Useful commands

Clone a repository

For the first time

When you clone a repository, Git automatically adds a shortcut called origin that points back to the “parent” repository. Just run:

git clone [e-mail]:host/path/to/repo.git    # Good ! Don't forget to add your ssh keys to your github to make it work.
or
git clone https://user@host/path/to/repo.git  # https but it's not safer than ssh.

Problems

When I run the clone command with the ssh, I have the following error:

fatal: Could not read from remote repository.  
Please make sure you have the correct access rights
and the repository exists

If you encounters problems with the ssh, do not directly go to the https ! Try adding your ssh key to your ssh agent by running:

$ ssh-add ~/.ssh/id_rsa  # For Mac, you need to add a -K option after the ssh-add command.

For more information check out this link.

The historic of the project is waaaaaay too big that the cloning it takes forever. Plus it might fill my drive !

I you feel that you don't need the full history right now, you can use the --depth <depth> option in order to truncated the history. It will create a shallow clone with the history truncated to the specified number of commits.

git clone --depth 5 [e-mail]:host/path/to/repo.git    # history truncated with only the 5 newest commits.

More can be found here

For a second time

You can achieve that by running the following command :

git clone --reference ../path/to/root/first/repo --dissociate [e-mail]:host/path/to/repo.git

The command use the --dissociate option. With this option, the new repository made will use the repository specify with the --reference option, to borrow objets from it, in order to reduce network transfer. Once the clone operation is done, the two repositories remain independent ! Either one can be deleted without impacting the other !

You can also us the --reference option without the --dissociate one. But keep in mind that it will create dependencies between the two repositories. see NOTE for the --shared option in this link.

Create a branch

$ git checkout develop
$ git branch sprint-X/feature-1
$ git push --set-upstream origin sprint-X/feature-1

Pulling work made on the remote

• If we want to explicitly create a merge commit when pulling :

  git pull --no-ff

• If you don't want to create a merge commit no matters what :

  git pull --ff-only

Note: It's a good practice to use the option --ff-only without thinking and then decide if we want to merge or rebase if an error shows up.

Rebase (instead of merge commit)

git pull --rebase origin master

• Explanation(s): Rebasing allow to keep a cleaner history for a repository.

• Warning: Use rebase only if the changes do not deserve a separate branch.

git rebase --abort

• Explanation(s): Cancel a current rebase.

Merge branches (instead of rebasing)

git merge origin/develop

Checkout a branch

$ git fetch origin
$ git checkout sprint-X/feature-1

Commiting (and pushing)

$ git status
$ git diff "Folder/feature-1/script.cs"`
$ git add . --all
$ git commit [-m "this is the commit message"]    # automatically call git gc --auto.
$ git push

Checking the graph of the repository

git log --graph   // Built-in
gitk              // Additional software

Acknowledgements

• Git Clone - Git documentation

• Adding ssh key to ssh agent - GitHub help

• StackOverflow - What is the difference between git merge and git merge --no-ff?

Ce(tte) œuvre est mise à disposition selon les termes de la Licence Creative Commons Attribution - Pas d’Utilisation Commerciale - Partage dans les Mêmes Conditions 4.0 International.