This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
Hey there 👋
So I work for as a Senior Test Automation Engineer, and have spent a large part of the last 3 years forming a test strategy, architecture, and processes for our UI component monorepo.
For those unaware about what a component library is, it’s essentially a repository full of reusable components for frameworks such as React / Vue etc, that can be used by consuming applications, to reduce duplication, and provide a more consistent UI experience for teams with multiple applications.
When we first started this repo, it was immediately clear there was a major lack of resources available to help QA engineers form reliable and efficient test strategy and architecture when working in monorepos, and we certainly found it an extremely daunting task!
Fortunately, we’re in a much better place, and have learned a lot from working in this space, and we’re keen to share our learnings to help others.
It’s highly likely this will be a multi-part series, as there’s a lot to cover, and I’ll put some of my initial ideas below, but I’m keen to hear from those in the space about what you’d like to hear about 😊
Potential topics include
How we measure quality through component, accessibility, and visual regression tests before our UI components are used in consuming applications.
The tooling we have in place to give engineers the visibility of their change, and ensure all impacted dependencies / dependants are tested.
How we focus on developer experience and efficient CI pipelines to allow us to “fail fast”
I’d love to hear your thoughts! If you’d like to see covered!
Subreddit
Post Details
- Posted
- 2 years ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/softwaretes...