Exploring History

As we saw in the previous lesson, 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.

nano mars.txt
cat mars.txt

Now, let’s see what we get.

git diff HEAD mars.txt

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.

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:

git diff HEAD~2 mars.txt

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. git show HEAD~2 mars.txt

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:

git diff f22b25e3233b4645dabd0d81e651fe074bd8e73b mars.txt

That’s the right answer, but typing out random 40-character strings is annoying, so Git lets us use just the first few characters:

git diff f22b25e mars.txt

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 accidentally overwrite our file:

nano mars.txt
cat mars.txt

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

git status

On branch master
Your branch is up to date with 'origin/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:   02-creating-repositories_instructor.ipynb
	modified:   03-tracking-changes_instructor.ipynb

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 checkout:

git checkout HEAD mars.txt
cat mars.txt

As you might guess from its name, git checkout checks out (i.e., 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 a commit identifier instead:

git checkout f22b25e mars.txt

cat mars.txt

git status

On branch master
Your branch is up to date with 'origin/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:   02-creating-repositories_instructor.ipynb
	modified:   03-tracking-changes_instructor.ipynb

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

Notice that the changes are on the staged area. Again, we can put things back the way they were by using git checkout:

git checkout HEAD mars.txt

$ git checkout f22b25e mars.txt

$ git checkout f22b25e

Note: checking out 'f22b25e'.

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 f22b25e Start notes on Mars as a base

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 get rid of. 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:

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

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.

$ 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 checkout HEAD venus.txt
$ cat venus.txt #this will print the contents of venus.txt to the screen

$ git log mars.txt

$ git log --patch mars.txt

$ git log --patch HEAD~3 *.txt

The material in this notebook is derived from the Software Carpentry lessons © Software Carpentry under the terms of the CC-BY 4.0 license.