Identifying a single Git workflow is a necessary step in ensuring rapid delivery. Software development teams encompass contributors from various backgrounds and experiences, and they’re likely to feel comfortable with a workflow they’ve used previously. Without a single workflow, a team’s development workflow could be chaotic and slow down cycle time.
Git workflows empower teams to determine roles and responsibilities, set boundaries, and identify areas of improvement.
A centralized Git workflow enables all team members to make changes directly to the main branch (sometimes called the master branch or default branch), with every change logged in a running history. A centralized workflow involves every contributor committing to the main branch without using any other branch. This strategy works well for small teams, because team members can communicate so that multiple developers aren’t contributing to the same piece of code simultaneously. Centralized workflow can be seamless if team members communicate well, but there are limitations. If multiple developers commit to the same branch, it’s challenging to find a stable moment to release changes. Consequently, developers must keep unstable changes local until they’re ready for release.
After developers apply a stash and solve any merge conflicts, they can just commit as usual without dealing with automatic merge commits, unless someone pushed their changes at the same time. Because this strategy is simple, it is well-suited for small teams, Git beginners, and projects that don’t get a lot of updates.
Every feature gets its own branch when developers commit to this workflow. Rather than commit directly to the main branch, developers create a branch, make changes, submit a merge request (or pull request), and then merge it into main.
Ideally, a feature branch should have a lifespan of a few hours. The longer the branch lives, the higher the risk to find integration conflicts when merging back to main. After all, at this scale, there are plenty of teams working on other branches and directly streaming changes to the main branch, incrementing entropy and chances of running into conflict with local changes.
This Git workflow has the benefit of keeping a clean main branch that isn’t polluted with unfinished features. Teams of any size can use this feature branching, because it permits multiple developers to work on the same feature simultaneously. Software that’s still in development sees the most benefit from feature branching, but this workflow can be used for more mature applications as well.
Trunk-based development facilitates concurrent development on a single branch called trunk. When developers are ready to push changes to the central repository, they’ll pull and rebase from it to update the working copy of the central branch. Successful trunk-based development requires a developer to resolve merge conflicts locally. Regularly updating the local branch reduces the impact of integration changes, because they’re spotted when they’re still small, avoiding merge hell.
Trunk-based development decreases the likelihood of merge conflicts and keeps code clean, because there are many frequent, small merges made each day. With continuous integration, a trunk-based workflow ensures fast feedback and a team-oriented approach to code ownership and development.
Personal branching is similar to feature branching, but rather than have a single branch per feature, it’s per developer. This approach works well if team members work on different features and bugs. Every user can merge back to the main branch whenever their work is done.
Personal branching has similar advantages as feature branching, and also benefits from having fewer branches, so branch management is easier. Personal branches can be used for bug fixes and other small changes, and they help developers innovate if they’re interested in experimenting. Personal branching is useful for long-running features that may not fit into a single release cycle. This strategy can work well for small teams in which every team member develops their own part of the application.
A forking approach to version control starts with a complete copy of the repository. Forking effectively creates a local copy of a Git repository and provides the ability to create a new collaboration structure. In other words, every developer in the team has two repositories: a local workspace and a remote repository.
This workflow is popular for projects that have multiple developers contributing to it, particularly open source projects. After all, keeping track and providing privileges to collaborate to a repository with thousands of contributors is difficult to maintain. If a maintainer enables contributors to try their changes on their forked copy, managing change proposals is easier and safer.
With a forking workflow, contributors can push changes to a server-side repository, decreasing the likelihood of including low-quality code and bugs into the source code. Only a project maintainer can integrate changes to the source code. When large teams collaborate on software development projects, forking enables secure, quality-driven development.
With GitFlow, the main branch should always be releasable to production, and there should never be untested or incomplete code on the main branch. When using this Git workflow, no one commits to the main branch but rather uses a develop branch with feature branches. When the develop branch is ready to go to production, a contributor creates a release branch where testing and bug fixing occur before being merged back to the develop branch. The release branch makes the code review process easier, because there’s a dedicated place to resolve conflicts when merging into the main branch. With this strategy, the main branch always reflects production.
The benefit of GitFlow as a Git production workflow is that it allows larger teams to work on complex software while still being able to quickly fix bugs in production. In addition, the release branch allows for a staging period where users can test the software in a staging environment before it’s released, which doesn’t hinder code development. Teams of any size can use GitFlow, but smaller teams may find one of the other strategies easier to use due to its complexity. When dealing with multiple environments and regular deployments, GitFlow may offer the workflow flexibility some teams require.
GitLab is more than just source code management or CI/CD. It is a full software development lifecycle & DevOps tool in a single application.Try GitLab Free