Yaroslav Grebnov

Using git

Repositories

Get the repository status

git status

Initialize a repository

git init

Initialize a bare repository

A bare repository contains only a .git directory. Bare repositories can be used on git servers and/or with git hooks.

git init --bare

Clone a remote repository

Clone branch targetBranch of the remoteRepository into targetDirectory:

git clone -b targetBranch remoteRepositoryUrl targetDirectory

Submodules

Fetch git submodules code

Presence of the .gitmodules file in the project root indicates that it has git submodules. An example of such file:

[submodule "submodule_name"]
  path = path from the project root
  url = submodule repository url
  branch = submodule repository branch

After cloning the parent repository, the git submodules directories are empty. In order to fetch their code:

git submodule init
git submodule update

The first command registeres submodules for their corresponding paths. The second one checks out the code.

Alternatively, the above two commands can be combined into one:

git submodule update --init

In order to checkout submodules inside submodules, a –recursive flag can be added to command:

git submodule update --init --recursive

Check code changes for submodules*

git diff --submodule=diff

More about git submodules: Git Submodules

Committing work

Stage all modified files for commit

git add .

Commit with staging files and adding a commit message

git commit -am "Commit message"

Save non-commited work to make pull

If you need to make a pull, but you have made some modifications you cannot commit for now, you can use a combination of git stash - git stash pop commands.

The process can be the following:

  • save non-commited work:
git stash

  • temporarily checkout another branch

  • checkout your working branch

  • apply your saved changes:

git stash pop

Discard all changes till the most recent commit

git reset --hard

Modify the most recent commit

If you need to modify the most recent commit without changing the commit message:

git commit --amend --no-edit

Branches

Get all branches

git branch -a

Checkout a targetBranch

git checkout targetBranch

Create a newBranch from the current branch

git checkout -b newBranch

Delete dangling local branches after deletion of remote ones

git remote prune origin

Remotes

Get all remotes

git remote -v

Add a remote

git remote add <remoteName> <remoteRepositoryUrl>

Modify remote repository url

Supposing we need to modify the remote (which is called origin) url from url1 to url2:

git remote set-url origin url2

Pull workingBranch from the sourceRemote

In order to have the local workingBranch synchronized with the one from the sourceRemote.

git pull sourceRemote workingBranch

Avoid redundant MERGE commit on pulling

Pull workingBranch from the sourceRemote with rebase:

git pull --rebase sourceRemote workingBranch

Push workingBranch to the targetRemote

git push targetRemote workingBranch

Push workingBranch to the targetRemote with setting an upstream

Usually an upstream is set on the first push of the branch to remote.

git push -u targetRemote workingBranch

Force push workingBranch to the targetRemote

Sometimes necessary after doing an interactive rebase, as remote and local versions of workingBranch are different and the local version cannot be pushed to targetRemote:

git push --force targetRemote workingBranch

Or a safer version:

git push --force-with-lease targetRemote workingBranch

.gitignore

Fixing directory still in repository after adding it to .gitignore

Supposing that a directory has been added to the repository and pushed to the remote. If you add this directory to .gitignore and push your changes, you will see that this directory is still present in the remote repository. In order to remove it, execute:

git rm -r --cached <directory_name>

and push your changes again.

← Jenkins
Using WSL →