Understanding Git branches

In the last article we talked about internal working of Git. We saw the actual meaning of git commits and how git uses them to track different versions of stuff in the repository.

One of the biggest things git offers is the ability to create branches. These are separate parallel paths carrying different versions of the repository, at times independent from each other. In this article we will look at what actually happens when you use branches.

A new branch

So far we have only had one branch,i.e, master. Go ahead and create a new branch using

> git branch canon
> git checkout canon

If we check the .git folder now, we have a new file called canon in refs

The refs folder basically tracks the different references throughout the repository. The HEAD file contains the path to this canon file as well. Now let's check the contents of the canon file.

We can see that this is the most recent commit in the canon branch. Thus the repository tree looks something like this.

New branch moves forward

Lets make some changes in testfile in the canon branch and commit it. We now have three new files in the objects folder

Using git cat-file, we can deduce the relations between each of these files

This is similar to what we saw in the last article, 5d... is the commit, 51... is the tree originating from the previous commit(f5...) and 23... is the blob where the actual content is present. The repository looks something like this.

Notice that the objects folder only contains this information and not the information of which branch this is on. That is maintained by the HEAD file. Since the HEAD points to the canon and that contains the new commit, the HEAD now points to the latest commit in canon.

Merging the new branch

Let's merge the new branch in the master branch

❯ git checkout master
❯ git merge canon

Now if we check the contents of master in refs, we get 5d6808021c9dfd7bbe88fb8ac450e572982f79b7, which is the latest commit that is merged from the new branch. Thus the repository looks something like this:

Thus we can see how git maintains the reference of the branches along with the contents of it. This is what makes git so powerful. Having a tree structure means git can navigate to any specific commit in any specific branch.

Thanks for reading!