April 2022 Summaries
2 posts from Felt
Filter
Month:
Year:
Post Summaries
Back to Blog
Hashrocket's post on building an Elixir CI with GitHub Actions has influenced many in the Elixir community by recommending a setup involving Elixir, cache restoration, formatting checks, Credo, Dialyzer, and ExUnit tests. However, the authors of the post have expanded this basic CI setup significantly by implementing a system that deploys a staging environment for every pull request (PR), allowing for real-time code testing and feedback, reducing friction in the PR review workflow. They also parallelize CI tasks to provide faster feedback, refactor CI boilerplate to streamline processes, and clear caches on retries to ensure reliability. Additionally, they incorporate extensive static analysis and code coverage checks, use Dependabot for dependency management, and are exploring further improvements like end-to-end tests and more powerful build machines. The overarching philosophy is to maximize feedback and efficiency while minimizing human oversight and error.
Apr 12, 2022
1,977 words in the original blog post.
Hashrocket's post "Build the Ultimate Elixir CI with GitHub Actions" has significantly influenced the Elixir community by providing a widely accepted setup for continuous integration (CI). While the basic setup includes steps like setting up Elixir, restoring cache, and running tests, the team at Felt has expanded on this by implementing more advanced practices. These include deploying a staging environment for every pull request (PR), parallelizing CI tasks for quicker feedback, and using static analysis tools to enhance code quality. They also created a custom GitHub Action to streamline repetitive setup tasks and implemented strategies to handle flaky builds by automatically clearing caches when necessary. Additional practices include reporting code coverage, leveraging Dependabot for dependency management, and planning future enhancements like end-to-end testing and more powerful build machines. The overarching goal is to maximize developer efficiency by providing fast, comprehensive feedback and reducing the cognitive load associated with PR reviews.
Apr 12, 2022
1,977 words in the original blog post.