How to Effectively Split Large Changes into Smaller, Reviewable Pieces

How to Effectively Split Large Changes into Smaller, Reviewable Pieces


In the world of software development, the process of code review stands as a critical checkpoint before changes are integrated into the main codebase. It ensures that code not only functions as expected but also adheres to the project’s coding standards and practices. However, reviewing large chunks of code can be a daunting task, both for the author and the reviewers. It increases the complexity of understanding the changes, identifying potential bugs, and assessing the impact on the overall system. This is where the strategy of splitting large changes into smaller, more manageable pieces comes into play.

Why Split Changes?

Splitting large changes into smaller ones has several advantages:

  • Improved Understandability: Smaller changes are easier to understand, making it easier for reviewers to provide valuable feedback.
  • Quicker Feedback Loop: Smaller, focused changes can be reviewed and merged faster, speeding up the development process.
  • Easier Bug Tracking: Identifying and fixing bugs becomes simpler when changes are isolated.
  • Less Risk of Merge Conflicts: Incremental updates reduce the likelihood of conflicting changes.

Identifying Logical Units

The first step in splitting up changes is identifying logical units within your code. A logical unit can be a new feature, a bug fix, a refactoring task, or any other discrete piece of work that contributes to the project. The key is to ensure that each unit is independent and can be understood, reviewed, and merged without relying on the other parts.

Example: Adding Authentication to an Application

Suppose you’re adding authentication to an existing application. Instead of submitting one large pull request that touches the database schema, backend logic, frontend components, and tests, you break it down as follows:

  1. Database Schema Update: Start with the database changes needed to support authentication, such as adding a new users table.
  2. Backend Logic for Authentication: Implement the backend logic for registering new users and authenticating existing ones.
  3. Frontend Changes: Update the frontend to include forms for registration and login.
  4. Integration Tests: Add integration tests to ensure everything works together as expected.
  5. Glue Step: After independently verifying each component, proceed to the final integration phase. This involves unhiding the frontend code from users by removing feature flags, ensuring the backend and frontend are fully connected and communicating as intended, and performing an end-to-end test to confirm the seamless operation of the authentication system.

Each of these steps can be submitted as a separate pull request, allowing for focused and efficient review.

Staging for Review

Once you’ve identified the logical units, the next step is staging each piece for review. Here’s how you can approach this:

Incremental Development

Develop each logical unit incrementally, ensuring that each piece is complete and testable on its own. This might mean you need to mock or stub out dependencies temporarily.

Draft Pull Requests

Use draft pull requests (if your version control system supports them) for work in progress. This allows you to get early feedback from your peers without the pressure of it being in a merge-ready state.

Continuous Integration (CI)

Leverage continuous integration to run tests against each change. This not only ensures that your incremental changes do not break existing functionality and adhere to coding standards but also deploys each change individually to a staging environment where they can be tested incrementally in a real-world environment that mirrors production. Then, once the functionality is all merged and sufficiently tested, it can be deployed to production for customers.

Descriptive Commit Messages and Documentation

For each piece you stage for review, ensure your commit messages and any accompanying documentation are clear and descriptive. This helps reviewers understand the context and intent behind your changes. A great way to enhance clarity and consistency across your project is by adopting Conventional Commit Messages. This standardized style format includes a type, an optional scope, and a succinct description of the change. For example:

  • feat(auth): add user registration endpoint clearly indicates the addition of a new feature related to user authentication.
  • fix(login): correct password validation error specifies a bug fix in the login functionality.
  • refactor(database): optimize user query performance highlights a database-related code refactoring to improve performance.

By using this convention, you not only make your commit history more readable and organized but also facilitate automatic versioning and changelog generation. This practice supports a smoother review process and enhances the overall documentation of your project’s development lifecycle.

Tools and Frameworks to Assist

Several tools and frameworks can assist in this process, enhancing both efficiency and effectiveness:

  • Git: Use branching and merging strategies to manage your changes effectively. Adopting a structured workflow, such as Gitflow, can further streamline this process.
  • GitHub/GitLab/Bitbucket: These platforms offer features like draft pull requests, inline comments, and CI/CD integrations to facilitate the review process.
  • Review Tools: Tools like Phabricator or Review Board can also be used to manage and review changes. These tools are designed to integrate with Git and other version control systems, offering a robust set of features for code review, including side-by-side diffs, inline comments, and automated testing results.

Utilizing Feature Flags

Feature flags can be incredibly helpful when you need to merge code into the main branch without exposing incomplete features to users. They allow you to toggle features on or off without multiple deployment cycles.


Splitting large changes into smaller, reviewable pieces is an art that can significantly enhance the code review process. It requires planning and a bit of discipline but pays off by making reviews more manageable and productive. By following the strategies outlined above, you can ensure that your contributions are not only easier to review but also maintain the quality and integrity of the codebase.

About PullRequest

HackerOne PullRequest is a platform for code review, built for teams of all sizes. We have a network of expert engineers enhanced by AI, to help you ship secure code, faster.

Learn more about PullRequest

PullRequest headshot
by PullRequest

March 27, 2024