What Is Git? - Definition From - TechTarget

How does Git work?

Because Git is a distributed system, it can be used with or without a central repository, unlike centralized version control systems that require a server or hosting service to maintain the primary repository. With Git, each user maintains a local copy, or clone, of the repository, including its entire history. As a result, each clone serves as a backup, eliminating any single point of failure. Most operations that users perform are carried out locally, so they're not impacted by network latency issues. Users can also work out whether or not they're connected to the network.

After installing Git on their systems, users can work with their repositors by entering commands at a terminal or by using a graphical user interface, such as Sourcetree or GitHub Desktop. They can then set up their own projects or join another project. To join a project, users clone the repository from another location to their desktops, where they can modify the existing files or add new ones. The files in a user's local repository are always in one of three states:

1. Modified. The user has modified one or more files but has not staged or committed those files. Changes occur in the working directory, which is the project folder that hosts the repository. Any files that are changed or added to this folder are considered to be in a modified state until they are staged.
2. Staged. The user has marked the new or modified files as being ready to commit to the repository. Staged files are added to the staging area, which is a logical, intermediate area for reviewing the files before committing them. The staging area is more a designation than a physical location. Staged files are tracked in a special file named index.
3. Committed. The user commits the staged files to the repository. Each commit operation represents a version of the repository that is assigned a unique checksum (Secure Hash Algorithm 1), which is used to reference the commit in subsequent operations.

After users commit their changes locally, they can push them out to a central repository or to other users' repositories. Users can also pull changes from a central repository or from other users' repositories. For example, a user might update several files, stage and commit the files, and then push them out to the team's central repository on GitHub. The other team members can then pull the updated files to their own systems to ensure they're working with the latest version. However, there is nothing in Git to prevent users from pushing and pulling changes directly from each other without relying on a central repository.