More advanced steps with git#
Creating branches#
Let us say we are not happy with our current version of the code. We would like
to go back to v1 and start fresh.
We use
$ git checkout v1
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short
* 476b980 2022-10-06 | Commit changes (HEAD, tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
$ ls
file1.txt
Now let’s create a new empty file with a new file name
$ touch file4.txt
$ git add file4.txt
$ git commit -m "A better code now?"
$ ls
file1.txt file4.txt
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short
* 1e19277 2022-10-07 | A better code now? (HEAD) [Romain Teyssier]
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
And another one
$ touch file5.txt
$ git add file5.txt
$ git commit -m "Yes it is a better code"
$ ls
file1.txt file4.txt file5.txt
Let’s go back again to our second version v2
$ git checkout v2
Warning: you are leaving 2 commits behind, not connected to
any of your branches:
6ed6b75 Yes it is a better code
1e19277 A better code now?
If you want to keep them by creating a new branch, this may be a good time
to do so with:
git branch <new-branch-name> 6ed6b75
HEAD is now at 41f8d80 Committing file3
You see that git is not happy because you didn’t create a branch for all these new commits. Indeed, you have created a new thread of commits that are in competition with what you did before. You have now 2 diverging code versions. Let’s follow git’s advice and create a new branch for these 2 new commits
$ git branch better_code 6ed6b75
We can see how many branches we have using the git branch command
$ git branch
* (HEAD detached at v2)
better_code
master
We can now see in the history of our repository all the branches using the
--all option as
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short --all
* 6ed6b75 2022-10-07 | Yes it is a better code (better_code) [Romain Teyssier]
* 1e19277 2022-10-07 | A better code now? [Romain Teyssier]
| * 41f8d80 2022-10-06 | Committing file3 (HEAD, tag: v2, master) [Romain Teyssier]
| * c6e6535 2022-10-06 | Committing file2 [Romain Teyssier]
|/
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
If you don’t use --all you only see the history of the branch you sit on,
namely here the master branch
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short
* 41f8d80 2022-10-06 | Committing file3 (HEAD, tag: v2, master) [Romain Teyssier]
* c6e6535 2022-10-06 | Committing file2 [Romain Teyssier]
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
If you examine the content of your repository, you see only the files in the
master branch.
$ ls
file1.txt file2.txt file3.txt
Let’s go back to the other branch
$ git checkout better_code
Previous HEAD position was 41f8d80 Committing file3
Switched to branch 'better_code'
We can check we are on the right branch using
$ git branch
* better_code
master
We only see the files of the second branch (our better version of the code).
$ ls
file1.txt file4.txt file5.txt
Without the --all option, we also only see the history of this branch.
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short
* 6ed6b75 2022-10-07 | Yes it is a better code (HEAD -> better_code) [Romain Teyssier]
* 1e19277 2022-10-07 | A better code now? [Romain Teyssier]
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
Merging branches#
Now that we have a better code, we want to import in this better branch what was
done in the master branch. In other words, we want to merge to work done in
these 2 diverging versions of the code. We can do this using the git merge
command.
First, let’s check we are in the correct branch using
$ git branch
* better_code
master
Second, let’s merge the master branch into the better_code branch.
$ git merge master -m "Merging previous work in better version of the code"
Merge made by the 'recursive' strategy.
file2.txt | 0
file3.txt | 0
2 files changed, 0 insertions(+), 0 deletions(-)
create mode 100644 file2.txt
create mode 100644 file3.txt
We can now have a look at the history using the --all option.
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short --all
* 8b3b89f 2022-10-07 | Merge branch 'master' into better_code Merging previous work in better version of the code (HEAD -> better_code) [Romain Teyssier]
|\
| * 41f8d80 2022-10-06 | Committing file3 (tag: v2, master) [Romain Teyssier]
| * c6e6535 2022-10-06 | Committing file2 [Romain Teyssier]
* | 6ed6b75 2022-10-07 | Yes it is a better code [Romain Teyssier]
* | 1e19277 2022-10-07 | A better code now? [Romain Teyssier]
|/
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
You can see how the branches are now converging back together. Let’s see what we have in our repository now.
$ ls
file1.txt file2.txt file3.txt file4.txt file5.txt
Wow! We have everything now.
Let’s now try something more difficult. We go back to the master branch and
modify directly the text inside file1.txt.
$ git checkout master
Edit file1.txt in order to obtain
$ cat file1.txt
This is my first file but I modified it again to match the better code version.
You can check that the file has been modified using
$ git status
You can also see the modifications using
$ git diff
Let’s commit these changes as usual.
$ git add file1.txt
$ git commit -m "Modify file1.txt"
Let’s go back to the better_code branch again and merge these changes from the
master branch.
$ git merge master -m "Trying to merge again"
Merge made by the 'recursive' strategy.
file1.txt | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Wow! It worked flawlessly.
Resolving conflicts#
Let’s try something more complex. Go in the better_code branch and edit
file1.txt so that it looks like
$ cat file1.txt
I went crazy!
Obviously you decided to make drastic changes here. Commit these changes to the
repository (using git add and git commit in sequence). Now go back to the
master branch. Edit file1.txt to make more sensible changes.
$ cat file1.txt
This is my first file but I modified it one more time to match the better code version.
Now go back again to the better_code branch and try and merge the master
branch.
$ git merge master
Auto-merging file1.txt
CONFLICT (content): Merge conflict in file1.txt
Automatic merge failed; fix conflicts and then commit the result.
Crap! We have a conflict between the two different commits. file1.txt now
looks like this:
$ cat file1.txt
<<<<<<< HEAD
I went crazy!
=======
This is my first file but I modified it one more time to match the better code version.
>>>>>>> master
You must now edit yourself this file and decide which version to use. Obviously, you want to use the correct version that looks like
$ cat file1.txt
This is my first file but I modified it one more time to match the better code version.
Now that the file has been properly edited, you can commit the fixed changes using as usual
$ git add file1.txt
$ git commit -m "Fixed conflict"
[better_code 1243e96] Fixed conflict
Pfew! That was stressful but it all went back to normal.
Now that you happy with your better version of the code, you can merge back
everything to the master branch.
$ git checkout master
Switched to branch 'master'
$ git merge better_code
Updating abc9ee2..1243e96
Fast-forward
file4.txt | 0
file5.txt | 0
2 files changed, 0 insertions(+), 0 deletions(-)
create mode 100644 file4.txt
create mode 100644 file5.txt
$ ls
file1.txt file2.txt file3.txt file4.txt file5.txt
Now the master and the better_code branches are identical. You can check the
history of the master branch.
$ git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short
* 1243e96 2022-10-07 | Fixed conflict (HEAD -> master, better_code) [Romain Teyssier]
|\
| * abc9ee2 2022-10-07 | One more modif to file1 [Romain Teyssier]
* | 1cba3e4 2022-10-07 | I went crazy [Romain Teyssier]
* | ce60f0f 2022-10-07 | Trying to merge again [Romain Teyssier]
|\|
| * 07557d4 2022-10-07 | Modify file1.txt [Romain Teyssier]
* | 8b3b89f 2022-10-07 | Merge branch 'master' into better_code This is necessary. [Romain Teyssier]
|\|
| * 41f8d80 2022-10-06 | Committing file3 (tag: v2) [Romain Teyssier]
| * c6e6535 2022-10-06 | Committing file2 [Romain Teyssier]
* | 6ed6b75 2022-10-07 | Yes it is a better code [Romain Teyssier]
* | 1e19277 2022-10-07 | A better code now? [Romain Teyssier]
|/
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
Cloning a repository#
When working in a large team of developers, it might be useful to duplicate the entire repository across different machines. git is ideally suited for this kind of collaborative development. For simplicity, we will duplicate the current repository on your own laptop, but pretend this is in fact a different developer writing code on a different computer.
First go up one level in your file system to see your directory mywork.
$ cd ..
$ ls
mywork
Now we will clone the repository using the git clone command.
$ git clone mywork cloned_work
Cloning into 'cloned_work'...
done.
$ cd cloned_work
$ ls
file1.txt file2.txt file3.txt file4.txt file5.txt
You can see that all the files of the original repository are here. Looking at the history of the cloned repository, now we get
git log --pretty=format:'%h %ad | %s%d [%an]' --graph --date=short
* 1243e96 2022-10-07 | Fixed conflict (HEAD -> master, origin/master, origin/better_code, origin/HEAD) [Romain Teyssier]
|\
| * abc9ee2 2022-10-07 | One more modif to file1 [Romain Teyssier]
* | 1cba3e4 2022-10-07 | I went crazy [Romain Teyssier]
* | ce60f0f 2022-10-07 | Trying to merge again [Romain Teyssier]
|\|
| * 07557d4 2022-10-07 | Modify file1.txt [Romain Teyssier]
* | 8b3b89f 2022-10-07 | Merge branch 'master' into better_code This is necessary. [Romain Teyssier]
|\|
| * 41f8d80 2022-10-06 | Committing file3 (tag: v2) [Romain Teyssier]
| * c6e6535 2022-10-06 | Committing file2 [Romain Teyssier]
* | 6ed6b75 2022-10-07 | Yes it is a better code [Romain Teyssier]
* | 1e19277 2022-10-07 | A better code now? [Romain Teyssier]
|/
* 476b980 2022-10-06 | Commit changes (tag: v1) [Romain Teyssier]
* c073d19 2022-10-06 | First commit [Romain Teyssier]
The first line is markedly different than the history of the original
repository, with now 3 new branches origin/master, origin/better_code and
origin/HEAD. These new branches are associated to the remote repository from
which the local one has been cloned. We can get the information regarding the
remote original repository using the git remote command.
$ git remote
origin
We can learn more about this remote repository named origin using
$ git remote show origin
* remote origin
Fetch URL: /Users/rt3504/mywork
Push URL: /Users/rt3504/mywork
HEAD branch: master
Remote branches:
better_code tracked
master tracked
Local branch configured for 'git pull':
master merges with remote master
Local ref configured for 'git push':
master pushes to master (up to date)
Let’s now make a change in the original repository.
$ cd ../mywork
$ echo "I have changed file5.txt" > file5.txt
$ git add file5.txt
$ git commit -m "Modify file5.txt in origin repository"
Let’s go back to the cloned repository. file5.txt is still empty.
$ cd ../cloned_work
$ cat file5.txt
Now let’s pull the changes in the remote repository into our cloned one.
$ git pull
remote: Enumerating objects: 5, done.
remote: Counting objects: 100% (5/5), done.
remote: Compressing objects: 100% (2/2), done.
remote: Total 3 (delta 1), reused 0 (delta 0), pack-reused 0
Unpacking objects: 100% (3/3), 287 bytes | 287.00 KiB/s, done.
From /Users/rt3504/mywork
1243e96..1cf5640 master -> origin/master
Updating 1243e96..1cf5640
Fast-forward
file5.txt | 1 +
1 file changed, 1 insertion(+)
We can check now that file5.txt is now modified as in the original repository.
$ cat file5.txt
I have changed file5.txt
When using git pull, you are in fact merging the remote branch with your local
branch, using under the hood git merge.
Using GitHub as origin repository#
Note that repository mywork does not have any origin repository. Usually,
true origin repository are located on a remote server, most of the time publicly
available pages like GitHub or BitBucket. Let’s first create a new project. Go
to the GitHub web page and make sure you are properly
logged in. Then press the button with a + sign. Choose New repository. Only
specify the name. I suggest you use se_git. Leave all other fields empty. Now
go to the repository mywork on your laptop and use the command
$ git remote add origin git@github.com:your_login/se_git.git
You can now push to the remote GitHub repository all your ongoing work using
$ git push --set-upstream origin master
If you now look at the GitHub website, you can see all your hard work listed there, including all the past history. Note that what we just did never occurs in practice. You will always create an empty project on GitHub or BitBucket first, and then clone it to your computer and start editing files there.
Let’s now learn how to change a file in your local repository and push it to the remote repository.
$ echo "I have changed file4.txt" > file4.txt
$ git add file4.txt
$ git commit -m "Modify file4 on my laptop"
[master ecac23d] Modify file4 on my laptop
1 file changed, 1 insertion(+)
$ git push
X11 forwarding request failed on channel 0
Enumerating objects: 5, done.
Counting objects: 100% (5/5), done.
Delta compression using up to 10 threads
Compressing objects: 100% (2/2), done.
Writing objects: 100% (3/3), 297 bytes | 297.00 KiB/s, done.
Total 3 (delta 1), reused 0 (delta 0), pack-reused 0
remote: Resolving deltas: 100% (1/1), completed with 1 local object.
To github.com:rteyssier/se_git.git
1cf5640..ecac23d master -> master
Check now the GitHub webpage. You will see your modifications there. Other collaborators can now directly clone your GitHub repository and contribute to your code. You are in business!
Using git rebase instead of git merge#
In large projects with multiple developers, merging different branches can result in very cumbersome histories with multiple diverging tracks converging back to the master branch with complex patterns.
git offers the rebase functionality to combine two branches along the same
trunk, one entire diverging branch after the other. Let’s try an example. Go
back to the repository mywork. Checkout branch better_code and create a new
file called file6.txt. Don’t forget to git add and git commit. Then
checkout branch master and create a new file called file7.txt. Again,
git add and git commit.
If you git log, you will see now 2 diverging branches like before. This time,
we will combine these 2 branches using
$ git rebase better_code
Successfully rebased and updated refs/heads/master.
If now you type git log, you see that the master branch has merged the
better_code branch in a single timeline. You don’t see any diverging and
converging path anymore. Commits made in parallel in the master branch appear
now after the better_code commits.
Use this online tool: https://learngitbranching.js.org/?NODEMO to build examples
that highlight the difference between merge and rebase.