How to Create a Base Git Repository Using a Bash Script

Last updated on 

Creating a "bare" Git repository is a common practice in development environments where multiple people collaborate on the same project, or when you need a centralized point to synchronize code without keeping a working directory. A "bare" repository doesn’t have a working directory associated with it; it only holds the version control data and history.

In this post, we will explore how to create a base repository using a simple Bash script.

The Script

Here is the script you can use to set up a "bare" Git repository:

#!/bin/bash

user="USER_HERE"
repository="api-site-app"
echo "[TASK $((++count))] Configuring git bare repository"
dir="/path/app/${repository}"
[ -d "${dir}" ] && echo "Message: Path ${dir} already exists."
[ -d "${dir}.git" ] && echo "Message: Path ${dir}.git already exists." && exit

mkdir -p ${dir}
mkdir -p ${dir}.git
pushd ${dir}.git > /dev/null 2>&1 
    git init --bare > /dev/null 2>&1
    echo "git --work-tree=${dir} --git-dir=${dir}.git checkout -f" > ${dir}.git/hooks/post-receive
    chmod +x ${dir}.git/hooks/post-receive
    chown -R ${user}:${user} {${dir},${dir}.git}
popd > /dev/null 2>&1
echo "[TASK $((++count))] Done!"

Script Breakdown

  1. Defining the user and repository name
    • The user variable should be set to the system user who will own the repository.
    • The repository variable defines the name of the repository to be created.
  2. Checking if the directory already exists
    • The script checks if the directory where the repository will be stored (/path/app/${repository}) exists. If so, it prints a message.
    • It also checks if the .git directory already exists, which would mean the repository was already initialized. If it does, the script exits.
  3. Creating the necessary directories
    • If the directories do not exist, the script creates the repository and .git directories, preparing them for Git initialization.
  4. Initializing the "bare" Git repository
    • Inside the .git directory, the git init --bare command is run to initialize a "bare" repository. This type of repository is central storage where code will be pushed, without a working tree (no code files).
  5. Configuring the post-receive hook
    • The script sets up a post-receive hook. A Git hook is a script that Git runs automatically during certain events. In this case, the post-receive hook is triggered after a git push to the repository.
    • The hook is configured to check out the files into the working directory when a new push is received.
  6. Adjusting permissions and ownership
    • The script changes permissions and assigns ownership of the files to the specified user, ensuring the repository and directories have the correct access rights.

Conclusion

This script is perfect for quickly setting up a "bare" Git repository, which can serve as a central repository for your applications. It creates the required structure, initializes Git, and configures the post-receive hook to automate the deployment of code after a push. This setup is ideal when you want to maintain centralized code on a server without the need for a working directory.

To use this script:

  1. Edit the user variable with the appropriate system username.
  2. Configure the path and repository name in the repository variable.
  3. Run the script on your server.

Now you will have a properly configured "bare" Git repository ready for use.