Illustration of a Git branching model.
July 16, 2020

What Your Git Branching Model Needs for Multiple Releases

Git at Scale
Branching

Selecting a Git branching model depends on the size of your team, your amount of digital assets, and your release process. It can be a challenge to choose one that can support your team now, and as they grow.

Let’s review different types of branching models in Git version control and how they can help support multiple releases.

What Is a Git Branching Model?

A Git branching model defines your branching strategy in Git. It determines when and how developers make changes and commit them back to your codebase.

Using a Git branching model can expedite the process of delivering feedback to developers. Git hosting solutions don’t come with a branching model out of the box. These Git branching models are branching patterns designed to help overcome challenges. But with Git, you have to build it!

Different Git Branching Models

Review Git branching models and their advantages and disadvantages.

Git Flow Branching Model

Git Flow is a Git branching model that uses two long-lived branches — main (sometimes known as the master branch) and development. The main branch is your production branch. Code in this branch should be the most stable. Alongside the main branch lives the development branch.

The development branch is used by developers to integrate all of their changes. Once the code is tested, it can be merged into the main branch. Depending on your branching strategy, you can have multiple types of supporting branches like feature branches, hotfixes, and release branches.

Git Flow Advantages

Keeping your development and main branches separate protects your production code, in theory. Because with this model, no matter where you are in development lifecycle, the main branch should be stable enough for a release. Because branches follow a standard pattern, developers know what each one is used for. But as you continue to add more branches, this becomes difficult to manage.

Git Flow Disadvantages

So, let’s say you recently merged a bunch of changes from the development branch into main. It’s tested and fails. Now developers have to go in and find what broke. But because changes are all promoted together from the development branch, it is almost impossible to track down the original change. Developers are lost in a sea of commits.

Using Git Flow at scale creates a complicated history because there is no relationship between branches. Many teams attempt to manage this with naming conventions documented on white boards and wikis. But developers often don’t reference external tools.

Because of Git Flows complexity, features can take days to merge, even if there aren’t many minimal merge conflicts. This can slow down your team, and even delay a release.

GitHub Flow Branching Model

GitHub Flow is considered to be a simplified model of Git Flow. Small teams use it because they don’t need to manage multiple environments and versions.

With GitHub Flow, you start with just one main branch. Developers check out a new branch to do their work. When the change is complete, a pull quest is opened, and work is reviewed and merged back into main. This Git branching model is most similar to trunk-based development.

GitHub Flow Advantages

Using a lightweight workflow like GitHub Flow can work well for CI/CD. You are only testing and automating changes to a single branch. Smaller teams can build, test, and deploy features fast. But because everyone is working off main branch, you need to implement tighter controls around code review and testing to gate changes.

Learn how to scale code reviews >>

GitHub Flow Disadvantages

Because all changes are merged into the main branch, you run the risk of production code being unstable. Before any other code can be integrated or worked on, developers need to spend time fixing any issues in the main branch.

As your teams grow, this model can cause costly merge conflicts. Everyone is merging into the same branch with no visibility into what other team members are working on.

To compensate, teams often use tagging to let developers know what is stable. These tags combined with complicated scripts attempt to show how branches are related. Overtime, these tags become out of date, making it more confusing. Is this tag current? Do you have everything you need for testing? Better send an email or slack to find out!

GitLab Flow Branching Model

GitLab Flow emphasizes feature-driven development. The main difference between Git Flow and GitLab Flow is the role of environments and issue tracking. Where Git Flow uses a main and development branch, GitHub Flow uses a staging and production environments, along with a main branch.

The main branch in GitLab Flow is still your base branch. Team members use feature branches to work on code. When work is complete, a pull request is created. Once approved, the changes are merged into main. Then they are pulled into environments for testing. Teams who use this model often have a staging area — pre-production environment — and production environment.

This model can keep the main branch stable because development work is separate from the main branch and production environments. Teams can then then cherry pick bugs, features, and fixes to move from main into production.

GitLab Flow Advantages

One of the biggest benefits of GitLab Flow is that immature code is kept separate. And unlike Git Flow, your history is cleaner. You can more easily find issues in the main branch even if you end up discovering them in production.

GitLab Flow does make the relationship between code and your issue tracker more apparent. This is because every change that is made starts with defining the issue. You can track the history. But you may still need scripting to help guide developers on where to merge on the main branch.

GitLab Flow Disadvantages

Because GitLab Flow has multiple environments, it can be complex to implement CI/CD. Moving code through multiple staging areas can potentially delay its release.

Because GitLab Flow does not track relationships between branches, how do you know which release needs changes? Now teams need to rely on naming conventions or external documentation.

Git Flow, GitHub Flow, and GitLab Flow are just three types of Git branching models.

These branching models work well for simple release schedules. As your teams and projects grow, you may need to support multiple releases. How can you adapt your strategy to scale? Can you use the same Git branching model?

Git Branching Strategy for Multiple Releases

Here’s how you can adapt your Git branching strategy for multiple releases.

Git Flow For Multiple Releases

With the Git Flow model, your last release version maps directly back to your main branch. It works well for a traditional release model with a regular cadence. If you are Agile, this may be tough. Because Git Flow uses several types of branches, you need to bring a lot together for CI/CD. This makes supporting multiple releases especially challenging.

GitHub Flow for Multiple Releases

GitHub Flow works with tags. Tags are set by the user. Although deployments can be automated using tags, they rely on developers to label correctly. Because GitHub Flow uses trunk-based development, it is really best for maintaining a single release.

GitLab Flow For Multiple Releases

Because GitLab Flow uses staging environments, you can keep development moving and release more frequently. But there can be issues when you start supporting multiple releases. Which feature needs to go in which release? Also, if you are testing in production, you need to trace changes back to their feature branch.

Regardless of your model, Git doesn’t make it clear what branches, versions, and environments need changes. With more assets, developers experience more delays. But there is a better way.

Git Branching Model From Perforce

To implement any Git branching model at scale requires a lot of resources. Large teams with a lot of assets and branches struggle to manage it all. So how can you gain traceability, help guide developers, and streamline your CI/CD pipeline?

Using Helix Core and Helix4Git — version control from Perforce — you can implement any Git branching model. But unlike other Git solutions, you can scale to support multiple releases and complex projects.

Helix4Git — What Your Git Branching Model Needs

Helix4Git allows teams to store Git assets in the powerful Perforce server. For developers, it looks just like Git (but faster). On the backend, you get a single source of truth for all of your digital assets including binaries, artifacts, artwork and more.

Developers can retrieve files faster. With your streamlined CI/CD pipeline you get 80% faster builds that take up less storage.

Helix Core — Branching With Perforce Streams

For those who need to scale and manage a lot of complexity, Git may simply not be enough. Perforce Streams can help you conquer challenges that come with supporting multiple releases, component based development, and more.

How Perforce Streams Works

When a developer creates a branch, Streams knows what files are available and developers can see which ones are being worked on (and by whom). Streams creates and shows the flow of change between branches.

Streams Branching Strategy for Multiple Releases

Your team no longer needs to rely on naming conventions and tags to determine how branches and releases are related. And forget about documenting branching strategies on a white board! Because with Streams, you always know where your code needs to go. Streams is flexible enough to adapt to your branching model and supports how you work.

You can combine Git + Streams + all your other digital assets into one pipeline. This all-in-one solution keeps your developers productive and your releases on time.

Learn more about Perforce + Git

CONTACT US

Want to learn more? Explore version control branching.