all held here. Here's what the contents of that directory look like:
We aren't going to go through what each directory and file does here. Most of it isn't human-readable, and is only intended to be interacted with using Git. Just know that this
.git/ sub-directory structure (or most of it) is required for git to do its thing.
If you wanted, you could even manually create
.git/ and all its contents, and git would still recognize it as a repository (Try it out if you have some free time! Which files are required? Which aren't?). This also means if you delete the
.git directory on an existing repository, all your branches and git history will be deleted as well. This can actually be useful if you're very early in a project and decide you want to re-start a repository to initialize from your current working files. Just
rm -rf .git and
git init all over again.
We often think about git storing historical data, while your working directory (everything outside
.git/) has the current files. It's important to remember that Git, or more specifically: the
.git directory, has all the data — including the most recent committed changes. In fact, you can even
git clone directly from the
.git directory, just like you would from a remote repository!
Did you know you can actually
push to and
pull from your local working repository? However, you might run into some issues. Watch what happens when we try to push the
master branch to a local working repository where
master is also checked out:
This is why you'll typically push and pull from a remote repository instead of from someone else's working repository. Most often, this remote repository setup is handled through a git repo service such as Bitbucket or GitHub, but if you wanted to set up your own, you'd use the
Recognize those files and directories? That's everything we saw before in the hidden
.git repository, except now they aren't covered by a hidden folder! Yes, they're...naked! 😮 So it shouldn't be surprising that this is known as a "bare repository". With a naked – um, I mean – bare repository, you're essentially telling git, "Nobody's working here. Just store the data, and let users push and pull from it."
Since nobody's working here, you don't really need anything outside the
.git directory. And if you don't need anything outside the
.git directory, why not just put all the
.git contents at the top level?
And finally...Is your
git init making TOO MUCH NOISE???
Stop all that terrible noisy racket with the
I hope this has helped shine a light on how
git init works and how a repository is born. For further reading, you may want to check out:
Stay tuned for the next article in the series where I cover the slightly-more-interesting command:
Continue the conversation