====== Ubuntu - GIT - Undo changes ======

Generally there are three ways of reverting changes:

  * checkout
  * revert
  * reset

----

===== Checkout =====

If you just need to revert specific files, you could run **git checkout** to retrieve an exact version.  In the below example, I wanted to revert the “app.rb” file so that it only contains "Some app work".

<code bash>
git log --oneline
1ab20e2 More experimentation
8d9b88d Experimenting with something
cb1aa50 Do some good work
d05eb2a Initial commit

ls
app.rb file1.rb README

cat app.rb
Some app work
experimenting with something
More experimentation

git checkout cb1a app.rb

cat app.rb
Some app work

git status
On branch master
Changes to be committed:
  (use "git reset HEAD <file>..." to unstage)
  
        modified:   app.rb

git commit -am "Not a good experiment, moving back to good code"
[master 5a39d14] Not a good experiment, moving back to good code
 1 file changed, 2 deletions(-)

git status
On branch master
nothing to commit, working directory clean

git log --oneline
5a39d14 Not a good experiment, moving back to good code
1ab20e2 More experimentation
8d9b88d Experimenting with something
cb1aa50 Do some good work
d05eb2a Initial commit
</code>

----

===== Revert =====

Revert will create a new commit undoing the changes made during a specific commit.

It remove an entire commit in your project history.

In this example, I'm going to undo the changes in last commit.

As you could see, the history of the revert is kept.

<code bash>
git log --oneline
d69cd9c modify second line
ab44e89 add second line
7a49968 first line

cat hello.rb
first line
modify second line

git revert d69c
[master 6769261] Revert "modify second line"
 1 file changed, 1 insertion(+), 1 deletion(-)

git log --oneline
6769261 Revert "modify second line"
d69cd9c modify second line
ab44e89 add second line
7a49968 first line
</code>

----

===== Reset =====

Unlike revert, reset will undo all subsequent commits.

It has the potential to cause issues, so only use this to undo local changes.

Most use reset to unstage files to match the most recent commit and perhaps create more focused commits/snapshots.

The working directory is unchanged unless the **"–hard"** option is set.

<code bash>
git log --oneline
d69cd9c modify second line
ab44e89 add second line
7a49968 first line

git status
On branch master
nothing to commit, working directory clean

git reset ab44
Unstaged changes after reset:
M        hello.rb

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 changed in working directory)

        modified:   hello.rb

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

git reset --hard ab44
HEAD is now at ab44e89 add second line

git status
On branch master
nothing to commit, working directory clean

git log --oneline
ab44e89 add second line
7a49968 first line
</code>

You could also reset to a tag.

<code bash>
git tag
v1.0
v1.1

git log --oneline
4f04459 add burrito
30ee499 add hamburger
e521e3a Release v1.0
ab44e89 add second line
7a49968 first line

git reset --hard v1.0
HEAD is now at e521e3a Release v1.0

git log --oneline
e521e3a Release v1.0
ab44e89 add second line
7a49968 first line

ls
hello.rb

git reset --hard v1.1
HEAD is now at 4f04459 add burrito

git log --oneline
4f04459 add burrito
30ee499 add hamburger
e521e3a Release v1.0
ab44e89 add second line
7a49968 first line

ls
burrito.rb hamburger.rb hello.rb
</code>