missing_class

Git

menu_book

Git

Version control systems (VCSs)

Modern VCSs let you easily answer questions like:

While other VCSs exist, Git is the de facto standard for version control. This XKCD comic captures Git’s reputation:

xkcd 1597

While Git admittedly has an ugly interface, its underlying design and ideas are beautiful. While an ugly interface has to be memorized, a beautiful design can be understood. For this reason, we give a bottom-up explanation of Git, starting with its data model and later covering the command-line interface. Once the data model is understood, the commands can be better understood, in terms of how they manipulate the underlying data model.

Git’s data model

Git has a well thought-out model that enables all the nice features of version control, like maintaining history, supporting branches, and enabling collaboration.

Snapshots

Git models the history of a collection of files and folders within some top-level directory as a series of snapshots. In Git terminology, a file is called a “blob”, and it’s just a bunch of bytes. A directory is called a “tree”, and it maps names to blobs or trees (so directories can contain other directories). A snapshot is the top-level tree that is being tracked. For example, we might have a tree as follows:

<root> (tree)
|
+- foo (tree)
|  |
|  + bar.txt (blob, contents = "hello world")
|
+- baz.txt (blob, contents = "git is wonderful")

The top-level tree contains two elements, a tree “foo” (that itself contains one element, a blob “bar.txt”), and a blob “baz.txt”.

Modeling history: relating snapshots

How should a version control system relate snapshots? One simple model would be to have a linear history. A history would be a list of snapshots in time-order. For many reasons, Git doesn’t use a simple model like this.

In Git, a history is a directed acyclic graph (DAG) of snapshots. All this means is that each snapshot in Git refers to a set of “parents”, the snapshots that preceded it. It’s a set of parents rather than a single parent (as would be the case in a linear history) because a snapshot might descend from multiple parents, for example due to combining (merging) two parallel branches of development.

Git calls these snapshots “commit”s. Visualizing a commit history might look something like this:

o <-- o <-- o <-- o
            ^  
             \
              --- o <-- o

In the ASCII art above, the os correspond to individual commits (snapshots). The arrows point to the parent of each commit (it’s a “comes before” relation, not “comes after”). After the third commit, the history branches into two separate branches. This might correspond to, for example, two separate features being developed in parallel, independently from each other. In the future, kkthese branches may be merged to create a new snapshot that incorporates both of the features, producing a new history that looks like this, with the newly created merge commit shown in bold:

o <-- o <-- o <-- o <---- <strong>o</strong>
            ^            /
             \          v
              --- o <-- o

Commits in Git are immutable. This doesn’t mean that mistakes can’t be corrected, however; it’s just that “edits” to the commit history are actually creating entirely new commits, and references (see below) are updated to point to the new ones.

Data model, as pseudocode

Git’s data model written down in pseudocode:

// a file is a bunch of bytes
type blob = array<byte>

// a directory contains named files and directories
type tree = map<string, tree | file>

// a commit has parents, metadata, and the top-level tree
type commit = struct {
    parent: array<commit>
    author: string
    message: string
    snapshot: tree
}

It’s a clean, simple model of history.

Objects and content-addressing

An “object” is a blob, tree, or commit:

type object = blob | tree | commit

In Git data store, all objects are content-addressed by their SHA-1 hash.

objects = map<string, object>

def store(object):
    id = sha1(object)
    objects[id] = object

def load(id):
    return objects[id]

Blobs, trees, and commits are unified in this way: they are all objects. When they reference other objects, they don’t actually contain them in their on-disk representation, but have a reference to them by their hash.

For example, the tree for the example directory structure above (visualized using git cat-file -p 698281bc680d1995c5f4caaf3359721a5a58d48d), looks like this:

100644 blob 4448adbf7ecd394f42ae135bbeed9676e894af85    baz.txt
040000 tree c68d233a33c5c06e0340e4c224f0afca87c8ce87    foo

The tree itself contains pointers to its contents, baz.txt (a blob) and foo (a tree). If we look at the contents addressed by the hash corresponding to baz.txt with git cat-file -p 4448adbf7ecd394f42ae135bbeed9676e894af85, we get the following:

git is wonderful

References

Now, all snapshots can be identified by their SHA-1 hash. That’s inconvenient, because humans aren’t good at remembering strings of 40 hexadecimal characters.

Git’s solution to this problem is human-readable names for SHA-1 hashes, called “references”. References are pointers to commits. Unlike objects, which are immutable, references are mutable (can be updated to point to a new commit). For example, the master reference usually points to the latest commit in the main branch of development.

references = map<string, string>

def update_reference(name, id):
    references[name] = id

def read_reference(name):
    return references[name]

def load_reference(name_or_id):
    if name_or_id in references:
        return load(references[name_or_id])
    else:
        return load(name_or_id)

With this, Git can use human-readable names like “master” to refer to a particular snapshot in the history, instead of a long hexadecimal string.

One detail is that we often want a notion of “where we currently are” in the history, so that when we take a new snapshot, we know what it is relative to (how we set the parents field of the commit). In Git, that “where we currently are” is a special reference called “HEAD”.

Repositories

Finally, we can define what (roughly) is a Git repository: it is the data objects and references.

On disk, all Git stores is objects and references: that’s all there is to Git’s data model. All git commands map to some manipulation of the commit DAG by adding objects and adding/updating references.

Whenever you’re typing in any command, think about what manipulation the command is making to the underlying graph data structure. Conversely, if you’re trying to make a particular kind of change to the commit DAG, e.g. “discard uncommitted changes and make the ‘master’ ref point to commit 5d83f9e”, there’s probably a command to do it (e.g. in this case, git checkout master; git reset --hard 5d83f9e).

Staging area

This is another concept that’s orthogonal to the data model, but it’s a part of the interface to create commits.

One way you might imagine implementing snapshotting as described above is have a “create snapshot” command that creates a new snapshot based on the current state of the working directory. Some version control tools work like this, but not Git. We want clean snapshots, and it might not always be ideal to make a snapshot from the current state. For example, imagine a scenario where you’ve implemented two separate features, and you want to create two separate commits, where the first introduces the first feature, and the next introduces the second feature. Or imagine a scenario where you have debugging print statements added all over your code, along with a bugfix; you want to commit the bugfix while discarding all the print statements.

Git accommodates such scenarios by allowing you to specify which modifications should be included in the next snapshot through a mechanism called the “staging area”.

Git command-line interface

To avoid duplicating information, we’re not going to explain the commands below in detail. See the highly recommended Pro Git for more information, or watch the lecture video.

Basics

The git init command initializes a new Git repository, with repository metadata being stored in the .git directory:

$ mkdir myproject
$ cd myproject
$ git init
Initialized empty Git repository in /home/missing-semester/myproject/.git/
$ git status
On branch master

No commits yet

nothing to commit (create/copy files and use "git add" to track)

How do we interpret this output? “No commits yet” basically means our version history is empty. Let’s fix that.

$ echo "hello, git" > hello.txt
$ git add hello.txt
$ git status
On branch master

No commits yet

Changes to be committed:
  (use "git rm --cached <file>..." to unstage)

        new file:   hello.txt

$ git commit -m 'Initial commit'
[master (root-commit) 4515d17] Initial commit
 1 file changed, 1 insertion(+)
 create mode 100644 hello.txt

With this, we’ve git added a file to the staging area, and then git commited that change, adding a simple commit message “Initial commit”. If we didn’t specify a -m option, Git would open our text editor to allow us type a commit message.

Now that we have a non-empty version history, we can visualize the history. Visualizing the history as a DAG can be especially helpful in understanding the current status of the repo and connecting it with your understanding of the Git data model.

The git log command visualizes history. By default, it shows a flattened version, which hides the graph structure. If you use a command like git log --all --graph --decorate, it will show you the full version history of the repository, visualized in graph form.

$ git log --all --graph --decorate
* commit 4515d17a167bdef0a91ee7d50d75b12c9c2652aa (HEAD -> master)
  Author: Missing Semester <missing-semester@mit.edu>
  Date:   Tue Jan 21 22:18:36 2020 -0500

      Initial commit

This doesn’t look all that graph-like, because it only contains a single node. Let’s make some more changes, author a new commit, and visualize the history once more.

$ echo "another line" >> hello.txt
$ git status
On branch master
Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git checkout -- <file>..." to discard changes in working directory)

        modified:   hello.txt

no changes added to commit (use "git add" and/or "git commit -a")
$ git add hello.txt
$ git status
On branch master
Changes to be committed:
  (use "git reset HEAD <file>..." to unstage)

        modified:   hello.txt

$ git commit -m 'Add a line'
[master 35f60a8] Add a line
 1 file changed, 1 insertion(+)

Now, if we visualize the history again, we’ll see some of the graph structure:

* commit 35f60a825be0106036dd2fbc7657598eb7b04c67 (HEAD -> master)
| Author: Missing Semester <missing-semester@mit.edu>
| Date:   Tue Jan 21 22:26:20 2020 -0500
|
|     Add a line
|
* commit 4515d17a167bdef0a91ee7d50d75b12c9c2652aa
  Author: Anish Athalye <me@anishathalye.com>
  Date:   Tue Jan 21 22:18:36 2020 -0500

      Initial commit

Also, note that it shows the current HEAD, along with the current branch (master).

We can look at old versions using the git checkout command.

$ git checkout 4515d17  # previous commit hash; yours will be different
Note: checking out '4515d17'.

You are in 'detached HEAD' state. You can look around, make experimental
changes and commit them, and you can discard any commits you make in this
state without impacting any branches by performing another checkout.

If you want to create a new branch to retain commits you create, you may
do so (now or later) by using -b with the checkout command again. Example:

  git checkout -b <new-branch-name>

HEAD is now at 4515d17 Initial commit
$ cat hello.txt
hello, git
$ git checkout master
Previous HEAD position was 4515d17 Initial commit
Switched to branch 'master'
$ cat hello.txt
hello, git
another line

Git can show you how files have evolved (differences, or diffs) using the git diff command:

$ git diff 4515d17 hello.txt
diff --git c/hello.txt w/hello.txt
index 94bab17..f0013b2 100644
--- c/hello.txt
+++ w/hello.txt
@@ -1 +1,2 @@
 hello, git
 +another line

Branching and merging

Branching allows you to “fork” version history. It can be helpful for working on independent features or bug fixes in parallel. The git branch command can be used to create new branches; git checkout -b <branch name> creates and branch and checks it out.

Merging is the opposite of branching: it allows you to combine forked version histories, e.g. merging a feature branch back into master. The git merge command is used for merging.

Remotes

Undo

Relative Refs

Relative refs are useful to refer to commits without having to type out the hash

Moving work around

Finding your way around

Advanced Git

Miscellaneous

Resources

Exercises

  1. If you don’t have any past experience with Git, either try reading the first couple chapters of Pro Git or go through a tutorial like Learn Git Branching. As you’re working through it, relate Git commands to the data model.
  2. Clone the repository for the class website. git clone https://github.com/missing-semester/missing-semester
    1. Explore the version history by visualizing it as a graph. git log --graph
    2. Who was the last person to modify README.md? (Hint: use git log with an argument) git log -n 1 README.md
    3. What was the commit message associated with the last modification to the collections: line of _config.yml? (Hint: use git blame and git show) 
      git blame -L /collections:/ _config.yml
git show a88b4eac
  3. One common mistake when learning Git is to commit large files that should not be managed by Git or adding sensitive information. Try adding a file to a repository, making some commits and then deleting that file from history (you may want to look at this).
  4. Clone some repository from GitHub, and modify one of its existing files. What happens when you do git stash? What do you see when running git log --all --oneline? Run git stash pop to undo what you did with git stash. In what scenario might this be useful?
  5. Like many command line tools, Git provides a configuration file (or dotfile) called ~/.gitconfig. Create an alias in ~/.gitconfig so that when you run git graph, you get the output of git log --all --graph --decorate --oneline.
  6. You can define global ignore patterns in ~/.gitignore_global after running git config --global core.excludesfile ~/.gitignore_global. Do this, and set up your global gitignore file to ignore OS-specific or editor-specific temporary files, like .DS_Store.
  7. Clone the repository for the class website, find a typo or some other improvement you can make, and submit a pull request on GitHub.

Edit this page.