Skip to main content

Modernizing Workflows with Pre-commit Hooks in Drupal

· 3 min read
Victor Jimenez
Victor Jimenez
Software Engineer & AI Agent Builder

When managing a Drupal distribution for a massive intergovernmental corporation with developers working across three continents, "code review" cannot be the first line of defense against poor syntax. By the time a senior architect is reviewing a Pull Request, it is too late to argue about indentations or missing docblocks.

In evaluating a custom upstream (CHG0099785), we noticed that CI pipelines were continuously failing due to trivial PHPCS (PHP CodeSniffer) violations, creating massive bottlenecks in the merge cycle.

The Cost of Late Detection​

If a developer submits a PR with a tab instead of two spaces, the CI pipeline still consumes 5 to 10 minutes to spin up the container, run the automated tests, and eventually fail on the syntax check.

Multiply 10 minutes of wait time by 20 developers committing 3 times a day, and the organization hemorrhaged hundreds of hours a week strictly waiting for Jenkins/GitLab to tell them they missed a semicolon.

Shift-Left: Enforcing the Edge​

To solve this, we moved the validation logic from the remote CI server directly onto the developer's local machine via automated Git Pre-Commit hooks.

1. Husky and Lint-Staged​

We integrated husky and lint-staged into the project's root package.json. Rather than relying on developers to manually run phpcs before committing, the hook intercepts the git commit command.

It specifically targets only the files currently staged in Git. Checking an entire legacy Drupal core takes minutes; checking the two files the developer just edited takes 400 milliseconds.

2. Auto-Fixing on the Fly​

The integration was configured not just to complain, but to assist. The pre-commit hook runs phpcbf (PHP Code Beautifier and Fixer) first.

If a developer uses double quotes instead of single quotes, phpcbf instantly corrects the file locally, stages the corrected file, and allows the commit to proceed silently. The developer never even notices the error.

3. Graceful Upstream Disablement​

An edge case arose when deploying the upstream distribution to downstream vendor agencies who didn't necessarily have the required binaries (like Node.js for Husky) installed locally.

We engineered the composer.json scaffolding scripts to intentionally disable or bypass these strict pre-commit hooks during downstream deployment building, ensuring the hooks strictly applied to first-party UI/UX developers while remaining invisible to the production build pipeline.

The ROI of Syntax Automation​

By shifting validation to the extreme left—right at the point of the commit—we dropped CI failure rates due to syntax by 98%. Senior architects spend their time reviewing business logic and security vectors, rather than leaving comments about line-length limits.

Need an Enterprise Architect to optimize your team's throughput? View my case studies at victorjimenezdev.github.io or connect with me on LinkedIn.

Looking for an Architect who doesn't just write code, but builds the AI systems that multiply your team's output? View my enterprise CMS case studies at victorjimenezdev.github.io or connect with me on LinkedIn.