Exploring History

Last updated on 2024-08-05 | Edit this page

Overview

Questions

How can I identify old versions of files?
How do I review my changes?
How can I recover old versions of files?

Objectives

Explain what the HEAD of a repository is and how to use it.
Identify and use Git commit numbers.
Compare various versions of tracked files.
Restore old versions of files.

As we saw in the previous episode, we can refer to commits by their identifiers. You can refer to the most recent commit of the working directory by using the identifier HEAD.

We’ve been adding one line at a time to mars.txt, so it’s easy to track our progress by looking, so let’s do that using our HEADs. Before we start, let’s make a change to mars.txt, adding yet another line.

BASH

$ nano mars.txt
$ cat mars.txt

OUTPUT

Cold and dry, but everything is my favorite color
The two moons may be a problem for Wolfman
But the Mummy will appreciate the lack of humidity
An ill-considered change

Now, let’s see what we get.

BASH

$ git diff HEAD mars.txt

OUTPUT

diff --git a/mars.txt b/mars.txt
index b36abfd..0848c8d 100644
--- a/mars.txt
+++ b/mars.txt
@@ -1,3 +1,4 @@
 Cold and dry, but everything is my favorite color
 The two moons may be a problem for Wolfman
 But the Mummy will appreciate the lack of humidity
+An ill-considered change.

which is the same as what you would get if you leave out HEAD (try it). The real goodness in all this is when you can refer to previous commits. We do that by adding ~1 (where “~” is “tilde”, pronounced [til-duh]) to refer to the commit one before HEAD.

BASH

$ git diff HEAD~1 mars.txt

If we want to see the differences between older commits we can use git diff again, but with the notation HEAD~1, HEAD~2, and so on, to refer to them:

BASH

$ git diff HEAD~3 mars.txt

OUTPUT

diff --git a/mars.txt b/mars.txt
index df0654a..b36abfd 100644
--- a/mars.txt
+++ b/mars.txt
@@ -1 +1,4 @@
 Cold and dry, but everything is my favorite color
+The two moons may be a problem for Wolfman
+But the Mummy will appreciate the lack of humidity
+An ill-considered change

We could also use git show which shows us what changes we made at an older commit as well as the commit message, rather than the differences between a commit and our working directory that we see by using git diff.

BASH

$ git show HEAD~3 mars.txt

OUTPUT

commit f22b25e3233b4645dabd0d81e651fe074bd8e73b
Author: Vlad Dracula <vlad@tran.sylvan.ia>
Date:   Thu Aug 22 09:51:46 2013 -0400

    Start notes on Mars as a base

diff --git a/mars.txt b/mars.txt
new file mode 100644
index 0000000..df0654a
--- /dev/null
+++ b/mars.txt
@@ -0,0 +1 @@
+Cold and dry, but everything is my favorite color

In this way, we can build up a chain of commits. The most recent end of the chain is referred to as HEAD; we can refer to previous commits using the ~ notation, so HEAD~1 means “the previous commit”, while HEAD~123 goes back 123 commits from where we are now.

We can also refer to commits using those long strings of digits and letters that git log displays. These are unique IDs for the changes, and “unique” really does mean unique: every change to any set of files on any computer has a unique 40-character identifier. Our first commit was given the ID f22b25e3233b4645dabd0d81e651fe074bd8e73b, so let’s try this:

BASH

$ git diff f22b25e3233b4645dabd0d81e651fe074bd8e73b mars.txt

OUTPUT

diff --git a/mars.txt b/mars.txt
index df0654a..93a3e13 100644
--- a/mars.txt
+++ b/mars.txt
@@ -1 +1,4 @@
 Cold and dry, but everything is my favorite color
+The two moons may be a problem for Wolfman
+But the Mummy will appreciate the lack of humidity
+An ill-considered change

That’s the right answer, but typing out random 40-character strings is annoying, so Git lets us use just the first few characters (typically seven for normal size projects):

BASH

$ git diff f22b25e mars.txt

OUTPUT

diff --git a/mars.txt b/mars.txt
index df0654a..93a3e13 100644
--- a/mars.txt
+++ b/mars.txt
@@ -1 +1,4 @@
 Cold and dry, but everything is my favorite color
+The two moons may be a problem for Wolfman
+But the Mummy will appreciate the lack of humidity
+An ill-considered change

All right! So we can save changes to files and see what we’ve changed. Now, how can we restore older versions of things? Let’s suppose we change our mind about the last update to mars.txt (the “ill-considered change”).

git status now tells us that the file has been changed, but those changes haven’t been staged:

BASH

$ git status

OUTPUT

On branch main
Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)

    modified:   mars.txt

no changes added to commit (use "git add" and/or "git commit -a")

We can put things back the way they were by using git restore:

BASH

$ git restore mars.txt
$ cat mars.txt

OUTPUT

Cold and dry, but everything is my favorite color
The two moons may be a problem for Wolfman
But the Mummy will appreciate the lack of humidity

As you might guess from its name, git restore restores an old version of a file. In this case, we’re telling Git that we want to recover the version of the file recorded in HEAD, which is the last saved commit. If we want to go back even further, we can use the ‘source’ flag -s to specify a commit identifier instead:

BASH

$ git restore -s f22b25e mars.txt

BASH

$ cat mars.txt

OUTPUT

Cold and dry, but everything is my favorite color

BASH

$ git status

OUTPUT

On branch main
Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)

    modified:   mars.txt

Again, we can put things back the way they were by using git restore:

BASH

$ git restore mars.txt

It’s important to remember that we must use the commit number that identifies the state of the repository before the change we’re trying to undo. A common mistake is to use the number of the commit in which we made the change we’re trying to discard. In the example below, we want to retrieve the state from before the most recent commit (HEAD~1), which is commit f22b25e:

So, to put it all together, here’s how Git works in cartoon form:

Simplifying the Common Case

If you read the output of git status carefully, you’ll see that it includes this hint:

OUTPUT

(use "git restore <file>..." to discard changes in working directory)

As it says, git restore without a version identifier restores files to the state saved in HEAD.

The fact that files can be reverted one by one tends to change the way people organize their work. If everything is in one large document, it’s hard (but not impossible) to undo changes to the introduction without also undoing changes made later to the conclusion. If the introduction and conclusion are stored in separate files, on the other hand, moving backward and forward in time becomes much easier.

Recovering Older Versions of a File

Jennifer has made changes to the Python script that she has been working on for weeks, and the modifications she made this morning “broke” the script and it no longer runs. She has spent ~ 1hr trying to fix it, with no luck…

Luckily, she has been keeping track of her project’s versions using Git! Which commands below will let her recover the last committed version of her Python script called data_cruncher.py?

$ git restore
$ git restore data_cruncher.py
$ git restore -s HEAD~1 data_cruncher.py
$ git restore -s <unique ID of last commit> data_cruncher.py
Both 2 and 4

Show me the solution

The answer is (5)-Both 2 and 4.

The restore command restores files from the repository, overwriting the files in your working directory. Answers 2 and 4 both restore the latest version in the repository of the file data_cruncher.py. Answer 2 doesn’t specify a commit, which means it automatically refers to the latest, whereas answer 4 uses the unique ID of the last commit, which would be the same as using HEAD instead.

Answer 3 gets the version of data_cruncher.py from the commit before HEAD, which is NOT what we wanted.

Answer 1 reports an error fatal: you must specify path(s) to restore: you haven’t specified which file(s) to restore. It’s a good idea to be specific about which files you mean,so you don’t accidentally restore more files than you need. If you do want to restore all previously committed files in your repository, you can use . to specify the current folder (and all subfolders), i.e., git restore ..

Reverting a Commit

Jennifer is collaborating with colleagues on her Python script. She realizes her last commit to the project’s repository contained an error, and wants to undo it. Jennifer wants to undo correctly so everyone in the project’s repository gets the correct change. The command git revert [erroneous commit ID] will create a new commit that reverses the erroneous commit.

The command git revert is different from git restore -s [commit ID] because git restore returns the files not yet committed within the local repository to a previous state, whereas git revert reverses changes committed to the local and project repositories.

Below are the right steps and explanations for Jennifer to use git revert, what is the missing command?

________ # Look at the git history of the project to find the commit ID
Copy the ID (the first few characters of the ID, e.g. 0b1d055).
git revert [commit ID]
Type in the new commit message.
Save and close

Show me the solution

The command git log lists project history with commit IDs.

The command git show HEAD shows changes made at the latest commit, and lists the commit ID; however, Jennifer should double-check it is the correct commit, and no one else has committed changes to the repository.

Understanding Workflow and History

What is the output of the last command in

BASH

$ cd planets
$ echo "Venus is beautiful and full of love" > venus.txt
$ git add venus.txt
$ echo "Venus is too hot to be suitable as a base" >> venus.txt
$ git commit -m "Comment on Venus as an unsuitable base"
$ git restore venus.txt
$ cat venus.txt #this will print the contents of venus.txt to the screen

OUTPUT

  Venus is too hot to be suitable as a base

OUTPUT
```
  Venus is beautiful and full of love
```

OUTPUT

  Venus is beautiful and full of love
  Venus is too hot to be suitable as a base

OUTPUT

  Error because you have changed venus.txt without committing the changes

Show me the solution

The answer is 2.

The command git add venus.txt places the current version of venus.txt into the staging area. The changes to the file from the second echo command are only applied to the working copy, not the version in the staging area.

So, when git commit -m "Comment on Venus as an unsuitable base" is executed, the version of venus.txt committed to the repository is the one from the staging area and has only one line.

At this time, the working copy still has the second line (and git status will show that the file is modified). However, git restore venus.txt replaces the working copy with the most recently committed version of venus.txt.

So, cat venus.txt will output

OUTPUT

Venus is beautiful and full of love.

Checking Understanding of `git diff`

Consider this command: git diff HEAD~9 mars.txt. What do you predict this command will do if you execute it? What happens when you do execute it? Why?

Try another command, git diff [ID] mars.txt, where [ID] is replaced with the unique identifier for your most recent commit. What do you think will happen, and what does happen?

Getting Rid of Staged Changes

git restore can be used to restore a previous commit when unstaged changes have been made, but will it also work for changes that have been staged but not committed? Make a change to mars.txt, add that change using git add, then use git restore to see if you can remove your change.

Show me the solution

After adding a change, git restore can not be used directly. Let’s look at the output of git status:

OUTPUT

On branch main
Changes to be committed:
  (use "git restore --staged <file>..." to unstage)

        modified:   mars.txt

Note that if you don’t have the same output you may either have forgotten to change the file, or you have added it and committed it.

Using the command git restore mars.txt now does not give an error, but it does not restore the file either. Git helpfully tells us that we need to use git restore --staged instead to unstage the file:

BASH

$ git restore --staged mars.txt

Now, git status gives us:

BASH

$ git status

OUTPUT

On branch main
Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)

        modified:   mars.txt

no changes added to commit (use "git add" and/or "git commit -a")

This means we can now use git restore to restore the file to the previous commit:

BASH

$ git restore mars.txt
$ git status

OUTPUT

On branch main
nothing to commit, working tree clean

Explore and Summarize Histories

Exploring history is an important part of Git, and often it is a challenge to find the right commit ID, especially if the commit is from several months ago.

Imagine the planets project has more than 50 files. You would like to find a commit that modifies some specific text in mars.txt. When you type git log, a very long list appeared. How can you narrow down the search?

Recall that the git diff command allows us to explore one specific file, e.g., git diff mars.txt. We can apply a similar idea here.

BASH

$ git log mars.txt

Unfortunately some of these commit messages are very ambiguous, e.g., update files. How can you search through these files?

Both git diff and git log are very useful and they summarize a different part of the history for you. Is it possible to combine both? Let’s try the following:

BASH

$ git log --patch mars.txt

You should get a long list of output, and you should be able to see both commit messages and the difference between each commit.

Question: What does the following command do?

BASH

$ git log --patch HEAD~9 *.txt

Key Points

git diff displays differences between commits.
git restore recovers old versions of files.

Exploring History

Overview

Questions

Objectives

BASH

OUTPUT

BASH

OUTPUT

BASH

BASH

OUTPUT

BASH

OUTPUT

BASH

OUTPUT

BASH

OUTPUT

BASH

OUTPUT

BASH

OUTPUT

BASH

BASH

OUTPUT

BASH

OUTPUT

BASH

Simplifying the Common Case

OUTPUT

Recovering Older Versions of a File

Show me the solution

Reverting a Commit

Show me the solution

Understanding Workflow and History

BASH

OUTPUT

OUTPUT

OUTPUT

OUTPUT

Show me the solution

OUTPUT

Checking Understanding of git diff

Getting Rid of Staged Changes

Show me the solution

OUTPUT

BASH

BASH

OUTPUT

BASH

OUTPUT

Explore and Summarize Histories

BASH

BASH

BASH

Key Points

Checking Understanding of `git diff`